INFECTION AND RESISTANCE
THE MACMILLAN COMPANY
NEW YORK • BOSTON • CHICAGO
DALLAS • ATLANTA • SAN FRANCISCO
MACMILLAN & CO.. LIMITED
LONDON • BOMBAY • CALCUTTA
MELBOURNE
THE MACMILLAN CO. OF CANADA, LTD-
TORONTO
INFECTION AND
RESISTANCE
AN EXPOSITION OF THE BIOLOGICAL PHENOMENA
UNDERLYING THE OCCURRENCE OF INFECTION
AND THE RECOVERY OF THE ANIMAL BODY
FROM INFECTIOUS DISEASE
BY
HANS ZINSSER, M.D.
Professor of Bacteriology at the College of Physicians and Surgeons, Columbia University,
and Bacteriologist to the Presbyterian Hospital, New York
Major, Medical Officer's Reserve Corps, U. S. A.
WITH A CHAPTER ON
COLLOIDS AND COLLOIDAL REACTIONS
BY
PROFESSOR STEWART W. YOUNG
Department of Chemistry, Stanford University
SECOND EDITION REVISED
gorfe
THE MACMILLAN COMPANY
1918
COPTBIGHT. 1914 AND 1918
BY THE MACMILLAN COMPANY
Set up and electrotyped Published October, 1914
Reprinted January 1916
Reprinted January. 1917
New Edition, Revised and Reset, May. 1918
TO
a. z.
THIS BOOK IS AFFECTIONATELY
DEDICATED B\ HIS
SON
387343
PREFACE
INFECTIOUS disease, biologically considered, is the reaction which
takes place between invading micro-organisms and their products, on
the one hand, and the cells and fluids of the animal's body on the
other. The disease is the product of two variable factors, each of
them to a certain extent amenable to analysis, and it is self-evident
that no true understanding of this branch of medicine is possible
without a knowledge of the biological principles which laboratory
study has revealed.
For the purpose of helping to render such knowledge easily ac-
cessible this book was written. While it is hoped that it may prove
useful to the practitioner and laboratory worker, it is intended pri-
marily for the undergraduate medical student. To many it will
seem that the subject in general and our method of treatment espe-
cially are too technical and difficult for this purpose. Our own ex-
perience contradicts this. During the past three years the writer has
had the opportunity to deliver lectures and to give laboratory courses
on this subject to medical students of 2d, 3d, and 4-th-year classes at
the Stanford and Columbia Universities. It has been a pleasant
experience to find the medical student eager for the opportunity to
obtain this knowledge and, under the present increased require-
ments for preliminary training at our best schools, fully capable of
assimilating it. It is not a good plan to attempt too extensively to
simplify material that, in its close analysis, presents complex phe-
nomena and intricate reasoning. For this reason no attempt has
been made to write an A B C of immunity as a quick road to com-
prehension. No true insight into any branch of medicine or, for that
matter, into any other science, can be attained without a certain
amount of labor; however the concepts of this subject are, indeed,
relatively simple after the first principles have been mastered, and
the writer has attempted, therefore, at the risk of seeming pedantic
in places, to treat the subject critically, separating strictly those
data which may be accepted as fact from those in which legitimate
differences of opinion prevail.
As far as was feasible every chapter has been written as a sepa-
rate unit. This has necessitated occasional repetition, but, it is
hoped, will add considerably to clearness of presentation in each indi-
vidual subject. Theories have been discussed with as little prejudice
as the possession of a personal opinion in many cases has permitted.
vii
viii PREFACE
The chapter on Colloids was written especially for the book by
Prof. Stewart W. Young, of Stanford University. Since so many
analogies between serum reactions and those taking place between
colloidal substances generally have been observed, it has seemed best
to devote this chapter entirely to the elucidation of the principles
governing colloidal reactions, so that its contents may be utilized
as explanatory of the many allusions made to colloids in the rest of
the text.
All available sources of information have been freely used. In
the large majority of cases we have had access to the original papers
and monographs. However, we acknowledge much aid from care-
ful reading of the admirable summaries, written by acknowledged
authorities, in the works edited by Kolle and Wassermann, and
by Kraus and Levaditi. Similar acknowledgment is made to
equally important sources in Weichhardt's Jahresbericht, the Bulle-
tins of the Pasteur Institute, and in such text-books as those of Paul
Theo. Miiller, Emery, Adami, Gideon Wells, Marx, Dieudonne, and
others. It is needless to acknowledge the use of such classics as that
of Metchnikoff or of the many critical writings of Bordet and of
Ehrlich — masters who have helped to shape the thoughts of all men
working in this field.
The writer takes pleasure in acknowledging many helpful sugges-
tions from his associates, Drs. Hopkins and Ottenberg, and much
aid, in the verification of references, from Mr. Walter Bliss, Fellow
in the Department of Bacteriology.
PREFACE FOE SECOND EDITION
THE revision for the second edition of this book has been made
under difficult conditions — since, for the greater part of the time
immediately preceding its issue, the writer was removed from the
facilities of libraries and journal files. There are, in consequence,
many spots, here and there, which have not been as thoroughly
scrutinized as we would wish. Nevertheless, all the important
changes necessitated by the lapse of time since its first printing
have been made and much new material added.
The chapters on anaphylaxis have been almost completely re-
written, as demanded by changes of view resulting from the work
of others, as well as our own. The Abderhalden reaction having
been proved to be an interesting camouflage, the material contained
in this section has been revised and the more recent work on
enzymes added. The development of conceptions of non-specific
serum and cellular reactions has been discussed especially in con-
nection with the recent important work of Jobling and Peterson.
A section of Immunity in Syphilis has been added, based on our
own studies with Hopkins and McBurney, and the chapter dealing
with specific therapy in various infections has been revised and
expanded. In addition to this, many minor alterations and com-
ments have been made, paragraphs omitted and inserted.
Throughout, there has been no attempt to make the book what is
generally known as a "practical manual." There are many books
which devote themselves particularly to the instruction of workers
in measurements and manipulations, and we believe now, as we did
before, that the present volume would prove most useful by present-
ing in detail the fundamental principles underlying the biology of
infectious diseases. Such treatment of the subject is, in our opinion,
indispensable as a preliminary training for those who deal with in-
fection in the clinic or the laboratory, and the material on which
the book is based is included, at present, in lectures required of
second year students at the Columbia School of Medicine and is a
prerequisite to further practical training in Serology.
HANS ZINSSER.
CONTEXTS
PAGB
CHAPTER I. — INFECTION AND THE PROBLEM or VIRULENCE ' . . * . 1
Scope of subject. Conception of infection. Attributes of pathogenic
microorganisms. Forms of infection. Influences of biological adapta-
tion. Classification of parasites on the basis of invasive properties.
Factors which determine the power to invade. Fluctuations in viru-
lence. How microorganisms defend themselves against destruction.
Serum fastness, arsenic fastness, capsule formation, inagglutinability,
etc. Resistance to Phagocytosis. Development of offensive properties
on the part of bacteria. Specificity of different infections. Chronic
septicaemia. "Sub-infection." Selective lodgment in tissues. Locali-
zation and generalization. Incubation time.
CHAPTER II. — BACTERIAL POISONS . ... . . . . . . . 28
Part played by bacterial poisons in clinical manifestations. Pto-
maines. Importance of ptomaines in disease. True toxins or exo-
toxins. Endotoxiris. Chemotactic bacterial extracts. Basic proper-
ties of true toxins. Other substances biologically similar to them.
Analogy to enzymes. Snake venoms. Incubation time of toxins.
Conception of antitoxins. Work of Vaughan. Researches of Fried -
berger. Absorption of toxins. Selective action of toxins. Distribu-
tion of tetanus poison. Causes underlying selective action in general.
Injury done during the excretion of toxins. Union of toxins with sus-
ceptible cells. Importance of cell lipoids.
CHAPTER III. — OUR KNOWLEDGE CONCERNING NATURAL IMMUNITY, AC-
QUIRED IMMUNITY AND ARTIFICIAL IMMUNITY -.• -, - . . 49
The struggle between the infectious agent and the defensive forces
of the body. External defenses. Skin secretions. Natural vs. arti-
ficially acquired immunity. Species immunity. Racial immunity. Dif-
ference between individuals. Inheritance of natural immunity. Im-
munity resulting from an attack of the disease. Jenner and smallpox.
Pasteur's work with chicken cholera. Active immunization. Passive
immunization. Pasteur's studies on anthrax. Different methods of con-
ferring active immunity. Methods of obtaining bacterial extracts. De-
velopment of our knowledge of passive immunization. Early attempts.
Behring and his collaborators. Ehrlich 's work on ricin. Snake venom.
Specificity.
CHAPTER IV. — THE MECHANISM OF NATURAL IMMUNITY AND THE PHE-
NOMENA FOLLOWING UPON ACTIVE IMMUNIZATION. ... 78
Investigations on problems of inflammation. Metchnikoff's earlier
studies. Concentration of attention upon the properties of the blood.
Grohman's work. Early opposition of cellular and humoral points
of view. Buchner. Nuttall. Earlier arguments brought forward by
the two schools. Behring 's summary of the situation at this time.
Phenomena following upon active immunization. Earlier theories.
Kxhaustion theory. Retention theory. Alkalinity theory. Osmotic
theory. Discovery of specific antibodies by Behring and collaborators.
Khrlich 's study on ricin. Antitoxins. Pfeiffer's discovery of lysins.
Agglutinins. Precipitiiis. Opsonins. Tropiiis. Conception of anti-
x CONTENTS
PAGE
bodies as a whole. Generalization of the facts discovered in the case
of bacteria. Haemolysins. Cytotoxins. Haemagglutinins. Precipitins
to unformed proteins. Conception of an antigen. Nature of antigens.
Analogy of immunity with drug tolerance.
Origin of antibodies.
CHAPTER V. — TOXIN AND ANTITOXIN 104
Nature of the reaction. Earlier views. Calmette's work on snake
poison. Filtration experiments. Morgenroth 's work on HCl-toxin modi-
fications. Ehrlich's ricin neutralization. Development of the neutrali-
zation ideas by Ehrlich and Behring. Conception of antitoxin unit.
Instability of toxin. Ehrldch 's experiments. The conceptions of
M.L.D., LO and L+ doses. Discrepancy between L0 and L+. Toxoids
and toxons. Method of partial absorption. Toxin spectrum. Opinions
of Arrhenius & Madsen. Bordet's opinion. The Danysz effect. THE
SIDE CHAIN THEORY. Early work of Knorr. Ehrlich's analogy of cell
with chemical substances. Analogy with ferments. Weigert's law of
overcompensation. Antibodies and cell receptors. Theoretical con-
clusions drawn from work with tetanus poison. Importance of lipoidal
substances in brain tissue.
CHAPTER VI. — BACTERICIDAL PROPERTIES OF BLOOD SERUM, CYTOLYSIS AND
SENSITIZATION .......... 134
The phenomenon of bacteriolysis and the bactericidal effect. Haemo-
lysis. The mechanism of cytolysis. Amboceptor or sensitizer? The
complement or alexin. Iso-antibodies. Discussion of views of Ehrlich
and Bordet. Multiplicity of antibodies. Multiplicity of complement
or alexin. Anti-antibodies. Neisser and Wechsberg phenomenon of
complement deviation. Quantitative relation between complement and
amboceptor. Conglutinins.
CHAPTER VII. — DEVELOPMENT OF OUR KNOWLEDGE CONCERNING COMPLE-
. MENT OR ALEXIN. COMPLEMENT FIXATION . . ... 168
Origin of alexin. Microcytase and Macrocytase. Anti-lysins. Alexin in
osdema fluids, etc. The question of alexin in the circulating blood.
Alexin and the thyroid. Alexin and the liver. Chemical nature of
complement and alexin. Cobra-lecithid. Enzyme-like nature of alexin.
Filtration of alexin. Complement or alexin splitting. Return to activ-
ity of inactivated alexin on standing. Inactivation by shaking.
ALEXIN FIXATION. Bordet-Gengou experiments. Theoretical explana-
tion of these facts. Albuminolysins of Gengou. Alexin fixation by
precipitates. Views of earlier writers. Nicoll 's view. Writer 's opinion.
Conception of ' ' Bordet-antibody. ' ' Nonspecific alexin fixation. Impor-
tance of lipoids. Fixation by unsensitized cells, substances in sus-
pension. Anticomplementary properties of serum.
CHAPTER VIII. — PRACTICAL APPLICATIONS OF COMPLEMENT-FIXATION
METHOD. THE WASSERMANN REACTION 198
Historical. Early work on monkeys. First use of human beings.
Theories of the Wassermann reaction. Methods of preparing antigen.
Titration of antigen. Titration of haemolytic sensitizer. Alexin titra-
tion. Performance of the test. Noguchi's modification. Modifica-
tions of Bauer, Stern and others. Results of test. Reliability. Spinal
fluid, etc. COMPLEMENT OR ALEXIN FIXATION FOR THE DETERMINATION
OF UNKNOWN PROTEIN. Neisser-Sachs method. Principles of the
method. Performance of test. COMPLEMENT-FIXATION TESTS IN DIAG-
NOSIS OF MALIGNANT NEOPLASMS. Historical. Von DungenTs method.
Results obtained. Complement-fixation in glanders. Complement-fixa-
tion in gonococcus infections. Complement-fixation in tuberculosis.
CONTENTS xi
PAGE
CHAPTER IX. — THE PHENOMENON OF AGGLUTINATION 218
Discovery. Applications of clinical methods. Clinical usefulness. Re-
lation to motility. Passive role of bacteria. Bordet 's discovery of
the importance of electrolytes. Nature of agglutinogen. Alterations
by heat. Alterations in agglutinability. Reasons for agglutinability.
Specificity. Biological relations between bacteria parallel to agglutin-
ins. Castellani's method of absorption. Normal agglutinins. Agglu-
tinoids. Inhibition zones. Bordet 's views. * ' Two-phase ' ' theory. Phys-
ical interpretation. The work of Neisser and Friedemann. Acid agglu-
tination. Iso-agglutinihs.
CHAPTER X. — THE PHENOMENA OF PRECIPITATION . * ... ,248
Discovery. Bacterial filtrates. Expansion of principle to proteins in
general. Nature of precipitinogen. Specificity. Quantitative rela-
tions in the reaction. Practical uses. Nuttall's studies on precipitins
and historical relationship. Forensic uses of the test. Performance
of the test as advised by Uhlenhuth. Influence of heat upon precipi-
tinogen. Organ specificity. Ehrlich 's view of the nature of the reac-
tion. Physical views of the reaction. Presence of precipitinogen and
precipitin in same serum. Analogy with colloids of known constitution.
CHAPTER XI. — PHAGOCYTOSIS. CHEMOTAXIS . . ." . . . . 272
Early investigations. Metchnikoff 's first studies. Phagocytosis in lower
animals. Its significance. Importance in the development from the
larva to the adult. Its importance in resorption of degenerated cells.
Varieties of phagocytosis. Giant cells. Leucocytosis in response to the
presence of bacteria. In the peritoneum. Phagocytosis in tuberculosis.
CHEMOTAXIS. Botanical studies. Early studies of Leber. Early studies
of Buchner. Methods. Theories of chemotaxis. Importance of surface
tension.
CHAPTER XII. — PHAGOCYTOSIS, Continued. THE RELATION OF THE LEUKO-
CYTES AND OF PHAGOCYTOSIS TO IMMUNITY . . . . . 296
Opsonins and tropins. Metchnikoff 's attempt to establish parallelism
between phagocytosis and resistance. Work of his pupils. Metchni-
koff 's interpretations. Origin of bactericidal substances from leuko-
cytes. ' ' Macrocytase " and microcytase. Metchnikoff 's interpretation
of the Pfeiffer phenomenon. Origin of alexin. Leukoeytic bactericidal
substances. Their nature. Leukoeytic ferments. Leukoprotease. Pet-
terson 's experiments. Leukoeytic extract of Hiss. Bordet 's views.
CHAPTER XIII. — PHAGOCYTOSIS, Continued. FACTORS DETERMINING PHAGO-
CYTOSIS H . * . . . 311
Opsonins. Tropins. Metchnikoff 's conception of stimulins. Work of
Denys and his pupils. Other early observations. Work of Wright.
Conception of opsonins definitely advanced. Analysis of opsonic action.
Normal and immune opsonins. Neufeld's opinions. Bacteriotropins,
Structure of opsonins. Specific absorption of opsonins. Heat stability
of immune opsonins. Relation to other anti-bodies. Relation to alexin.
Variations in leukocytes as a factor in opsonic measurements. Resistance
to opsonic action on the part of bacteria. Relation to virulence.
CHAPTER XIV. — PHAGOCYTOSIS, Continued. OPSONIC INDEX AXD VACCINE
THERAPY . . . . • • ...'.". . 328
Wright 's work on typhoid immunization. Development of technique for
measuring phagocytio activity. The phagocytic index. Opsonic index.
Dilution method. Simon and Lamar's method. Accuracy of opsonic
index. Wright 's work on the staphylococcus infections. Relation of
opsonic index to clinical conditions. Negative phase. Summation of
negative phase. Summation of positive phase. Clinical value of op-
xii CONTENTS
PAGE
sonic index estimations. Opsonins and tuberculosis. Treatment by auto-
inoculation. The value of opsonic index determinations. The value of
vaccine therapy. Prophylaxis. Different types of infection and the
logic of vaccine therapy in each type. The production and standardiza-
tion of vaccines. The tuberculins.
CHAPTER XV. — ANAPHYLAXIS. FUNDAMENTAL FACTS . . . . .. . 358
The relation of immunity to hypersusceptibility. Various kinds of
hypersusceptibility. Historical development of our knowledge of these
phenomena. The work of Richet and others. The phenomenon of
Arthus. The phenomenon of Theobald Smith. Experimental produc-
tion of the anaphylactic state. Laws governing the condition as at
first determined. Symptoms of experimental anaphylaxis in guinea
pigs. Autopsy findings and causes of death. Changes in blood pres-
sure. Changes in temperature. Leucopenia. Diminution of comple-
ment. Symptoms in rabbits and dogs. Variation in different animals.
Anaphylactic antigen. Specificity of anaphylactic reaction. Quantita-
tive relations. Variations depending upon method of administration.
Anti-anaphylactic state. Prevention of anaphylaxis by drugs. Passive
sensitization. Conditions governing its accomplishment. Quantitative
studies of Doerr and Russ.
CHAPTER XVI. — ANAPHYLAXIS Continued. FURTHER DEVELOPMENT AND
THEORETICAL CONSIDERATIONS 384
Theory of Gay and Southard. Besredka 's theory. Gradual develop-
ment of the antigen-antibody conception. Quantitative work. Identity
of sensitizing and toxic substances. Idea of sessile receptors. Ana-
phylaxis and precipitins. The work of Vaughan. Diminution of alexin
during anaphylactic shock. Toxic substances obtained by action of
active serum. Friedberger 's "anaphylatoxin. " Obtained from pre-
cipitates. Obtained from bacteria. Is the mechanism of anaphylaxis
intravascular or cellular? Precipitins and albuminolysins. Writer's
opinion. THE MECHANISM OF ANTI-ANAPHYLAXIS. Nature of ana-
phylactic poison. Peptone shock. PHENOMENA CLOSELY RELATED TO
ANAPHYLAXIS. Toxicity of normal serum. Toxin hypersusceptibility.
CHAPTER XVII. — ANAPHYLAXIS Continued. BACTERIAL ANAPHYLAXIS AND
ITS BEARING ON THE PROBLEMS OF INFECTIOUS DISEASE . . 412
Early work on sensitization with bacterial protein. Technique for
sensitizing with bacteria. Revision of our ideas of " endo-toxin. "
Vaughan 's work on toxic protein split-products. Friedberger 's ana-
phylatoxin. Methods of production. Quantitative proportions which
must be observed. Time and temperature conditions. Bearing of this
work upon our understanding of infectious disease. Friedberger 's
interpretation. Bacterial toxemia. Is the bacterial antigen the matrix
for the poison?
CHAPTER XVIII. — ANAPHYLAXIS Continued. THE CLINICAL SIGNIFICANCE
OF ANAPHYLAXIS * • • 428
Serum sickness. Accelerated reactions and immediate reactions. Meth-
ods of avoiding anaphylaxis in antitoxin injections. Anaphylaxis and
bacterial vaccines. Asthma and hay fever. Sensitiveness to contact
with certain animals. Possible anaphylactic reason for eclampsia.
Sympathetic ophthalmia. Diagnostic reactions. Tuberculin reaction.
Luetin reaction. Discussion of tuberculin reaction. Experimental ana-
phylaxis with tuberculin. Diagnostic use of anaphylaxis.
CHAPTER XIX. — THERAPEUTIC IMMUNIZATION IN MAN . ... * 448
THERAPEUTIC USE OF DIPHTHERIA ANTITOXIN. Statistical results.
Amounts to be injected. Amount of antitoxin normally present in the
CONTENTS xiii
PAGE
human blood serum. PRACTICAL CONSIDERATIONS CONNECTED WITH
DIPHTHERIA ANTITOXIN PRODUCTION AND STANDARDIZATION. Toxin pro-
duction. LO and L+ doses. Methods of determination. Production of
antitoxin. Standardization of antitoxin, II. S. Hygienic Laboratory
method. Chemical concentration of antitoxic serum. ACTIVE IMMUNIZA-
TION IN DIPHTHERIA WITH MIXTURES OF TOXIN AND ANTITOXON. Behr-
ing's work. Use of the method. Results obtained. INTRACUTANEOUS
METHOD OF DETERMINING TOXIN AND ANTITOXIN VALUES. Principles of
the method. Uses. Application of the method to the determination of anti-
toxin in human beings. TETANUS ANTITOXIN AND ITS STANDARDIZA-
TION. Determination of the unit. ANTITOXIN AGAINST SNAKE POISON.
Calmette 's work. Differences between cobra and rattlesnake poison.
Production of antiserum. PASSIVE IMMUNIZATION ix DISEASES CAUSED
BY BACTERIA WHICH Do NOT FORM SOLUBLE TOXINS. General considera-
tion of principles involved. Difficulties. Serum treatment of epidemic
meningitis. Work of Kolle and Wassermann. Experiments of Joch-
mann. Flexner and Jobling 's experiments. Results. Present methods.
Streptococcus antiserum. Differences between various races of strepto-
cocci. Marmorek 's serum. Work of Aronson, Tavel, Van de Velde and
others. Probable manner of action. Serum treatment in pneumonia.
Neuf eld 's work. Recent experiments and methods of Cole. Serum treat-
ment of typhoid fever. Earlier experiments. Attempts to produce anti-
endotoxin. Principles involved. Immunization with trypsin digested
bacteria. Immunization with sensitized bacteria. Prospects of success.
Serum treatment of plague. Yersin's attempts. Kolle and Martini's
serum. Work of British Plague Commission. Lustig's serum. General
results obtained. FACTS CONCERNING ACTIVE PROPHYLACTIC IMMUNIZA-
TION IN MAN. General principles. Typhoid and Paratyphoid vaccina-
tion. Earlier history. Work of Wright, Kolle, and others. Russell's
report of vaccination in the United States army. Statistics. Work of
Metchnikoff and Besredka. Therapeutic vaccine treatment in typhoid
fever. Prophylactic immunization against cholera. Methods. Results.
Plague vaccination. Difficulties. Methods. Results. Smallpox vaccina-
tion. Rabies. Principles and methods of application. INFECTION AND
IMMUNITY IN POLIOMYELITIS. Epidemiology. Relation to Animals.
IMMUNITY IN SYPHILIS. Superinfection. Transmission to Animals.
Infection without Disease. INFLUENCE OF INJECTIONS OF NON-SPECIFIC
SUBSTANCES UPON INFECTIOUS DISEASES.
CHAPTER XX. — SERUM ENZYMES. LEUKOCYTIC ENZYMES. ABDERHALDEN
REACTION. PHYSICAL PRINCIPLES IN SERUM REACTION. MEIO-
STAGMIN AND EPIPHANIN REACTIONS. COLLOIDAL GOLD REACTION. 523
CHAPTER XXI. — COLLOIDS, by Professor Stewart W. Young, Stanford Uni-
versity, California . . . . ... . . . 543
Introduction. Definition. Reversible and irreversible colloids. Sta-
bility of colloidal systems. Physical properties of colloids. Form and
size. Osmotic pressure. Rate of settlement. Brownian movement.
Electrical properties of colloids. Surface tension. Chemical properties
of colloids. Flocculation of colloids by electrolytes. Salts and acid
electrolytes. Influence of concentration. Difference in sensitiveness to
electrolytes. Explanation of phenomenon. The "zone-phenomenon."
Mutual reactions of colloids. Mutual flocculation. Protective action.
Theories of interaction. The preparation of colloid solutions. Applica-
tions to biology. Living tissues as colloids. Agglutination of bacteria.
Analogy to colloid phenomenon. Electrical charge Carried by bacteria.
Sensitiveness to light. Danysz phenomena. Conclusions.
INFECTION AND RESISTANCE
CHAPTER I
INFECTION AND THE PROBLEM OF VIRULENCE
THE early history of our knowledge of infectious disease is that
of fermentation. It was a philosopher, Robert Boyle, writing in the
17th century, who prophesied that the problem of infectious dis-
ease would be solved by him who elucidated the nature of fermenta-
tion. His prediction was fulfilled 200 years later by the train of
investigations begun by Cagniard-Latour and by Schwann, and car-
ried to a brilliant culmination by Pasteur. It was the discovery of
the living nature of ferments and the specific nature of the various
micro-organisms which caused the several forms of fermentation,
and especially of putrefaction, which made possible rational investi-
gations in the field of infectious disease and led by analogy, first to
logical speculation — then to actual experimental proof of the etiolog-
ical relationship between the minute forms of life and the com-
municable diseases.
It is not much more than 50 years since Pollender described the
anthrax bacillus in the blood and spleens of animals dead of this
disease. In this short period the large number of maladies of ani-
mals and human beings caused by micro-organisms belonging both
to the varieties spoken of as bacteria and to those classified as
protozoa has necessitated the segregation of this branch of knowl-
edge into a separate chapter.
The period of etiological investigation is now approaching its
maturity. The causative agents of most of the more common infec-
tious diseases have been discovered, and the biology of many of the
pathogenic micro-organisms has been thoroughly studied both in
their artificial cultures and in the infected animal body. In spite
of a considerable accumulation of facts, however, the science of
immunity, that is, the study of the defensive powers of the living
animal body against infection, is still in its infancy, and the practi-
cal therapeutic successes based on this science are disappointingly out
of proportion to the really large amount of detailed knowledge of
cellular and serum reactions at our disposal.
The study of putrefaction and of fermentation — though furnish-
ing the basic analogy from which the first impulse was obtained —
1
2 A . ... :t : .INFECTION AND RESISTANCE
presented after all a problem infinitely more simple than that of the
infection of living tissues with bacteria. For, given any organic
material containing suitable nutritive constituents, with favorable
environmental conditions of moisture and temperature, and spon-
taneously or experimentally inoculated with germs of a proper
species, and the phenomena which ensued were merely those of bac-
terial growth, in which an active part was played by the bacteria
only, the dead organic materials serving simply as a passive men-
struum for these activities.
During the earlier days of the development of bacteriology,
therefore, when the attention of investigators was concentrated prij
marily upon the discovery of the specific causal agents of various in-
fectious diseases, it seemed that the simple bringing together of
pathogenic germ and susceptible subject should suffice for the ac-
complishment of an infection. We have learned, however, that the
process is much more involved, and that, fortunately for the sur-
vival of the higher animals and man, the conditions which determine
infection are intimately dependent upon a variety of secondary
modifying factors.
Throughout nature bacteria are abundant, and the environment
of man and animals, the outer integuments of skin and hair, and the
mucous membranes of the conjunctive, the intestinal and respira-
tory tracts, are constantly inhabited by a thriving bacterial flora.
The distribution of certain species in definite localities is often suffi-
ciently constant to be regarded as a normal condition. Thus the
Bacillus xerosis is a characteristic inhabitant of the conjunctiva,
certain cocci and spirilla are always present in the mouth and
pharynx, as is Doderlein's bacillus in the vagina. The fact that
bacilli of the colon group are invariably present in the bowels of ani-
mals and man from the first few days or hours after birth has even
been interpreted by some investigators as a physiologically beneficial
condition. In the course of ordinary existence, therefore, and much
more so during the course of accidental exposure to individuals in
whom infection is present, the bodies of the higher animals are in
intimate contact, not only with ordinarily harmless bacteria (sapro-
phytes), but also with many varieties of the micro-organisms spoken
of as "pathogenic" or disease-producing. Perfectly normal indi-
viduals have, then, on occasion, been found to harbor diphtheria
bacilli in nose and pharynx, meningococci have been found in simi-
lar localities, and tetanus bacilli, the bacillus of malignant edema,
the Welch bacillus, and other distinctly pathogenic germs have been
isolated from the intestinal contents of individuals who showed no
evidence of disease. In fact, the problem of the so-called bacillus car-
riers— persons who, though themselves apparently well for the time
being, harbor within their bodies and distribute to their environ-
ment bacteria capable of causing disease in others — is, as we shall
THE PROBLEM OF VIRULENCE 8
see, now recognized as one of the most important difficulties of sani-
tary prophylaxis. In the case of typhoid fever this is particularly
true, for it is now well known that a perfectly healthy individual may
harbor typhoid bacilli in the gall-bladder for years and constitute,
through all this time, a constant focus of danger to the public health.
The accomplishment of an infection, then, is not determined
merely by the fact that a micro-organism of a pathogenic species
finds lodgment in or upon the body of a susceptible individual, but
it is further necessary that the invading germ shall be capable of
maintaining itself, multiplying and functionating within the new
environment. An infection, then, or an infectious disease, is the
product of the two factors, invading germ and invaded subject, each
factor itself influenced by a number of secondary modifying circum-
stances, and both influenced materially by such fortuitous conditions
as the number or dose of the infecting bacteria, their path of en-
trance into the body, and the environmental conditions under which
the struggle is maintained.
We have in truth, then, a battle of two opposed forces, the result
of which is infectious disease. And it is the systematic analysis of
these forces in their variable conditions, and the laws which govern
them, which constitutes the science of immunity. It is the initial
skirmish between the two which determines whether or not a foot-
hold shall be gained upon the body of the subject and an infection
thus established, and it is the balance between them which decides
the eventual outcome of recovery or death. And though it is un-
fortunately true that much of the knowledge gained by such studies
has yielded no direct therapeutic results, the facts that have been
revealed are fundamental to the pathology of infectious disease and
as essential to the clinical understanding of these maladies as is the
knowledge of the mechanism of the circulation, the chemistry of
metabolism, or the structural changes of the tissues to the compre-
hension of other pathological conditions.
And from this point of view the study of infectious diseases can
be made an eminently logical one, in that, knowing the criteria
which govern the infection of a human being with a given germ,
knowing the probable path of entrance, manner of distribution, and
biological activities of the micro-organism, and the peculiarities of
the mechanism of resistance set in motion in the body by this par-
ticular infection, definite clinical deductions can often be made.
One of the most fundamental facts, immediately apparent on
considering the problems of infection, is the phenomenon that among
the innumerable varieties of bacteria and protozoa present in nature
there is a very limited group which is capable of becoming parasitic
upon the body of higher animals, and among these a still smaller
proportion which is capable of being "pathogenic" or causing dis-
ease. We have used the terms pathogenic and non-pathogenic as
4 INFECTION AND RESISTANCE
practically synonymous respectively with "parasitic" and "sapro-
phytic." But, as we shall see, although as a rule a micro-organism
must be parasitic to possess pathogenic powers, some of the true
saprophytes or so-called half-saprophytes may, be pathogenic under
certain conditions, and the terms do not cover each other absolutely.
It is reasonable to suppose that all micro-organisms were origi-
nally in the condition which we designate by the term "saprophytic."
By this term we imply that these germs maintain themselves only
upon dead organic matter and do not thrive in or upon the living
animal tissues. The class of saprophytes is widely distributed and
constitutes, of course, the most important group of bacteria in na-
ture, since upon the activities of these germs depends the unlocking
of nitrogen and carbon from the organic complexes in the dead bod-
ies and waste products of animals and plants. Such bacteria if
strictly saprophytic, that is, entirely unable to maintain themselves
upon living tissues, have little importance as producers of disease,
or, expressed in technical terms, have little "pathogenicity." Nev-
ertheless, there are cases in which strict saprophytes may cause dis-
ease by lodging upon and growing in animal tissues which have
been killed by other causes, so-called necrotic areas ; and these, still
being in relation with the body as a whole through the blood and
lymph channels, furnish an area of saprophytic growth from
which products of putrefaction or even bacterial poisons may be
absorbed. While, as a rule, the disease following the invasion
of necrotic tissue — such as gangrenous amputation stumps, old
unhealed sinuses, diabetically gangrenous areas, etc., may be caused
by a large variety of saprophytic bacteria, there are a few very
important and specifically pathogenic bacteria which are, strictly
speaking, saprophytes. Thus tbe form of meat poisoning caused
by the Bacillus botulinus is due entirely to the poison formed by
this bacillus outside of the body within the substance of the dead
foodstuff, and disease ensues as the result of subsequent ingestion of
this poison with the food. In the same way the tetanus bacillus and,
less strictly speaking, the diphtheria bacillus, at least in its ordinary
mode of attack, are rather closer to the class of saprophytes than to
that of the parasites, since neither of these bacteria, under usual
•circumstances, invades the substance of the tissues beyond the point
•of initial lodgment, causing disease only by the production of
specific poisons, a condition known as "toxemia" or intoxication.
The tetanus bacillus, moreover, is not usually capable of maintain-
ing itself and multiplying even at the point of initial lodgment
unless the tissues have been injured by trauma or irritated by the
presence of foreign bodies. Bacteria of such characteristics, there-
fore, though pathogenic — that is, incitant of disease — remain never-
theless essentially saprophytes living upon the dead animal tissues,
not invading the living cells or body fluids. It is true that investi-
THE PROBLEM OF VIRULENCE 5
gations of Frosch 1 have shown that diphtheria bacilli may often be
found in blood and organs of diphtheritic patients, and tetanus
bacilli have occasionally been found in the spleen. However, such
distribution is not necessary for the production of disease by these
bacteria, and the essential point remains that they may cause violent,
often fatal, disease without truly departing from their saprophytic
mode of life upon dead tissues. Between the saprophytes and the
true parasites or invaders of living tissue many transitions occur,
and the condition of parasitism is probably a form of specific adap-
tation.
How such transition may be biologically developed is probably
well illustrated by the investigations of Italian bacteriologists upon
tetanus bacilli.2 Tarozzi 3 inoculated guinea pigs and rabbits with
tetanus spores subcutaneously and found that these spores were
rapidly transported to the liver, spleen, and kidneys, where they
could maintain a latent existence for as long as 51 days. If during
this period trauma or any injury of the organs was practiced which
led to the formation of necrotic tissue the spores would develop upon
this basis and cause acute or chronic tetanus. Canfora,4 continuing
these studies, likewise found that tetanus spores inoculated under the
skin are rapidly distributed throughout the circulation. If no
trauma has taken place at the point of inoculation the locally lodged
spores may be rapidly destroyed, probably by phagocytosis. In the
circulation they appear to be less rapidly eliminated and may be
present for from ten to thirteen days. If, during this period, there
is produced a small wound, blood clot, or necrotic area in the body —
this may serve as a focus for development and tetanus may ensue.
After ten or more days the spores disappear from the blood, but may
then take up a latent existence in some of the organs — as stated by
Tarozzi. Apart from their importance as constituting a sort of
. transitional condition between pure saprophytism and parasitism,
these investigations would seem to have much bearing upon the so-
called cases of acryptogenic tetanus."
True infection, that is, the invasion of one species by individuals
of another, and the ability of the latter to multiply and functionate
within the cell complexes of the former, is a process quite out of
keeping with the ordinary plans of nature, throughout which there
seems to be a distinct opposition to the colonization and functiona-
tion of one living being within the living substance of another.
Thus, as Bail 5 has pointed out, a mass of frogs' eggs will remain
1 Froseh. Zeitsdir. f. Hyg., Vol. 13, 1893.
2Belfanti, quoted from Canfora, Centralblt. f. Bact., I. Orig. Vol. 45,
1908.
3 Tarozzi. Centralblt. f. Bact., Orig. Vol. 38, 1905.
4 Canfora. Centralblt. f. Bact., Orig. Vol. 45, 1908.
5 Bail. "Das Problem der Bakt. Infection." Klinkhardt, Leipzig, 1911.
6 INFECTION AND RESISTANCE
entirely iminvaded while alive, though the water surrounding it may
swarm with bacteria of many varieties, but when by some accident
such a mass of eggs ceases to live, it immediately falls prey to bac-
terial infection. The same point is illustrated by the rapidity with
which intestinal bacteria will spread throughout the body after
death, when during life they have remained confined to the lumen
of the intestine, or, at most, get into the portal circulation, to be de-
stroyed in the liver. By the living cell, therefore, an opposition is
offered to invasion by bacteria, a vital function which Bail has at-
tempted to make clearer by formulating it as a law, referring to it as
"Das Gesetz der Lebensundurchdringlichkeit." Upon what cell func-
tion this vital resistance to invasion depends is to a large extent a
mystery. It would seem to rest in principle upon the fact that the
invading cell meets the invaded one under conditions peculiarly
adapted to the activities of the latter, and is overcome before condi-
tions suitable for its own activities have been established. The con-
ditions here are not unlike those observed in the case of digestive
enzymes, a comparison which becomes more than an illustrative anal-
ogy when we consider that apart from the mere mechanical disturb-
ance created by the presence of bacteria as foreign bodies the struggle
between invader and tissue is largely one of enzyme against enzyme.
Thus, for instance, the gastric juice does not act upon the mucous
membrane of the stomach during life — but after death, at autopsy,
partial digestion of this membrane by the pepsin is often seen.
Whenever this vital resistance or opposition is overcome, and
micro-organisms enter the tissues or cells, an abnormal process is
taking place, and this process is, strictly defined, infection. Never-
theless, it is by no means necessary that such infection should al-
ways be accompanied by manifestations of disease. It is true that,
in most cases, the natural resistance is such that a struggle ensues
by which the invader is destroyed or thrown off, or in which the
invaded subject is functionally injured or even killed, and the ac-
companying evidences of such a struggle constitute what we know
as infectious disease. ' But there are special cases, cases of adapta-
tion, biologically speaking, in which neither invader nor host is seri-
ously harmed.6 In the field of protozoology, especially, there are
many examples of true parasites, that is, invaders truly maintaining
their metabolism at the expense of the tissues and body substances
of the host, which do not ar,ouse reactions sufficiently vigorous to be
termed "disease." Thus the Trypanosoma Lewisi may be found in
the blood of rats 7 without noticeably affecting the health of the ani-
mals, and other protozoa have similarly been found in organs and
blood stream of a number of other apparently healthy animals. Al-
though such conditions have been frequently spoken of as "infection
6 See also Bail, loc. cit.
7 Doflein. "Die Protozoen ais Krankheitserreger."
THE PROBLEM OF VIRULENCE 7
without infectious disease," the distinction is probably one of degree
only — there being some reaction on the part of the host even in the
mildest cases, if only in the weakening by withdrawal of body sub-
stance, which distinguishes the infected from the uninfected animal.
In other cases there may even be advantage to the host, following the
infection, to the detriment of the invading micro-organism, a phe-
nomenon most clearly illustrated by the invasion of the root hairs of
leguminous plants by the Nitrogen-fixing " root-tubercle" bacilli, a
condition in which, as Fischer says, the plant may be regarded as
parasitic upon the bacteria.
The actual harm resulting from the infection must, to a large
extent, depend upon the degree of adaptation to the new conditions
of life possible on the part both of the invader and of the host. If
the invader can acquire resistance to the defensive properties of the
host, and the latter can be similarly adapted to the harmful effects of
the invader, a prolonged condition of infection might ensue, a sort
of truce without manifestations of the disease. Although this is con-
ceivable, such mutual adaptation is probably very rare in human
disease.
In cases of so-called chronic septicemia in which bacteria may
be again and again isolated by blood culture from the circulation it
is more than likely that the organisms are constantly present, not
because they multiply or maintain themselves within the circulation,
but rather because they are being continuously discharged into the
blood from an established focus in the tissues — as, for instance, on a
heart valve. We have examined the serum of patients with subacute
and chronic septicemia (endocarditis), and often found powerful
opsonic action against the invading germs even when the patient's
own serum and leukocytes were used in the tests, evidence that the
bacteria were probably being successfully disposed of after they had
gained entrance into the blood stream. In rabbits, too, in our ex-
perience and in that of Miss Gilbert of this laboratory, it would
seem that protracted septicemia is present only when secondary foci
have been established from which the bacteria are constantly being
discharged into the blood. This we believe is rather the rule and
the establishment of a balance within the blood stream an exception.
When bacteria do succeed in withstanding successfully the opposing
forces active within the circulating blood their rapid accumulation,
the collapse of the defensive mechanism, and death of the patient are
probably the most common course.
The point of view which we have expressed in the preceding
paragraph has been impressed upon us with particular insistence by
the observation of certain cases of bacteriemia following infections of
the middle ear, mastoid processes, and thromboses of adjacent veins.
In such cases it appears that the blood may be flooded with bacteria
which, nevertheless, disappear after the focus of infection has been
8 INFECTION AND RESISTANCE
removed. We have recently had the opportunity to observe this in
a case of septicemia caused by Streptococcus mucosus, in which blood
culture plates showed very numerous colonies, and in which recovery
followed promptly upon complete excision of the thrombosed veins.
It would seem to us, therefore, that bacteriemia offers a rather better
prognosis than was formerly supposed, at least in cases in which the
focus is surgically accessible.
The same principle is illustrated in the ordinary clinical course of
typhoid fever in the human being. Here the disease begins as a
bacteriemia. Very rapidly, usually within two weeks, the bacteria
disappear from the blood stream and a high serum immunity is
established in the patient. Nevertheless, the bacteria remain actively
growing within definite foci in the tissues, where they are to a cer-
tain extent protected or inaccessible to the defensive powers so suc-
cessfully active in the blood stream. At any rate the patient remains
diseased and the bacteria can be isolated from the spleen, gall-blad-
der, and intestines at a stage when they are no longer present in the
blood stream, and during which a measurement of the bactericidal
and opsonic powers of the patient will reveal a serum immunity
much higher than normal. Just why the organisms are protected
from these influences in the tissues we do not know.
On the other hand, it is nevertheless true that a certain amount
of actual adaptation between the bacteria and the body may take
place and contribute to the chronicity of an infection. This seems
to be shown especially by the experiments of Walker and others,
which are referred to in other places, in which it was found that bac-
teria grown on immune sera gain a certain amount of resistance
against the injurious properties of these substances, and evidence
more directly bearing upon the question is furnished by the studies
on the typhoid carrier state in rabbits made by Chirolanza,8 Black-
stein,9 Johnston, 10 and recently by Gay and Claypole.11 The last-
named writers found that they could regularly produce the typhoid
carrier state in these animals if they first cultivated the typhoid
bacilli upon a medium containing defibrinated rabbits' blood. Even
in these cases, however, it is not at all improbable that the typhoid
bacilli establish a permanent focus from which they are discharged
into the blood stream.
An infectious disease, therefore, may be interpreted as the result
of parasitism in which no such mutual adaptation has taken place,
and in which the invasion of the host by the micro-organism is
marked by a struggle, the local and systemic manifestations of which
constitute the disease. The disease is an evidence of conflict be-
8 Chirolanza. Ztschr. f. Hyg., Vol. 62, 1909.
9 Blackstein. Bull. Johns Hop. Hosp., 1891.
10 Johnston. Journ. Med. Res., 27, 1912.
11 Gay and Claypole. Arch, of Int. Med.., 12, 1913.
THE PROBLEM OF VIRULENCE 9
tween the two forces, mild and locally limited if the protective pow-
ers far outweigh the invasive powers of the micro-organisms, violent
if the balance is reversed. This conception is probably a correct one
in the case of the large majority of diseases — those in which invasion
is accompanied by more or less rapid and violent inflammatory and
other reactions. In diseases like leprosy, tuberculosis, and a few
others of the more chronic infections it is also possible that extensive
invasion of the body depends, not so much upon the active invasive
powers of the micro-organism, powers which we will attempt to
analyze presently, but rather upon the fact that for reasons of in-
solubility and lack of irritating properties on the part of the invader
no reaction is set up at first, and the invasion, though progressive,
elicits no violent symptoms and no energetic opposition. The in-
vader therefore progresses unopposed, becoming an incitant of dis-
turbed bodily functions to the degree of actual disease only when it
has gained a foothold in some organ and begun to proliferate, or
has multiplied in such numbers that the cumulative effect of its
toxic powers becomes manifest.
Such a conception would assign the slow and gradual but pro-
gressively invasive powers of such diseases as tuberculosis, leprosy,
and syphilis in which systemic symptoms are manifest only after
the disease has gained an extensive foothold, to the lack of acute
physiological reaction resulting from the presence of the invading
micro-organism. In the case of such infections as those caused by
some of the yeasts or blastomyces we have seen foci of blastomycotic
lodgment in the kidney and other organs surrounding which there
was neither an accumulation of mobile cells — (leukocytes or lympho-
cytes)— nor any evidence of cloudy swelling or other injury, by
poisons, of adjacent parenchyma cells. Here, as in tuberculosis 01
leprosy, the reaction induced by the presence of the micro-organisms
is slow and gradual — expressed in an eventual fixed tissue-cell reac-
tion and giant-cell formation — similar to that induced by insoluble
foreign bodies. And it may well be that the progressive ability to
multiply without arousing the invaded body to rapid and powerful
reaction may account for the .prolonged period of apparent well-
being in the early stages of such infections and permit the invaders
to pervade the body so extensively.
This point of view has been, we believe, most clearly expressed
by Theobald Smith.12 Bacteria may lack invasive or pathogenic
properties and be, therefore, immediately destroyed after gaining
entrance to the host. They may be powerfully invasive and because
of lack of adaptation arouse a violent defensive reaction on the part
of the host. "There is another type of parasite," Smith says, "which
may dispense largely with both offensive and defensive processes.
We can conceive of this type as exerting a metabolic activity approx-
12 Theobald Smith. Journ. of A. M. A., May, 1913, Vol. 60.
10 INFECTION AND RESISTANCE
imating so closely to that of the host that the latter reacts but slightly
and then only after a long period of stimulation." Into this class
he places the syphilis spirochseta and, in a somewhat modified sense,
the tubercle bacillus.
We have seen, then, that a micro-organism may be pathogenic
and still be saprophytic in its mode of life. In order that this can
occur, however, it is necessary that it should possess the power of
producing at the place of lodgment a poison or toxin which can be
absorbed and cause disease. The condition wrhich ensues is not, prop-
erly speaking, an infection, but rather a "toxemia" differing from
the toxemias resulting from the ingestion of drugs or other poisons
only in so far as the toxins are manufactured at some point of bac-
terial lodgment within the body of the victim. Typical tetanus and
diphtheria, for instance, can be produced as readily by ingestion of
the bacteria-free culture filtrates as by inoculation with the bacteria
themselves. And although these bacteria may, on occasion, become
invasive and thereby satisfy the criteria of true infection, this is not
necessary for their pathogenicity.
In the large majority of bacterial diseases, however, it is neces-
sary that the germs shall be capable of producing a true infection
before they can become pathogenic, and it is our task therefore to
attempt to analyze those bacterial attributes upon which the invasive
power or virulence may be said to depend.
In the realm of infectious micro-organisms a wide range of cul-
tural variations is encountered which indicates that some of these
germs have adapted themselves very closely to the specific environ-
mental conditions found in the living animal body, while others can
take up with ease and under the simplest cultural conditions a purely
saprophytic existence.
Many pathogenic micro-organisms have so far defied all attempts
at cultivation in artificial media. These we cannot use for examples
since it may well be that the failure of attempts in many of them
may hinge upon such simple alterations of method as the exclusion
of oxygen, the addition of fresh tissue, or the supplying of amino-
acids, which have made possible the cultivation of the spirochaeta
pallida and the leprosy bacillus. But among those which we can
cultivate there are many which require for successful cultivation
the production of artificial conditions simulating closely those ob-
taining in the living body. Thus malarial plasmodia can be made to
multiply only if furnished with uninjured human red blood cells,
within which they can develop. The gonococcus requires, in its
first cultures outside the body, a medium containing human protein ;
and the hemophile bacteria, among them the influenza bacillus, re-
quire hemoglobin. Other organisms like pneumococci, many strep-
tococci, diphtheria bacilli, and many others, though easily grown on
artificial media, are still fastidious in their requirements and develop
THE PROBLEM OF VIRULENCE 11
sparsely or not at all unless definite conditions of nutrient materials,
temperature, reaction, and osmotic pressure are observed. On the
other hand, typhoid, anthrax, and dysentery bacilli, staphylococci
and numerous other pathogenic germs grow easily and luxuriantly
on the simplest laboratory media and within a wide range of envi-
ronmental variations.
Biologically considered, we could arrange the scale of adaptation
to parasitic conditions on this basis and it would seem, a priori, that
those bacteria which had thus adapted themselves most closely to the
living body should be the most infectious. There is not, however,
such parallelism, since many of the most powerfully invasive or
virulent germs, for instance, the anthrax bacillus, have retained
their capacity for saprophytic life to the fullest extent. It is more
logical, therefore, to classify parasites, not according to their ability
to revert to saprophytic conditions, but rather, as Bail 13 has done it,
on the basis of their relative powers of invading the living body.
His classification, of course, implies that the position of each micro-
organism in this scale must be determined with reference to a given
animal species, since a germ which is highly infectious ("parasitic"
in Bail's sense) for one species may be a "half -parasite" or even a
pure saprophyte for another.
Briefly reviewed, his classification is as follows:
I. Pure Saprophytes. — (Necroparasites, superficial parasites, or
external parasites.)
Micro-organisms which under no circumstances can be made to
develop within the living tissues of a given animal. This does not
exclude their pathogenicity for this animal, since, like the diphtheria
or tetanus bacillus, they may develop and produce toxins on the
basis of a localized area of dead tissues.
II. Pure Parasites. — Organisms like the anthrax bacillus or
the bacilli of the hemorrhagic septicemia group which, implanted in
small quantity in an animal, will rapidly gain a foothold, thrive,
and spread throughout the body.
III. Half parasites, organisms which may be infectious if in-
troduced into the animal body, but, not possessing this invasive
power to the same degree as the preceding class, require the inocula-
tion of considerable quantities, often a special mode or path of in-
oculation, or even possibly a preliminary reduction of the local and
general resistance of the infected individual in order that they may
multiply and become generalized. This class includes the large
majority of the bacteria pathogenic for man.
This property of invasive power is spoken of as virulence in
contradistinction to toxicity — the latter implying merely the abil-
ity to produce poisons, and not necessarily being associated with the
power to invade.
13 Bail. LOG. cit.
18 INFECTION AND RESISTANCE
In order that a micro-organism may be a true parasite in Bail's
sense — or invasive — for any given species of animal it must of
course possess certain basic cultural attributes which enable it to
grow in the environment furnished by the host. For instance, a
micro-organism which does not grow at temperatures below 37.5° C.
cannot very well become parasitic upon cold-blooded animals. An
excellent illustration of this influence of body temperature upon
the invasive powers of bacteria is furnished by the different races
of acid-fast bacilli which invade the bodies of man and of birds.
The avian tubercle bacillus, for instance, is non-pathogenic for man
and in cultures will not develop at temperatures below 40° C.,
which is about the body temperature of most birds. The human
tubercle bacillus, on the other hand, is non-pathogenic for birds and
ceases to grow in artificial cultures when the temperature is raised
above 40° to 41° C. This is merely one of a number of examples
which might be cited to demonstrate the necessity of simple cultural
adaptation, as it influences the property of virulence. Again, it is
probable that in order to develop in the animal body it is necessary
that a micro-organism shall be capable of developing without free
oxygen. While this point is not definitely certain, it is not probable
that any of the virulent bacteria can be strict aerobes. As a matter
of experience none of the pathogenic bacteria at present known are
absolute aerobes — though many of them grow better in artificial cul-
ture when oxygen is freely present than when it is absent.
Furthermore, the conditions encountered by bacteria as they
enter the animal body will vary considerably according to the path
by which they gain entrance. Organisms entering by the intestinal
canal are subjected to conditions of acidity or alkalinity, the action
of digestive juices, of bile, and to competition with other intestinal
bacteria, forces to which many pathogenic germs will succumb, while
others may survive there and thrive. Those entering into the tissues
by way of the skin and mucous membrane, on the other hand, en-
counter an immediately mobilized protective mechanism which, suc-
cessfully resisted by some of them, might easily and quickly dispose
of small quantities of other bacteria more resistant to conditions in
the bowel. It is but natural for this reason that the accomplishment
of an infection by any given germ must depend to a great extent
upon its gaining entrance to the body by the path best adapted to its
peculiar requirements.
The mechanical protection afforded by the coverings of skin and
mucous membranes is as a rule sufficient to prevent the penetration
of any bacteria which by chance may have found lodgment upon
them. In the case of the most usual pyogenic cocci and many bacilli
such protection is probably absolute, and a distinct break of con-
tinuity, such as a bruise or a wound, even though this may be too
small to attract attention, is necessary for successful infection. In
THE PROBLEM OF VIRULENCE 13
the case of a very limited number of diseases infection seems to take
place even through the unbroken skin, and the method, often spoken
of as the vaccination method of Kolle, employed in many instances
when it is desired to produce experimental plague infection in rats
or guinea pigs, consists in merely rubbing a small amount of cul-
tural material into a shaven area of the skin. However, in this
case, as well as in other instances where mere massage of bacteria
into unbroken skin has led to successful* inoculation, it is more
than likely that success has depended upon either microscopic
lesions or possibly the violent introduction of the organisms into the
sebaceous glands, the sweat glands, or hair follicles. The defense
of intact mucous membranes, however, is by no means impervious.
While many organisms can be implanted upon mucous membranes
with impunity, there are a number of others that can cause local
inflammations upon these and can further pass through them into the
deeper tissues and thence into the general system. Thus gonorrhea
is ordinarily a disease of implantation upon a mucous membrane,
and diphtheria bacilli and streptococci give rise to localized disease
on the pharyngeal and nasal mucosse, the latter not infrequently
penetrating from the initial point of lodgment upon the mucosa into
the deeper tissues and the circulation, causing a condition of "septi-
cemia" or "bacteriemia." For the experimental determination of
the penetrative power of organisms through mucous membranes the
conjunctiva has been a favorite test object, and it has been shown
that plague 14 and glanders,15 as well as hydrophobia, may be trans-
mitted by simple instillation of infectious material into the unin-
jured conjunctival sac. In the case of hydrophobia 16 it is related
that in Paris a young man contracted hydrophobia by rubbing his
eyes with a finger contaminated with the saliva of a rabid dog. In
the case of syphilis, though often claimed, there is no positive proof
to show that infection may take place through the uninjured sur-
faces. It has been definitely shown, however, that tubercle bacilli 17
may pass into the lymphatics through the intestinal mucosa without
there being any traceable injuries on this membrane.
It may well be, however, that even without the existence of
demonstrable morphological lesions penetrability by micro-organisms
may presuppose local physiological or functional injury, such as con-
gestion or catarrhal inflammation.
Thus it is seen that the mechanical obstacle to the entrance of
micro-organisms offered by skin and mucous membranes, though
important and not to be underestimated, is by no means a perfect
safeguard.
14 Germ. Plague Com. Arb. a. d. kais. Gesundheitsamte, Vol. 16, 1899.
15 Conte. Rev. veterin., Vol. 18, 1893.
16 Galtier. Compt. rend, de la soc. biol, 1890.
17Bartel. Wien. Klinikhandt, 1906-1907.
14 INFECTION AND RESISTANCE
However, it is only very definite species of micro-organisms
which can cause disease at all when introduced into the body by
these paths. For, although the rubbing of plague bacilli into the
skin, or the inoculation of a cut surface with streptococcal or glan-
ders bacilli, will rapidly lead to progressive infection, similar inocu-
lation with the typhoid bacillus or the cholera spirillum would lead
to no such result. And, though the swallowing of pus cocci, pneu-
mococci, and a number of other micro-organisms would be entirely
without effect, similar ingestion of the typhoid and cholera organism
would usually result in typical infection.
The path of introduction, therefore, is an important considera-
tion in determining whether or not a given micro-organism may give
rise to disease. It is necessary that the manner of gaining entrance
be suited to the cultural and other peculiarities of the germ in ques-
tion. In the case of cholera, for instance, the spirillum which causes
this disease is peculiarly susceptible to the deeper defences residing
in the body fluids and cells, and cutaneous infection by the small
numbers of bacteria likely to be introduced in this way would
promptly be checked by these agencies. In the intestinal mucosa,
however, the cholera spirillum finds conditions most favorable for
rapid multiplication, and the disease is caused by the inflammation
and destruction of the mucous and submucous tissues by the poison-
ous substances emanating from the large numbers of cholera spirilla
which die and are disintegrated, as well as by the absorption of these
poisons into the circulation. The bacteria themselves, however, never
gain a permanent foothold within the blood or other organs. In the
case of typhoid fever the conditions are somewhat similar, although
here, during the earlier weeks of the disease, we have an actual
penetration of the bacilli into the circulation. This, however, prob-
ably takes place only after intraintestinal proliferation has taken
place, which then, on the injured mucosa, represents a dose out of
all proportion great when compared with the quantities that would
spontaneously come into contact with the external surface of the
body.
This leads us to another important factor concerning the invad-
ing forces, in the determination of successful infection, namely, that
of the quantity introduced or the dosage.
In order to cause infection, even when the bacteria are of the
variety known to produce disease or "pathogenic," and are brought
into contact with the body by a path suitable to their peculiar re-
quirements, the initial quantity introduced must be sufficiently large
to preclude complete annihilation by the first onslaught of the de-
fensive powers of the body. It is plain, therefore, that in the case
of bacteria weak, in power to cause disease, given the subject of in-
fection and his defences as a constant, the quantities to be introduced
must be larger than in the case of micro-organisms of violent disease-
THE PROBLEM OF VIRULENCE 15
producing properties. The dosage necessary to cause infection,
therefore, is in inverse proportion to that property of bacteria spoken
of as their "virulence." Thus we measure the degree of the so-called
virulence of bacteria by determining the smallest quantity, measured
by dilution of platinum loops or by fractions of agar slant cultures
(both very inexact methods), which will still cause infection and
death in susceptible animals of a standard weight. In the case of
micro-organisms of extreme virulence, such as the anthrax bacillus
or bacilli of the hemorrhagic septicemia group, the inoculation of a
very small number of bacteria may suffice to initiate infection. In-
deed, it has been claimed for the anthrax bacillus that the injection
of a ' single bacterium will produce fatal disease in a susceptible
animal. The inverse relation existing between the degree of viru-
lence and the number of bacteria inoculated is well illustrated by the
experiments of Webb, Williams, and Barber,18 carried out upon
white mice with anthrax, by the method of inoculation devised by
Barber.19 This technique consists in picking up single organisms with
a capillary pipette under microscopic control, from a very thin
emulsion of bacteria and injecting directly from the pipette through
a needle puncture in the skin. While requiring a considerable de-
gree of skill, the method, when successful, permits an actual accurate
count of injected bacteria instead of the merely approximate esti-
mate which can be made by consecutive dilutions of thicker emul-
sions. In their experiments with anthrax in white mice Webb, Wil-
liams, and Barber found that the inoculation of a single thread of
anthrax bacilli (3 to 6 individuals) taken directly from the blood
of a dead animal (that is, in the most virulent condition) would
regularly cause death, and it was impossible for this reason to
immunize with such bacilli. On the other hand, if taken from
12-hour agar cultures of the same strain such small quantities
would often fail to kill. The brief period of growth under
artificial conditions had sufficiently lessened the virulence of the
bacilli so that 2, 3, and more threads could be injected with-
out harm. And after several generations of such cultivation as
many as 27 and more threads could be inoculated with im-
punity.
Another example of the measurement of relative degrees of
virulence, by a method more commonly employed, may be illustrated
as follows : The problem in which this particular measurement was
used consisted in the comparison of the virulence of two strains of
pneumococcus, one (^2) successively passed through white mice, the
other (Ni) kept alive for several weeks on serum-agar. To accom-
plish this graded quantities of 18-hour broth cultures of the twa
18 Webb, Williams, and Barber. Jour. Med. Res., 1909, Vol. XV.
19 Barber. Kansas Univ. Science Bulletin, March, 1907.
16 INFECTION AND RESISTANCE
strains were injected into mice of approximately the same weight as
follows:20
Ni
Result
N2
Result
0.1 c. c.
0.05 c. c.
0.02 c. c.
0.01 c. c.
= dead 24 hrs.
= lives
= lives
= lives
0.1 c. c.
0.05c. c.
0.02 c. c.
0.01 c. c.
= dead 24 hrs.
= dead 24 hrs.
= dead 24 hrs.
= lives
This example further illustrates another important fact in con-
nection with the problem of infection — namely, that within the
same species of bacteria different races of strains may exhibit widely
varying degrees of virulence. This has been known since the days
of Pasteur, and it is indeed of great importance in the immunization
of animals that weakly virulent strains of a given micro-organism
may be used to produce a gradual immunity against the same species
of bacteria in their fully virulent condition. Though observed in
almost all species of bacteria such variations are especially notice-
able in the cases of streptococci and pneumococci — organisms in
which no two strains may be alike in infectiousness, and in which
the injection of some strains into susceptible animals may produce
no result whatever, while other strains will kill if administered in
the smallest measurable quantities. To a large extent these fluctua-
tions of virulence appear to represent degrees of adaptation on the
part of the bacteria to the conditions met with in the living body;
and the ease with which such variations can often be artificially pro-
duced would seem to furnish another proof that the property of in-
fectiousness is a biological attribute of relatively recent acquisition.
For, although no general statement of absolute accuracy can be made,
it is a fairly uniform rule that races of pathogenic bacteria gain in
virulence as they are passed through successive animals of the same
species, and lose in virulence as they are preserved upon media
under conditions of artificial cultivation.
Further showing this ability to rapidly adapt themselves is the
observation that passage through animals of a certain species will
enhance the virulence for this species, but often reduce it for animals
of another kind. Among the earliest observations on this point are
those of Pasteur 21 in his work on rabies. He found that the virus
of hydrophobia when successively passed through rabbits gained in
virulence until a degree of maximum infectiousness was attained
20 For making such accurate measurements we have recently found very
useful the Precision syringe described by Terry, Jour, of Inf. Dis., Vol. 13,
1913.
21 Pasteur and Thuillier, Compt. rend, de I'acad. des. sc., Vol. XII, 1883.
THE PROBLEM OF VIRULENCE 17
beyond which it could no longer be enhanced. After only three
passages through monkeys, however, the virulence of this "virus
fixe" for rabbits was reduced almost to extinction. His experience
with swine plague was similar. Swine plague bacilli successively
passed through rabbits and pigeons gained enormously in virulence
for these animals respectively, but lost in virulence for hogs.
There are numerous methods by which the virulence of micro-
organisms can be attenuated by laboratory manipulations, and since
many of them are of great importance in the active immunization of
animals we will reserve their detailed discussion until we come to
consider the methods of immunization themselves. Suffice it to say
in this place that most methods of attenuation consist in subjecting
the bacteria, in artificial culture, to deleterious influences, either of
unfavorably high temperature, exposure to light or harmful chemi-
cal agents, or allowing them to remain in prolonged contact with the
products of their own metabolism by infrequent transplantation. As
a rule the attenuation which inevitably follows any form of arti-
ficial cultivation in the case of bacteria like streptococci or pneumo-
cocci can be delayed by preserving them in media containing sera
or tissues. In the case of the pneumococcus, for instance, one of
the best methods of conserving virulence in storage is to keep them
either in a soft rabbit-serum-agar mixture, as practiced by Wads-
worth, or, better still, to store them within the spleen of a mouse
dead of pneumococcus infection, as recommended by Neufeld. The
mouse is autopsied and the spleen kept in the dark and cold in a des-
iccator, under sterile precautions. This, again, as well as the en-
hancement of virulence on passage through the same species of ani-
mal— or the reduction of virulence for one species by passage through
another — shows that such fluctuations are dependent upon a very
delicate biological adaptation.
It is interesting, moreover, to look upon this process of adapta-
tion as a sort of immunization of the bacteria against the defensive
powers of the host, a conception early suggested by Welch. For just
as the animal body may become more resistant to the offensive
weapons of the invaders, so it is reasonable to suppose that the bac-
terial body may gradually develop increased resistance to the de-
fensive mechanism of the host. And this, if it occurs, would of
course lead to an increase of its invasive power or virulence. The
increase of virulence by passage through animals would alone lead
us to suspect that such acquired resistance to destructive agents on
the part of the bacteria might be responsible for the enhancement,
but additional evidence pointing in this direction has been brought
by experiments in which it was shown that bacteria cultivated in the
serum of immune animals not only gained in resistance to destruction
by the serum constituents, but at the same time were rendered more
highly pathogenic. Experiments of this kind were carried out by
18 INFECTION AND RESISTANCE
Sawtchenko,22 by Danysz,23 and by Walker.24 The results of Wal-
ker are especially instructive. He worked with a typhoid bacillus
which he cultivated for a number of generations upon the serum of a
typhoid-immune animal, and found that after such treatment the
organism had gained in virulence and lost in agglutinability by im-
mune serum, and that a larger amount of specific immune serum was
necessary to protect animals against it than sufficed for protection
against normal typhoid strains not thus cultivated. We will refer
to these results more in detail in a later chapter, since the conception
will be easier to grasp when we have Considered more fully the
mechanism of defence at the disposal of the animal body.
That this power of gaining resistance against deleterious influ-
ences on the part of bacteria is not confined to their resistance to the
animal defences alone is well shown by the experiments of Danysz 25
upon the immunization of anthrax bacilli against arsenic. In in-
oculating series of 50 tubes containing arsenic dilutions (ranging
from 1 to 10,000 to 1 to 200) with anthrax bacilli Danysz found
that up to 1 to 5,000 the arsenic increased the growth of the bacilli ;
in concentrations higher than this growth was inhibited. By grad-
ually progressive cultivation of the organisms in increasing concen-
trations of arsenic he finally succeeded in obtaining growth in solu-
tions five times more concentrated than those in which they would
develop at first.
It is intensely interesting also that Danysz found, both in the
case of his serum-resistant and arsenic-resistant strains, that, as
they became less sensitive to the deleterious effects of these agencies,
they were altered morphologically in that they developed capsules.
Similar in significance to this is the very important observation that
certain strains of spirochseta pallida may acquire resistance against
salvarsan or "606." 26 These so-called arsenic-fast strains are ap-
parently unaffected by the injection of this preparation into the
patient.
. * The experiments of Danysz were probably the first to call atten-
tion to the possible relationship of bacterial capsule formation to
virulence, and this particular phase of the subject has since then
been extensively studied. It is a matter of common observation that
micro-organisms like the pneumococcus, the anthrax bacillus, some
streptococci, and a number of other germs which are capable of pro-
ducing capsules under suitable conditions are most virulent in the
capsulated stage. As the strains are passed through animals and
their virulence increases their ability to form capsules becomes more
22 Sawtchenko. Ann. Past., Vol. 11, 1897.
23 Danysz. Ann. Past., 14, 1900.
24 Walker. Jour, of Path, and Pact., Vol. 8, 1903.
25 Danysz. Loc. cit.
26 Oppenheim. Wien. kl. Woch., 23, 1910, No. 37.
THE PROBLEM OF VIRULENCE 19
•and more apparent — whereas the diminution of virulence which
takes place on artificial media is accompanied by a gradual loss -of
capsule formation. Organisms like the Friedlander bacillus which
retain their ability to form capsules almost indefinitely in artificial
culture moreover do not lose their virulence to any great extent as
long as this property is preserved. It is also well known that cap-
sulated bacteria are peculiarly insusceptible to the ordinary aggluti-
nating powers of specific immune sera. This has been noticed, not
only in the case of heavily capsulated bacteria like those of the Fried-
lander group or the streptococcus mucosus, but, in the case of plague
bacilli — where capsulation is usually present only in cultures taken
directly from the animal body and cultivated at 37° C., Shiba-
yama 27 has found a direct relation between non-agglutinability and
a slimy condition of the cultures. Cultures kept at 5° to 8° C. in
the ice-chest were easily aggiutinable and lacked the slimy property.
Cultures kept at 37.5° C. were slimy and thready in consistency and
were not as easily agglutinated by the same immune serum.
Forges 28 later showed that inagglutinable, capsulated bacteria can
be made amenable to the agglutinating action of the serum — which
we may assume to indicate vulnerability by the serum if the capsule
is previously destroyed by heating at 80° C. for about 15 minutes in
!/4 normal acid.
Against the cellular defences, the leukocytes, capsulated bacteria
seem also to be more resistant than are the non-capsulated. This
has been especially studied by Gruber and Futaki,29 who find that a
capsulated bacillus is rarely taken up by a phagocyte even when
these cells are apparently normal and able to take up the uncap-
sulated organisms. They go so far as to claim that, in the case of
anthrax in rabbits, the development or absence of a capsule deter-
mines whether or not infection can take place. The same conclusion
is reached in similar studies by Preisz,30 who does not believe that
anthrax bacilli can ever cause infection unless they possess the
power of forming capsules. All this experimental evidence points
strongly toward a probable direct relationship between capsule for-
mation and virulence, in the sense that a thickening of the ectoplasm
may in some way protect the bacteria from the destructive forces
aimed at them by the cells and fluids of the invaded body.
As a matter of fact, even when no distinct capsule is visible, it
is nevertheless possible that ectoplasmic changes may take place.
This phase of the subject has been thoroughly discussed by a number
of writers, more especially by Eisenberg.31 It appears that many
27 Shibayama. Centralbl. f. Bact., Orig. Vols. 38, 1905, and 42, 1906.
28 Forge's. Wien. klin. Woch., p. 691, 1905.
29 Gruber and Futaki. Munch, med. Woch., 6, 1906.
30 Preisz. Centralbl f. Bakt., Vol. 49, 1909.
81 Eisenberg. Centralbl. f. Bakt., I, 45, 1908, p. 638.
20 INFECTION AND RESISTANCE
bacteria, in which true capsule formation has not been observed, may
show swelling or enlargement under conditions in which their of-
fensive activities in the infected animal body are called into play.32
Radziewsky33 has noticed such swelling of B. coli in fatal guinea-
pig infections, and spoken of it as "one of the characteristic signs
of infectiousness." Kisskalt 34 has described the same thing in the
case of streptococci, and Eisenberg interprets this as signifying an
ectoplasmic hypertrophy comparable in principle to capsule forma-
tion. He looks upon the ectoplasmic zone as a protective layer, and
calls attention to the observation of Liesenberg and Zopf)35 who
showed that capsulated strains of leukonostoc mesenteroides will
withstand 85° C., a temperature at which uncapsulated forms are
rapidly killed.
There is a considerable amount of evidence, then, which seems
to indicate that the development of a capsule is at least one im-
portant method by which the bacteria can protect themselves against
the onslaught of the defences of the invaded animal body and in so
doing become more virulent.
It is not likely, however, that this merely passive increase of the
resistance to injury on the part of the bacteria accounts for the
entire train of phenomena included in an enhancement of virulence.
It has been suggested by a number of observers that definite active .--
offensive characteristics distinguish the virulent from the avirulent
bacteria, in that the former may secrete, within the living body,
substances by which the destructive powers of serum and leukocytes
are neutralized or held at bay. A very definite suggestion of such a
possibility we find expressed in the now classical paper of Salmon
and Smith 36 on hog cholera immunity, published in 1886. They
say: ". . .. the germs of such maladies are only able to multiply in
the body of the individual attacked, because of a poisonous principle
or substance which is produced during the multiplication of these
germs." 37 Bouchard formulated such a theory in 1893 by speaking
of the "produits secretes par les microbes pathogeniques," substances
which he found in cultures of virulent bacteria, and which seemed
to reenf orce the invasive powers of the germs. Kruse 38 also within
the same year developed a similar idea. He assumed that bacteria A
may secrete enzyme-like substances which paralyze the destructive ^Si-
properties of animal serum, and in this way gain the power to
12 These forms Bail has spoken of as "thierische Bazillen."
33 Radziewsky. Zeitschr. f. Hyg., Vol. 34.
54 Kisskalt. Cited after Eisenberg, loc. cit.
35 Liesenberg and Zopf. Centralbl. f. Bakt., Vol. XII, 1892.
36 Salmon and Smith. Proc. Biol. Soc., Washington, D. C., Ill, 1884.
6, p. 29.
37 A typewritten copy of this paper was kindly put at my disposal by
Prof. Theobald Smith.
38 Kruse. Ziegler>s Beitrage, Vol. XII, 1893.
THE PROBLEM OF VIRULENCE 21
invade. As a matter of fact we have learned, since that time, that
staphylococci may secrete soluble substances, "leukocidins," which
injure white blood cells, and that many bacteria produce similar
poisons, "hsemotoxins," which specifically injure red blood cells —
thereby causing anaemia and reducing the resistance of the host.
However, the correlation and further elaboration of these thoughts
of Salmon and Smith, of Bouchard and of Krnse was left to Bail,39
in what is known as his "aggressin theory." Bail maintains on the
basis of careful experimentation that virulent bacteria can produce
within the animal body substances which he calls "aggressins," upon
which depend their invasive powers or virulence. These substances
are secreted only under stress of the struggle against the unusual
defences, are not demonstrable in test-tube cultures, and are in
themselves, according to Bail, entirely non-toxic.
He obtains these aggressins by injecting virulent bacteria into the
peritoneal cavity of a guinea pig and immediately after death re-
moving the exudate. This he centrifugalizes, removes the bacteria
and cells, and sterilizes the supernatant liquid by the addition of
small quantities of chloroform. The action of the exudates in which
aggressins have been produced by the bacteria is the following:
(We take this tabulation from Bail's own paper on typhoid and
cholera aggressins in the Archiv fur Hygiene, Vol. 52, p. 342.)
1. Sublethal doses of typhoid bacilli or cholera spirilla become
lethal when the aggressin is injected with them.
2. Lethal doses of bacilli which ordinarily would cause a slow
infection only cause a rapid and severe infection when aggressins
are added.
3. The addition of aggressin neutralizes the bacteria-destroy-
ing power of immune serum in the peritoneal cavity of a guinea pig.
4. The injection of aggressin alone produces subsequent im-
munity.
It is impossible to discuss with completeness the arguments ad-
vanced for and against the correctness of Bail's views until we have
described in detail the mechanism of protection at the disposal of
animals. But the main objection brought against this theory is that
of Wassermann and Citron,40 who claim that all these properties of
the aggressive exudates can be explained by the fact that they con-
tain extracts of the bacteria (endotoxins), which, injected with
a sublethal dose of bacteria, merely enhance their action in the same
way that this would have been accomplished by the injection of
additional dead bacterial bodies. It will require much further work
before this point is settled, and the problem is peculiarly involved and
39 Ball. Archiv f. Hijg., Vols. 52 and 53, 1905 ; Folio serologica, Vol. 7,
1911.
40 Wassermann and Citron. Deutsche med. Woch.,, Vol. 31, 28, 1905.
22 INFECTION AND RESISTANCE
difficult. However, the recent work of Rosenow 41 on pneumonococci
seems to bring some reinforcement to the ranks of those who main-
tain the existence of a special offensive substance at the command of
virulent bacteria. Rosenow extracted pneumococci grown on serum
broth and found that such extracts when made from virulent strains
would protect avirulent strains from engulfment by phagocytes. The
non-virulent strains left in these extracts for 24 hours became viru-
lent. He believes, therefore, that the virulence of pneumococci de-
pends largely upon the possession of these substances which he calls
"virulins," and which in function at least are conceived as very
similar to the "aggressins."
Recent results obtained by the writer 42 with Dwyer seem to
indicate that anaphylatoxins produced from the typhoid bacillus
possess some of the properties claimed for his aggressin by Bail.
It is not impossible that the "aggressins" obtained by him were of
this nature.
Virulence, then, may be analyzed into two main attributes : one a
purely passive property of resistance or self-preservation on the
part of the bacteria, perhaps morphologically expressed in ectoplas-
mic hypertrophy and capsule formation; the other an actively of-
fensive weapon in the form of substances of the nature of the "ag-
gressins" of Bail or the "virulins" of Rosenow. The extent of our
present knowledge of details does not warrant a statement of the
case in more definite terms.
From the facts we have discussed in the preceding paragraphs
it now becomes manifest that the elements which determine the
nature of an infectious disease are twofold. On the one hand
each variety of infectious germs possesses certain biological and
chemical attributes which are specific and peculiar to itself ; by these
its predilection for path of entrance and mode of attack is de-
termined, and upon these depends the nature of the reaction called
forth in the animal body. On the other hand the degree of infec-
tion in each case, the severity of the reaction and the ultimate out-
come are determined by the balance which is struck between the
virulence of the entering germ and the protective mechanism opposed
to it.
The specific properties of each micro-organism are the factors
^vhich account for the clinical uniformity (within definite limits)
-which is observed in the maladies produced in different individuals
iby the same species of bacteria. Thus a severe typhoid fever is, in
•essential characteristics, entirely similar to a mild case — since in
"both instances the path of entrance, through the intestine, is the
same, the distribution of the germs after entrance differs only in
degree, and the reactions, local and systemic, which are called forth
41 Rosenow. Jour, of Inf. Dis., Vol. 4, 1907.
4? Zinsser and Dwyer. Proc. Soc. Exp. Biol. and Med., Feb., 1914.
THE PROBLEM OF VIRULENCE 23
are alike. And cases of this disease in general differ as a class from
the maladies caused by, let us say, the group of clinical conditions
resulting from anthrax infection, where entrance is through the
skin, and generalized infection of the blood ensues without definite
or regular localization in any given organ. Again, a localized
staphylococcus abscess will differ materially from an equally local-
ized focus of tuberculosis, because the chemical constituents of these
bacteria respectively call forth each a characteristic response on the
part of the defensive mechanism.
Such specificity of the various micro-organisms may of course
be due partly to their mode of attack and distribution, and partly,
as we shall see, to the pharmacological action of the poisonous prod-
ucts given out by them.
That both factors contribute seems beyond doubt; but recent
work, especially that of Friedberger, which is fully discussed in
another place (see p. 415), seems to show that clinical differences
Depend much less than was formerly supposed upon specificity of the
intracellular poisons, and much more upon distribution and localized
accumulation of the germs, conditions which are determined rather
by the mode and extent of invasion than by chemical differences of
poison production. This problem, rather difficult to discuss on the
limited basis of the facts so far outlined, will become clearer as we
proceed, but we need only refer at present to the essential clinical
uniformity of the various forms of septicemia, where organisms
freely circulate in the blood — with often a focus of distribution on a
heart valve — conditions in which it is rarely possible to determine
the species of the responsible germ except by blood culture. Or,
again, as Friedberger 43 points out, there is great similarity between
the ordinary pneumococcus pneumonia and that caused by the Fried-
lander bacillus. In both cases the distribution and mode of attack of
the bacteria are essentially the same, though the micro-organisms'
themselves are biologically very dissimilar.
One and the same micro-organism, on the other hand, may cause
entirely different clinical conditions, and here the type of infection
depends purely on the degree of invasion possible in the given case —
that is, the balance between virulence and resistance. A germ may
enter the body and cause an inflammatory reaction at the point of
entrance, the process remaining purely localized. In such cases the
defensive forces have been so efficient, the invasive properties of the
germ so relatively weak, that progression beyond the point of en-
trance is prevented and the resultant disease takes the form merely
of a localized abscess. This is the case when a healthy individual
is infected with an attenuated organism or by one whose species'
characteristics do not include a powerful invasive property. Thus
streptococci, if entering the tissues of a normal subject in small
43 Friedberger. Deutsche med. Woch., No. 11, 1911.
24 INFECTION AND RESISTANCE
numbers or in attenuated form, may produce a purely localized in-
fection, and ordinarily non-pathogenic germs like proteus, subtilis,
or colon bacilli may produce localized abscesses in weak and debili-
tated individuals, though implanted upon a healthy subject they
would be rapidly disposed of without gaining even a preliminary
foothold. Such tendency to localization is the common form of in-
fection in the case of a number of germs. It is the most usual type
of staphylococcus infection, for instance, in which the degree of
virulence of the strains ordinarily met is such that the balance struck
by them with the average defensive powers of man results in localiza-
tion. However, the same micro-organism, enhanced in virulence, or
gaining entrance in unusual numbers in a weakened individual, may
rapidly spread from the point of inoculation, at first by contiguity,
then by invasion of the blood and lymph channels, and become
generalized.
When organisms become generalized and circulate in the blood
the resulting condition is spoken of as septicemia or bacteriemia.
This is the form of infection commonly caused by streptococci,
bacilli of the hemorrhagic septicemia group, anthrax bacilli, and
many others. It implies a powerful invasive property and always
constitutes a condition of great- gravity when persistent. We are
learning of recent years, however, that in many infectious diseases
formerly regarded as purely localized a temporary entrance of the
bacteria into the circulation is a usual occurrence. Thus Fraeiikel 44
has shown that lobar pneumonia is almost always accompanied dur-
ing the acute stages of the disease by pneumococcus septicemia, and
in typhoid fever we now know that the organisms circulate freely in
the blood during the first two weeks of the disease, and often longer
than this.
In these and other conditions the bacteria may be gradually de-
stroyed and disappear from the blood stream as the immunity of the
subject increases. In other cases the bacterial activities may be
partially checked, the process becoming slower and more chronic.
This is especially often the case when micro-organisms after entrance
to the circulation have found a secondary lodgment upon a heart
valve, from which a continuously renewed supply of bacteria can
be given off to the blood. A special form of such "malignant endo-
carditis" caused by the Streptococcus viridans is particularly apt to-
take this chronic course.
The presence of bacteria in the blood is not, therefore, as for-
merly supposed, an invariably fatal condition.
Adami's recent work would indicate, moreover, that bacteria may
normally enter the portal or even the general circulation from the
intestine during health. This condition of "sub-infection," as he
calls it, is more fully discussed on p. 234. That colon and other in-
44FraenkeL V. Ley den Festschr., 1902.
THE PROBLEM OF VIRULENCE 25
testinal bacteria may often penetrate into the portal circulation is
indicated by the occasional occurrence of colon bacillus abscesses
after trauma of the liver. In most septicemias, however, caused by
virulent bacteria the invasion of the blood stream persists, rapid
multiplication occurs and leads to death.
From the circulation the bacteria may gain lodgment in various
organs and cause the formation of Secondary abscesses. This condi-
tion is known as ^pyemia," and may be caused by almost any bac-
teria which are capable^ of producing septicemia. Thus staphylo-
cocci, streptococci, or pneumococci may lodge in bones, joints, brain,
or kidneys, in fact in any organ in which they can gain a foothold.
However, there are evidences of distinct tissue predilections on the
part of certain germs. Thus the virus of rabies and that of polio-
myelitis, though to some extent universally distributed, seem espe-
cially to concentrate in the nervous system; cholera spirilla and
dysentery bacilli appear to find conditions most favorable for de-
velopment in the intestinal mucosa; amebic abscesses are most
common in the liver; gonococcus infections when generalized find
secondary localization with particular frequency on heart valves and
joints; leprosy bacilli have a predilection for the nerve sheaths; and
glanders bacilli injected into the peritoneum of a male guinea pig
localize with such regularity in the testicles that the experiment has
diagnostic value (Strauss test). Conversely it is only explicable on
the assumption of such selective lodgment that tubercle bacilli, even
though otherwise universally distributed through the body, will be
absent from striped muscle tissue, and rare in the walls of the stom-
ach. Such selection — as far as we can account for it at all, seems to
depend upon the varying cultural conditions encountered by the
germs in different organs.
On the other hand, localization may also be dependent upon
accidental conditions such as trauma. Infections in which the en-
trance of bacteria is coincident with injury — as in the case, for in-
stance, of compound fractures — will be able to spread throughout
the injured region much more easily than they could enter the
healthy tissue. In fact, it is well known that local tissue injury at
the point of inoculation favors infection since it furnishes a rich
substratum for growth in the form of dead cells or blood clot and
interferes with the accomplishment of a normal protective reaction.
In cases in which bacteria are circulating in the blood mechanical
injury may create a focus of reduced resistance on which the in-
vaders can gain a foothold. It is in this way perhaps that, among
other things, we can explain tuberculosis of joints or bones which
present a history of injury preceding the development of the infec-
tion— or the pleurisy and lobular pneumonias which have been known
to ensue upon the fracture of a rib.
It is also possible that bacteria may be distributed in various
26 INFECTION AND RESISTANCE
organs directly from the initial focus by embolism or by the massive
invasion of a blood vessel. It is by such breaking into a vein that
Weigert explains the generalization of miliary tuberculosis.
The inflammatory reaction which usually ensues at the point of
entrance of bacteria is merely a result of the local struggle between
invader and tissues, and the violence -of this reaction is in a large
measure an indication of the resistance of the infected subject.
When, for instance, a streptococcus of moderate virulence gains
lodgment in the skin of a healthy individual the rapid mobilization
of leukocytic and other defences may prevent further invasion by
the bacteria and lead to a struggle which is clinically evidenced by
severe local symptoms. Did the virulence of the streptococci far
overbalance the powers of resistance the local struggle might be
reduced to a minimum, the infection progressing without any, or
with but a slight local, reaction. The fact that pneumococci lodging
in the human lung ordinarily cause lobar pneumonia is merely an •
evidence of a considerable degree of resistance to these germs on the
part of the average human being. Pneumococci introduced into the
pulmonary alveoli of very susceptible animals (rabbits) may pass
directly through into the circulation, causing fatal septicemia with-
out leading to a more than mild and temporary reaction in the lungs
themselves. If, as in Wadsworth's 45 experiments, the rabbits are
partially immunized — that is, their resistance increased before the
pulmonary inoculation is carried out — a violent local reaction, anal-
ogous to lobar pneumonia, may follow, the severity of the reaction
at the portal of entry being manifestly an evidence of more energetic
opposition to further penetration of the bacteria.
The entrance of bacteria into the deeper tissues, and even the
circulation, without any, or with but slight, local evidences of infec-
tion at the point of entrance is by no means rare. The innocent
appearance of the site of the entrance of the bacteria in generalized
streptococcus infection is a common surgical observation, and a strep-
coccus-infected wound of the hand or leg in a patient dying of septi-
cemia may appear but slightly inflamed and edematous and incom-
parably milder in appearance than a staphylococcus boil with which
the patient is walking about and suffering hardly any systemic dis-
turbance.
Between the time of entrance of the bacteria into the body and
the first appearance of symptoms of disease there is always a definite
interval which is spoken of as "incubation time." This period is
made up of two definite divisions — one the time necessary for
growth, distribution, and accumulation of the bacteria, the other the
time necessary for the action of the toxin or poison which may be
secreted. The latter, the incubation time of the toxin, is a subject
which is still unclear in many of its phases, and will be discussed
45 Wadsworth. Am. Jour, of the Med. Sc., Vol. 27, 1904.
THE PROBLEM OF VIRULENCE 27
in the following chapter (see p. 37). The former, however, is easily
comprehended, in fact, is to be expected. For the small number of
bacteria which gain entrance to the tissues in spontaneous infection
is entirely inadequate in itself to produce symptoms. It is neces-
sary that multiplication shall take place until the bacteria have ac-
cumulated in number sufficient to cause noticeable physiological
disturbance. That the interval necessary for this must vary accord-
ing to the number of bacteria originally introduced, the virulence of
these, and the specific resistance of the patient goes without saying.
Von Pirquet and Schick have suggested also that the incubation
time may correspond roughly to the interval during which the sub-
ject is becoming "allergic" or hypersusceptible to the bacteria or
virus. This will be discussed at greater length in the chapter on
anaphylaxis.46
But within the limits of the variations introduced by these fac-
tors the incubation time of each infectious disease — if spontaneously
acquired — is sufficiently uniform to be characteristic. Thus the pri-
mary lesion in syphilis follows the inoculation after an interval of
two to three weeks, rabies follows inoculation with street virus after
about four to six weeks, the period being somewhat dependent on the
location of the bite ; typhoid fever takes about two weeks to develop ;
gonorrhea about five to seven days; small-pox about two weeks;
yellow fever three to five days; and scarlet fever and diphtheria
about two to six days. In general, it may be stated that within the
limits observed for each particular infection the shorter the incuba-
tion time the more severe is the infection. Thus if tetanus follows
inoculation with the tetanus bacillus within seven days the prognosis
is far more grave than when the incubation time has occupied two
or three weeks. And if localized and general symptoms follow rap-
idly (within twenty-four to forty-eight hours) after a streptococcus
infection it is likely that the process is a very severe and virulent
one.
46 Von Pirquet u. Schick. Wien. kl. Woch., 16, 1903, pp. 758 and 1244.
CHAPTER II
BACTEKIAL POISONS
WHEN bacteria have gained a foothold anywhere within the
animal body the local and general disturbances which follow, in all
but the mildest and most trifling cases, are such that we cannot
account for them solely on the basis of mechanical injury.
It may well be that the obstruction of capillaries and lymphatics
and the pressure upon parenchyma cells, always incident to inflam-
matory reactions, contribute materially to local destruction, and
thereby indirectly to systemic effects. However, even in diseases
like anthrax, in which the body of the victim after death is found
flooded throughout with masses of bacteria, these factors cannot fully
explain the clinical manifestations. And such cases, indeed, ara
extreme examples, since, in the large majority of bacterial diseases,
the illness resulting in the patient is severe out of all proportion to
the extent of the tissue area invaded.
Moreover, all infections, if at all severe, whatever their nature
or localization, give rise to fever, and this symptom alone, if care-
fully observed from hour to hour, may be sufficiently characteristic
to indicate the specific micro-organism which is causing the illness.
With this there occur alterations of the blood picture, either a
numerical increase of white blood cells (leukocytosis) or a change in
the relative proportions of the different kinds of leukocytes — or
again an anemia caused by the destruction of red cells. There may
also be degenerative changes in parenchyma cells of organs far re-
moved from the actual site of bacterial lodgment. All these facts
indicate very definitely that, apart from localized tissue destruction
or purely mechanical interference with function by capillary ob-
struction or pressure, there is at the same time an absorption of
poisonous substances emanating from the bacteria.
From the earliest days of logical investigation into the nature
of infectious disease, as soon, in fact, as cultural methods had been
introduced, bacteria were studied with the purpose of throwing light
upon this phase of their activity. As a result of such investigations
Selmi,1 in 1885, described certain basic toxic substances which he
obtained from putrefying human cadavers and for which he sug-
gested the designation "ptomain" (from rrrw/xa — dead body). These
1 Selmi. Cited from Hammarsten, "Textbook of Phyaol. Chem.," p. 16.
BACTERIAL POISONS 29
poisons were later more extensively studied by Brieger,2 Gautier,3
Griffiths,4 and others, and it was at first surmised that the formation
of such substances in the infected animal might be held responsible
for the toxemic manifestations which accompany bacterial disease.5
This, as we shall see, is not the case. Ptomains are probably not
formed in traceable quantity in the living tissues and are not in any
way identical with the specific bacterial poisons which are respon-
sible for the toxemia of infectious diseases. Nevertheless, they have
some pathogenic significance, since they are invariably products of
the proteolysis caused by bacteria and can give rise to illness when
ingested with putrefying foodstuffs. It is important, therefore, that
we discuss them briefly and consider their fundamental distinction
from the true bacterial poisons.
Whenever dead organic material, meat, fish, vegetable refuse,
etc., is left to itself under suitable conditions of moisture and tem-
perature, putrefaction sets in. As a result of bacterial growth the
protein is broken up and among the intermediate products of such
proteolysis ptomains appear. Chemically 6 T 8 these substances are
basic nitrogenous compounds which may or may not contain oxygen.
Because of their basic and often highly toxic properties they have
been spoken of as "animal alkaloids." Many of them contain only
C, H, and !N", and are ammonia substitution products. (See
Yaughan and Novy, loc. cit., p. 248.) Thus some of the simpler ones
are:
Methylamin==(CH3) KH2
Dimethylamin— ( CH3 ) 2 NH
Trimethylamin==(CH3)3 N
Among those somewhat more complex are:
Putrescin= KS2— CH2— CH2— CH2— CH2— KH2
and Cadaverin=NH2— CH2— CH2— CH2 — CH2— CH2— KE2
Samuely classifies the ptomains according to their nitrogen con-
tents as follows:
1. Those with one nitrogen atom (CgH^N) (C8H13!N")
(C10H1BN)
2. Those with two nitrogen atoms such as putrescin (C4H12N2)
and cadaverin (C5H14N2) and
2Brieger. "Die Ptomaine," Berlin, 1885; Virchow's Archiv., Vols. 112
and 115; Berl klin. Woch., 1887, 1888.
3 Gautier. Cited after Pick, Bull, de I'acad. de med., 1886.
4 Griffiths. Compt. Rend, de I'acad. des sc., Vol. 113.
5 For a historical outline of our knowledge of these poisons, as well as
for a thorough treatment of their nature, see Vaughan and Novy, "Cellular
Toxins."
6 For a discussion of the chemistry of the ptomains see Vaughan and
Novy, "Cellular Toxins," Lea Bros., Philadelphia, 1902.
7 Also Samuely in Oppenheimer's "Handbuch der Biochemie," Vol. I,
pp. 794 et seq.
8 See also Wells, "Chemical Pathology," Saunders, Phila., 1907.
30 INFECTION AND RESISTANCE
3. Those with three nitrogen atoms such as methyl guanidin
(C2H7N,).
4. Finally there is an important group which contains oxygen,
such as the substance sepsin (C5H14N2O2) obtained by Faust from
putrefying yeast cells.
They are not in all cases protein cleavage products, since bodies
of the cholin group, cholin, neurin, and muscarin, the two last named
highly toxic, are lecithin derivatives, and Samuely points out that
other lipoid cleavage products, always present in decomposing tis-
sues, may well contribute to ptomain production in the presence of
a source of nitrogen. It is interesting to note also that the vegetable
poison muscarin, isolated by Schmiedeberg from mushrooms, is
chemically identical with a toxic base found by Brieger in decom-
posing fish.
The ptomains are not poisonous in every case. The chemically
simpler ones like methylamin, di- and trimethylamin possess little
or no toxicity. Others chemically more complex — like cadaverin
and putrescin — may be capable merely of causing local necrosis,
while sepsin, closely related to cadaverin in chemical constitution,
but containing oxygen, is a powerful poison which acts violently
upon the intestinal blood vessels, causing capillary dilatation, con-
gestion, and diapedesis.9 The presence of oxygen seems indeed to be
necessary for the development of strong toxicity (Brieger, Vaughan,
and ]N~ovy). Again, the lecithin derivative, cholin, is but weakly
toxic, while neurin is exceedingly poisonous. In putrefying mix-
tures these toxic bodies appear on or about the fifth or seventh day
after putrefaction sets in, and disappear, by further cleavage, more
or less rapidly, yielding less complex nitrogenous substances that are
non-toxic.
With the limited knowledge regarding bacteria and infectious
diseases at the disposal of the earlier investigators it was but natural
that the discovery of ptomains in cultures of putrefactive bacteria
aroused the suspicion that these bodies were responsible for the
toxemia of infectious disease.
The search for poisonous substances in pure cultures of patho-
genic bacteria was, therefore, assiduously taken up by Brieger and
his pupils, and, in truth, ptomains were actually found as products
of some of the disease-producing micro-organisms, just as they had
been found in the mixed cultures involved in the putrefaction of
meat. Thus cadaverin was found in cultures of the cholera spiril-
lum, another nitrogenous poison, typhotoxin, in those of typhoid
bacilli, and still another in tetanus cultures, all of them producing
more or less severe illness when injected into animals.
In spite of this evidence, however, we have been forced to con-
clude that the ptomains cannot properly be held responsible for bac-
9 Meyer and Gottlieb. "Experim. Pharmakologie," 2d ed., p. 262.
BACTERIAL POISONS 31
terial toxemia as manifested in disease. In the first place it is
doubtful whether ptomains, in noticeable quantity, are ever produced
within the living infected body. Then, again, potent ptomains are
produced in culture by many bacteria having absolutely no patho-
genic power, while highly pathogenic bacteria may produce little or
no ptomains. Ptomain production, moreover, is not specific, since
the same ptomains may be produced by many different bacteria or
mixtures of bacteria, provided the conditions of nutrient materials
and temperature are favorable for growth. We cannot therefore
account for bacterial toxemia, in which the poison produced by an
individual species is characteristic and invariably the same, under
varying cultural and environmental conditions, by the production of
ptomains. And even when ptomains are produced in culture fluids
by pathogenic bacteria their physiological action is usually quite
different from that of the poisons produced by the same micro-organ-
isms in the infected subject.
Briefly summarized, therefore, the ptomains are poisons elab-
orated by all bacteria that are capable of producing protein cleavage,
if planted on suitable nutrient materials under conditions favoring
growth. The matrix of these poisons is the protein nutriment ; they
are not products of intracellular metabolism specifically characteris-
tic of the bacteria which produce them.
Their importance in the production of disease, therefore, is really
an indirect one. They may cause disease if putrid meat or other
material is ingested, and with it preformed ptomains, which may be
taken in and further elaborated by continued putrefaction in the
intestines. This form of meat poisoning, without bacteriological in-
vestigation, may be difficult to distinguish from' such bacterial forms
of meat poisoning as those caused by the Gartner bacillus or the
bacillus botulinus. Novy 10 believes that true ptomain poisoning of
this kind is rather less frequent than formerly supposed. However,
in such cases as those of Vaughan, who isolated a poisonous ptomain
"tyrotoxicon" from cheese and milk, their importance seems rea-
sonably certain. It is also probable that certain forms of auto-
intoxication may be caused by the production in the intestinal ca-
nal of ptomains resulting from bacterial putrefaction incident to
faulty digestive conditions. It is the antagonism to such intes-
tinal putrefaction by the acid production of the bacillus Bul-
garicus which is probably the basic cause of any favorable thera-
peutic effects which have attended the soured milk therapy of
Metchnikoff. Again the growth of saprophytes in necrotic tissues
such as gangrenous extremities in diabetes or amputation stumps,
may lead to the formation of ptomains which, after absorption, can
cause disease. In all such cases the process is one determined by the
bacterial putrefaction of dead organic materials, and the absorbed
10 Novy in Osier's "Modern Medicine," Vol. 1, p. 223.
32 INFECTION AND RESISTANCE
poisons are not true bacterial toxins, since they do not emanate
specifically from the cell substance of the micro-organisms but rather
represent incidental cleavage products of the nutrient materials.
Therefore, also, the ptomains are unspecific — their formation a com-
mon attribute of a large variety of saprophytic organisms, their
production, as to quantity and kind, primarily dependent upon
the nature of the nutrient materials on which the bacteria are
grown.
In contradistinction to the ptomains, the specific bacterial poi-
sons, in the technical meaning of the term, are substances which are
characteristic for each individual species of bacteria and truly the
products of bacterial metabolism in that they emanate from the cell
itself, either as a secretion or excretion during cell life, or as an
inherent element of the cytoplasm liberated after death (or possibly
as a cleavage product of the disintegrating bacterial protein).11
They are dependent upon the nature of the culture medium only in
so far as this favors or retards the normal development of the micro-
organisms. While, therefore, a diphtheria bacillus undoubtedly pro-
duces the largest quantities of its specific poison on bouillon suitably
prepared for this particular purpose, it will also, in smaller amount,
produce qualitatively the same poison on all media on which its
growth is free and uninhibited, even on a medium such as that of
Uschinsky, which is entirely devoid of proteins. The toxins are,
therefore, elements of intracellul&r metabolism, permanently or
transiently constituent parts of the cell body.
A specific bacterial toxin was first obtained from the diphtheria
bacillus by Roux and Yersin12 in 1889. They discovered that if
diphtheria bacilli were grown on veal broth and the cultures filtered
through porcelain candles, after seven days at 37.5° C. the filtrates
were highly toxic, producing the same symptoms and autopsy find-
ings in rabbits, guinea pigs and birds which followed the injection
of the living bacilli themselves. The poison was therefore a soluble
product of the bacteria during the period of their vigorous growth,
apparently given up by them to the culture' fluid. Very soon after
this, in 1891, Kitasato 13 discovered a similar specific toxin in cul-
ture filtrates of the tetanus bacillus, and it was the hope of bacteri-
ologists that analogous poisons could be determined for all patho-
genic bacteria.
This hope, however, has been disappointed. It was soon found
that cultures of cholera spirilla, typhoid bacilli, and many other
germs did not yield toxic filtrates of this kind but that the poisons
in these cases seemed to be firmly bound to the bacterial bodies dur-
11 In connection with this read the discussion on anaphylaxis in chapter
XVII, p. 413.
12 Roux and Yersin. Ann. de I'Inst. Pasteur, Vol. 2, 1889.
13 Kitasato. Zeitschr. f. Hyg., 1891, Vol. 10.
BACTERIAL POISONS 33
ing life, and given tip to the surrounding media only after death
and disintegration of the cells.
Pfeiffer 14 was the first one to formulate this conception in his
studies upon cholera poisons. He found that when cholera spirilla
were grown upon broth and filtered after 6 or 7 days, the filtrate was
hut slightly toxic, but that, in this case, unlike the conditions pre-
vailing in diphtheria and tetanus cultures, the residue of bacterial
cell bodies, even after they had been killed by chloroform, thymol,
or drying, were powerfully poisonous.
We have then two main classes of specific bacterial poisons. One
— typified by diphtheria and tetanus poisons — is produced during
the period of energetic growth by the living bacteria, is given off to
the surrounding culture fluid as a secretion or excretion, and can be
obtained in bacteria-free filtrates at a time when few, if any, of the
micro-organisms have died or disintegrated. These are spoken of as
"true toxins" or "exotoxins."
The other group — typified by the cholera poisons as described by
Pfeiffer — is apparently an intracellular, constituent part of the bac-
terial body — not given off during life and not, therefore, obtained in
filtrates of young living cultures. If the cultures are preserved until
cell death has taken place and the dead bodies have been extracted
by the culture fluid, the filtrate becomes gradually more toxic. The
bodies of such bacteria are in themselves powerfully toxic when
injected, dead or alive. These poisons for obvious reasons Pfeiffer
has named the "endotoxins" since he regarded them as specific and
definite substances, present as such in the living bacterial cell.
In addition to the endotoxins the bacterial protein contains sub-
stances which attract and lead to the accumulation of leukocytes. In
other words, they exert a positive chemotactic influence. This was
first observed in 1884 by Leber,15 who induced the formation of pus
hy injecting dead staphylococcus cultures, and, later, found that the
same effect resulted from the injection of alcoholic extracts of
staphylococci. These chemotaxis-inducing substances were later
particularly studied by Buchner. Buchner 16 extracted them from
many varieties of bacteria, independent of pathogenicity. Although
there are quantitative differences, all bacteria seem to contain such
substances, and Buchner believed the chemotactic property to be a
general attribute of the bacterial protoplasm. He speaks of his ex-
tracts as bacterial proteins.
The true toxins or exotoxins, then, appear to be products of liv-
ing bacteria given off from these very much as are the ferments and
enzymes by which micro-organisms cause cleavage of carbohydrates
or proteins — and indeed the French school, from the first, compared
14 Pfeiffer. Zeitschr /. Hyg., Vol. II, 1892.
15 Leber. "Tiber die Entziindung," Leipzig, 1884.
16 Buchner. Berl. klin. Woch., 1890.
34 INFECTION AND RESISTANCE
these toxins to enzymes, with which, as we shall see, they have much
in common. The endotoxins — on the other hand — at least as con-
ceived by Pfeiffer, are structural ingredients of the bacterial proto-
plasm which are toxic when brought into solution as the cells break
up.
Concerning the accuracy of this conception, however, much doubt
has recently arisen, as a result of researches which will be discussed
below.
These two types of poison, moreover, differ from each other not
only in mode of origin but in biological characteristics far more
fundamental than this.
The discovery of diphtheria toxin by Roux and Yersin was fol-
lowed by diligent investigations into the toxic properties of all
known pathogenic bacteria, and it was soon found that a few only
of these germs could produce poisons biologically similar to that
found in diphtheria cultures. It was in the course of investigations
of this kind, indeed, that Pfeiffer, failing to discover an exotoxin in
cultures of cholera and other germs, formulated his endotoxin theory.
The list of true toxin or exotoxin producers, then, is short.
Among the more important are, in addition to the diphtheria and
tetanus bacilli — which have been mentioned above — the Bacillus
botulinus,17 the Bacillus pyocyaneus,18 and that of symptomatic an-
thrax.19 It has also been claimed that similar toxins are formed by
the cholera spirillum (Brau and Denier),20 by the dysentery bacillus
of the Shiga-Kruse type (Kraus and Doerr) 21 and the Bacillus ty-
phosus (Arima).22 In the cases of the three last-named organisms,
however, the secretion of a true exotoxin has not been accepted as a
fact by all observers. Indeed, even though such substances may pos-
sibly be produced by these bacteria in small amounts it is not likely,
in the light of our present knowledge, that they play more than a sec-
ondary role in the toxemic manifestations of cholera, dysentery, and
typhoid, the important poisons in these cases being those derived
from the bacterial cell bodies.
Similar in essential properties to the true exotoxins also are the
erythrocyte poisons (hemotoxins) produced by many bacteria which
cause hemolysis of red cells, and the leukocyte-destroying poison
(leukocydin) which is a product of the Staphylococcus aureus.
All of these atrue bacterial toxins" or exotoxins, apart from sim-
ilarity of origin, as soluble secretions of the living bacteria, possess
certain common biological characteristics which sharply differentiate
17 Kempner. Zeitschr. f. Hyg., Vol. 26, 1897.
18 Wassermann. Zeitschr. f. Hyg., Vol. 22, 1896.
19 Grassberger and Schattenfroh. Wien Deuticke, 1904.
20 Brau and Denier. Ann. de I'Inst. Past., Vol. 20, 1906.
21 Kraus and Doerr. Wien kl. Woch., 42, 1905.
22 Arima. Centralbl. f. Bakt., I, Vol. 63, 1912.
BACTERIAL POISONS 35
them from the "endotoxins." These characteristics they share with
a number of non-bacterial substances such as the vegetable poisons
ricin, crotin, and abrin, with animal poisons like snake venom and
spider poison (arachnolysin), and, in certain important respects,
with the substances spoken of as enzymes.
Thus the bacterial true toxins are not biologically unique sub-
stances. Both in themselves and in regard to the reactions they
elicit when injected into the animal body, they share certain cardinal
properties with analogous substances derived from the higher plants
and from animals. And it is important to recognize at once that we
are dealing here, as in other phases of the study of bacterial immu-
nity, with broad biological laws, which find application not only in
bacteriology, but in general pathology and in the phenomena of pro-
tein metabolism in general. It so happens that these phenomena
have been studied and are most easily elucidated in connection with
bacteria. But their general significance must not be lost sight of.
The cardinal characteristic which unites all of these substances
into a single well-defined biological group is their property of in-
ducing the formation of antitoxins when injected into animals. This
property is so important and its thorough comprehension so essential
that we may be permitted to digress briefly in order to make it clear.
As we shall see, in subsequent chapters, all substances which lead
to the formation of specifically reacting antibodies in the treated
animal are spoken of as "antigens" or "antibody-inducing sub-
stances." The class of "antigens" is a large one, including all
known proteins, and possibly some of the higher proteid split prod-
ucts, and protein-lipoid combinations, though the "antigenic" prop-
erties of the last two are still in controversy. But among this
large group of substances it is only the bacterial true toxins (exo-
toxins), obtained in broth filtrates of living cultures, together
with the vegetable poisons and other substances we have classified
with them above, which induce in the blood of the treated animal a
neutralizing antibody — (antitoxin) — which inhibits quantity for
quantity the activity of the injected toxin or vegetable or animal
poison. This property of eliciting the production of antitoxin in the
animal body alone separates these substances sharply from all other
antigens, toxic or otherwise, and, in this respect, they differ sharply
from the so-called "endotoxins" against which no antitoxins can be
produced.
As an important secondary characteristic of this group of sub-
stances we may regard their chemically indefinable nature. In the
case of none of them have we any definite knowledge of chemical
constitution except in so far as it has been hitherto impossible to
separate them from the protein molecule. The intensive chemical
study of the toxins has universally resulted in failure to obtain a
protein-free product which has the characteristic toxic properties of
36 INFECTION AND RESISTANCE
the original filtrate, or its antitoxin-inducing power. Concerning
the methods which have been employed in the study of the chemistry
of these substances we will have more to say in another place.23 It
is safe to summarize all this work for our present purposes, by stat-
ing that, whatever the method employed, until now all of the prep-
arations obtained have given one or another of the protein type-
reactions, and that none of them can be positively accepted as pro-
tein-free. The results here obtained have been entirely analogous to
those obtained in similar investigations upon enzymes. (See also
discussion of antigens, chapter 4.)
The analogy with enzymes is indeed a striking one and noted by
the first investigators of a true toxin, Roux and Yersin. Biolog-
ically, of course, we have the cardinal similarity in that the injec-
tion of toxins into animals induces the production of antitoxin, and
treatment with enzymes induces specific and neutralizing anti-en-
zymes. In addition to this, they are alike in their susceptibility to
heat (both being destroyed when in solution by temperatures over
80° C.), in their gradual deterioration on standing, and their mys-
terious activity in small quantities upon disproportionately larger
masses of the substances they attack. There is, however, one impor-
tant difference between the two in their mode of action. For, while
the toxins are apparently bound or neutralized by the tissues they
attack, the action of an enzyme seems rather to be a process in which
the enzyme unites with the substance it acts upon, is released as the
result is attained, and freed for further action, without noticeable
loss of quantity. Such catalytic properties have not yet been satis-
factorily demonstrated for the bacterial toxins. However, there are
other modifying factors which may account for lack of similarity in
this respect, and in all other important points the two classes of sub-
stances are closely analogous.
The property of heat sensitiveness, which is a characteristic of
bacterial exotoxins and enzymes, is shared with them by all of the
substances mentioned above except snake venoms. Snake venoms
are not destroyed completely until the temperature is raised to 75°-
80° C. The earlier contention of Leclainche and Yallee, that the
toxin of symptomatic anthrax possessed similar heat stability has
been satisfactorily refuted by Grassberger and Schattenfroh,24 who
.find that heating it to 50° C. for an hour completely destroys it.
There is another important attribute of the true toxin which
(deserves discussion, though we are by no means in a position to offer
any satisfactory explanation for it. We refer to the incubation time
which elapses between the administration of a toxin and the occur-
23 An extensive and authoritative summary of this phase of the subject is
that of E. Pick in "Kolle u. Wassermann Handbuch," etc., 2d ed., Vol. 1.
24 Grassberger and Schattenfroh. "tiber das Rauschbrandgift, etc.,"
Wien. Deuticke, 1904.
BACTERIAL POISONS 37
rence of symptoms. Here again snake poisons form an exception —
since local manifestations may appear within an extremely short
period after the injection of the venom or as the result of a snake
bite. However, in the case of all other toxins there is a definite
lapse of time between the entrance of the poison and the first symp-
toms, local or general. This interval is longer when small doses are
given — shorter when the doses are large — but is never entirely elim-
inated— even when many times the fatal dose is given.
In the case of tetanus poison, for instance, injections into a horse
may not cause symptoms for as long as four or five days. In mice,
animals that are extremely susceptible, the incubation time may be
shortened from 36 to 12 hours if we inject 3,600 lethal doses, but,
in any case, whatever the dose, this interval cannot be shortened
below 8 or 9 hours.25 Many attempts have been made to explain
this. Ehrlich, as we shall see, assumes that the action of a poison
depends upon two occurrences: one, the union of the poison with
the vulnerable cell, the other the gradual injury of the cell by the
toxic atom groups in the poison molecule. The time necessary for
the institution of this process, he believes, explains the interval.
Eichet has suggested that the toxin itself may not be potent until
acted upon by the body of the recipient and transformed into a
potent form. His views are more directly related to the phenom-
enon of anaphylaxis and are discussed in another section. De Waele
has recently advanced a theory which implies that the incubation
time represents the period necessary for the gradual concentration
of the poisons in the vulnerable tissues, a process which depends
either upon chemical affinities or solubility of the toxins in the cell
lipoids. A little at a time would then be absorbed by the vulnerable
cells as they come in contact with the poison, through the circulation,
and the symptoms would not appear until a definite intracellular
concentration had been attained. His views are so closely bound up
with the theories on the selective action of the toxins upon individual
tissues and organs that they will be rendered clear as we proceed
with a discussion of the latter.
The majority of pathogenic bacteria do not, as we have seen, pro-
duce true toxins or exotoxins. Cultures of cholera spirilla, plague
bacilli, and of many other bacteria do not yield toxic filtrates until
the cultures have been allowed to stand for prolonged periods during
which extraction and possibly autolysis have occurred. In these
cases, moreover, definite toxic properties can be demonstrated in the
dead cell bodies or in extracts prepared by various methods. In no
case, however, is the injection of these "endotoxins" followed by the
production of antitoxins. It was very natural to suppose that in
micro-organisms of this class the toxic principle might be present in
the form of a preformed intracellular poison which could be ex-
25 De Waele. Zeitschr. /. 7mm., Vol. 4, 1910.
38 INFECTION AND RESISTANCE
tracted or which became free as cell-death occurred and disintegra-
tion ensued.
It was assumed that, when bacteria entered the animal body and
were destroyed by the action of the serum or cells, these endotoxins
were liberated and poisoning resulted. The very protective action
of the serum, which prevented the extension of the infectious in-
vasion, by limiting bacterial growth, was thus looked upon as the
agency by whicji the endotoxins, toxalburnins, were set free. Ex-
periments by Radziewsky and others, in which it was shown that large
doses of bacteria injected into immunized animals were violently
toxic and more rapidly fatal than corresponding amounts injected into
normal animals, were taken to mean that in the immune animals a
more powerfully cell-destroying property of the serum led to a more
rapid liberation of the endotoxins.
This was the conception of Pfeiffer and, in more recent theoret-
ical discussions, that of Wolff-Eisner. Its essential features con-
sisted in the assumption that the poisons were preformed and were
contained within the cell body as such, and that they were specific for
each micro-organism, determining to a certain extent its pathogenic
properties. Thus typhoid endotoxin, cholera endotoxin, or dysen-
tery endotoxin was supposed each to possess its own particular
pharmacological properties by which the clinical manifestations of
the respective diseases were partially determined.
It is chiefly the work of Vaughan 26 which has begun to throw
doubt upon Pfeiffer's original views, in that Vaughan has shown
that all proteins, bacterial or otherwise, would yield, upon cleavage
with alkalinized alcohol, toxic split products which possessed many
of the pharmacological properties of the so-called endotoxins. In
fact, Vaughan succeeded in producing, in animals, fever and other
symptoms which are generally associated with infection, merely by
injecting into them graded quantities of his toxic split products.
Following Vaughan, Friedberger succeeded in showing that
toxic substances similar to Vaughan' s split products are formed
when bacteria of various species are subjected to the action of nor-
mal or immune sera, and that such poisons were pharmacologically
alike and produced with equal ease from pathogenic and non-patho-
genic micro-organisms. These phenomena are discussed in greater
detail in our section on bacterial anaphylaxis. It is necessary, how-
ever, to point out in this place the uncertainty in which these re-
searches have left the conception of endotoxins. They suggest that
the toxic effects following upon the introduction of pathogenic bac-
teria into the animal body are not due to endotoxins, but are rather
the result of the action of toxic cleavage products formed in the re-
action between blood plasma and bacterial cell. These split products
26 For a complete discussion of Vaughan's work see Vaughan, "Protein
Split Products," Lea & Febiger, Phila. and N. Y., 1913.
BACTERIAL POISONS 39
are not conceived as specific for individual bacteria but may be
formed from all bacterial proteins, both the pathogenic and the non-
pathogenic. The differences in pathogenicity between bacteria of
this class would then depend entirely upon their powers to invade —
not at all upon their possession of individually peculiar cell poisons.
The differences in clinical course and toxemic manifestations would
be taken to depend entirely upon the accumulation and the distribu-
tion of the invading germs, and the consequently variable energy in
the production of the toxic split products from them. Considerable
experimental evidence has accumulated in favor of this point of
view. We will reserve a consideration of this for a later chapter.
The entire question of "endotoxins," or rather the problem of
the mechanism by which such bacteria as typhoid bacilli, plague
bacilli (and other organisms which do not produce exotoxins) poison
the animal body, must be subjected to experimental revision. In
addition to the idea of toxic split products in the sense of Vaughan
and Friedberger, there are other alternatives. Jobling and Petersen
have shown that bacteria injected into the circulation may absorb
lipoidal substances which ordinarily act as anti-enzymes. In con-
sequence of this, serum protease may be liberated to act upon the
plasma itself, and produce toxic substances.
Again, it is well known that the bacterial cells are relatively
poor in coagulable protein, and we have shown with one of our stu-
dents (Aronovitch) that primary and secondary proteoses may be
obtained in considerable quantities in bacterial extracts. It is not
impossible that these in themselves may have toxic functions when
liberated, without further splitting. This particular subject finds a
more extensive discussion in the chapter on bacterial anaphylaxis.
In order to do injury to the infected individual the bacterial
poisons must be produced in such locations that they can easily enter
the physiological interior of the body. None of the poisons that
have been so far investigated can produce injury when introduced
into the alimentary canal. In this location they are, as a rule, de-
stroyed, or they pass through without doing harm. Neither diph-
theria toxin nor tetanus toxin will produce symptoms when intro-
duced intra-intestinally.27 28 29 30 Even cholera poison does not pass
through the uninjured intestinal wall. Kruse 31 assumes, and Kolle
and Schiirmann 32 seem to agree with him, that the absorption of
cholera poison does not occur until the intestinal wall has been in-
27 Meyer and Gottlieb. "Exp. Pharmakol.," Urban & Schwartzenberg,
Berlin, 1911.
28 Ransom. Deutsche med. Woch., No. 8, 1898.
29 Neneki. Centralbl. f. Bakt., Vol. 23, 1898.
30 Carriere. Ann. de I'Inst. Past., Vol. 13, 1899.
31 Kruse. "Allgemeine Mikrobiologie," Vogel, Leipzig, 1910, p. 934.
32 Kolle and Sclmrmann in "Kolle u. Wassermann Haudbuch," 2d Ed.?
Vol. 4.
40 INFECTION AND RESISTANCE
jured by the actual growth of the living bacteria. Kruse calls atten-
tion to experiments by Burgers in which enormous quantities of
cholera poison, i. e., 200 cultures of dead or living cholera bacilli,
could be administered to healthy guinea pigs and rabbits by mouth
without harm in spite of the fact that these animals are definitely
susceptible to the poisons and although the poisons are not injured
by the intestinal ferments. It is likely therefore that the absorption
of poison begins only after the bacteria have extensively invaded the
intestinal mucosa and, by injuring tissue, have opened paths for
absorption. In the case of diphtheria probably a similar condition
exists in that the localized injury to the mucous membrane at the
point of lodgment of the primary infection prepares a portal of
entry. The poison of the Bacillus botulinus alone seems to form an
exception to this rule,33 since this substance, though apparently a
true bacterial toxin, is absorbed directly from the intestinal canal.
With most bacteria this problem does not arise, since the poisons are
elaborated within the tissues, where resorption is a necessary result.
Like alkaloids and other organic as well as inorganic drugs, the
action of many bacterial poisons is largely selective. Most of these
poisons may excite inflammatory reactions if concentrated in any
part of the body, but, in addition to this, there is a specific distribu-
tion after introduction which indicates that the poison goes into
selective relationship with certain tissues and cells. This fact is most
clearly illustrated by the bacterial hemotoxins which specifically in-
jure the red blood cells of the infected individual and by such sub-
stances as the leukocidin produced by the Staphylococcus aureus,, a
poison which directly and visibly injures the white blood cells. Here
the action is specifically aimed at a well-defined variety of body cell.
In considering this problem in connection with infectious disease,
it is of great importance to distinguish between selective injury by the
poisons transported through the body by the lymph, blood, and other
channels, on the one hand, and the selective lodgment of the micro-
organisms themselves on the other. The latter may occasionally de-
pend on local cultural advantages for the particular bacteria in one
organ or another, but may just as often be determined by the peculiar
manner of entrance to the body which is most suitable for lodgment
of the germs in question, and the degree of local resistance at the
point of entrance, which determines whether or not the infection shall
be locally limited or permitted to invade beyond this point. In the
case of a disease like acute anterior poliomyelitis, where our knowl-
edge of the microorganisms which cause the disease is yet in its in-
fancy, it is impossible to decide whether the injuries noted in the
motor areas of the cord and medulla are due to toxins or the lodg-
ment of the germs themselves. In the case of rabies it seems reason-
ably sure that the microorganisms themselves select the nervous sys-
33 Madsen in "Kraus u. Levaditi, etc.," Vol. 1.
BACTERIAL POISONS 41
tern. In such instances as the injury of the motor areas by tetanus
poison, that of certain peripheral nerves by diphtheria toxin, or even
the characteristic lesions of post-syphilitic maladies like tabes, we
can be reasonably sure that we are dealing with the specific action
of the poisons, independent of actual localized growth of the infec-
tious agents.
Diphtheria toxin, after distribution through the body, may act
upon many different tissues, as is evident by degenerations in the
heart muscle, liver, and kidney, and the petechial hemorrhages in
serous surfaces. In addition to this general action, however, there is
a very marked selection of certain nerve centers. By Meyer and
Gottlieb 34 diphtheria toxin is classed as a specific vascular poison.
Its action results in a rapid sinking of the blood pressure with final
cardiac death in spite of artificial respiration. These manifestations
seem to have a central origin, with particular action upon the vagi
and the phrenic nerves. Apparently also the localization of the
diphtheritic lesion may influence the selection of individual nerves,
the most concentrated action taking place upon the nerves whose
endings are distributed in this particular region, for, as Meyer and
Ransom 35 have shown, this poison, like tetanus toxin, may be ab-
sorbed into the nerves directly through the nerve endings. An in-
teresting selective action also of diphtheria poison is the apparently
specific alteration of the suprarenal glands which is regularly no-
ticed, as enlargement and congestion, in diphtheria-infected guinea
pigs, and which has been associated by many workers with the char-
acteristic drop in blood pressure which accompanies all severe cases
of the disease. Abramow 36 has studied this lesion particularly, and
believes that it consists in a degeneration and final disappearance of
the chromafnn substance and of the medullary cells. He believes
that this, together with degeneration of the heart muscle itself, is of
great importance in causing the characteristic vascular failure.
In botulinus poisoning there is, as Marinesco 37 and Kempner
and Pollack 38 have shown, a direct effect upon the cells of the an-
terior horns with degenerative changes in the I^issl granules.
Tetanus poison, which has been studied extensively by pharma-
cologists, shows a very marked affinity for the nervous system, as, in
fact, the symptoms of tetanus indicate. Indeed, while many of the
bacterial poisons are distributed by the blood stream to the point of
final attack, in tetanus the absorption of the toxin from the lesion or
34 Meyer and Gottlieb. "Pharmacology Trans. Halsey," Lippincott, 1914,
p. 556.
35 Meyer and Ransom. Arch, de pharmacodyn., Vol. 15, 1905, also Meyer,
Berl klin. Woch., 25 and 26, 1909, also Arch. f. exp. Path. u. Ther., Vol.
60, 1909.
36 Abramow. Zeitschr. f. Imm., Vol. 15, 1912.
37 Marinesco. Compt. rend, de la soc. de biol., Vol. 3, 1896.
38 Kempner and Pollack. Deutsche med. Woch., 32, 1897.
42 INFECTION AND RESISTANCE
the point of injection takes place entirely by the path of the nerves.
That this method of poison distribution might be, among others,
an important one was suggested as early as 1892 by Bruschettini,39
who found tetanus toxin in the nerves but not in the adjacent muscle
and other tissues surrounding the point of subcutaneous injection.
Similar results were obtained subsequently by Hans Meyer, whose
experiments were confirmed and extended by Marie and Morax.40
Finally Meyer and Ransom41 furnished complete proof that the
poison was absorbed from the blood and tissues by the peripheral
nerve endings alone and was transported centripetally only by the
paths of the neurons. The experimental facts elicited may be sum-
marized as -follows :
1. When tetanus toxin is injected into the thigh muscles of a
guinea pig the poison is found at first only in the sciatic nerve of the
same side and in the blood. (The determination of poison was made
by injecting macerations of the respective tissues into mice.) If ex-
amination was delayed until the symptoms had become generalized,
the poison was found in the opposite sciatic, but the muscle bundles,
fat, etc., from the vicinity of the injection area were poison-free.42
2. When a nerve is cut poison absorption ceases as soon as axis
cylinder degeneration has set in.
3. If the nerve is cut before the poison is injected the distal
end contains poison, the proximal end does not. This again shows
that the nerve absorbs the toxin not from its capillaries but solely
through the end organs.
4. If a nerve which already contains poison is severed, toxin
will disappear rapidly from the proximal end, since it no longer
obtains a renewed supply from the periphery.
5. If antitoxin is injected into the nerve, above the point of
injection, it will successfully bar the way for the ascending toxin.
6. Severing of the spinal cord prevents the passage of the poison
from below upward.
These facts ascertained in the case of tetanus find their parallel in
the phenomena of the distribution of rabic virus 43 as well as in that
of poliomyelitis, in both of which there seems to be a progressive cen-
tripetal transportation through the nerves. However, in these condi-
tions we are probably dealing not with a poison but with a living virus
and, though analogous, the conditions are not entirely comparable.
From the practical point of view these facts regarding tetanus
39 Bruschettini. Riforma medica, 1892.
40 Marie and Morax. Ann. de I'Inst. Past., 1902.
41 Meyer and Ransom. Archiv f exp. Path. u. Pharm., 49, 1903.
42 In view of our discussion of the importance of fats in the absorption
of tetanus toxin, it seems inconsistent that the toxin does not concentrate
in fatty as well as in nervous tissues. This Meyer explains by the inactive
and poorly vascularized condition of the fat tissues.
43 Di Vestea and Zagari. Fortschr. d. Med., Vol. 6, 1888.
BACTERIAL POISONS 43
may explain the frequent failure of therapeutic success attending
the injection of tetanus antitoxin after the symptoms of the disease
have set in, since in such cases the poison is already distributed to
the nerves and is largely inaccessible to the antitoxin. They also
have pointed a way toward a more hopeful therapy, namely, the
method of injecting the antiserum directly into the nerves about the
point of injury. It is not surprising, however, in view of the stated
facts, that even this is unsuccessful when done at too late a time,
after a considerable amount of poison has already passed above the
point of injection to the spinal centers.
Such selective action on the part of the bacterial poisons is en-
tirely analogous to the similar specific action of alkaloids, narcotics,
and other drugs. In order that the poison may act upon a cell we
must, of course, assume that it has either chemical or physical affin-
ity for this cell. The problem, as many writers have pointed out, is
.strongly analogous to that of tissue staining. A dye must be able to
form a chemical Tinion with the cell or it must be soluble in the cell
substance in order to stain it. The chemical difference between cells
is a delicate one and not often definable by our present methods. We
can obtain an insight into the principles probably underlying selec-
tive action only by inference from the relation between the chemical
constitution of drugs or their physical properties, solubility, etc., and
their respective tissue affinities. These problems are difficult and, to a
large extent, obscure. They cannot be directly investigated upon
bacterial poisons since these are themselves of chemically unknown
nature. But the study of drugs of known constitution has revealed
certain definite relations of this kind which have furnished analogies
from which the general principles of selection in bacterial poisons
•can be surmised.
It is a well-known fact to pharmacologists that there is a definite
relation between chemical structure and toxicity. Fraenkel 44 ex-
presses it as follows : "By the addition of identical atom groups in
an identical manner, similarly acting substances are obtained." He
cites the well-known example of curare ; whichever the path by which
this poison is injected it leaves intact the tissues with which it comes
in contact, but after general distribution acts specifically upon the
nerve endings. It had been discovered by Brown and Eraser 45 that
by introducing methyl radicles (CH3) into molecules of various alka-
loids, strychnin, morphin, atropin, and others, substances were ob-
tained which paralyzed nerve endings, and this irrespective of their
previous physiological action. It appears that the combination of
four methyl radicles attached to the nitrogen atom (quaternary bases)
44 Sigmund Fraenkel. "Arzneimittel Synthese," 2d Ed., Springer, Ber-
lin, 1906.
45 Brown and Fraser. Trans. Royal Soc. of Edinburgh, 25, 1868, cited
from Fraenkel.
44 INFECTION AND RESISTANCE
universally possesses this paralyzing action. Tertiary bases on the
other hand lack this property.
CH3
OH3
CH3.
CH3.
"Ammonium base" ''Tertiary base"
Quaternary
Subsequently Bohm 46 47 discovered that curare contains two bases
— the one, ucurin," is slightly toxic and is a tertiary base; the other,
which possesses the typical curare action, "curarin," is an "ammo-
nium base." By "methylizing" curin, curarin could be obtained.
From these and other examples it is clear that in a certain num-
ber of cases actual chemical affinity must play a part in toxic action ;
on the other hand, there are many cases in which toxic action seems
to depend merely upon physical conditions such as solubilities.
Meyer and Overton's well-known theory of narcosis maintains that
certain narcotics exert their action by passing out of blood and
lymph solution into solution by the fat-like, lipoidal substances
(lecithin, cholestrin, etc.) contained in the nerve cells, because the
latter are better solvents for them than is the blood plasma. This
theory of Meyer and Overton has stimulated much investigation and
speculation, and it is not unlikely that it is valid in the case of many
narcotics, although it does not explain the action of narcotics in gen-
eral; for Dickson notes that chloral hydrate, for instance, is more
soluble in water than in oils, and some narcotic drugs like alcohol
exert definite action on proteins and are oxidized in the body. These
are pharmacological questions of which we cannot speak with author-
ity. We wish merely to point out that the action of poisons upon the
body may depend in some cases upon mere physical or mechanical
relationship between the two.48 49
As regards bacterial poisons the union between poison and sus-
ceptible cell is extremely firm and difficult to dissociate in many in-
stances, and this points to the possibility that, in these cases at least,
true chemical union takes place rather than merely a loose combina-
tion like that of the solution of one substance in another. Further-
more, the complete inactivation of some poisons by mixture with the
46 Bohm. Arch, de Pliarm., cited from Fraenkel.
47 See also Dickson, "A Manual of Pharmacology," E. Arnold, London,
1912.
48 Ivar Bang. "Biochemie der Lipoide," Bergmann, Wiesbaden, 1911.
49 Meyer and Gottlieb. "Experimented Pharmakologie," 2d Ed., Urban.
& Schwartzenberg, Berlin, 1911.
BACTERIAL POISONS 45
cells of tissues capable of binding them would likewise point to more
than mere physical union. Nevertheless, it does not by any means
exclude the thought that the poisons may, in fact, go into selective
relationship with special cells because of physical properties, such as
solubility in the lipoidal cell membranes,50 51 and may subsequently
be bound chemically or destroyed by oxidation or enzymotic hydroly-
sis after such entrance. In such a case the actual specificity would
yet depend on purely physical properties.
In addition to the specific physical and chemical affinities be-
tween the poisons by certain cells there are probably also certain
fortuitous factors connected with the distribution and local accumu-
lation of the poisons which have some weight in determining the
location of injury. For the specific selection is not absolutely strict
and there are probably few parenchyma cells in the body that are
entirely insusceptible to injury if the poisons are sufficiently con-
centrated upon them. Thus, to cite an analog}7" from the toxicology
of non-bacterial poisons, in lead poisoning, as Meyer and Gottlieb
point out, the paralysis of the extensors of the arm occurs chiefly in
adults who use these muscles in the exercise of their professions
(painters, type-setters), while in children and in animals, in which
no such selective use of particular muscle groups is habitual, lead
paralyses are atypical, attacking legs as well as arms. It is not un-
likely that the frequent injury of the heart muscle by bacterial poi-
sons or the irregular parenchymatous changes in various organs is
determined by analogous fortuitous factors, in that functional activ-
ity and increased metabolism may predispose to injury.
Bacterial poisons also may produce their lesions in the course of
excretion. This seems likely in the case of typhoid poisons in which
we have often seen bloody diarrhea in rabbits within a few hours
after intravenous injection of powerfully toxic culture filtrates. In
connection with the dysentery bacillus Flexner and Sweet 52 have
studied the conditions carefully. They succeeded in showing first
that the introduction of the dysentery poison into the lumen of the
intestine does no harm and that the toxin is slowly destroyed by peptic
and tryptic digestion. They concluded that probably no absorption
of the poison through the uninjured intestinal mucosa takes place.
They then showed that the toxin after intravenous administration
is excreted by the intestine and that the inflammatory reactions and
injury of the mucosa are incident to this act of elimination.
Whether or not the kidneys are injured in the same way it -is
difficult to decide. In many infectious diseases, of course, the bac-
50 For Overtoil's theory of osmosis see R. Hober, "Physikalische Chemie
der Zelle u. Gewebe," Leipzig, Engelmann, 1911.
51 Compare also, regarding this entire question, the discussion in P. Th.
Miiller, "Vorlesungen iiber Immunitat, etc.," Fischer, Jena, 1910.
52 Flexner and Sweet. Jour, of Exp. Med., Vol. 8, 1906.
46 INFECTION AND RESISTANCE
teria themselves pass through the kidney into the urine, and renal
injury may result from the actual presence of the bacteria in the
kidney; however, renal injury may also occur without this, and it
is not at all impossible that the conditions here are similar to those
just described for the intestine.
All the facts which we have considered indicate that, although
most bacterial poisons can injure many different tissues, yet in some
cases there is a particular susceptibility on the part of an individual
tissue which is independent of accidental factors and seems to be
due to specific chemical or physical affinity. It seems even that in
tetanus, botulismus, and a few other conditions there is a differential
selection of particular areas within a tissue like the nervous system,
just as this occurs in the case of certain drugs. As stated above we
have no satisfactory scientific explanation for this, but a great deal
of work has been done to show that the bacterial poisons actually
unite with and are taken up by the susceptible tissues.
Indirectly, proof of this has been brought by the demonstration
of the rapid disappearance of various toxins from the blood streams
of susceptible animals and their persistence in the circulation of
animals insusceptible to them. Thus Donitz 53 has shown that
tetanus toxin injected into the blood stream of a susceptible animal
rapidly diminishes in quantity, and Knorr,54 in similar experiments,
showed .that the demonstrable disappearance of such toxins out of
the blood stream is synchronous with the appearance of symptoms, a
fact which excludes disappearance by excretion. Conversely Asa-
kawa 55 showed that in pigeons, which are but slightly susceptible,
tetanus poison could be demonstrated in blood, liver, spleen, kidneys,
and muscles six days after injection, but not in the brain, showing
that in this organ, at least, there must have been either a union or a
destruction of the poison. Similar to these results are those of
Metchnikoff,56 who found the poison unchanged after two months in
the circulation of insusceptible animals (lizards).
Direct evidence of union between susceptible tissues and poison
has been furnished by the experiments of Wassermann and Takaki,57
who showed that the brain and cord tissues of rabbits and guinea
pigs, mixed with tetanus toxin before injection, served to neutralize
its harmful effects. And it appears that the toxin-neutralizing prop-
erty of the brain substances of various animals is proportionate to
their individual susceptibility to the poison. Thus Metchnikoff 58
not only confirmed the results of Wassermann and Takaki for rab-
63 Donitz. Deutsche med. Woch., No. 27, 1897.
54 Knorr. Fortschr. der Medizin, 1897, No. 17, and Munch, med. Woch.,
1898, Nos. 11 and. 12.
55 Asakawa. Centralbl. f. Bakt., Vol. 24, pp. 166 and 234.
56 Metchnikoff. "L'Imnmnite dans les maladies Infect.," Paris.
57 Wassermann and Takaki. Berl. Tclin. Woch., 1898, No. 1.
58 Metchnikoff. Ann. de I'Inst. Past., 1898, p. 81.
BACTERIAL POISONS 47
bits and guinea pigs, but showed further that the brains of chickens,
animals that are but moderately susceptible, possess a correspond-
ingly slighter neutralizing power, and, further, that brain tissues of
entirely insusceptible cold-blooded animals, turtles and frogs, pos-
sess absolutely no neutralizing properties.
The original interpretation by Wassermann of these facts was
based on the assumption that the poison was bound to the brain tissue
just as it is bound to antitoxin. Experiments by Besredka 59 have
cast some doubt upon this. This worker's experiments seem to indi-
cate that a brain emulsion which has been saturated with the toxin
can be rendered capable of absorbing more toxin if tetanus antitoxin
is mixed with it. In other words, the affinity of the antitoxin for the
toxin is stronger than that of the brain substance for the poison, and
that the union toxin-brain tissue is very easily dissociated ; as indeed
it should if tae union were purely a physical one depending on solu-
bility.
After it had been shown that the poisons which acted specifically
upon certain cells were actually taken up by these cells, a number of
attempts were made to determine chemically the tissue element which
united with the poisons. Xoguchi 60 showed that cholesterin and
alcoholic extracts of blood serum neutralized tetanolysin. The same
thing was later shown by Miiller,61 and Landsteiner 62 showed that
ether extracts of red blood cells likewise neutralized this poison.
In a later study by Landsteiner and von Eisler 63 the relation of the
tissue lipoids to various toxic substances was still more definitely
established. They studied first the various hemolysins and found
that extraction of blood cells with ether rendered the stromata less
capable of binding the hemolytic substances. The same thing they
showed for bacteriolysins, in the latter case demonstrating at the
same time that the ether extracts of bacterial bodies possessed slight
binding properties for the bactericidal substances of the serum. These
experiments have, of course, a merely indirect significance in the
present connection, since they do not deal with the type of poisons
we have discussed. However, Landsteiner and von Eisler also
worked with tetanus toxin and found that the treatment of the brain
substance of guinea pigs with ether, by taking out lipoidal sub-
stances, considerably reduces the power of this tissue to bind and
neutralize the tetanus poisons.
Takaki,64 who investigated these relations in great detail, iso-
lated an alcohol-soluble element, cerebron, from nerve tissues, a sub-
59 Besredka. Ann. Past., 1903, p. 138.
60 Noguchi. Univ. Pa. Med. Bull, Nov., 1902.
61 Miiller. Centralbl. f. Bakt., Vol. 34, 1903.
62 Landsteiner. Wien. kl. Bundschau, 13, 1905.
63 Landsteiner and von Eisler. Centralbl. f. Bakt., 39, p. 318, 1905.
64 Takaki. Beitr. zur chem. Phys. u. Path., 11, No. 19, 1908.
48 INFECTION AND RESISTANCE
stance to which he ascribes the toxin-binding properties. Overton
and Bang 65 found, furthermore, that cholesterin and lecithin inhibit
the action of cobra venom, a poison which is in so many ways similar
to those produced by bacteria. Taking into consideration all avail-
able evidence, we are forced to admit that the lipoids seem to play an
important role in determining the selective action of the nervous sys-
tem by the bacterial poisons. It may not, of course, be an influence
depending merely upon the solubility of the harmful substances in
the lipoids themselves. For, as Bang expresses it, "the lipoids pos-
sess to a high degree the property of altering by their presence the
solubilities of other bodies/' and it is quite possible that in the tis-
sues they are present as lipoid-protein combinations. Their action
in determining the solubility of toxins in a given cell may therefore
be a purely indirect one.
It is of some interest in this connection to recall the experiments
of De Waele,66 which bring out another clear analogy between alka-
loids and bacterial poisons in their relation to lecithin. He found
that the addition of small quantities of lecithin increases the activity
of both toxins and alkaloids in the animal body, whereas larger
amounts inhibit both.
65 See Ivar Bang, "Biochemie der Lipoide," Bergmann, Wiesbaden, 1911.
66 De Waele. Zeitschr. f. Immunit., Vol. 3, 1909, p. 504.
CHAPTER III
OUK KNOWLEDGE CONCEKNING NATURAL IMMU-
NITY, ACQUIEED IMMUNITY, AND AKTIFICIAL
IMMUNIZATION
NATURAL RESISTANCE AGAINST INFECTION
IN the preceding chapters we have confined ourselves largely to
the consideration of those properties of the bacteria which determine
their ability to infect. In this discussion, however, we have repeat-
edly emphasized the fact that every infectious disease is the result of
a struggle between two variable factors — the pathogenic powers of
the bacteria on the one hand, and the resistance of the subject on the
other, each of these again modified by variations in the conditions
under which the struggle takes place. Thus a given micro-organism
may be capable of causing fatal infection in one individual but may
be only moderately virulent or even entirely innocuous for another.
Conversely the same individual may be highly susceptible to one va-
riety of bacteria, but highly resistant to others. Even in reactions
with one and the same micro-organism, the susceptibility or resist-
ance of the individual may be determined by variations in the physi-
ological state or by the environmental conditions under which the
two factors — invader and invaded — are brought together. There-
fore, the conceptions "resistance," "immunity," and its opposite
"susceptibility," are relative terms which can never be properly dis-
cussed without careful consideration of all modifying conditions
which influence them.
The science of immunity deals with a detailed analysis of these
variables. Its ultimate practical aim is the determination of meth-
ods by which an original susceptibility can be transformed into re-
sistance or even immunity. And the rational method of approach-
ing this subject consists in a careful study of the conditions of sus-
ceptibility and immunity as they exist naturally in the animal king-
dom.
The mere fact that both animals and man are in constant con-
tact with infectious micro-organisms, many of them in a high state
of virulence, indicates in itself that the animal disposes normally
over a defensive mechanism of considerable efficiency.
To a certain extent, of course, this escape from harm is due to
49
50 INFECTION AND RESISTANCE
the external defences of skin and mucous membrane which, in the
healthy state, mechanically prevent the entrance of the micro-organ-
isms into the body. For we have seen, in another place, that few of
the bacteria can pass through the uninjured surfaces. Moreover,
added to this, there is some protection in the bactericidal properties
of the secretions. An example of this is the inhibitory power exer-\
cised by the acidity of the normal gastric juice upon the choler^r '
spirillum. In order to infect the intestinal canal of guinea pigs
with these organisms Koch found it necessary to neutralize the gas-
tric juice with sodium carbonate solutions, and other observers have
found it necessary to inject directly into the duodenum. But even
after entrance into the animal tissues a second line of defence is
normally encountered by all invading germs which tend to inhibit
their further progress more or less perfectly. This active opposition
to the bacteria after their entrance is expressed chiefly in the anti-\
bacterial (bactericidal) activity of the blood serum, and the pha-/
gocytic powers of leukocytes and other cells. To a certain extent
these forces are active against all bacteria in all animals, but they
may vary in different species, races, or even individuals in potency
against any given infectious agent, and, to a certain extent, varia-
tions in resistance may be referable to this. The analysis of these
forces, both in the normal and in the artificially immunized animal,
forms the substance of the systematic discussions which are to fol-
low, and, for the present, we will confine ourselves to an examination
of the facts that have been gathered regarding the actual differences
in normal resistance or "Natural Immunity" between various spe-
cies of animals.
And if we glance over the list of diseases to which different spe-
cies and races of animals are victim, it is immediately evident that
some animals are never spontaneously infected with many of the)
micro-organisms that cause extensive and fatal ravages in others.
Also, within the same race or species, an epidemic sweeping through
a community will kill many individuals and leave others unscathed.)
Such differences point to variations in the defensive mechanism,
since the invader in these cases is the same. We speak, therefore, of
Natural Immunity which is an attribute of species, that which,
within the same species, is racial, and that which, within the same
race, is individual. And the attempts to discover the causes under-
lying such differences in natural resistance have elucidated many of
the fundamental principles of immunity in general.
Instances of natural immunity which appear to depend on spe-
cies are common. We have pointed out, above, that in order to make
infection at all possible, it is necessary that the invading germ shall
find suitable cultural conditions in the body of the host. It is this
simple principle which probably explains the fact that bacteria whick
cause disease in warm-blooded animals cannot, as a rule, cause di/
NATURAL IMMUNITY 51
ease in those that are cold-blooded, and vice versa. Thus frequent
attempts to produce anthrax in turtles, frogs, and other cold-blooded
species have failed. Also among warm-blooded animals differences^
in body temperature have been shown to influence susceptibility.
Thus avian tuberculosis does not develop in mammals, nor do the
human and bovine types of tubercle bacilli infect birds. And this is
probably due to the fact that the avian bacillus has become adapted
to growth at from 40° to 45° C., about the normal temperature of
birds, while the mammalian bacilli cease to grow when the tempera-
ture is raised above 40° C. Another observation which clearly illus-
trates the influence of body temperature upon susceptibility is that
made by Gibier l upon anthrax. Frogs are ordinarily resistant to
this disease. When they are kept in water at 35° C. a fatal infec-
tion can be produced. Nuttall's 2 experiments with plague infection
in lizards illustrate the same point. Kept at 16° C., no infection
could take place. Warmed to 26° C., they could be readily infected.
It is ordinarily assumed that these results are explicable upon the
basis of purely cultural and temperature considerations. And this,
indeed, is most likely. It is possible, however, that an additional
factor involved in this may be the lowering of the general resistance
of cold-blooded animals when warmed, just as warm-blooded animals
can be rendered susceptible by chilling.
It is for similar simple cultural reasons, possibly, that diseases
which occur spontaneously in carnivora do not occur in purely
herbivorous animals. The relative resistance of dogs to anthrax
and to tuberculosis may possibly be accounted for in this way.
However, there are many micro-organisms which infect easily
both carnivorous and herbivorous animals, and it may well be that
the frequently cited cases we have mentioned above depend on fac-
tors more complicated than mere cultural conditions incident to
metabolic differences. In most cases of species resistance, indeed,
simple nutritional conditions alone do not serve as valid explana-
tions.
Species resistance may be so perfect that it amounts to an ab-
solute immunity. This is apparently so in the cases cited above,
namely the immunity of the cold-blooded species to certain diseases
of warm-blooded animals. However, such examples are exceptional.
When we are dealing with diseases of warm-blooded animals only,
natural resistance, in all but a limited number of cases, is sufficient
only to prevent the spontaneous occurrence of the particular disease,
or to prevent infection when experimental inoculation with moderate
doses is practiced upon normal animals. In most of these, cases,
however, when the dose experimentally administered is excessive, or
the resistance is lowered artificially, by chilling or by any other
1 Gibier. Compt. rend, de I'acad. des sc., Vol. 94, 1882.
2 Nuttall. Centralbl. f. Bdkt., Vol. 22, 1897.
52 INFECTION AND RESISTANCE
form of local or general injury, infection can be accomplished. In
the case of protozoan diseases species adaptation is much more rigid
and parasites that infect one species are very often restricted en-
tirely to that class, heing unable to infect any other animal, even
though no striking difference in temperature or metabolism exists.
We may convey the clearest conception of all such species differ-
ences by a tabulation of some of the more important infectious dis-
eases of man with a statement in each case concerning its transmissi-
bility to animals, as follows :
Tuberculosis, human type, spontaneously infects man. It is very
often observed in monkeys kept in captivity. Cattle, swine, and
sheep are probably never spontaneously infected; guinea pigs are
highly susceptible to experimental inoculation. Cattle, swine, sheep,
and rabbits are relatively very resistant to experimental infection.
Dogs and goats are still more so. Birds seem to be entirely refrac-
tory.
Tuberculosis, Bovine Type. — Spontaneous infection occurs in do-
mestic animals, chiefly cattle ; it is less frequent in sheep, hogs, and
horses; it has been reported in dogs and goats. In man infection
does occur, but only a small percentage of human tuberculosis is of
the bovine type, and these cases are almost exclusively in children.
In tabulating 1,042 cases which have been carefully studied, Park
and Krumwiede 3 report the following figures :
Cases of Tuberculosis in Man (1042)
Over 16 years
Human type 677, bovine type 9.
5 years to 16 years
Human type 99, bovine type 33.
Under 5 years
Human type 161, bovine type 59.
The large majority of bovine infections were abdominal or in-
volved cervical lymph nodes.
Experimental infection is successful in rabbits and guinea pigs,
both of these animals succumbing more rapidly to this than to the
human bacillus. In fact, the relative resistance of rabbits to the
human bacillus is such that rabbit inoculation is one of the most
important methods of differentiating between the two types. Birds
are refractory.
Tuberculosis of the avian type occurs spontaneously in birds. It
may be experimentally produced in rabbits (Strauss and Gamaleia).
Injected into cattle it causes a local reaction only.
Tuberculosis of cold-blooded animals is not transferable to warm-
blooded animals.
Syphilis spontaneously occurs in man only. It can be inoculated
3 Park and Krumwiede. Jour, of Med. Res., Vol. 23, 1910.
NATURAL IMMUNITY 53
chimpanzees, in which primary and secondary lesions develop,
corresponding mildly to human syphilis. Primary lesions can be
^produced in lower monkeys. It can be transferred by intratesticular
inoculations to rabbits.
Gonococcus infection occurs spontaneously in man only. No
typical lesions can be produced in experimentally inoculated ani-
mals, though death can be caused by large doses, probably by toxic
action.
Influenza bacillus spontaneously infects man only. Experi-
mental infection is partly successful in monkeys only. (Pfeiffer
and Beck, Deut. med. Woch., 1893.)
Glanders. — Spontaneous infection occurs in horses and mules;
less frequently in sheep, goats, and camels. This disease, like plague,
may be regarded as primarily a disease of animals, but man may be
infected by direct or indirect contact with the diseased animal. All
domestic animals may be infected experimentally with ease, except
cattle and rats, in which cases large doses are necessary. Birds show
local reactions only. (Wladimiroff — in "Kolle und Wassermann
Handbuch," Vol. 5, 2d Ed.)
Plague occurs spontaneously chiefly in man and in rats. It has
also been found in California ground squirrels and in hogs during
plague epidemics in Hong Kong. It is highly infectious for guinea
pigs and white rats — slightly less so for mice ; rabbits are much less
susceptible than guinea pigs. Dogs, cats, and cattle are relatively
resistant. Birds appear to be immune. Cold-blooded animals are
immune unless artificially warmed. (See above.)
Malta fever occurs spontaneously in man and in goats. It is
pathogenic for all mammals, but it is not fatal for lower animals
when the organisms are directly cultivated out of the human body.
Diphtheria occurs spontaneously in man only. Experimental in-
oculation is fatal in guinea pigs, rabbits, dogs, cats, and birds. Rats
and mice are highly resistant. The typical pseudomembranous in-
flammation can be produced in susceptible animals only after pre-
vious injury of the mucous membrane, and then it rarely shows any
tendency to spread.
Tetanus is spontaneous in man, horses, cattle, and sheep. It ia
found rarely in dogs and goats. Birds are highly resistant to ex-
perimental inoculation.
Anthrax is primarily a spontaneous infection of cattle, sheep,
and horses ; it occurs in man largely through direct or indirect contact
with these animals. Guinea pigs, rabbits, and white mice are very
susceptible to experimental inoculation. Rats and hogs are less sus-
ceptible, and dogs are relatively resistant, though they can be regu-
larly killed by moderate doses intravenously injected. Birds and
cold-blooded animals are highly resistant.
Asiatic cholera develops spontaneously in man only. Rabbits
54 INFECTION AND RESISTANCE
and guinea pigs can be killed by injections of cultures, but die prob-
ably of toxemia. In rabbits a cholera-like condition has been pro-
duced by injection of the spirilla into the duodenum after ligation
of the common bile duct. (Nikati and Rietsch, Ref. in Deut. med.
Woch.; Vol. II, 1884, p. 613.) Ordinarily no multiplication takes
place in the animal body. Pigeons are insusceptible, a fact which
helps to distinguish this organism from Spirillum metchnikovi and
other similar bird-pathogenic spirilla.
Typhoid fever occurs spontaneously in man only. It has recently
been produced in a mild form in chimpanzees. Animals are suscep-
tible to the endotoxins and can therefore be killed by injections of
bacilli and extracts, but the organism is not invasive as in the case of
the lower animals. Typhoid septicemia can be produced in rabbits
by inoculating them with especially virulent cultures of the bacilli,
or cultures previously grown on rabbit-blood agar (Gay). The ty-
phoid-carrier state may ensue for considerable periods in such ani-
mals.
Pneumococcus infection in various forms occurs spontaneously
in man. Rabbits, mice, and guinea pigs are highly susceptible.
Rats, dogs, cats, cattle, and sheep are relatively resistant.
Staphylococcus and streptococcus infections may occur in almost
all of the warm-blooded animals, chiefly as abscess producers. In
horses a severe form of pleuropneumonia is caused by them.
Leprosy occurs spontaneously in man only. Lesions simulating
human leprosy have been produced in monkeys by inoculation, and
partially successful experiments have been made upon the Japanese
dancing mouse. Other animals are immune.
Scarlet fever occurs spontaneously in man only. Monkeys may
possibly be susceptible, though not all observers have been successful
in such experiments. (Draper and Handford, Journ. of Exp. Med.,
Vol. 17, 1913.) Landsteiner and Levaditi (Ann. Past., Vol. 25,
1911) have succeeded in producing the disease in the chimpanzee,
though they failed with lower monkeys.
Small-pox occurs spontaneously in man only. It is probably iden-
tical with cow-pox. (See reasons for this assumption given by Ha-
cius as cited by Paul in "Kraus and Levaditi Handbuch," etc., Vol.
1.) It can be experimentally produced in monkeys.
• Measles develops spontaneously only in man. Macacus rhesus
ias been successfully inoculated by Anderson and Goldberger (U. S.
Pub. Health Reports, 26, 1911). Other animals are immune.
Typhus fever occurs in man only. Experimentally it has been
produced in chimpanzees, Macacus,, Cercopithecus, Ateles, and My-
cetes monkeys. Anderson has succeeded in producing temperature
reactions in guinea pigs by injecting blood from typhus patients or
from other similarly infected guinea pigs. More exact information
concerning this disease will probably be available soon, if the re-
NATURAL IMMUNITY 55
ported cultivation of the organism of the disease by Plotz is authen-
ticated.
Yellow fever up to the present has been observed in man only.
Poliomyelitis is spontaneous in man only. Can be transmitted to
monkeys and — in a doubtful form — to rabbits. No other animals
are known to be susceptible.
The above represents an incomplete tabulation of the variations
in susceptibility in the animal kingdom for infections which occur
spontaneously in man. They will illustrate sufficiently, however,
the facts of variable species susceptibility as we have stated them.
We might, with equal profit, tabulate the infections occurring spon-
taneously in any single species of animal and show how variable
would be their pathogenic powers for other animals and for man.
Thus man is immune to the organism which causes cattle plague,
and to that of chicken cholera, and probably to many other diseases
peculiar to animals, though, of course, in the case of infections of
the human being we are entirely dependent for such information
upon observed immunity to spontaneous infection, and upon a few
instances of accidental inoculation.
In regard, also, to differences of susceptibility between various
races, within the same species, many interesting facts have been ob-
served. Thus gray mice are, as a rule, more resistant to strepto-
coccus and pneumococcus infection than are wnite mice. Algerian
sheep are said to be more resistant to anthrax than are European
sheep. Of black rats inoculated by Miiller 4 with anthrax over 79
per cent, survived, while of white rats similarly inoculated only 14
per cent, survived.
In man, too, racial differences are marked. The extraordinary
susceptibility of the negro to tuberculosis is familiar to all American
physicians, and it is well known that Eskimos transported to tem-
perate climates and civilized conditions are particularly prone to
contract this disease. Small-pox is considered a relatively mild
disease in Mexico. Dr. James Carroll 5 stated that whites are more
susceptible to yellow fever than are negroes, and that among the
latter those living nearest the equator are less susceptible than are
the more northern races. There seems to be no doubt about the
actual occurrence of such racial differences, although, as Hahn 6
very justly points out, many instances formerly regarded as racial
differences of susceptibility may have been simulated by racial,
or often religious, differences of custom that influence sanitary con-
ditions, and consequently the incidence of epidemic disease.
Apart from the explanations furnished in a few instances by
4 Miiller. Fortschr. der Med., 1893. Cited from Sobernheim, in "Kolle
u. Wassermann Handbuch," 2d Ed., Vol. 3.
5 Carroll in "Mense, Tropenkrankheiten," Vol. 2, p. 124.
6 Hahn in "Kolle und Wassermann's Handbuch," Vol. 1.
56 INFECTION AND RESISTANCE
gross physiological differences such as body temperature, the factors
determining species resistance are largely a mystery, and in the
matter of racial variations, of course, we have no instances in which
such very obvious physiological factors play a part. In attempting
to find causes for differences of resistance or susceptibility in gen-
eral, the nature of the problem makes it necessary for us to examine
it from a number of different points of view. A micro-organism
may be infectious for a given species of animal more than for
another, because of special adaptation to the conditions, nutritive
and otherwise, encountered in the tissues of these animals. Such
adaptation is illustrated in the experience of Pasteur with "rouget"
and with rabies, where passage through one variety of animal en-
hanced the virulence for this species but reduced it for others; and
the same thing is easily demonstrated in the laboratory with so many
bacteria that it may be accepted as a principle underlying enhance-
ments of virulence in general. This adaptation implies that, to a
certain extent, the part played by the animal body in determining
its own susceptibility is passive. Gonococcus, for instance, infec-
tious for man only, requires human protein for growth, at least in
its first generations outside the body. Its ability to cause disease
in man may be largely dependent upon its cultural need of human
protein. The resistance of other animals to this disease, then, is, in
part, due to their failure to supply proper nutriment. This, as Kolle
points out, is analogous to Atrepsie, a term used by Ehrlich, in
speaking of the insusceptibility of one species to cancerous growths-
originating in another.
Again, "adaptation" on the part of the bacteria may imply, not
only an increased ability to meet altered cultural conditions, but an
actual acquisition of greater offensive or invasive powers with which
to meet the particular defences opposed to it by the given animal.
Thus the increased virulence of typhoid bacilli after cultivation in
immune sera would point toward an increased ability to survive
under the adverse conditions encountered in the animal body. An
organism may possibly acquire particular infectiousness for one
species to the exclusion of others, by a succession of spontaneous
inoculations — comparable to the experimental passage of the micro-
organism through animals of the same species. This is especially
probable in diseases such as gonorrhea, syphilis, and some others
where infection is usually direct from one person to another. And
it is these diseases particularly in which infectiousness is rather
strictly limited to the human species.
Regarding the matter purely from the point of view of the ani-
mal body and the factors which determine its powers to ward off a
given infection, we may justly assume that natural resistance may
be largely a matter of inheritance. Whether this is to be interpreted
as purely an instance of survival of the fittest or whether immunity
NATURAL IMMUNITY 57
acquired by an individual can be wholly or in part transmitted to
the offspring is an open question — at present in the same state of
unclearness as are other questions relating to the transmissibility of
acquired characteristics. However this may be, there are a number
of facts available which indicate that inheritance plays an important
part. It is apparent in the case of many diseases afflicting human
beings that infection takes a milder course in those races among
which it has long been endemic — whereas the same disease, suddenly
introduced among a new people, is relatively more severe and spreads
more rapidly. This seems to be the case with yellow fever and tuber-
culosis, and in measles and small-pox, too, the principle seems
to hold good. Syphilis when first described authentically — as epi-
demically sweeping through Europe toward the close of the 15th
century — appears to have been a far more acute and violent disease
than it is among us to-day. It may well be that this depends upon
a gradual elimination (elimination in this case, especially as far as
reproduction is concerned) of those individuals that are fortuitously
more susceptible and, by natural selection, a higher racial resistance
is gradually developed. Whether or not direct inheritance of the
individually acquired immunity can be considered at all as a con-
tributing factor is difficult to decide. That immunity can be trans-
mitted from mother to offspring was observed by Chauveau 7 as
early as 1888. Lambs thrown by anthrax-immune ewes possessed a
higher resistance against this infection than did the lambs of normal
ewes. The extensive experiments of Ehrlich,8 carried out chiefly
upon mice with the vegetable poisons ricin and abrin, showed that in
these cases immunity may be transmitted from mother to offspring,
but depends upon a passive transfer of the specific antitoxins both
by the blood and the milk of the mother. The sperm of the father
did not seem to have anything to do with inherited resistance, since
no immunity followed in the offspring when immunized males were
paired with normal females. From the complete absence of im-
munity in the second generation (grandchildren) of the immunized
female, and from the short duration (2 to 3 months) of its per-
sistence, he concluded that the ovum itself had no influence, but that
the entire phenomenon was attributable to a passive transference of
antitoxins from mother to child during gestation and lactation. He
interpreted, in the same sense, Chauveau's anthrax experiments, and
similar experiments of Thomas9 and Kitasato 10 with symptomatic
anthrax, suggesting that, here also, a transept of antibodies -from
mother to offspring had taken place. The experiments of Ehrlich
permit of no doubt as to the validity of his conclusions. However,
7 Chauveau. Ann. Pasteur, 1888.
8 Ehrlich. Zeitschr. f. Hyg., 1892, Vol. 12.
9 Thomas. Compt. rend, de I'acad. des sc., Vol. 94, cited by Ehrlich, loc. cit.
10 Kitasato. Cited by Ehrlich, loc. cit.
58 INFECTION AND RESISTANCE
we must remember that they were carried out with antitoxic im-
munity only, in which the resistance is purely dependent upon the
circulating antibody and is never, even in actively immunized indi-
viduals, a permanent state. In immunity such as that acquired
against typhoid fever, plague, cholera, and other diseases after re-
covery from an attack, the individual remains relatively resistant
long after the demonstrable antibodies have disappeared from
the circulation, and we must assume that this permanent re-
sistance depends upon a physiological alteration — inexplicable for
the present, but surely residing in the body cells. In such cases
it is by no means certain that there may not be a very slight, but
through generations gradually accumulating, inheritance of im-
munity. At any rate the experiments of Ehrlich do not disprove
such a possibility. Moreover, in this connection it must not
be forgotten that natural immunity, unlike acquired immunity,
cannot be passively transferred from one' animal to another, and
implies therefore a fundamental cellular difference rather than
a condition depending merely upon antibodies circulating in the
l)lood.
For this last reason also it has been unsatisfactory to attempt
explanations of natural immunity purely upon grounds of bacteri-
cidal and other properties of the blood serum. These points we will
take up at greater length when we discuss the mechanism of resist-
ance in general.
An important observation upon the inheritance of serum prop-
erties is that which has been made by Ottenberg and Epstein n in
connection with the iso-agglutinins. We shall see in another section
that the blood serum of one human being will often possess the
property of agglutinating the human blood cells of another indi-
vidual. These iso-agglutinating constituents of the serum are ap-
parently transmitted from parents to offspring. Yon Dungern and
Hirschfeld,12 in studying these iso-agglutinins in 72 families, upon
348 people, not only confirmed the observations of the preceding
workers, but showed that such inheritance follows Mendelian laws.
Not only is this of great biological interest, but it is of great im-
portance in connection with our present discussion in showing that
such properties as agglutinating powers of serum can be influenced
by inheritance from the father as well as from the mother.
The individual differences in resistance which unquestionably
exist among members of the same species and races are very difficult
to explain, but, as far as we can tell anything about them at all, they
seem to depend upon variation in what is popularly spoken of as
"general condition." The laboratory animals with which most ex-
perimentation is done, rabbits and guinea pigs, if healthy, show very
11 Ottenberg and Epstein. Proceedings of the N. Y. Path. Soc., 1908.
12 Von Dungern and Hirschfeld. Zeitschr. f. Immunitats., Vol. 4, 1910.
NATURAL IMMUNITY 59
slight individual variations. In fact, the astonishing uniformity of
reaction on the part of guinea pigs of similar age and weight against
measured quantities of bacterial toxins has alone made it possible
to utilize these animals in the standardization of antitoxins. Pneu-
mococcus and streptococcus cultures can be measured with reason-
able accuracy upon white mice of approximately uniform weight,
and the same animals are relatively uniform in their reactions to
identical amounts of tetanus poison. Many other examples might
be cited which make it clear that healthy animals of the same species,
kept under the same conditions, fed upon the same food, and of ap-
proximately the same age and weight, differ but slightly from each
other in reaction to the same infectious agent. This would indicate
that the individual differences in resistance displayed so plainly by
human beings are due, not to any fundamental individual variations,
but rather to such fortuitous factors as nutrition, metabolic fluctua-
tions, temporary physical depression, fatigue, or chilling. A person
suffering from functional impairment of any kind is more likely
to permit the invasion of a pathogenic micro-organism than is a per-
fectly healthy well-nourished individual of the same species.
Most of these facts we know from the accumulated experience of
clinicians who also have given us much valuable information con-
cerning the susceptibility to infection on the part of chronically dis-
eased persons, especially diabetics and nephritics. In the case of a
few of these influences, chilling and fatigue, experimental data on
animals are available. It is, however, extremely difficult to analyze
the causes underlying such depression of resistance. For instance,
with fatigue or chilling there may be temporary congestion of mucous
surfaces, due to vasomotor influences, which alter the secretions on
mucous surfaces, or interfere with the normal mobilization of
leukocytes, permitting penetration of bacteria where ordinarily
they would have been held back. Our ignorance is nowhere more
clearly illustrated than in the fact that we know practically nothing
concerning the relation between a thorough chilling and the acquisi-
tion of what is spoken of as a common acold." We can only assume
that there is interference in some way with the normal bactericidal
and phagocytic mechanisms,, making possible the penetration and
lodgment of small quantities of bacteria, ordinarily destroyed imme-
diately after entrance or prevented from entering at all.
Of course we must except those individual differences of sus-
ceptibility which may be dependent upon inheritance. We know,
for instance, that in such diseases as diphtheria, where resistance
depends upon antitoxins circulating in the blood, there may be a
passive immunity, conferred from mother to offspring, which lasts
for several weeks or months after birth. It is important to remem-
ber such a possibility in the selection of guinea pigs for diphtheria
antitoxin standardization, as Anderson has pointed out. Whether
60 INFECTION AND RESISTANCE
or not a tendency to tuberculosis can be inherited is still an open
question. In most cases it is more than probable that the supposedly
inherited tendency to tuberculosis is not really an inherited sus-
ceptibility, but rather an actual infection acquired during childhood
from the parents. Cornet and Kossel,13 who have recently sum-
marized the statistics dealing with this problem, have come to the
conclusion that this factor, namely, infection from the parents,
probably is the cause of the greater frequency of tuberculosis among
children of tuberculous parents, and that there is no definite proof
of inherited susceptibility.
ACQUIRED IMMUNITY AND IMMUNIZATION
We have outlined in the preceding pages the differences in sus-
ceptibility to various diseases apparent among different species of
animals, and have noted that the degree of resistance of some animals
to infection with germs rapidly fatal to others is often sufficiently
well-marked to be termed "immunity." Such immunity, because it
is a natural biological attribute of the species, as much a character-
istic property as are its anatomical or physiological properties, has
been spoken of as "Natural Immunity."
It is a matter of common knowledge, however, that among
species of animals readily susceptible to certain infections resistance,
or even extreme resistance, i. e., immunity, may be acquired by an
attack of the disease. Thus human beings who have recovered from
plague, small-pox, typhoid fever, cholera, the exanthemata, mumps,
typhus, yellow fever, and a number of other conditions do not ordi-
narily contract the disease a second time. In some of these condi-
tions, notably cholera, plague, typhoid fever, and small-pox, the rule
is almost invariable. In others, such as measles, scarlet fever, and
mumps, a second attack may occur, though it is rare.
The following table briefly indicates infectious diseases in which
permanent immunity follows an attack :
Infectious Diseases in Which One Attack Conveys Lasting Immunity
Plague.
Typhoid — second attack rare — about 2.4 per cent. (Curschmann).
Cholera.
Small-pox — second attack very rare.
Chicken-pox — second attack very rare.
Scarlet fever — second attack about 0.7 per cent.
Measles — second attack uncommon, but less rare than scarlatina.
Yellow fever.
Typhus fever.
Syphilis — reinfection rare, though more common than formerly supposed.
Mumps-second attack rare (Kraus).
Poliomyelitis.
13 Cornet and Kossel in "Kolle u. Wassermann," Vol. 5, 2d Ed.
ACQUIRED IMMUNITY 61
No lasting immunity is conferred by one attack in:
Infection with the Pyogenic cocci
Gonorrhea
Pneumonia
Influenza
Glanders
Dengue fever
Diphtheria in general protection, second attack In 0.9
per cent, cases-0.01 antitoxin unit per c. c. of circu-
lating blood protects.
Recurrent fever
Tetanus
Erysipelas
Beri beri
Tuberculosis
There is another group of diseases in which the immunological
conditions after infection are as yet not clear — namely, protozoan
infections like malaria and trypanosomiasis, and treponema diseases
like syphilis. In these conditions reinfection seems to be impossible
only so long as the individual still harbors the microorganism, but no
lasting immunity is conferred. We have discussed these conditions
in extenso in the section on syphilis immunity, (see p. 508).
These observations actually form the point of departure of that
entire branch of medical science which devotes itself to the study of
resistance to infection, serum diagnosis, and specific therapy, and
it will be seen that all the facts that have been gathered upon these
subjects are the fruits of detailed analysis of this phenomenon of
acquired immunity.
Its occurrence in many instances nas been so striking that ancient
observers, long before the birth of rational medicine, referred to it,
and often drew from it conclusions of great hygienic importance.
Thucydides, in the second book of his account of the Peloponnesian
Wars, in describing the plague at Athens, notes the apparent safety
from reinfection of those who had recovered, suggesting the possibil-
ity of their being therefrom immune against disease in general. The
literature of the Middle Ages and of earlier modern times contains
numerous further references which indicate that acquired resistance
was clinically recognized as a result of recovery from many diseases.
The phenomenon was not only observed, but put to practical utiliza-
tion by the ancients of China and India. Thus the practice of inocu-
lating children with small-pox material from the active pustules of
patients, or making them sleep in beds or wear the shirts of sufferers
was a dangerous practice but logical, on the reasoning that the disease
conveyed to a person in full health and good condition would probably
take a mild course and confer immunity, while the naturally acquired
disease, contracted often because of the weak and debilitated condi-
tion of the individual, would be more apt to end fatally.
62 INFECTION AND RESISTANCE
Such methods, though barbaric and eventually unjustified by the
naturally high mortality incident upon them, were actually brought
to Europe from the East, and for a time practiced in European
countries.
The first great advance which bridged the gap between the obser-
vations regarding naturally acquired immunity and rational experi-
mental immunization was made by Edward Jenner. It has been no-
ticed before Jenner began his w.ork that milkmaids and others who
had contracted cow-pox in the course of their occupations were usually
spared when a small-pox epidemic occurred in their community. Spo-
radic attempts had been made to put this observation to practical use,
but no one with sufficient intelligence, persistence, and training had
taken up the matter seriously. Jenner, interested by the reports of this
nature and by his own observations, was especially impressed by the
similarity between the local manifestations of small-pox, cow-pox, and
a disease of horses spoken of as "grease." Though at first disinclined
to identify small-pox with cox-pox (at present the prevailing opinion
is that the second is an attenuated form of the former), Jenner
thoroughly investigated cases of alleged protection by cow-pox, a claim
which before this had been hardly more than a rumor, and finally,
with the encouragement of John Hunter, proceeded to the vaccina-
tion of human beings with cow-pox, testing the result by subsequent
inoculation of the same individual with small-pox. His report to the
Royal Society in 1796 and his subsequent publications incorporate
the results of these experiments by means of which the practice
of vaccination against small-pox was introduced and the virtual
eradication of the disease from civilized communities was attained.
The principles underlying small-pox vaccination are extremely
simple. The attenuated virus after inoculation incites a mild and lo-
calized form of the disease, from which the subject recovers rapidly
and completely. The recovery implies the mobilization of certain pro-
tective forces and a specific physiological alteration of the body in
such a way that a permanently, or at least prolongedly, increased re-
sistance against the disease remains. In consequence, if the indi-
vidual is subsequently exposed to spontaneous infection with this dis-
ease, his acquired specific resistance suffices to prevent invasion by
the virus. This is merely an artificial imitation of the conditions
which obtain when an individual recovers from an attack of a disease
and is rendered immune thereby. In this case, however, the attenua-
tion of the virus has eliminated the dangers attendant upon an actual
attack. The immunity thus conferred is probably never as perfect nor
as lasting as that following a seizure of the disease in its unattenuated
form; however, it suffices, as a rule, to prevent spontaneous infec-
tion which is never as severe a test as experimental inoculation.
In contrast to the "Natural Immunity" which is an inherited at-
tribute of race or species, we speak of such increased resistance in a
ACQUIRED IMMUNITY 63
member of an originally susceptible race as "Acquired Immunity."
When the immunity has been attained as the result of an attack of the
disease itself it is spoken of as "Naturally or Spontaneously Acquired
Immunity" When produced by some form of treatment with the
virus of the disease, altered in such a way that an actual attack is
averted, we speak of it as "Artificially Acquired Immunity."
The premises of Jenner's reasoning were valid as his experiments
were convincing. But knowledge regarding infectious disease and
its causation by living germs was not developed until almost one
hundred years later, by the work chiefly of Pasteur. For this reason
no direct continuation of Jenner's work appeared until Pasteur
made his communication upon Chicken Cholera to the Parisian
Academy of Medicine in 1880. Though his investigations differed
entirely from those of Jenner both in method and the nature of the
disease with which they dealt, Pasteur recognized the similarity of
the fundamental principles underlying both discoveries.
His observations took origin in a purely accidental occurrence.
Cultures of chicken cholera which had been allowed to stand without
transplantation and under aerobic conditions for periods of several
months were found to have diminished in virulence. Inoculated into
chickens, they failed to kill, giving rise in many cases to localized
lesions only. It occurred to Pasteur that inoculation with such an
attenuated culture might protect against subsequent infection with
fully virulent strains and, indeed, experimental investigation of this
idea proved to be correct. He developed a method of "vaccination"
against chicken cholera which consisted in injecting -first a very
much attenuated culture of the organism (premier vaccin), and,
after 12 or 14 days, another less perfectly attenuated (deuxieme
vaccin), since he observed that a single inoculation was often in-
sufficient to confer protection. After two inoculations a degree of
immunity could be attained which sufficed to protect against spon-
taneous infection as well as against experimental inoculation with
doses of the virulent germs, fatal for untreated animals.
These experiments, simple as they are, constitute the beginnings
of the science of Immunity, since, for the first time, an investigator
working with a pure culture of a pathogenic microorganism had
succeeded, in planned and purposeful experiments, in conferring
artificial immunity. The path was now clearly indicated and the
years immediately following were fruitful in the development of
many methods by which pathogenic bacteria may be attenuated and
changed in such a way that they can be used to confer immunity
without causing more than a transient and harmless reaction in the
subject. Most of the earlier discoveries of this kind came from
Pasteur himself and from members of his school.
Since in all these methods the inoculated animal attains its in-
creased resistance by reason of the activities of its own tissues, these
64 INFECTION AND RESISTANCE
processes are spoken of as "Active Immunization/' ~No protective
factor is conferred directly. The disease itself is inoculated, though
in an altered form, and the subsequent immunity is purely the result
of the physiological reaction occurring as the subject struggles against
and overcomes the injected virus, bacteria, or their products. Such
"Active Immunization," we shall see, is in contrast to "Passive
Immunization," a procedure in which the serum of an actively im-
munized animal is injected into another, carrying with it certain
substances by which protection is conferred. The recipient here is
passively protected by products of the active reaction which has
taken place in the body of the donor.
After his success in active immunization against chicken cholera
Pasteur applied the principles here learned to experiments upon the
protection of animals against anthrax. This problem was fraught
with considerable difficulty because of the great virulence of the
anthrax bacillus. However, successful attenuation was attained by
a method which depended upon the cultivation of anthrax cultures at
temperatures above the optimum for its growth. Toussaint 14 had
shown that the resistance of sheep could be increased if they were
inoculated with blood from animals dead of anthrax after this had
been heated to 55° C. for 10 minutes. Toussaint's idea had been
that by heating the blood in this way the bacteria themselves were
killed. Pasteur 15 showed, however, that this was not the case, but
that what actually occurred was a reduction of the virulence of the
strain by the exposure to heat. As a matter of fact, moreover, the
method of Toussaint did not furnish a reliable means of attenuating
anthrax, and Pasteur succeeded in developing a far more satis-
factory procedure on which he based a practical method for the pro-
tective vaccination of sheep and cattle.
His method was as follows : 16 Virulent anthrax bacilli were cul-
tivated at 42° to 43° C. on neutral chicken bouillon (Sobernheim
states that horse or beef broth — or even agar — answers the same pur-
pose). Cultivated under these conditions a gradual and progressive
reduction of virulence occurs. After about 12 days of such cultiva-
tion the culture as a rule no longer kills rabbits, but is still virulent
for guinea pigs and mice. After twenty-four or more days the
virulence for rabbits and guinea pigs is lost and mice only can be
killed with it. The latter — the most fully attenuated strain — was
called premier vaccin by Pasteur, and, in the immunization of cattle
or sheep, is first injected. After 10 or 12 days the stronger deuxieme
vaccin is administered. This is the method which Pasteur used in
his now classical experiments at Pouilly-le-Fort, in which he con-
14 Toussaint. Compt. rend, de I'acad. des sc., 1880.
15 Pasteur, Chamberland and Roux. Compt. rend, de Vacad. des sc., Vol.
91, 1881.
16 Cited from Sobernheim. "Kraus und Levaditi Handbueh der Technik,
etc,," Vol. 1, 1909.
ACQUIRED IMMUNITY 65
vinced a hostile audience of the efficacy of his immunization. Sheep
were protected in the manner indicated, and 14 days after the last
injection a fully virulent culture was inoculated and the animals
found capable of successfully resisting it.
In the train of this work many other methods of producing active
immunity have been devised — all of them of considerable theoretical
interest and many of them practically adapted to some special case.
We may conveniently classify these methods as follows :
I. IMMUNIZATION WITH LIVING BUT ATTENUATED CULTURES
(1) Methods in which the attenuation is obtained by heating.
This is the method of Toussaint as outlined above, in which anthrax
blood was heated to 55° C. for 10 minutes, and is probably the least
efficient or reliable method for the attenuation of the anthrax bacillus.
It has been applied to rabies by Babes (cited from Kraus in "Kraus
u. Levaditi Handbuch, etc.," Vol. 1, p. 708), who attenuated the
virus by heating to 58° C. for periods varying from 2 to 40 min-
utes.
(2) Attenuation by prolonged cultivation of the bacteria at tem-
peratures above the optimum for their growth. This is illustrated by
Pasteur's anthrax immunization as described in the preceding para-
graphs.
(3) Attenuation by passage through animals. Examples of this
are Pasteur's experiments with the arouget" organism, in which pas-
sage through rabbits diminished the virulence for hogs. The attenu-
ation of rabic virus by passage through monkeys is another instance,
and Jennerian vaccination is also an example of this, although here
the attenuation by passage through cattle is attained naturally and
not by experimental procedures. Based *on the same principle is
Behring's 17 method 18 of immunizing cattle against tuberculosis by
inoculating them with tubercle bacilli of the human type.
(4) Attenuation by prolonged growth of bacteria on artificial
media in the presence of their own metabolic products. This is the
method first employed by Pasteur in chicken cholera, as described
above, and is applicable to many organisms, such as pneumococci,
streptococci, and others. In fact, it is difficult to maintain the viru-
lence of many of these bacteria unless special methods of cultivation
or passage through animals are practiced. Pasteur believed that free
access of oxygen to the cultures increases the rapidity of the attenua-
tion.
(5) Attenuation by drying. The classical example for this
method is the Pasteur method of prophylactic immunization against
17Behring. "Therapie der Gegenwart," April, 1907.
18 See also Romer, "Kraus u. Levaditi Handbuch," 1st Suppl., p. 310.
66 INFECTION AND RESISTANCE
rabies. Rabbits are inoculated with virus fixe, and their spinal
cords dried for varying periods in bottles containing KOH at a tem-
perature of about 25° C. The virus grows progressively weaker
with each day of drying. Greater details concerning this method are
given in another place (see page 489).
(6) Attenuation by the use of chemicals. — Chamberland and
Roux 19 succeeded in attenuating anthrax by growing it in the
presence of various antiseptics. They used carbolic acid 1 to 600,
bichromate of potassium 1 to 1,500 and sulphuric acid 1 to 200,
and found that, after a short time of cultivation under such condi-
tions, the bacilli lost their ability to form spores and became avirulent
for sheep. Behring 20 and others have applied this method to the
attenuation of diphtheria toxin; Behring adds terchlorid of iodin,
Roux potassium iodid — iodin solutions. The principle, of course,
is not exactly the same in the last cases, since here the attenuation
is not of the bacteria themselves, but rather of the toxin.
(7) Attenuation by cultivation under pressure. This method
is difficult to apply, and has no striking advantages over other pro-
cedures. It was employed by Chauveau 21 for the attenuation of
anthrax. He succeeded in accomplishing this by cultivation of
anthrax bacilli at 28-39° C. at a pressure of 8 atmospheres.
II. ACTIVE IMMUNIZATION WITH FULLY VIEULENT CULTURES IN
SUBLETHAL AMOUNTS
The original methods of Pasteur carried out with chicken cholera
and anthrax were aimed particularly at diminution of virulence,
since these organisms, as isolated from the diseased animal, are so
extremely infectious that it would be very difficult — (in the case of
many animals, impossible) — to inoculate with the unattenuated
germs without producing fatal disease. However, in the case of
many other infections it has been found feasible to inoculate normal
animals with the fully virulent germs in such small quantities that
the body can successfully overcome them, and, in doing so, acquire
specific resistance. It is obvious that this method is more easily
carried out with the organisms which Bail terms "half parasites"
than with organisms as highly infectious as anthrax. Ferran 22
applied this method both to animals and to human beings with broth
cultures of cholera spirilla. Hogyes 23 has introduced a similar
procedure for immunization against rabies by injecting dilutions of
19 Chamberland and Roux. Compt. rend de I'acad. des sc., 96, 1882.
20 Behring and Wernicke. Zeitschr. f. Hyg., 12, 1892.
21 Chauveau. Compt. rend, de I'acad. des sc., Vol. 98, 1884.
22 Ferran, Compt. rend, de I'acad. des sc., 1885.
23 Hogyes. "Lyssa Nothnagels Handbuch, etc.," Vienna, 1897.
ACQUIRED IMMUNITY 67
fully virulent rabic virus, beginning with a dilution of 1 to 10,000
and rapidly working up to a dilution of 1 to 10. In tuberculosis
immunization with fully virulent cultures in small amounts has been
attempted by Webb, Williams, and Barber,24 using the Barber method
of isolation, and giving a single micro-organism at the first injec-
tion. That such a method is feasible, if carried out with sufficient
care, even with the most virulent germs, was demonstrated by the
same workers. They succeeded in immunizing animals against
anthrax (with cultures kept 12 hours on agar) 25 by injecting a
single thread (3 to 6 bacilli) as the first dose, and then gradually
increasing the amount.
In the general laboratory immunization of animals treatment
with virulent bacteria in sublethal doses is of considerable value and
frequently employed.
It would seem that possibly this method or some modification of
it will be found to have very definite advantages over methods in
which either attenuated or dead bacteria are employed. Bail's work
upon the aggressins and upon anti-aggressin immunity (see chapter
I, page 21) has opened the possibility that virulent bacteria pro-
vide, within the living body, specific aggressive substances which are
not produced in the test tube. If this proves to be true, and the
question is by no means settled, it may be necessary in such cases
to immunize with organisms which are in a condition capable of
producing these aggressins. Sublethal doses of fully virulent or-
ganisms would furnish these conditions more perfectly than at-
tenuated avirulent strains, in which the invasive (aggressive) power
is considerably diminished.
The methods of active immunization so far described differ from
those which are to follow in that the preceding were all based upon
the use of living bacteria or virus, whereas the methods to be de-
scribed below depend upon the treatment of animals with dead bac-
teria or bacterial products. It is well to call attention in this place
to the fact that a number of recent investigations seem to point to
the greater efficiency of immunization with living germs. This
method has recently given hopeful results in the case of plague in
the hands of Strong;26 and Metchnikoff and Besredka,27 in their
attempt to vaccinate chimpanzees against typhoid fever, make the
statement that vaccination with dead typhoid bacilli or autolysates
does not confer adequate protection, but that this can be attained by
treatment with small doses of the living bacilli.
24 Webb, Williams, and Barber. Jour, of Med. Res., Vol. 15, 1909.
25 This was not possible where the organisms were taken directly from
the blood of a dead mouse. In such cases even a single thread caused fatal
disease.
26 Strong. Jour, of Med. Res., May, 1908.
27 Metchnikoff and Besredka. Ann. Past., Vol. 25, 1911.
68 INFECTION AND RESISTANCE
In speaking of this subject it is well to mention recent ob-
servations upon immunization with "sensitized" bacteria,28 although
this necessitates anticipatory reference to subjects not so far dis-
cussed. It is a matter of common experience in laboratories that
rabbits and other animals will withstand relatively large amounts of
pathogenic bacteria if these are first treated with heated specific
immune serum (sensitized). This is probably due to the fact that
such "sensitized" micro-organisms are very rapidly taken up by
phagocytes. In spite of the phagocytosis, immunity is developed.
MetchnikofF and Besredka, in the communication alluded to above,
state that typhoid vaccination with unaltered living bacilli is efficient,
but is attended by severe local and general reactions. If the living
bacilli are first "sensitized" no such severe reaction occurs and im-
munization is nevertheless successful. The recent work of Gay points
in the same direction, and it is at least possible that by the practice of
sensitization we may be able to employ living unattenuated organisms
for purposes of immunization more extensively than we have in the
past.
III. ACTIVE IMMUNIZATION WITH DEAD BACTERIA, AND BACTERIAL
EXTRACTS
This method is the one most extensively practiced in the labora-
tory immunization of animals. It is usual in most experiments of
this kind to inject dead organisms once or twice before living bac-
teria are administered. High degrees of resistance can in some
instances be attained by progressively increasing doses of dead cul-
tures only. This method is not only useful in experimental work,
but is clinically employed in the active immunization of human
beings as introduced by Wright and as applied, before Wright, to
tuberculosis (tuberculin treatment). But it is very probable that
"the immunity so attained is not entirely comparable to the immunity
following an attack of a disease, nor even that produced by the in-
jection of living bacteria.
The method employed for killing the bacteria is of considerable
importance since, both by excessive heating as well as by too vigorous
•chemical treatment, the immunizing properties of the bacterial
protein may be destroyed. In employing heat it is a safe rule never
to expose the bacteria for prolonged periods to temperatures which
considerably exceed the thermal death point. As a rule, heating non-
spore-forming bacteria to a temperature of from 65° to 70° C. for
28 Refer to p. 159 and the discussion of the conception of "sensitization"
which follows.
ACQUIRED IMMUNITY 69
thirty minutes will suffice to kill them without too radically altering
the immunizing properties of the protein constituents.29
If the temperature is not raised above 60° C., and this is ad-
vised by many workers, the suspensions must be carefully controlled
by cultural tests before they are used, at least for the treatment of
human beings. As we shall see in a later section, the best results have
been obtained when heating was not carried beyond 53° to 55° C.
When bacterial death is to be accomplished by chemicals the
antiseptics most commonly used are carbolic acid (0.5 per cent.),
toluol (removed before use of vaccine by filtration or evaporation),
chloroform, and formaldehyd (1 per cent.).
Pfeiffer, who was one of the first to practice the immunization
of animals with dead bacteria on an extensive scale, believed that,
in the case of bacteria which were toxic by reason of their intra-
cellular constituents (endotoxins), the injection of the cell protein
itself, whether dead or alive, was the sole essential for successful
immunization. The method developed by Kolle 30 and by Pfeiffer
and Marx 31 for the prophylactic immunization of human beings
against cholera depends upon the injection of cholera cultures
emulsified in salt solution, killed by exposure to. 5 8° C. for one hour,
and further insured against contamination by the addition of 0.5
per cent, phenol. The application of this method to other diseases,
both prophylactically and therapeutically, is more fully discussed
in another place. (See chapter XIX.)
Since the essential point in such immunization is the introduc-
tion of the bacterial protein, it is often customary to inject bacterial
extracts instead of the whole cells. This has been especially desir-
able in the case of such insoluble micro-organisms as the tubercle
bacillus, where the injection of the whole dead organism produces
localized reactions similar to those caused by the living bacteria.32
Thus "Old Tuberculin," as commonly used, is a glycerin-broth ex-
tract of tubercle bacilli. The method has been extensively used and
a variety of procedures have been devised for bacterial extraction.
These have included simple autolysis of the bacterial bodies in alka-
line broth, shaking in salt solution in mechanical shakers, trituration
with salt or sand, trituration after freezing, digestion with proteo-
lytic enzymes, and extraction by pressure in a Buchner press.
We may mention some of the more important methods for pre-
29 In a subsequent chapter (p. 258) we shall see that the physical changes
produced in an antigen by heat result in differences in the antibodies formed
after animal inoculation. This point has practical significance in the present
connection. See also the chapter on agglutinins, the work of Joos there dis-
cussed, and Friedberger and Moreschi, Centralbl. f. Bakt., 1905, Vol. 39.
30 Kolle. Deutsche med. Woch., 1897, p. 4.
31 Pfeiffer and Marx. Deutsche med. Woch., 1898.
32 Prudden and Hodenpyl. N. Y. Med. Journal, 1891.
70 INFECTION AND RESISTANCE
paring bacterial extracts for purposes of immunization and antigen
production in general as follows:
A. Extraction of Bacteria by Permitting Them to Remain for
Prolonged Periods in Fluid Media
The bacteria may be grown upon slightly alkaline bouillon and
kept at incubator temperature for one to two months. They are
then filtered through Berkefeldt or other suitable filters. This is the
common method of producing antigen for precipitin reactions, in fact
the method employed by Kraus in the discovery of the bacterial pre-
cipitins. It is by no means certain whether the antigens prepared in
this way represent simple extractions or autolytic products of the
bacteria; probably both processes take place. The antigenic value
of the fluids obtained in this way is never very great. From such
filtrates Brieger and Mayer, Pick, and others have attempted to
obtain the antigen in a purified form by chemical precipitation.
Pick 33 precipitates the bouillon filtrate by saturation with ammonium
sulphate; the precipitate is redissolved in water and again precipi-
tated with ammonium sulphate and the resultant precipitate dried
on a filter. It is then dissolved in water and precipitated with
alcohol. The sticky substance which comes down represents the
antigen.
Suitable extracts can occasionally be obtained also by emulsify-
ing agar cultures in physiological salt solution and allowing them
to stand for twenty-four hours or more at incubator temperature.
In our own experience we have found this method rather inefficient
for yielding strong extracts. More efficient extraction is usually ob-
tained when the bacteria are suspended in alkaline fluids such as
N
— sodium hydrate. Lustig and Galleotti digest the bacterial mass
for 24 hours with 1 per cent. NaOH, then precipitate with am-
monium sulphate, dry in vacuo and pulverize.34
Recently, also, Uhlenhuth 35 has employed the proprietary prep-
aration aantiformin" 36 for the production of antigens. This
33 Pick. "Hoffmeister's Beitrage, etc.," Vol. 1, 1902. For an extensive
discussion of the various methods employed for the production of bacterial
antigens by chemical methods see Pick in Kraus und Levaditi, etc., Vol. 1, and
the same author in Kolle u. Wassermann, etc., 2nd Ed., Vol. 1.
34 See Pick. LOG. cit.
55 Uhlenhuth. Centralbl. f. Bakt., I, Ref. Vol. 42, Beilage, p. 62.
36 "Antiformin" is a substance largely employed for the cleansing of pipes
and vats of organic matter because of its powerfully solvent action. Its
value in concentrating tubercle bacilli out of sputum and other mixtures
depends upon its power to dissolve the tissue elements and all bacteria except
those that are acid-fast. Rosenau ("Preventive Medicine and Hygiene,"
ACQUIRED IMMUNITY 71
substance thoroughly dissolves all but the acid-fast bacteria when
used in concentrations of 2.5 per cent. Since it is alkaline it is
necessary to neutralize it with hydrochloric or sulphuric acid before
use.
For the preparation of antigen from pneumococci Neufeld 37 has
utilized the solvent action upon these organisms of bile. He adds
the bile and broth cultures just as this is done in the diagnostic
"bile test" (0.1-0.2 c. c. of fresh bile to a broth culture; sodium
taurocholate solution can also be used). Many bacteria can also be
broken up by emulsifying them in 17 per cent, salt solution and
allowing them to stand for some time in this medium and then
diluting this to 0.85 per cent.
B. Extraction by Mechanical Methods
One of the most useful methods for obtaining extracts of bacteria
within a relatively short time is that which Besredka 38 has applied
mainly for the preparation of typhoid (endotoxin), 24-hour agar
cultures washed up in very small quantities of physiological salt
solution, killed by heat at 60-65° C. and dried in vacuo. The dried
mass is mixed with a measured quantity of dry salt and the mix-
ture thoroughly triturated in a mortar for a considerable time.
While triturating distilled water is added in small quantities until
the fluid represents a 0.85 per cent, salt solution. This is allowed
to stand for anywhere from a few hours to a week, and the bacteria
are then removed by centrifugalization. This method has been modi-
fied by many observers and gives good results whenever thorough
trituration is practiced. It is also probable that the exposure to the
hypertonic salt solution in the earlier stages of the trituration may aid
considerably in breaking up the bacteria.
Trituration after freezing is a method which has yielded excel-
lent results in the hands of Macfadyen and others. This requires a
rather complicated piece of machinery originally described by Mac-
fadyen and Rowland. The principle of this is one of mechanical
trituration in a steel cylinder which is surrounded by an ice-brine
mixture so that the bacteria and sand may be kept frozen during the
process.
Appleton, 1913, p. 1020) gives its composition as follows: "Antiformin
consists of equal parts of liquor sodae chlorinate of the British Pharmacopoeia
and a 15 per cent, solution of caustic soda. The formula for liquor sodae
chlorinate is as follows:
Sodium carbonate 600
Chlorinated lime 400
Distilled water 4,000"
37Neufeld. Zeitschr. f. Hyg., Vol. 34, 1900.
38 Besredka. Ann. de I'Inst. Past., 19-20, 1905, 190C.
72 INFECTION AND RESISTANCE
Mechanical trituration is also the principle of the production of
the new tuberculins as advised by Koch.
One of the earliest methods of obtaining bacterial substances
by mechanical means was that used by Buchner and Hahn 39 in the
production of their "plasmines." The bacteria were grown in quan-
tity on large agar surfaces, the moist bacterial masses triturated
together with quartz and were then subjected to high pressure in an
especially constructed press spoken of as the "Buchner press."
Mechanical breaking up and extraction of the bacteria also under-
lies in principle the use of the variously constructed shaking ma-
chines. There are many models of such machines on the market, all
of them designed to accomplish prolonged agitation of bacterial emul-
sions. In many cases the apparatus can be placed inside of an incu-
bator and shaking carried on at 37.5° C. The bacteria are suspended
for this purpose in distilled water salt solution, weak alkali, or in
serum, and glass beads or sand may be added to aid in their mechan-
ical injury. Shaking must be continued for 24 hours or more in
order to give good results.
IV. ACTIVE IMMUNIZATION WITH BACTERIAL PRODUCTS ( TOXINS)
As soon as the investigations of Roux and Yersin had shown that
in some diseases, at least, the injury sustained by the infected animal
was largely due to the soluble toxins produced by the bacteria, it
was logical to attempt to immunize animals with such products.
Probably the first attempts in this direction were those made by
Salmon and Smith in hog cholera. The experiments of these writers
have attained much historical importance since they represent the
first purposeful attempt to immunize animals with the products of
bacterial metabolism. In the actual experiment, however, the im-
munization practiced by Salmon and Smith was probably a combina-
tion of immunization by bacterial products and by dead bacteria.
Nevertheless, the thought of immunization with bacterial products
was the underlying one in their experiments. Working with the
hog cholera bacillus which they had recently discovered they im-
munized pigeons in the following way: The bacilli were grown in
broth for two weeks, and the cultures were killed by exposure to 58°
to 60° C. for several hours. One and one-fifth cubic centimeter of
this culture liquid was then injected into pigeons, and after three
such injections the inoculated pigeons withstood, without harm, doses
of the bacilli which rapidly killed untreated animals. Salmon and
Smith 40 stated distinctly in their conclusions that : "Immunity may
39 Buchner and Hahn. Munch, med. Woch., 1897.
40 Salmon and Smith on "A New Method of Producing Immunity from
Contagious Disease," Proc. Biol Soc., Wash., D. C., Ill, 1884, 6, p. 29,
printed Feb. 22, 1886.
ACQUIRED IMMUNITY 73
be produced by introducing into the animal body such chemical
products (results of bacterial growth in culture fluids) that have been
produced in the laboratory.41
Similar attempts to immunize rabbits against certain forms of
septicemia by the injection of culture filtrates were made by Cham-
berland and Roux 42 in 1888,43 and the same investigators applied
this method to anthrax immunization just prior to the discovery of
diphtheria toxin by Roux. However, neither in hog 44 cholera 45
nor in the other infections upon which this method was first tried do
the bacteria produce a true soluble toxin, and the immunization
which was accomplished depended probably upon the injection of
bacterial extracts. Nevertheless, these attempts had shown the way
in a new direction, and bore immediate fruit in the investigations of
Brieger and Fraenkel,46 and more especially in those of Beh-
ring 47 48 49 and his collaborators. Fraenkel, though following the
method of injecting filtered diphtheria culture fluids, came to the
erroneous conclusion that the toxin and the immunizing substances in
the cultures were not identical (loc. cit., p. 1135).50 The degree of
immunity obtained in his experiments, moreover, was a slight one
only.
Behring, in his first work, in collaboration with Kitasato, suc-
ceeded in immunizing animals with culture filtrates and with pleural
exudates of diphtheritic animals. Similar results were accomplished
with tetanus. Since the publication of these results — especially in
consequence of the epoch-making discovery of passive immunization,
to which they were the immediate guides, toxin immunization has
been investigated and accomplished in all cases in which a true solu-
ble toxin can be demonstrated. It has accordingly been carried
out with the exotoxins of pyocyaneus 51 bacilli, the bacilli of
symptomatic anthrax 52 and botulinus,53 the specific leukocidins 54
41 For a copy of the original paper 'iy Salmon and Smith I am indebted
to Professor Theobald Smith.
42 Chamberland and Roux. Ann. Past., Vol. 1, 1888.
43 Op. Cit., Vol. 2, 1889.
44Joest in "Kolle u. Wassermann ^andbuch, etc.," Vol. 3, p. 632.
45 Karlinski. Zeitschr. f. Hyg., Vol. 28, 1898.
46 Brieger and Fraenkel. Berl. kl. Woch., 1890, Nos. 11, 12 and 49.
47 Behring and Kitasato. Deutsche med. Woch., No. 49, 1890.
48 Behring. Deutsche med. Woch., No. 50, 1890.
49 Behring and Wernicke. Zeitschr. f. Hyg., 1892.
50 De Schweinitz, indeed, who further studied hog cholera immunization
(Medic. News, 1892; Centralbl. f. Bakt., Vol. 20, 1896), claimed that in
sterilized milk the bacillus produced "enzymes" with which immunization
could be accomplished.
51 Wassermann. Zeitschr. f. Hyq., Vol. 22.
52Kempner. Zeitschr. f. Hyg., Vol. 23, 1897.
53 Grassberger and Shattenfroh. Deuticke, Wien, 1904.
54 Denys and Van der Velde. "La Cellule," Vol. 2, 1895.
74 INFECTION AND RESISTANCE
produced by staphylococci, and with various bacterial hemolytic poi-
sons (tetanolysin and other bacterial hemotoxins). The result of all
this work has been the very important determination that susceptible
animals may be actively immunized both against the effects of the
toxin alone, as well as against the virulent bacteria themselves, by
systematic treatment with culture filtrates containing the toxins.
Since in many cases the effects of the toxins were so powerful that
their attenuation was desirable, Behring and others have advised the
addition of iodinterchlorid and other chemicals to the first in-
jections.
PASSIVE IMMUNIZATION
In the logical development of the fundamental facts regarding
immunization, with attention focused early on the blood and body
fluids as the probable carriers of immunity, it was but a rational
step from active immunization to the conception that such acquired
immunity might be transferred from a treated to a normal animal
by injecting blood from the former into the latter. This was prob-
ably the underlying thought of Toussaint's 55 early work with an-
thrax, in which he heated anthrax blood to 55° C. and injected it
into other animals, wrongly believing that the bacteria had been
killed by the heating. The method of Toussaint, however, was vague
in its conception, and in no way constitutes an example of true
passive immunization. The beginning was made in a purposeful
and clearly conceived way by Richet and Hericourt.56
These investigators actively immunized dogs against stapkyio-
cocci, and then attempted to transfer the immunity to normal rabbits
by injecting defibrinated blood from the immune dogs. Their suc-
cess was a partial one only, for reasons that we will discuss directly.
Reasoning similar to that of Richet and Hericourt was applied by
Babes and Lepp 57 to rabies immunization. When the blood of
rabies-immune dogs was injected into normal dogs and rabbits, and
these inoculated with rabies several days later, the treated animals
regularly survived the controls, but in one dog only was the occur-
rence of rabies absolutely prevented. Since their animals were nof
experimentally inoculated, but subjected to the more uncertain
method of allowing them to be bitten by a mad dog, and since the
series included 4 animals only (2 treated and 2 controls), Babes and
Lepp were unable to draw definite conclusions. The establishment of
passive immunization as a proved scientific fact was finally accom-
55 Toussaint. Compt. rend, de I'acad. des sc., 1880.
56 Richet and Hericourt. Compt. rend, de I'acad. des sc., 1888, Vol. 107,
p. 750.
67 Babes and Lepp. Ann. Past., Vol. 3, 1889.
ACQUIRED IMMUNITY 75
plished in 1890-1892 by Behring and Kitasato,58 and by Bebring
and Wernicke. Tbe results of this work — tbe direct outcome of
their success in actively immunizing with soluble toxins, is sum-
marized in their first paper as follows: "The blood of tetanus-
immune rabbits possesses tetanus-poison-destroying properties ; these
properties are demonstrable in the extravascular blood and in the
serum obtained from this ; these properties are of so lasting a nature
that they remain active in the bodies of other animals, so that one
is enabled to obtain positive therapeutic results by transfusing the
blood or injecting the serum. These tetanus-poison-destroying prop-
erties are absent from the blood of non-immune animals, and when
the tetanus poison is inoculated into normal animals it can be dem-
onstrated as such in the blood and other fluids of these animals after
death."
With these researches begins the therapeutically practicable
method of passive immunization which is now in such widespread
and successful use in the treatment of diphtheria, in the prophylactic
treatment of tetanus, and, to a less common and less successful de-
gree, in the treatment of dysentery, typhoid fever (Besredka),
plague, and a number of other bacterial diseases. The same method
has been successful in the treatment of various diseases of domestic
animals. The principle was also applied by Ehrlich 59 to ricin and
crotin immunity, in the formulation of which he succeeded in work-
ing out passive immunization on a quantitative basis, showing that
the degree of immunity in such cases could be directly referred to
the amounts of the specific antitoxin present in the blood of the im-
munized animal. Calmette,60 and Physalix and Bertrand,61 then
succeeded in producing passive immunization against snake venoms.
To summarize the success of passive immunization in general we
may say that it has achieved its greatest usefulness in the case of
those diseases in which the pathogenesis depends upon a true exo-
toxin — which, as we have mentioned before, leads to the formation
of an antitoxin in the immunized animal. In these cases the passive
immunization is accomplished by the transfer of the antitoxins from
the treated to the normal animal.
In the case of bacterial infections in which no true toxin is
formed — where no antitoxin results and the immunity depends, as
we shall see, upon an enhancement of the bactericidal and phagocytic
properties of the blood and the cells, passive immunization has not
been a practical therapeutic success. The probable reasons for this
cannot be properly discussed until we have examined more closely
into the mechanism by which the immune animal is protected after
58 Behring and Kitasato. Deutsche med. Woch., No. 49, 1890.
59 Ehrlich. Deutsche med. Woch, 1891; Fortschr. d. Med., p. 41, 1897.
60 Calmette. Compt. rend, de la soc. de biol., 1894.
61 Physalix and Bertrand. Compt. rend, de la soc. de biol., 1894.
76 INFECTION AND RESISTANCE
specific treatment with bacteria or their products. The moderately
beneficial effects of the various antiplague sera and the limited suc-
cess attending the use of antistaphylococcus, antistreptococcus, and
antipneumococcus sera probably depend, as recent work tends to
show, not upon the direct action of antitoxic bodies, but rather upon
the indirect opsonic action 62 1(i3 64 65 which renders the bacteria
more easily amenable to phagocytic action. These points we shall
discuss at greater length in a succeeding chapter.
SPECIFICITY
In speaking of methods of immunization in the preceding sec-
tions we have frequently employed the terms ''specific" and ''spe-
cificity," without sufficiently defining them. It will be necessary to
explain them since the principle of specificity is at the same time one
of the most important and one of the most mysterious of the phe-
nomena of immunity. When an individual has recovered from an
attack of typhoid he is thereafter immune to typhoid — but to no
other disease — similarly with plague, cholera, small-pox, etc. The
same principle governs artificial immunization. Vaccination against
anthrax protects against anthrax only — and active or passive im-
munization in any of the infectious diseases produces a resistance
which is "specifically" aimed only at the particular infectious agent
writh which the original active immunity was produced. The prin-
ciple points to an exquisite chemical difference between the protein
substances which constitute the bacterial cell bodies or their meta-
bolic products. For although by chemical methods we can detect no
differences between them — yet the reactions of immunity are sharply
differentiating. When we come to consider the antibodies which
specifically precipitate the substances by which they are incited we
shall see that the delicacy and consequent differential value of these
reactions far outstrip any known chemical methods, and it is upon
this principle indeed — inexplicable as it is — that the great diagnostic
value which these reactions have attained depends. The conception
of the specificity of the causes of infectious disease, as well as that
of the specificity of toxins, has become so common and self-evident
to us that we are too apt to forget how fundamental to progress the
establishment of this fact was in the early days of bacteriological
research. When, in 1878, Koch published his treatise on the "Eti-
ology of Wound Infections" specificity was not generally accepted,
and the supposed metamorphosis of bacterial species, as asserted by
62Neufeld. Deutsche med. Woch., No. 11, 1897.
63 Neufeld and Rimpau. Deutsche med. Woch., No. 40, 1904.
64 Bail and Kleinhans. Zeitschr. f. Imm., Vol. 12, 1912.
65 Weil. Zeitschr. f. Hyg., 75, 1913.
ACQUIRED IMMUNITY 77
Hallier and others,66 had first to be scientifically refuted by Cohn,
Koch, and their pupils, before it could be assumed that a given in-
fectious disease was always the result of infection with a definite
and constant species of bacteria. The same applied to the specificity
of toxins — and rational investigations into the reaction of the animal
body against bacterial poisons was not possible until the works of
Roux and Yersin on diphtheria and that of Kitasato on tetanus had
differentiated between the true, specific bacterial poisons and the
unspecific ptomains and "sepsins" of Selmi, Nencki, and Brieger.
66 Hallier. Cited from Behring, "Bekampfung der Infektionskrank-
heiten," Leipzig, 1894.
CHAPTER IV
THE MECHANISM OF NATURAL IMMUNITY AND THE
PHENOMENA FOLLOWING UPON ACTIVE
IMMUNIZATION
ANTIBODIES AND ANTIGENS. THE ORIGIN OF ANTIBODIES
THE MECHANISM OF NATURAL IMMUNITY
PASTEUK'S work on active immunization was carried out in the
later seventies and the early eighties. During and immediately after
this time it was very natural that the attention of investigators
should have concentrated upon the elucidation of the causes under-
lying both the natural resistance against bacteria observed in animals
and man, and the changes which during active immunization were
fundamentally responsible for the acquisition of resistance.
It was easily determined that there were no anatomically and
physiologically determinable differences between the various mam-
malia which could account for the observed striking variations of
susceptibility, nor could gross anatomical or histological changes
be noted in an animal which had been artificially immunized. Mor-
phologically such an animal was indistinguishable both in the size
and appearance of its organs, and in the arrangement and structure
of its cells from any other individual of the same species not sub-
jected to treatment.
It was a natural development of the investigations brought to
bear upon this problem that attention should, for a time, be concen-
trated upon the phenomena of inflammation, processes which were
regularly associated with infections of all kinds and seemed indeed
to represent a sort of local expression of tissue resistance to the
invading micro-organisms.
It was in the course of investigations upon the nature of inflam-
mation that Metchnikoff first became interested in problems of re-
sistance. In 1883 he presented a paper at the Naturalists' Congress
at Odessa, in which he referred the absorption of dead or foreign
corpuscular elements in the bodies of invertebrates to a process of
intracellular digestion carried out by phagocytic cells. As early as
1874 Panum had suggested that possibly resistance against invading
micro-organisms might be due to a similar intracellular destruction,
78
PHENOMENA FOLLOWING IMMUNIZATION 79
and Metchnikoff, soon after his first communication, extended his
phagocytic studies to phenomena of infection. His first investiga-
tions concerned themselves with an infectious disease caused by a
form of yeast in a small crustacean — the daphnia or water flea. ' He
showed that recovery or death from the disease depended upon the
completeness with which the invading micro-organisms were taken
up by the white blood cells found in the body cavity of the daphnia.
Immediately subsequent studies, carried out with the aid of numer-
ous pupils, embraced an extensive material throughout the animal
kingdom in which he attempted to show parallelism between natural
immunity and the phagocytic activities mobilized by the body against
the invading germs.
Meanwhile studies along another path were in progress. It had
been observed many years before this by the physician, Hunter, that
the shed blood of animals was not as easily subject to putrefactive
change as wTere many other organic substances. Similar observations
by Traube and, in 1881, by Lord Lister a (the latter reported at a
time when Pasteur's experiments were reaping their first practical re-
sults) further stimulated investigation of the blood as the possible
seat of the increased antibacterial property. For, indeed, these
observations seemed to imply that by resisting decomposition, even
when inoculated with putrefying material, the blood must possess
definite means of inhibiting or even destroying the putrefactive
bacteria.
In 1884, in a dissertation submitted at Dorpat, Grohman 2 stated
that cell-free blood plasma inhibited the growth of micro-organisms.
But Grohman was unable to determine actual bacterial destruction.
Similar, but inconclusive, observations were published by Von Fo-
dor 3 in 1887. In 1888, however, Nuttall,4 who was investigating
the validity of the phagocytic theory of Metchnikoff, experimentally
determined that normal blood possessed the property of killing bac-
teria— a property now spoken of as "bactericidal" power. The atti-
tude taken by Nuttall, and others of the Fliigge school, toward
Metchnikoff' s opinions was one of doubt as to the fundamental sig-
nificance of phagocytosis in determining resistance. They argued
that Metchnikoff had not yet proved that living bacteria were taken
up by the phagocytic cell, and that the action of these cells might
therefore be interpreted as merely a process of removal of the dead
bacteria, after these had been killed by other influences. USTuttall,
accordingly, repeated some of MetchnikofFs experiments on anthrax
in frogs and rabbits, essentially confirmed the basic observations,
but showed also that the cell-free defibrinated blood of these and
1 Lister. Trans. Intern. Med. Congress, London, 1881.
2 Grohman. Cited from Lubarsch, Centralbl. /. Bakt., 6, 1889.
3 Fodor. Deutsche med. Woch., No. 34, 1887.
* Nuttall. Zeitschr. f. Hyg., Vol. 4, 1888.
80 INFECTION AND RESISTANCE
other animals possessed definite bacteria-destroying properties (bac-
tericidal power) for many different micro-organisms. He detected
similar properties in pleural exudates, pericardial fluids, and aqueous
humor, and determined that this property was "inactivated" or de-
stroyed when the fluids were heated to 55° C. for 10 minutes or
longer. Buchner 5 then confirmed NuttalPs results and showed
further that the bactericidal property resided, not only in defibri-
nated blood, peptone blood, and plasma, but was present also in the
serum obtained after clotting. He applied the term "Alexin" to this
active constituent of the blood — likening its action to that of a
ferment.
The immediate theoretical result of these discoveries was an at-
tempt, begun by Fliigge's school, to base natural as well as acquired
resistance upon the bactericidal properties of the blood and body
fluids in general. For the observations of Nuttall and Buchner were
soon extended to peritoneal and other exudates by Stern,6 and to
ascitic fluids by Prudden.7 By these two groups, that of Fliigge-
Nuttall-Buchner on the one hand, and that of Metchnikoff on the
other, there were founded the two schools of immunity — the humoral
and the cellular, both originating in attempts to explain natural
immunity, and later extending to problems of acquired resistance.
And it is to the diligent and ingenious intellectual and experimental
conflict between these schools that we owe much of the knowledge we
now possess concerning the phenomena of immunity. A bridge be-
tween them was early established when Buchner himself — (even
before Metchnikoff) — suggested the possible leukocytic origin of the
bactericidal serum constituent (alexin). The later work of Denys,
of Gruber and Futaki, of Wright, of Neufeld, of Bail, and of others
has demonstrated, as was to be expected, the inadequacy of either
point of view by itself, and the intimate interdependence of the
humoral and the cellular processes.
As concerns the relation of bactericidal serum effects and natural
immunity, it could be unquestionably shown by Nuttall, Buchner,
Kissen,8 and their immediate followers that the blood of most ani-
mals possessed bactericidal properties against many micro-organisms,
their experiments being so planned that the participation of leuko-
cytes could be absolutely excluded. However, a parallelism between
bactericidal power and the degree of natural resistance could not
be established. Lubarsch,9 writing during the early periods of the
controversy, stated that "he would regard the (purely humoral) 10
5 Buchner. Centralbl f.'Bakt., 1889.
6 Stern. Zeitschr. f. klin. Med., Vol. 18 (cited from Hahn).
7 Pnidden. Med. Eec., Jan., 1890.
8Nissen. Zeitschr. /. Hyg., Vol. 6.
9 Lubarsch. Centralbl. f. Bakt., Vol. 6, 1889.
*° Bracketed phrase our own.
PHENOMENA FOLLOWING IMMUNIZATION 81
experiments of Nuttall as decisively contradicting the phagocytic
theory if the bactericidal action of the blood (for anthrax bacilli)
could be shown to be more potent in immune than in suscep-
tible animals." Metchnikoff 1 1 himself, taking this point of view,
called attention to the fact that the blood serum of rabbits, animals
that are highly susceptible to anthrax, is more powerfully bactericidal
for these micro-organisms than is the blood of dogs or even that of
immunized calves, both of which are much more resistant than are
rabbits. Nuttall answered this by reporting that the blood of an-
thrax-immunized calves is actually more powerfully bactericidal than
is that of normal calves. Although this argument of JSTuttall was
perfectly valid in principle, it exerted little influence on opinions
at this time because anthrax happens as a matter of fact to
belong to that group of infections in which bactericidal protection
is actually secondary to phagocytic, and Lubarsch could show that
the differences observed by Nuttall were often less than those
obtaining between specimens of blood taken from individual normal
rabbits.
Lubarsch himself, then, in carefully planned experiments,
showed that rabbits and cats could be killed with quantities of anthrax
bacilli far less than the number which the extravascular blood of
these animals can destroy. He concluded that the resistance, in these
cases at least, is certainly not parallel writh the bactericidal properties
of the blood, and suggested the possibility that the intravascular blood
does not possess bactericidal power to the same degree in which it is
possessed by the extravascular plasma or serum. This point, first
raised by Lubarsch — namely, the possibility of a difference between
the intravascular blood and the extravascular blood serum or plasma
in bactericidal functions — soon became one of the focal points of
the controversy, since Metchnikoff, admitting the bactericidal power
of the shed blood, assumed that this was purely the result of sub-
stances given off by the leukocytic cell-bodies after extravascular
injury.
The Metchnikoff school defended its premise by the dual method
of attempting on the one hand to establish a parallelism between
phagocytic activity and natural resistance, and, on the other hand,
by showing that the cell-free blood serum of naturally resistant ani-
mals often furnished an excellent culture medium for the bacteria
in question. Thus Wagner showed that anthrax bacilli grow well
in the blood of fowls at 42° C., and Metchnikoff himself called at-
tention to the fact that pigeons' blood is an excellent medium for the
cultivation of the Pfeiffer bacillus, whereas the living pigeon is en-
tirely insusceptible to influenzal infection. Arguments based on
such observations, however, have lost much of their original weight,
for we have since then learned more about the delicate quantitative
11 Metchnikoff. Virch. Archiv, Vol. 97, 1884.
82 INFECTION AND RESISTANCE
conditions and the difficulties of accurate measurements obtaining in
experiments upon in vitro bactericidal phenomena. For although a
specimen of the blood of a naturally immune animal may be capable
of destroying a considerable number of bacteria of a given species, the
implantation of such a specimen with a slight excess of the bacteria
would soon exhaust the active serum constituents and profuse growth
could then take place. Furthermore, the conditions of temperature
established in cultural experiments lead rapidly to a deterioration
of the alexin necessary for bactericidal action, and any bacteria
remaining alive at the end of a number of hours would then have un-
opposed opportunity to multiply.
The attempts to establish parallelism between phagocytic activity
and natural immunity, though somewhat more successful than the
analogous efforts of the humoral school, nevertheless also failed to
furnish complete explanation for existing conditions, and, as we shall
see, no adequate generalizations could be made until later years re-
vealed the close cooperation between cells and fluids. We must post-
pone any attempts to do justice to this phase of the problem, there-
fore, until we are in a position to discuss the question of phagocytosis
on the basis of a fuller knowledge of the phenomena which influence
it.
The clear thinking and unprejudiced logic brought to bear upon
this controversy by some of the great bacteriologists of this time are
nowhere more instructively illustrated than in a short introduction
published by v. Behring 12 to his second article on diphtheria. He
says: "Neither deduction nor theorizing can at present decide
whether a compromise will be found in the future between the two
hypotheses (humoral and cellular), or whether the one or the other
alone will be found correct. As yet the opinions of many experiment-
ing bacteriologists are in direct opposition in this respect. Mean-
while, for the purposes of medical advancement and therapeutic suc-
cess it is not necessary to await a decision of this question. ... It is
indeed of advantage to the cause if the struggle against infection is
undertaken from the most varied points of view; attempts to make
proselytes for a dogma have never led to progress. In this sense I will
try to summarize those experimental results which support the hu-
moral point of view without attempting particularly to detract from
the importance of opinions which I do not share."
THE PHENOMENA FOLLOWING UPON ACTIVE IMMUNIZATION
The cellular and humoral points of view, formulated largely upon
the facts of natural immunity, were equally applied, almost from
the beginning, to the explanation of active immunization. The light
12 v. Behring. Zeitschr. f. Hyg., Vol. 12, 1892,
PHENOMENA FOLLOWING IMMUNIZATION 83
thrown upon these phenomena by the efforts of both schools rapidly
led to a complete abandonment of those earlier theories of immunity
which had conceived the acquired resistance of animals against bac-
teria as a purely passive development in the body of conditions
unfavorable for bacterial gro /vth.
Among these earlier theories, now of historical interest only, are
the "Exhaustion Theory" of Pasteur and the "Retention Theory" of
Nencki,13 Chauveau, and others.
Pasteur's views, defended for a time also by Garre, held that the
growth of any given variety of bacteria in the animal body exhausted
certain specific nutritive substances necessary for this growth. Sub-
sequent lodgment in the same body was impossible owing to the
absence of proper nutrient material. It is interesting to note, as
Kolle 14 points out, that this theory is in principle very similar to
the "Atrepsie" idea of Ehrlich advanced in explanation of species
immunity to cancer.
The hypotheses of Chauveau, of Nencki, and others were the
converse of those of Pasteur. They were based purely on inference,
assuming that conditions occurring in the test tube could be applied
also to those existing in the animal body. Baumann 15 had shown
that, among other things, phenol was produced as a result of bacterial
putrefaction. Nencki had noticed the inhibition of bacteria in
culture by the products of their own metabolism. Wernicke,16 too,
had demonstrated the presence of phenol, phenylacetate, skatol, and
other aromatic compounds harmful to bacteria in putrefying mix-
tures. The reasoning which formulated the so-called "Retention The-
ory," therefore, was the following : Bacteria growing in the animal
body produce certain substances peculiar to their own metabolism,
which eventually lead to inhibition of their growth. By the retention
of these products the animal is rendered immune. Chauveau's adher-
ence to this theory was largely based on the fact that he had observed
immunity in the lambs born of Algerian ewes which had recovered
from anthrax shortly before or during parturition. He explained
this by a transference of the retention products from mother to off-
spring. As a matter of fact the observation could just as well have
been utilized as support for the Exhaustion Theory.
Both the theory of "Exhaustion" as well as that of "Retention"
could not long withstand experimental criticism. .Theories which
were not so easily disproved and which have given rise to much in-
vestigation are the "Alkalinity Theory," first formulated by v. Beh-
13 Nencki. Jour. f. prakt. Chem., May, 1879, cited from Sirotinin,
Zeitschr. f. Ilyg., Vol. 4, 1888.
14 Kolle in "Kolle u. Wassermann Handbuch," 2d Ed., Vol. 1.
15 Baumann. Zeitschr. f. physiol. Chem., Vol. 1.
10 Wernicke. Virch. Archiv, Vol. 78.
84 INFECTION AND RESISTANCE
ring,17 and the "Osmotic Theory" of Baumgarten.18 In the iormer
an attempt was made to demonstrate a parallelism between blood alka-
linity and bactericidal action — the latter was based on the supposi-
tion that the destruction of bacteria in the body was largely due to
harmful osmotic conditions. Neither of these theories was long
seriously maintained. Behring himself took an active part in the
subsequent development of our present views. Baumgarten 10 still
clings to his own opinion in a modified way, in that he maintains
that the only effect produced by specific antibodies upon cells — bac-
terial or otherwise — is that they change the permeability of the cell
membranes and render them more vulnerable to osmotic injury.
However crude or vague these theories may seem to us now, it
must not be forgotten that they were conceived at a time when no
knowledge had been gained regarding specific "antibodies." The
phagocytic powers to which Metchnikoff attributed natural immunity
and the bactericidal powers of the blood, regarded in the same light
by the Fliigge school, were general properties possessed by many
animals toward many different micro-organisms. That immuniza-
tion could specifically increase these functions toward the particular
micro-organisms used for treatment seemed indicated by the experi-
ments of Nuttall in which higher bactericidal power was found in
the blood of anthrax-immune calves than in that of normal animals.
However, no definite and conclusive work on the specific increase of
measurable serum or cell properties was available.
This great advance, giving new energy and pointing out new
paths of investigation, carne in 1890-1892 with the publication of the
work of Behring and his collaborators, Kitasato and Wernicke, on
immunity to diphtheria and tetanus. As we have indicated in a pre-
ceding paragraph, the fundamentally important points of this work
were as follows :
1. The establishment of the fact that animals may be actively
immunized with products of bacterial metabolism — true toxins or
exotoxins.
2. The discovery that such active immunity was dependent upon
specific antibodies formed in the treated animal and circulating
freely in the blood ; and,
3. That, by the transfer of the blood or the blood serum contain-
ing these specific antibodies other normal animals could be passively
protected — not prophylactically only, but even after active disease
had set in.
These observations were rapidly confirmed for tetanus by Tiz-
zoni and Cattani, and by Yaillard, and, similar but less successful
attempts at passive immunization were made in other diseases by
17 v. Behring. Centralbl f. Jclin. Med., 1888, No. 38.
18 Baumgarten. Berl. klin. Woch., 1899, 1900.
19 Baumgarten. ILehrbuch, etc., 1912.
PHENOMENA FOLLOWING IMMUNIZATION 85
Foa, Emmerich, Bouchard, and many others. The discovery of
passive immunization established the fact of specific alteration of
the blood by active immunization, and represented, for the time,
a distinct triumph for the humoral hypothesis.
Summarizing the knowledge of immunity as it stood at the close
of this period, Behring says : "In the case of natural immunity no
generally applicable explanation has as yet been found. (By this he
referred to the lack of complete parallelism between natural im-
munity and either the bactericidal or the phagocytic activities.) For
artificial immunization, however, it has now been shown, in a number
of carefully studied infections, that we can surely attribute it to
properties of the cell-free blood."
Within a very short time after Behring and Kitasato's first paper
Ehrlich 20 demonstrated that the principle discovered by them was
not limited to bacterial poisons. He was investigating immuniza-
tion against ricin in mice, and showed that here, too, the blood of
the immune animals contained a body which would antagonize the
toxic action of ricin, and which, injected into normal mice, would
passively protect them. He spoke of this blood constituent as "anti-
ricin."
It is natural that extensive generalization followed these discov-
eries. However, while it was found that the blood of all actively
immunized animals possessed a certain degree of protective power
for normal individuals, it was soon shown that this was not due in
all cases to antagonism to the bacterial poisons on the part of the
immune blood serum. In immunity to the Vibrio Metchnikovi — in
pneumococcus and cholera immunity — Sanarelli,21 Isaeff,22 Pfeiffer
and Wassermann,23 and a number of others showed that here, unlike
diphtheria and tetanus, the protective power of the immune serum
did not rest on "antitoxic" properties, but rather on antagonism to
the bacteria themselves. It soon became definitely established that
antitoxic immunity resulted only in the cases of those bacteria in
which a true soluble exotoxin was produced, and where the disease
following infection was primarily due to the absorption of these
poisons. The antibodies incited in the blood of toxin-immune ani-
mals were therefore spoken of by Behring and Ehrlich as "anti-
toxins" and their action — after a number of false hypotheses — was
finally recognized as a direct neutralization of the bacterial poisons.
The strict specificity of these antibodies was, from the first, clear
to v. Behring, who observed that diphtheria-immune serum and
tetanus-immune serum acted each upon its respective toxin only. It
was recognized at the same time that the passive immunity produced
20 Ehrlich. Deutsche med. Woch., No. 32, 1891.
21 Sanarelli. Ann. Past., Vol. 7, 1893.
22 Isaeff. Ibid, and Zeitschr. f. Hyg., Vol. 16, 1894.
23 Pfeiffer and Wassermann. Zeitschr. f. Hyg., Vol. 14. 1893.
86 INFECTION AND RESISTANCE
by injecting antitoxic sera is almost immediately established ; that, by
proportionately increasing the amount of antitoxin, immunity can
be produced against any amount of toxin; and that this passive or
transferred immunity is of relatively short duration.
The antitoxins, then, as we shall see in the more detailed analysis
of their action (in chapter V), are specific poison-neutralizing anti-
bodies formed in the blood of animals immunized with a true bac-
terial toxin or exotoxin — conferring resistance or immunity, not by
influencing the bacteria, but by rendering innocuous the specific bac-
terial poisons.
The therapeutic successes of passive immunization achieved with
tetanus and diphtheria very naturally led to a careful inquiry into
the antitoxic properties of the blood of animals immunized with all
known pathogenic bacteria and bacterial products, and with many
poisons of animal and vegetable origin.
Contrary to earlier expectations, however, the list of bacteria
against which antitoxic immunity can be achieved has remained rela-
tively small, limited in fact, as we have previously stated, to those
species which produce a soluble exotoxin. The inciting of a specific
neutralizing antibody (antitoxin), however, is also a property of
many other substances of proteid nature which are for this reason
classified biologically with the true toxins or exotoxins. In fact,
the one absolutely constant attribute which defines our conception
of the "true toxins" and the substances classified with them is their
antitoxin-inciting power. We classify a bacterial product as a
"toxin" or "exotoxin" only if it incites a neutralizing "antitoxin"
in the serum of an immunized animal.
The first discovery of a non-bacterial antitoxin-stimulating sub-
stance was, as we have stated, that of ricin by Ehrlich,24 1891, and
this was soon followed by similar determinations for abrin and robin
— other vegetable poisons. In 1894 Calmette,25 and Physalix and
IBertrand 26i extended the principle to poisons of animal origin by
demonstrating antitoxin formation against snake poison. And that
similar specific neutralizing bodies were formed in response to im-
munization with ferments was shown in 1900 by Morgenroth.27
The more important individual substances which may be bio-
logically grouped together because of their property of inciting a
specific antitoxin (or toxin-neutralizing body) in the blood of im-
munized animals may be tabulated as follows :
Diphtheria toxin — (loc. cit. Behring & Wernicke).
Tetanus toxin — (loc. cit. Behring & Kitasato).
24 Ehrlich. Deutsche med. Woch., 1891; Fortschr. d. Med., 1891, 1897.
25 Calmette. Ann. Past., Vol. 8, 1894.
28 Physalix and Bertrand. Compt. rend, de la soc. de biol., 1894.
27 Morgenroth. Centralbl. f. Bakt., 26, 1899.
PHENOMENA FOLLOWING IMMUNIZATION 87
The Toxin of the Bacillus of Symptomatic Anthrax — (Grassberger &
Shattenfroh, Munch. Med. Woch., 1900, 1901 and 10 c. cit.).
The Toxin of the Bacillus Botulinus — (Kempner, Zeitschr. f. Hyg.,VoL 26,
1897).
The Toxin of the Bacillus Pyocyaneus — (Wassermann, Zeitschr. f. Hyg.,
Vol. 22, 1896).
The Toxin of the Dysentery Bacillus (?) Shiga-Kruse type — (Kraus u.
Doerr, Wien. klin. Woch., 1905).
The leukocyte poison of the Staphylococcus pyogenes aureus, Leucocidin —
(Denys & Van de Velde, La cellule, 1895).
The Hemolytic Poisons of Various Bacteria (see Pribram in "Kraus und
Levaditi Handbuch," Vol. II, p. 223).
Proteolytic Ferments of the Hog Cholera Bacillus (De Schweintz, Medical
News, 1892).
The Toxin of the Cholera Spirillum (?) Brau& Denier, Compt. rend, de Vacad.
des sc., 1906, Kraus, Centralbl. f. Bakt., 1906, and Wien. klin. Woch., 1906).
Ricin — (Ehrlich, loc. cit.).
Abrin — (Ehrlich, loc. cit.).
Krotin — (Ehrlich, loc. cit.).
Snake venom — (Calmette, loc. cit.).
Spider poison— (Sachs, " Hoffmeister's Beitrage," 1902, and Ehrlich, "Ge-
sammelte Arbeiten," etc.).
Lab. enzyme — (Morgenroth, loc. cit.).
Pepsin— (Sachs, Fortschr. d. Med., 1902).
Trypsin— (Achalme, Ann. Past., 1901).
Leukocytic ferments Leukoprotease — (Jochmann & Miiller, Munch, med.
Woch., 1906). »
The period of investigation which was initiated by the discovery
of the specific antitoxins was replete with efforts to determine true
toxins and, consequently, antitoxic immunity for all pathogenic
bacteria. We have already mentioned that in many cases these
efforts were futile — the bacteria in question being found to secrete
no exotoxin and the immunity established against them developing
without the formation of demonstrable antitoxin. Metchnikoff 29
showed this to be the case with hog cholera as early as 1892, and the
investigations of Sanarelli, IsaefT, and Pfeiffer and Wassermann
pointed in the same direction.
Perhaps the clearest definition of the conditions prevailing dur-
ing immunization of animals with non-toxin-forming bacteria was
that formulated at this time by Pfeiffer. The importance of the bac-
tericidal power of serum, as discussed before this by Fliigge, Nut-
tall, and others, had dealt largely with variations of this general
property in relation to natural immunity, but had failed to recognize
clearly a specific increase in these powers during active immuniza-
28 This list includes all the important antitoxin-inciting substances. For
a more complete tabulation see Wassermann in "Kolle u. Wassermann Hand-
buch, etc.," Vol. IV, 1st Ed., p. 498. Our own list is adapted from the one
there ^iven.
29 Metchnikoff. Ann. Past., 1892.
88 INFECTION AND RESISTANCE
tion. Pfeiffer with Wassermann 30 had studied the pathogenicity of
cholera spirilla for guinea pigs, and had come to the conclusion that
the animals died of toxemia (and not of bacteriemia, as claimed by
Gruber and Wiener), and that this toxemia was due to the liberation
of poisons from the dead bodies of cholera vibrios, killed by the
serum of the infected animals. Pfeiffer 31 now showed that the in
jection of cholera spirilla killed with chloroform brought about a tox-
emia identical with that following inoculation with living cultures.
He further determined that the resistance of animals against cholera
was due to the bactericidal effects of the serum, which killed the
injected cholera spirilla, and not to any poison-neutralizing property.
Isaeff,32 one of Pfeiffer's pupils, continuing this work, expresses
his own and Pfeiffer's conceptions as follows: "Guinea pigs vac-
cinated against cholera, in spite of high immunity to infection with
living spirilla, do not develop any immunity to cholera [endo]33
toxins. The blood of immunized guinea pigs possesses no antitoxic
properties. The maximal dose of cholera 'toxin' which immunized
guinea pigs can withstand is not higher than that which can be borne
by normal animals, and but slightly higher than the maximal dose
of living spirilla, which they can survive. The blood of cholera-vac-
cinated guinea pigs possesses strong specific protective powers. The
same specific immunizing properties are demonstrable in the blood
of cholera convalescents toward the end of the third week of the
disease."
The path was thus cleared for a definite conception of cholera
immunity, and this was formulated, in their next communication,
by Pfeiffer and Isaeff.34 35 36 In this paper they showed that the
cholera spirilla injected into the peritoneum of a cholera-immune
guinea pig were subjected to a rapid dissolution, a process which
could be observed by taking small quantities of exudate out of the
peritoneum, at varying intervals, with capillary pipettes. ]STo such
dissolution occurred in normal pigs or with normal serum. But the
same rapid swelling, granulation, and, finally, dissolution occurred
when the spirilla were injected into the peritoneal cavity of a nor-
mal guinea pig, together with the serum of an immunized animal.
The process took place apparently without the cooperation of the
leukocytes or other cells, and was absolutely specific. For instance,
no "lysis" occurred when the vibrios "Nordhafen," "Massauah," and
other cholera-like organisms were injected into cholera-immune pigs,
30 Pfeiffer and Wassermann. Zeitschr. f. Hyg., Vol. 14, 1893; also
Pfeiffer, Zeitschr. f. Hyg., Vol. 16, 1894.
31 Gruber and Wiener. Archiv /. Hyg., Vol. 15, 1893.
2 Isaeff. Zeitschr. f. Hyg., Vol. 16, 1894.
83 Bracketed word our own.
34 Pfeiffer and Isaeff. Zeitschr. f. Hyg., Vol. 17, 1894.
35 Pfeiffer. Ibid., Vol. 18, 1894.
36 Pfeiffer and Isaeff. Deutsche med. Woch., No. 13, 1894.
PHENOMENA FOLLOWING IMMUNIZATION 89
but took place regularly when true cholera strains, from - various
sources, were used in the experiment. The immunity of cholera-
treated animals, therefore, was found to be an antibacterial and not
an antitoxic one. Cholera spirilla introduced into a normal animal
were permitted to multiply and accumulate until a sufficient number
were present to furnish, upon cell death, a fatal dose of poison. In
immunized animals the small quantities of bacteria first introduced
succumbed rapidly to the lytic properties of the serum and accumu-
lation was prevented.
By these experiments, now commonly spoken of as the "Pfeiffer
Phenomenon/' it was definitely proved that active immunization
with bacteria incites in the serum of the treated animal a potent in-
crease of bactericidal properties — an increase which is entirely spe-
cific in that the bactericidal power toward bacteria other than those
employed in the immunization does not exceed the normal. The
immunity in these cases, then, is not antitoxic, but rather "antibac-
terial" and depends on the development, in the immune sera, of anti-
bodies quite distinct from the "antitoxins." These immune serum
constituents were spoken of by Pfeiffer as "bacteriolysins" or "spe-
cific bactericidal substances."
Not long after the discovery of the specific bacteriolysins another
property of immune sera was described by Gruber and Durham.37
They had been studying bacteriolytic phenomena with colon and
cholera organisms, and noticed that these bacteria were rapidly ag-
glomerated and gathered in small clumps when emulsified in homolo-
gous immune serum. Similar clumping had indeed been described
before. Metchnikoff, Isaeff, Washburn, and Charrin and Eoger had
described it on various occasions, but had not recognized it as a
specific property of immune serum.38 Gruber and Durham studied
it carefully, determined that it was present to a degree roughly pro-
portionate to the degree of immunization attained, and that its
specificity was such that it could be utilized for bacterial differen-
tiation. They believed that the substances in the immune serum
responsible for this agglutination were independent of other serum
constituents and applied to them the term "agglutinins."
The problems of immunization had now considerably expanded
and the nature of the new serum reactions was assiduously studied.
Primarily the phenomenon of agglutination was regarded as a part
of the struggle of the body against the living bacteria and Gruber
himself believed that it depended upon a swelling or "klebrig wer-
den" of the micro-organisms which tended to cause their sticking
together, and rendered them more readily amenable to the action of
the bactericidal powers of the serum. Bordet,39 however, early con-
37 Gruber and Durham. Munch, med. Woch., 1896.
ss ;por references see chapter on Agglutinins.
39 Bordet. Ann. Past., 1896.
90 INFECTION AND RESISTANCE
ceived the process as a physical phenomenon in which the bacteria
themselves were entirely passive, and, indeed, Widal 40 soon demon-
strated that bacteria killed by heat were equally as agglutinable as
the living germs.
This naturally suggested that the reaction between specific agglu-
tinating serum and bacteria was based on individual peculiarities of
the bacterial proteins, and it occurred to Kraus,41 accordingly, to
investigate whether or not the immune sera would cause any sort of
reaction when mixed with the dissolved body substances of homolo-
gous bacteria. Working at first with cholera and plague, he pre-
pared solutions of bacterial proteins, both by allowing broth cultures
to stand for varying periods and by emulsifying agar cultures in
alkaline broth. The extracts were then filtered through Pukal filters
to remove the bacterial bodies. When the sera of immunized ani-
mals were added to these clear filtrates — cholera serum to cholera
filtrate, and plague serum to plague filtrate, slight turbidity devel-
oped and was followed within twenty-four hours by the formation of
small flakes. In other words, it was found that the mixture of a
clear filtrate of a bacterial culture with the serum of an animal
immunized against these bacteria resulted in the formation of a
precipitate. The reaction was found to be as strictly specific as that
of agglutination.
Although, from the beginning, Paltauf 42 attempted to associate
the phenomena of agglutination and precipitation, the property of
precipitating homologous culture filtrates was attributed by Kraus
and others to specific antibodies in the immune sera, distinct and
independent of those previously described, and spoke of them as
"preciptiins."
The discovery of the various "antibodies" so far discussed re-
sulted from the study of the direct action of blood serum upon bac-
teria and bacterial products. This did not, however, completely de-
flect the attention of investigators from the unquestionable impor-
tance of phagocytosis in the defence of animals against bacterial in-
vasion. Metchnikoff and his school continued diligently to pursue
this other phase of the study of immunity and, although the increas-
ing knowledge of serum antibodies continued to strengthen the prem-
ises of the purely humoral point of view, it had still to be admitted
that in some diseases — particularly anthrax and the pyogenic coccus
infections, phagocytosis must largely be held responsible for recov-
ery. It was found, moreover, by the later investigations of Denys,
Wright, Neufeld, and others that phagocytosis in immunized animals
was far more extensive and efficient than in normal ones, and that
40 Widal. La semaine medicale, No. 5, 1897.
41 Kraus. Wien. klin. Woch., No. 32, 1897.
42 Paltauf. "Discussion of Kraus' Paper," Wien. kl Woch., No. 18, 1897.
p. 431.
PHENOMENA FOLLOWING IMMUNIZATION 91
this depended on specific constituents of the immune serum which
rendered the bacteria more amenable to the phagocytic action of the
cells. These further antibodies we will discuss in a subsequent chap-
ter, under the terms "opsonins" and "bacteriotropins," designations
applied to them by their discoverers.
We have thus reviewed briefly the various specific properties
which develop in the serum of an animal when it is systematically
treated (actively immunized) with bacteria or bacterial products.
These serum activities have been attributed to the development in
the serum of substances which we speak of as "antibodies."
In our discussion of the first of these antibodies, antitoxin, we
call attention to the fact that the principle discovered in the case of
bacterial toxins was rapidly extended to vegetable poisons, snake
venom, spider poison and enzymes. It was found that the power of
inciting antitoxins when injected into animals was an attribute be-
longing to a large group of substances in nature, and not limited to
bacteria alone. A similar generalization of conception has been pos-
sible with other antibodies. Specific lysins, agglutinins, and pre-
cipitins may be produced by the treatment of animals with many
substances not of bacterial nature.
The first observation of this kind was made almost simultaneously
by Bordet 43 and by Belfanti and Carbone.44 They observed that
the serum of an animal that had been treated with the red cells of
another species acquired the power of laking these cells. That the
normal serum of one species is often toxic to, and causes the laking
of, the erythrocytes of another species is an observation that dates
back to the earliest experiments on transfusion, and had been studied
in considerable detail by Landois as early as 1875. The phenom-
enon possesses much interest in its bearing on the problems of ana-
phylaxis and will be discussed more particularly in that connection.
We mention it in this place to show that, like bactericidal bodies,
"hemolytic" (erythrocyte laking) properties may be present in nor-
mal sera, though irregularly and by no means occurring in every
species of animal. Incidentally it may be stated that this is true
also of agglutinins and of opsonins which may be found in consider-
able amounts in normal sera. Of precipitins, however, this does not
seem to be true.
By the work of Bordet it was found that "hemolysins" could be
specifically 45 incited in an animal by systematically treating it with
43 Bordet. Ann. Past., Vol. 12, 1898.
44 Belfanti and Carbone. Giorn. della R. Acad. di Torino, July, 1898.
45 By the use of the word specific in this case we imply that an animal
immunized with any given variety of red blood cells will form hemolysins
for this variety only. Thus an animal treated with ox blood will form ox
blood hemolysins only, and his serum, though strongly hemolytic for ox
blood, will not lake sheep cells, dog cells, human cells, etc.
92 INFECTION AND RESISTANCE
the red blood cells of another species. Apart from the great interest
attaching to this discovery in itself, it has had a very profound influ-
ence upon investigations on immunity generally, since it has fur-
nished a method of studying lysis far more simple and easily con-
trolled than is the analogous phenomenon of bacteriolysis. And
since, in fundamental principles, bacteriolysis and hemolysis are
essentially alike, much of our knowledge regarding the former has
been arrived at by experiments upon the latter. The specific hemo-
lysins, then, are antibodies formed in response to "immunization"
with red blood cells, analogous to the similarly produced "bacterio-
lysins." Because both of these antibodies exert definite injury upon
cells, we speak of them by the group names of "cytolysins" or "cyto-
toxic" substances.
The discovery of hemolysins naturally suggested the use of other
cells, and the following years brought forth many reports of further
specific cytotoxins. In 1899, Metchnikoff,46 and very soon after-
ward Landsteiner,47 described specific "spermotoxins" which ap-
peared in the blood of animals treated with spermatozoa. Von Dun-
gern 48 obtained analogous substances by injecting ciliated epithe-
lium from the trachea. Neisser and Wechsberg 49 produced "leuko-
toxiri' by injecting leukocytes; Delezenne 50 produced "neurotoxin"
and "hepatotoxin" and Sunnont,51 pancreas cytotoxin. Subsequent
years have added to these " gastro-toxin' (Bolton),52 thymotoxin
(Slatineau),53 adrenal cytotoxin (Gilder sleeve),54 placentar cyto-
toxin (Frank),55 corpus luteum cytotoxin (Miller),56 and a number
of others. In fact, as Roessle 57 puts it, in a review of the literature,
there is no organ in the body for which it has not been claimed that
specific cytotoxins can be formed by the injection of homologous
macerated tissues.
Recent critical study of these organ-cytotoxins has revealed, how-
ever, that the specificity of a serum produced with the tissues of one
organ is not strictly limited to this organ alone, and that the serum
may injure other organs as well. It is true, indeed, that there are
certain cells and tissues in the body such as the spermatozoa, the
tissues of the testicles, the ovary, the lens of the eye, and, possibly,
46 Metchnikoff. Ann. Past., Vol. 13, 1899.
47 Landsteiner. CentralU. f. Bakt., Vol. 25, p. 549, 1899.
48 Von Dungern. Munch, med. Woch., p. 1228, 1899.
49 Neisser and Wechsberg. Zeitschr. f. Hyg., Vol. 36, 1901.
50 Delezenne. Ann. Past., 1900; Compt. rend, de Vacad. des sc., 1900.
51 Surmo nt. Compt. rend, de la soc. de biol, 1901.
52 Bolton. Lancet, 1908.
53 Slatineau. Cited from Roessle, loc. cit.
54 Gildersleeve. Cited after Roessle.
55 Frank. Jour. Exp. Med., 1907.
56 Miller. CentralU. f. Bakt., 47, 1908.
"Roessle. "Lubarsch und Ostertag," Vol. 13, 1909.
PHENOMENA FOLLOWING IMMUNIZATION 93
the placenta which have chemical characteristics so well defined and
individual that the cytotoxic sera induced by them have definite
organ specificity. The same to a more limited extent seems true
of kidney substance (Pearce). In most cases, however, in which
originally a specific cytotoxin was claimed, it has been possible to
show subsequently that the apparently selective injury was due not to
organ specificity alone but to the fact that the injection of tissue-
macerates, even when sufficiently freed from blood, induced the
formation of considerable amounts of hemagglutinins and hemol-
ysins.
Pearce 58 expresses it as follows : " ... it is evident that the
cells or the various organs of the body, while differing in morphology
and function, have certain (receptor) characteristics in common,
and that one type of cell may therefore produce antibodies affecting
several cells of differing morphology, but with like (receptor)
groups. This is shown by the sera prepared from washed liver,
Ividney, pancreas, and adrenal, all of which may agglutinate and
hemolyze red blood cells and may cause degenerative changes also
in the liver and the kidneys. Some of these cytotoxic sera have no
effect upon organs for which they are supposed to have a morpho-
logical affinity, but exert a powerful lytic influence upon other cells.
Aside from nephrotoxin, which has a distinct injurious action upon
renal epithelium, the various cytotoxins studied (kidney, liver, pan-
creas, and adrenal) have no specific action in the morphological
sense."
This opinion seems to be in harmony with that of most observers
who have studied the problem recently, at least as regards most of
the organ cytotoxins. Much of the promised light upon pathological
processes — looked for when cytotoxins were first studied, has faded,
moreover, since it has been found that cytotoxins cannot be produced
by injection into an animal of cells, tissues, or fluids from its own
body. "Autocytotoxins" in general cannot be produced, a question
discussed at greater length in the chapter on lysis, in connection
with Ehrlich's work on the isolysins.
The work outlined in the preceding paragraphs had thus ex-
tended the principles of antitoxin and lysin production beyond the
scope of pure bacteriology, and had shown them to possess the sig-
nificance of general biological laws. Similar generalization was soon
attained in the case of the agglutinins and in that of the precipitins.
In the former, the nature of the reaction limited it to observations
upon cells in suspension, and, in connection with the earlier experi-
ments upon hemolysis it was soon discovered that the erythrocytes
were often clumped before lysis could take place, when brought to-
gether with a hemolytic serum of moderate or feeble potency, or
when solution, for other reasons, was delayed.
58 Pearce. Jour, of Med. Res., N. Sv Vol. 7, 1914, p. 13.
94 INFECTION AND RESISTANCE
The first observations on the general significance of the precip-
itin reaction we owe to Tschistovitch 59'and to Bordet.60 Tschisto-
vitch was studying the toxic action of eel serum upon rabbits. This
serum, as Kossel 61 had shown, is toxic for rabbits and possesses the
property of causing hemolysis of rabbit erythrocytes. Its similarity
to ricin, in this respect, stimulated attempts to produce an antitoxic
substance against eel serum, even as Ehrlich had produced an an-
tiricin. In the course of such experiments Tschistovitch observed
that, when eel serum was mixed with the serum of a rabbit which
had received several injections of this substance, the mixture became
rapidly opalescent and soon a flocculent precipitate was formed.
Coincident with this discovery Bordet made a similar observation.
He had injected chicken blood into rabbits in the course of experi-
ments upon hemagglutination. He found that the serum of the rab-
bits so treated acquired the property not only of producing hemolysis
and hemagglutination of chicken cells, but also of giving a precipi-
tate if mixed with chicken serum.62 Soon after this precipitins were
produced by injecting rabbits with milk (Bordet), egg albumen
(Ehrlich, Uhlenhuth), and many other substances, and the speci-
ficity of such reactions was demonstrated by Fish,63 Wassermann and
Schiitze,64 Uhlenhuth, and many others.
It is apparent from the preceding paragraphs that the discovery
of specific antitoxins merely constituted the first step in the formu-
lation of a fundamentally important biological law. There is, then,
a large group of substances of animal and vegetable origin which
call forth the formation of specific reacting bodies when injected
into animals. In order to elicit this response it is necessary that
these substances shall penetrate to the physiological interior of the
body in a relatively unchanged condition. For this reason any form
of injection, subcutaneous, intravenous, or into a serous cavity, is
followed, with regularity, by antibody formation, whereas feeding
or other means of intraintestinal administration is negative in result,
unless abnormal conditions prevail which permit entrance into the
blood before the digestive enzymes have decomposed the ingested
materials.
The substances with which antibody-formation may be induced
are collectively spoken of as "antigens,"
59 Tschistovitch. Cited by Bordet, loc. cit., and also Ann. Past., 13, 1899.
60 Bordet. Ann. Past., Vol. 13, 1899.
61 Kossel. Berl klin. Wocn., No. 7, 1898.
62 This, we know now, was due to the fact that the blood cells injected
were not washed free of chicken serum. Thus chicken serum precipitin was
formed as well as were hemagglutinin and hemolysin.
63 Fish. St. Louis Med. Cour., 1900. Cited from Uhlenhuth.
64 Wassermann and Schutze. Deutsche med. Woch., No. 30, 1900.
Vereinsbeilage.
PHENOMENA FOLLOWING IMMUNIZATION 95
Antigens are all substances which, injected into the animal body, induce
specific antibody formation. They form a large group in nature and are
chemically proteins; indeed, we may say that all known proteins may act as
antigens. Whether or not this term may also include lipoid-protein com-
binations, lipoids or the higher protein derivatives is as yet uncertain and
need not in the present connection concern us.
We may divide antigenic substances into two main classes.1 One of these
comprises all of those substances of bacterial, animal or vegetable origin
which, injected into the animal body, give rise to specific neutralizing or
antitoxic properties in the blood of the injected animal. These are the bac-
terial exotoxins, the snake venoms, some powerful vegetable poisons and proteo-
lytic and other enzymes of animals and plants. They are all substances which
are powerfully active — some of them strongly toxic to the living animal,
others true enzymes or ferments. Indeed all of them possess properties which
at least suggest our placing them into the class of enzymes in general. The
number of such substances known is limited. The reaction they call forth in
the animal body seems aimed directly at the specific neutralization of their
respective activities, and is so unique and different from that induced by other
antigens that it would be convenient had we another term like "antitoxinogen"
to set them apart by themselves.
The other class of antigens comprises all proteins which are inactive,
showing in themselves neither toxic nor enzyme-like properties. Introduced
into the animal body parenterally, they call forth a response of a nature
entirely unlike that of the antitoxins, and which as far as we can fathom
its purpose seems aimed merely at the assimilation or the removal of the
infected substance. For the cells of the animal cannot utilize the foreign
protein as such, and thus it is only foreign proteins injected into an animal
that act antigenically, and no antibodies are formed when homologous material
is injected.
This large group, composed of all formed and unformed substances in
nature in which a protein structure is involved, does not induce the formation
of anything like the neutralizing antitoxins spoken of above. The antibodies
appearing in animals treated with such substances have been spoken of as
cytolysins or cytotoxins — precipitins — and in the case of formed antigens
like bacteria or blood cells — agglutinins and opsonins. As we shall see in
another section, it is our opinion that all these various antibodies are iden-
tical in structure and significance.
We must not forget, however, that the observation of antibodies in the
circulating blood is but one of the changes that have taken place in the
treated animal. Much has been made of this phase of the problem because
serum antibodies are readily studied in vitro; but their origin of course must
be sought in the body cell, in which the original and most profound changes
must necessarily have taken place during such treatment, changes the nature
of which are to a large extent still a mystery, but on which ultimately de-
pend the important physiological difference between treated and untreated
animals. For such changes — whether we refer to those immediately under
discussion, namely, those of allergy or anaphylaxis, or whether we think of
the so-called immunity remaining after attacks of many diseases — remain
present long after the circulating antibodies have disappeared and must there-
fore be regarded as associated with profound alterations in the ultimate tissue
unit, the body cell.
1 See also Zinsser, "The More Recent Developments in the Study of Ana-
phylactic Phenomena." Arch, of Int. Med., Vol. XVI, 1915, pp. 223-256.
Harvey Lecture.
96 INFECTION AND RESISTANCE
Since the phenomenon of antibody formation is not at all limited
to bacteria or bacterial derivatives, it cannot be looked upon merely as
a mechanism existing for the primary purpose of protecting the body
against infectious disease. This latter function is important, indeed,
but is probably incidental to the broader significance of the processes.
In the course of normal existence substances which are not di-
rectly assimilable as such — foreign proteins, for instance — do not
penetrate directly into the blood and tissues. Taken into the ali-
mentary canal, they are first hydrolized into peptons, albumoses,
polypeptids, and probably amino-acids before absorption, to be recon-
structed from these cleavage products ("Bausteine" is Abderhalden's
expression for the amino-acids) into protein biologically identical
with that of the tissues. Digestive and other accidents, however, on
numerous occasions during life permit the direct entrance of these
materials unchanged or insufficiently changed into the circulation.
It is probably by the action of digestive powers of the serum — or, in
the case of 'the entrance of undissolved foreign particles, by the
activity of the phagocytic cells — that such substances are then dis-
posed ot and assimilated. For each particular variety of substance
(antigen) a specific mechanism is called into play, and when this
mechanism is repeatedly called upon — as in successive injections of
foreign proteins — this mechanism, whatever it may consist of, is en-
hanced in efficiency — i. e., increased in quantity. How this increase
of specific antibodies is theoretically conceived we will discuss later
in connection with Ehrlich's side-chain theory.
The phenomena of antibody formation against bacteria on this
basis may be taken to constitute, then, a mechanism for the digestion
and disposal of a foreign protein which has penetrated into the tis-
sues and, because of its living state, increases within the body by
multiplication, furnishing progressive stimulation to the antibody-
producing function. Infectious disease, therefore, from this point
of view may be looked upon as an invasion of the body by a living
foreign protein which must be assimilated and disposed of; which,
in some cases, has a primary toxicity per se ; and which is variously
distributed among the organs and tissues according to the biological
peculiarities of the particular microorganism in question. This
general conception will become more clear as we analyze the phe-
nomena associated with the individual antibodies. It is, of course,
quite plausible as far as it refers to the phagocytic functions, or even
bacteriolytic and cytolytic phenomena. It has been less clear in
connection with the agglutinins and precipitins in which a direct de-
fensive or bacteria-destroying value is not apparent. However, in
our discussions of these phenomena we will have occasion to point, out
many reasons for assuming that, even in these phenomena, there are
features which fall into direct correlation with the views we have
just expressed.
PHENOMENA FOLLOWING IMMUNIZATION 97
The substances which possess antigenic properties — that is, which
give rise to antibody production — with the exception of a few isolated
and contested cases, are all of them protein in nature. Well-trained
chemists have exerted themselves to purify antigenic substances,
in attempts to determine the particular fractions of the complex
protein molecule upon which the antigenic properties depend. In
the course of such work a number of men claim to have obtained a
truly antigenic substance which no longer gave protein reactions.
The instance most frequently cited is Jacoby's 65 announcement of
a protein-free ricin. Jacoby worked with an apparently very impure
"Ausgangsmaterial" consisting of commercial ricin, which he di-
gested for five weeks in trypsin solution. At the end of this time he
obtained a ricin which still possessed the properties of the original
castor-bean extract, but no longer gave protein reactions. His "puri-
fied ricin," however, was quickly destroyed by further trypsin diges-
tion, and more recent work by Osborne, Mendel, and Harris 66
appears to have fully refuted Jacoby's results. They found the
purified ricin identical with the coagulable albumin of the castor
bean, and found that tryptic digestion destroys the characteristic
ricin properties.
Less easily refuted have been the careful experiments of Ford 6T
upon the active principle of a mushroom (Amanita phalloides) and
upon that of the poison-ivy plant — (Rhus toxicodendron) . These
substances, he claims, are non-protein. In the case of Amanita
phalloides Abel and Ford 68 have shown it to be a glucosid, and
similar structure has been claimed for Rhus by Syme.69 Yet with
both of these substances Ford has succeeded in producing specific
antitoxins. Rabe 70 has questioned these with Amanita phalloides.
He believes that the poison with which Ford worked is not a glucosid,
but is of protein nature. In the case of Rhus, however, Ford's con-
clusions have not, to our knowledge, been challenged.
With these and a few other less important exceptions, however,
observers have uniformly concluded that antigenic property and pro-
tein structure are inseparably associated. All procedures by which
proteins have been hydrolized into their simpler fractions, chemical
splitting, tryptic or peptic digestion have in every case resulted in a
simultaneous loss of protein reaction and antigenic property..
Many attempts have also been made to show a relation between
antigenic properties and the lipoid constituents of cells. These en-
deavors were obviously stimulated by the observation that many li-
65 Jacoby. Arch. f. exp. Path. u. Pharm., Vol. 46, 1901.
66 Osborne, Mendel, and Harris. Am. Jour, of PhysioL, 1905, Vol. 14.
67 Ford. Jour, of Inf. Dis., Vol. 3, 1906 ; Vol. 4, 1907.
68 Abel and Ford. Jour. Biol. Chem., 1907,
69 Syme. Johns Hopkins Thesis, 1906.
70 Rabe. Zeitschr. f. exp. Path. u. Therap., Vol. 9, 1911.
98 INFECTION AND RESISTANCE
poids are capable of binding antibodies in miro, and that, in nervous
tissues, toxin fixation was in some way related to the richness in
lipoids of these structures. Bang and Forsmann 71 accordingly
treated animals with ether extracts of red blood cells — claiming that
this resulted in the production of hemolysins. And these results
have been confirmed by Landsteiner and Dautwitz.72 The latter,
however, suggest that the hemolysin production may have been in-
duced, not by the lipoidal substances in solution, but by other anti-
genie substances which had gone into colloidal suspension in the
ether extracts. Much similar research on the antigenic nature of
lipoids has been done, but, after reviewing this very thoroughly,
Landsteiner comes to the conclusion that no definite proof of the anti-
genic nature of any pure lipoid has so far been presented. The prob-
lem is experimentally complicated by the fact that, as Landsteiner 7a
suggests, the antigen may often be present as a lipoid-protein com-
bination, and as such go into solution or fine emulsion in the organic
solvents ; also the lipoids possess the curious property of altering the
solubilities of proteins and other substances by their presence.
Summarizing our present knowledge of the chemical nature of
antigens, then, we must conclude that, with the exception of Ford's
glucosids, no protein-free antigens have been thus far demonstrated.
In the light of this fact it is all the more remarkable that antigen-
antibody reactions are specific. For we possess no chemical methods
by which one variety of protein can be distinguished from another.
And yet the serum antibodies produced with each species of bacteria
react with this species only — and the hemolysins, agglutinins, or
precipitins produced by the injection of bacterial, cellular, or serum
proteins react respectively only with the particular variety employed
in their production. This indicates that each of these antigens — of
almost unlimited number — must possess a chemical structure indi-
vidually characteristic and different from all the others. It is by
means of the biological reactions, indeed, that we can detect protein
in dilutions far beyond the reaction-sensitiveness of chemical tests
and can distinguish between varieties of protein when the chemical
methods will indicate only protein in general. Our knowledge of
the chemical constitution of protein has not yet advanced to a point
at which specificity can be based upon definite variations of chemical
structure, and the complexity of the problem is such that it does not
seem likely that we can hope in the near future to attain such knowl-
edge. We can merely accept it as a fact that the antibody produced
with one protein differs materially from that produced with another,
71 Bang and Forsmann. Hofm. Beitr., 1906; Centralbl f. Bakt., 40, 1906.
72 Landsteiner and Dautwitz. Hofm. Beitr., 9, 1907.
73 Landsteiner. "Wirken Lipoide als Antigene?" Weichardt's Jahresbe-
richt, Vol. 6, 1910.
PHENOMENA FOLLOWING IMMUNIZATION 99
and that this is a definite indication that the antigen in one case must
be chemically different from that in another.
The range of such variations is apparently enormous. For each
variety of bacteria or plant, each species of animal, and to a certain
extent each individual of the species, possesses certain special anti-
genie characteristics peculiar to itself. In general there is an under-
lying antigenic similarity which is peculiar to the species. This is
true of bacteria and, in the case of animal and vegetable proteins, an
antibody produced with material from an individual of a certain
species will react with the protein derived from this species in gen-
eral. However, that there are also antigenic differences between in-
dividuals within the same species is indicated by Ehrlich's experi-
ments on the antibodies produced by injecting the blood cells of one
goat into another. And we have further indicated that within the
same animal different organs may possess individual antigenic char-
acteristics. Added to this we know that certain special organs like
the testicle, the lens, and some others contain antigens which are
peculiar to this variety of organ, irrespective of species — a condition
spoken of as "organ specificity." Thus an antibody produced by
injections of the testicular substance of one animal will react with
testicular protein from many different species — the specificity here
depending upon the organ and not upon the zoological relationship.
It is clear, therefore, that there are more different varieties of
protein, biologically distinguishable, than there are species of living
beings in nature. As Abderhalden74 has recently pointed out, this
is a conception which it is a little difficult to grasp chemically, since
in breaking up different proteins into their "building stones" (Bau-
steine) we encounter again and again the same 20 amino-acids. By
a simple arithmetical consideration, however, he shows that merely
by combining these twenty amino-acids in different groupings an
.enormous number of isomeric but varying compounds can be formed
— even without assuming the additional possibility of quantitative
variations. He reasons that 3 "Bausteine" — A, B, and C — could
form 6 different structures, A B C, A C B, B C A, B A C, C A B,
C B A. Similarly 4 could form 26, and finally 20 could form 2, 432,
902, 008, 176, 640, 000 different compounds.75
The analogy between the active immunization of animals with
the various antigens and certain chemically well-defined poisons,
alkaloids, etc., is so obvious that it has led to much speculation as to
a possible similarity in the physiological mechanisms of the two phe-
nomena. As a matter of fact the acquired tolerance for such sub-
stances as morphin, atropin, and other alkaloids is not really anal-
ogous to the physiological reactions which follow the treatment of an-
74 Abderhalden. Munch, med. Woch., No. 43, 1913.
75 "We have not repeated the arithmetical labor and take Abderhalden's
word for it.
100 INFECTION AND RESISTANCE
imals with bacterial and other proteins, for whatever toxic properties
there are in the latter are, as we shall see later, rather the results of
the interaction of these injected substances and the reaction products
supplied by the cells and fluids of the body. It is at least probable
in the light of our modern conception that such protein antigens are
not toxic per se, in the native state. This, however, will receive
detailed consideration in succeeding sections. The analogy of drug
tolerance, however, to the acquired immunity against true bacterial
toxins and vegetable poisons like ricin, crotin, and others is a strik-
ing one, since in both classes of poisons there is a gradually devel-
oped tolerance for substances toxic in the native state and often very
similar in physiological effects (strychnin and tetanus toxin, etc.).
In the case of the toxins, however, there is a development of im-
munity by actual neutralization of the poisonous principle brought
about by a specific antibody, which circulates in the blood of im-
munized animals and man — the process following, within certain
limits, the law of multiple proportions. In the case of morphin
and other alkaloids no such neutralizing antibodies have as yet been
demonstrated.76 Whereas toxin immunity is passively transferable
from one animal to another with the blood serum, and, in vitro, the
mixture of the toxin with the immune serum brings about a neutrali-
zation of the poison, no such phenomena have been observed, as a
general rule, in the case of the alkaloids. We say "as a general
rule" since an exception is recorded in the observations of Fleisch-
mann,77 who claims to have found antagonistic action to atropin in
the blood of normal rabbits, this power being absent from the blood
of rabbits that had thyroid hypertrophies and were, in consequence,
atropin-susceptible. Other observations of a similar significance
have been made by Physalix and Contejean 78 on curare, but have
not been confirmed, and the investigations of all other workers on
this subject have had negative results. It seems from available evi-
dence that tolerance (immunity) against drugs is due to cellular
rather than to serum antagonism.
THE ORIGIN OF ANTIBODIES
The tissue cell, as the ultimate functional unit, must, of course,
})G looked upon as the source from which originate the various pro-
tective constituents of normal and immune sera ; and, though per-
haps unrecognizable by the coarse tests of morphological investiga-
tions, it is in the cells that changes must take place primarily when
76 Hans Meyer and Gottlieb. "Exp. Pharm.," 2d Ed., Neban & Schwart-
zenberg, Berlin, 1911, p. 517.
77 Fleischmann. Archiv f. exp. Path. u. Pharm., 62, 1910, cited from
Meyer and Gottlieb, loc. cit.
78 Physalix and Contejean. Cited from Meyer and Gottlieb.
PHENOMENA FOLLOWING
the animal body is subjected to any one of the processes spoken of as
immunization. The exact location of the antibody-forming cells and
tissues, in spite of much investigation, is not at all clear, though
many data seem to point to the lymphatic organs, the spleen, and the
bone marrow as particularly concerned with this process.
Thus Pfeiffer and Marx '9 exsanguinated animals five days after
injections of dead cholera spirilla and found that at this time bac-
teriolytic antibodies were more concentrated in the spleen than in
the blood serum itself. Wassermann's 80 analogous experiments with
typhoid bacilli seemed to show a higher antibody content in spleen,
bone marrow, thymus, and lymph nodes than was present in the
blood at an early period of immunization. Although these investiga-
tions, as well as many others of Castellani,81 seem, therefore, to indi-
cate a particular association of the special lymphatic organs with
antibody formation,82 extirpation of the spleen 83 before immuniza-
tion has not prevented animals from responding to injections of bac-
teria and red blood cells with sharp antibody production. The ex-
periments of Deutsch,84 in which reduction of antibody formation
resulted in animals in which splenectomy was practiced three or four
days after immunization was begun, can hardly be accepted as a con-
clusion, in the writer's opinion at least, since any severe operation or
interference with the normal functions of an animal during the
severe physiological strain of active immunization would naturally
lead to a less perfect response. That the resistance of animals and
man to infection with bacteria is not noticeably diminished by sple-
nectomy, moreover, has been variously shown. In unpublished ex-
periments by the writer splenectomized guinea pigs showed no differ-
ence from normal animals in regard to their susceptibility to tuber-
culosis. And though these and similar experiments of other workers
with various bacteria are not entirely devoid of interest, their nega-
tive results as a matter of fact have no great significance, since our
knowledge concerning the true function of the spleen is very incom-
plete, and it is not impossible that on removal of this organ other
elements of the lymphatic system may take over its function in part
or as a whole.
Removal of the spleen has not been an extremely unusual pro-
cedure in surgery, and there -is no evidence to show that patients so
treated have been abnormally susceptible to infection thereafter.
Yet, as we have seen, there seems to be an early concentration of
antibodies in the lymphatic organs in the course of immunization,
and it may well be that an association between the process and these
79 Pfeiffer and Marx. Zeitschr. f. Hyg., Vol. 27, 1898.
80 Wassermann. Berl. klin. Woch., p. 209, 1898.
81 Castellani. Zeitschr. f. Hyg., Vol. 37, 1901.
82 Pfeiffer and Marx. Loc. cit.
83 I. Levin. Jour. Med. Bes., Vol. 8, 1902.
84 Deutsch. Ann. de I'Inst. Past., Vol. 13, 1899.
KHJtM mFECTION AND RESISTANCE
tissues exists which cannot be experimentally demonstrated with
absolute certainty.
It is no less likely, however, that similar functions are exerted
by the cells of other organs. In fact, it is more than probable that
antibodies may be formed anywhere in the body — and that the local-
ity of their production is largely dependent upon the locality in which
the antigen is concentrated. Wassermann and Citron 85 demonstrated
this by injecting typhoid bacilli into rabbits intraperitoneally, in-
travenously, and intrapleurally, and nine days afterward determining
the comparative bactericidal strength of blood serum and of aleuronat
exudates of pleura and peritoneum in each of the three animals.
Their results showed that the bactericidal titre of the intravenously
inoculated animal was highest in the blood serum, while that of the
intraperitoneally and intrapleurally inoculated animals was highest
in peritoneal and pleural exudates respectively. Such experiments
point to the possibility of a "local" immunity, that is, a production
of antibodies directly by the cells with which the antigen comes into
contact in the most concentrated and direct manner. And, indeed,
another isolated experiment of the same authors, alone successful of
a series of similar attempts, would point in the same direction.
Typhoid bacilli were injected subcutaneously into the ear of a rabbit
and the ear immediately ligated at its base and kept so for several
hours. After nine days the bactericidal titre of the blood serum was
determined and the ear amputated. An immediate and rapid drop
of antibody contents occurred after the amputation — indicating that
the chief source of antibody function had been removed. More strik-
ing examples of the same thing are to be seen in the experiments of
Romer,86 who instilled abrin into a rabbit's eye and found that the
retina of the eye developed an antitoxic power against abrin which
protected mice against many times the fatal dose, while that of the
other eye remained practically inactive.
From these facts, as well as from other observations, it is at least
reasonable to believe that antibody formation is by no means a func-
tion of special organs and that many cells throughout the body may
take part in the process. It is of especial importance to consider this
in connection with the possible effects of the treatment of infections
by means of bacterial vaccines. If the focus of the infection can
possibly become also a local source of antibody production then such
treatment may well seem rationally founded, even in generalized
acute infections in which no logical basis for such treatment would
exist, were the production of antibodies a task for specialized organs
like spleen and bone marrow only. The therapeutic phases of this
problem are more extensively considered in a later chapter.
It is in this fact also that we must seek the explanation of the
85 Wassermann and Citron. Zeitschr. f. Hyg., Vol. 50, 1905.
86 Romer. Arch. f. OphthaL, 52, 1901.
PHENOMENA FOLLOWING IMMUNIZATION 103
apparent local immunity which occurs in certain infections of the
skin. Thus it frequently happens that successive crops of boils may
afflict different parts of a patient's skin — new ones arising as old
ones heal, showing that the process of the limitation and healing of
the infected foci is not due to any increase of generalized resistance,
but rather to local causes. In the same way, in erysipelas, the process
extends along the edges while the original central area of infection
is returning to the normal state, and it rarely occurs in adults that
the erysipelatous process extends back into the originally infected
area.87 From these localized laboratories of antibody formation, of
course, distribution to the circulation probably takes place and the
complete cure of the patient must await a sufficient concentration of
these in the body as a whole before further local foci cease to arise.
That the fixed tissue cells of any part of the body can and do
take an active part in the local reaction against the invasion of bac-
teria and other foreign materials is histologically evident. When
a more or less insoluble foreign body — a thread of lint, paraffin,
agar-agar, or other material — is deposited in the subcutaneous tissues
anywhere in the body, and is accompanied by acute infection with
bacteria, there is a characteristic tissue reaction which results in the
surrounding of the foreign particle by multinucleated cells spoken
of as giant cells. In the case of foreign bodies such as those men-
tioned the process is purely one of local ingestion of the particle
which later, if the material remains absolutely insoluble, results in
encapsulation by connective tissue. If soluble, however, there may
be an eventual digestion of the foreign material by the cell with a
subsequent degeneration or splitting up of the giant cell and a return
to normal. This also occurs in the case of such infections as those
due to yeasts or blastomyces, in which, as the writer has seen, the
apparent lack of liberation of toxic products gives rise to a purely
local giant-cell reaction, adjacent tissue cells remaining undegener-
ated and apparently unaffected. In the case of infection with bac-
teria like the bacillus of tuberculosis, the leprosy bacillus, that of
rhinoscleroma, and a few others the purely local picture of giant-
cell phagocytosis is complicated by secondary reactions arising prob-
ably from the liberation of toxic products from the living or dead
invaders which both stimulate specific cell reactions and call forth
cell degeneration in adjacent tissues, frequently giving the individual
infection a diagnostically characteristic appearance.
87 In children erysipelas not infrequently returns within a few days over a
recently healed area.
CHAPTER V
TOXIN AND ANTITOXIN
THE EEACTION BETWEEN TOXIN AND ANTITOXIN
(EHBLICH'S ANALYSIS)
THE TOXIN-ANTITOXIN REACTION
WHEN Behring and his collaborators, Kitasato and Wernicke,
had definitely shown that the cell-free blood serum of animals im-
munized with tetanus and diphtheria toxins respectively possessed
the power to protect other animals of the same and different species
against the poisons, it became of the utmost importance to deter-
mine, if possible, the mechanism by which the "antitoxic" effect was
attained. The earlier opinion, expressed by Behring himself, held
that in all probability the toxin was directly injured or destroyed by
the action of the antitoxic serum.- That this assumption was incor-
rect was soon demonstrated by the experiments of Eoux and Vail-
lard 1 and by those of Buchner.2 The worl: of the former investiga-
tors showed that the mixtures of tetanus toxin and antitoxin, meas-
ured in such proportions that they were harmless for normal guinea
pigs, could still be found toxic for animals weakened by preliminary
inoculation with other bacteria. Buchner claimed in analogous ex-
periments that similar mixtures, harmless for mice, could still show
toxicity for guinea pigs. He inferred from this that the nature of
the cell reactions of different animal species influenced the antitoxic
effect. Both investigations led the workers to conclude that the pro-
tective action of antitoxin was not due to a direct effect upon the
poison but was potent by acting upon the tissue cells of the animal by
protecting these from subsequent harm by the toxin. Their concep-
tion implied an indirect protective function on the part of the anti-
toxin, not due to any direct reaction between it and the poison.
That this explanation, too, was faulty was made evident by a
number of investigations which took advantage of the peculiar dif-
ferences in resistance to temperature between certain toxins and
their specific antitoxins.
1 Roux and Vaillard. Ann. de I'Inst. Past., 1894.
2 Buchner. Munch, med. Woch., p. 427, 1893.
104
TOXIN AND ANTITOXIN 105
In 1894 Calmette 3 and Physalix and Bertrand 4 had indepen-
dently succeeded in obtaining an antitoxin against snake poison. In
the course of further study of these bodies Calmette5 determined
that the venoms of certain varieties of snakes, the naja and cobra,
would remain potent even when subjected to 100° C. for a very
short time. In contrast to this the antitoxins to these poisons were
destroyed at much lower temperatures. Now when mixtures of
the two substances, so proportioned that their injection into ani-.
mals was innocuous, were heated to 68° C. for considerable
periods, toxia properties again became evident, a demonstration that
the toxin had not been destroyed, but had remained neutral only in
the presence of the intact antitoxins. These experiments were con-
firmed by Wassermann,6 who found that similar conditions pre-
vailed in the combination between pyocyaneus toxin and antitoxin.
The filtration experiments of Martin and Cherry 7 are not con-
vincing since they may be taken as indicating either neutralization
or toxin destruction. These workers subjected mixtures of snake
poison and its specific antitoxin to filtration through gelatin filters,
under pressure. Under the experimental conditions thus estab-
lished the presumably smaller toxin molecule was allowed to pass
through the filter while the larger antitoxin molecule was held back.
They showed that if filtered soon after the ingredients have been put
together most of the toxin still passes through, but that, as this inter-
val is prolonged, less and less comes through, presumably because of
the union of the smaller toxin to the larger antitoxin molecule. The
chief value of these experiments lies in their proof of the element of
time as an important factor in the toxin-antitoxin union.
In his experiments on snake venom just recorded, Calmette in-
terpreted the restitution of toxicity after the heating of neutral mix-
tures of cobra neurotoxin and its antitoxin as evidence "qu'il ne
s'etait pas forme aucune combinaison de ces deux substances ou que
la combinaison realisee etait, au moins, tres instable." Later experi-
ments of Martin and Cherry seemed for a time to contradict this con-
clusion. Observations by them, analogous to those of Calmette, but
carried out with the poison of an Australian snake, seemed to indi-
cate that when the toxin and antitoxin were allowed to remain to-
gether for a sufficiently long time no restitution of toxicity could be
obtained by heating. Apparently the application of heat to such mix-
tures merely prevented the further union of antitoxin with any toxin
that was not yet bound at the time that the heat was applied. Accord-
3 Calmette. Compt. rend, de la soc. de biol., 1894.
4 Physalix and Bertrand. Compt. rend, de la soc. de biol., 1894.
5 Calmette. Ann. Past., 1895.
6 Wassermann. Zeitschr. f. Hyg., 22, 1896.
7 Martin and Cherry. Proc. of the Royal Soc., Vol. 63, 1898.
106 INFECTION AND RESISTANCE
ingly Morgenrotli 8 again examined these relations and found that the
addition of a small amount of hydrochloric acid to mixtures of snake
poison and the antitoxin resulted in the dissociation of their union.
To mixtures of the venom lysin and its antitoxin, neutralized and
even overneutralized so that they were perfectly innocuous to suscep-
tible animals he added hydrochloric acid until the total concentra-
tion amounted to N/18. By this method a toxin-HCl modification
was produced which was dissociated from its union with the anti-
toxin and was extremely resistant to heat. In such a mixture of
toxin and antitoxin to which hydrochloric acid had been added, heat-
ing at 100° C. in a water bath for 30 minutes destroyed the ther-
molabile antitoxin and, after neutralization, undiminished toxic
properties could again be demonstrated by animal inoculation.
These researches and other similar ones of Morgenroth, then,
form a satisfactory confirmation of the original experiments of Cal-
mette and seem to show, beyond possibility of contradiction, that the
inhibition of harmful properties of any true toxin, after mixture
with its antitoxin, does not depend upon toxin destruction. But
while Calmette interpreted the facts as pointing toward a failure of
union of the two substances, Morgenroth's work is not incompatible
with the conception of a neutralization of one by the other in the
chemical sense. These experiments of Morgenroth are of great the-
oretical importance moreover in that they have shown that dissocia-
tion of a toxin-antitoxin complex can occur.
The nature of such neutralizations in regard to quantitative rela-
tions, speed of action, and relative concentrations, becomes apparent
partly from experiments like those mentioned above, but more espe-
cially from those carried out by Ehrlich with ricin and antiricin, ex-
periments which were primarily planned to demonstrate that the
reaction between a toxin and its antibody is a direct one, not depend-
ent upon intervention of the body cells, as at first supposed.
It had been shown by Kobert and Stillmarck that ricin, the
powerfully poisonous principle of Ricinus communis (castor oil
bean) would agglutinate the red blood cells of a number of animals.
Ehrlich recognized from the beginning how closely analogous the
neutralization of ricin by antiricin was to that of diphtheria toxin by
its antitoxin. The former reaction furnished him with a simple
method of test tube experimentation in that the agglutinating effects
of ricin upon rabbits' corpuscles could be directly inhibited by the
preliminary addition of antiricin. A visible reaction was thus avail-
able, which, of course, excluded absolutely the participation of the
tissue cells in the antigen-antibody neutralization, and in which
careful quantitative measurements were possible.
Ehrlich 9 determined by means of this method that the neutral
8 Morgenroth. Berl. kl Woch., No. 50, 1905, p. 1550.
9 Ehrlich. Fortschr. d. Med., Vol. 15, p. 41, 1897.
TOXIN AND ANTITOXIN 107
ization was accelerated by moderate heat and by concentration of the
reagents and, most important of all, that the reaction followed
roughly the law of multiple proportions, characteristics, all of them,
which were entirely analogous to chemical reactions in general.
When he added 0.3, 0.5, 0.75, 0.1, etc., cubic centimeters of serum
from a ricin-immune goat to constant quantities of ricin, and then
added rabbit cells, the hemagglutinating properties of the ricin were
inhibited in direct proportion to the amount of antiricin mixed with
it. And his test tube experiments were further found to represent
with much accuracy the occurrences which took place within the
animal body. For, similar mixtures injected into mice were toxic
in direct proportion to the balance of ricin and antiricin established
in the injected material.
Although the views of Ehrlich and his' followers have great im-
portance in connection with the union of antigens and their anti-
bodies in general, these ideas were worked out by him most elab-
orately in connection with his efforts to arrive at a practicable and
accurate method of establishing a standard of strength for diphtheria
antitoxin, and it is essential that we consider this work in detail.
The earlier attempts to standardize diphtheria antitoxin by the
use of living cultures (Roux and Behring) were soon abandoned,
since it was found that the accurate establishment of fixed lethal
doses of the culture was not possible. When the facts, just recorded,
concerning the interaction and quantitative relations of the soluble
toxins and their respective antitoxins came to light, Behring intro-
duced the standardization of the curative sera by the use of toxins,
both in the case of tetanus and in that of diphtheria. In order to
do this consistently he established for diphtheria poison an arbi-
trary toxin unit which he defined as the amount of any given diph-
theria filtrate sufficient to cause death in a guinea pig of 250
grams, and, borrowing the terms from chemical nomenclature, he
designated as a "normal" diphtheria poison one which contained
100 such units in one cubic centimeter. (D T !N", M250 = diph-
theria toxin normal, Meerschweinchen 250 grams.)
Together with Ehrlich, Behring then established an antitoxin
unit (I-E, Immunitats Einheit). They designated as a "normal"
antitoxic serum one "which contained in one cubic centimeter one
antitoxic unit" (I-E), and state further, "of this serum 0.1 c. c.
neutralizes 1 c. c. of the Behring normal toxin." (Conf. Madsen in
"Kraus u. Levaditi Handbuch," II, p. 94.) Alterations were subse-
quently made in this scale of standards and Ehrlich later desig-
nated as an antitoxin unit a quantity of an antitoxin which
completely neutralized 100 lethal doses (for guinea pigs of
250 grams) of a poison at that time in his possession. The unit
of diphtheria antitoxin at present in use therefore may be defined as
a quantity of serum sufficient to protect a guinea pig of 250 grams
108
INFECTION AND RESISTANCE
against 100 times the fatal dose (M L D, minima dosis lethalis)
of toxin. Since the methods of antitoxin standardization employed
at present in the United States were worked out by Rosenau 10 along
the lines of Ehrlich's method, and the standard is based on the one
introduced by Ehrlich, the antitoxin unit as employed in this country
is identical with the one spoken of in the following paragraphs.
In measuring the neutralizing value of antitoxin for toxin, then,
since both substances are chemically unknown and no purely chem-
ical indicator of neutralization is available, it was necessary to select
a susceptible animal by means of which excess
of toxin, in mixtures of the two, could be de-
tected. As the standard test animal guinea
pigs of 250 grams were chosen, and improve-
ments in the methods of measurement were
introduced, in that the toxin and antitoxin,
instead of being separately injected as hereto-
fore, were mixed, allowed to stand for 15 to
30 minutes, and then injected together sub-
cutaneously.
By means of this technique Ehrlich set
out to examine a large number of toxins and
their antibodies and obtained results which,
aside from their practical value, have had an
important influence upon the development of
the knowledge of antigen-antibody reactions.
These investigations were considerably
complicated by the fact that neither the diphtheria toxin nor the anti-
toxin is very stable and deterioration occurs unless special methods of
preservation are employed. Since the antitoxin, however, is much
less unstable than the toxin, the former is employed in order to pre-
serve the standard, and is preserved in sealed U tubes (see Fig-
ure) with anhydrous phosphoric acid. Kept in this way, in black,
light-proof boxes, and at low temperature, it may be preserved for
months without appreciable loss of value and may be renewed by ac-
curate comparative measurements from time to time. This is carried
out for the United States, at the present time, by the Government
Hygienic Laboratories at Washington.
Preservation of the toxin is much more difficult, and it is in
connection with the investigation of the instability of the toxin that
Ehrlich gained his first insight into the nature of the reaction. He
measured, in a number of toxic filtrates, the minimal lethal dose
for guinea pigs of 250 grams, establishing a time limit for death
in order to obtain more accurate comparisons. He designated as the
10 Rosenau. U. S. P. H. & Mar. Hosp. Service. Hygienic Laboratory BulL>
21, April, 1905.
TUBE FOR THE PRESERVA-
TION or THE STAND-
ARD ANTITOXIN.
Taken from Eosenau, U.
S. Hygienic Labora-
tory Bulletin, No. 21,
1905, p. 53.
TOXIN AND ANTITOXIN 109
new M L D or "T" (that is: toxic unit) the quantity of toxin
which will kill a guinea pig of the designated weight in from 4 to 5
days. He then determined for a number of poisons the exact quan-
tity just neutralized by one antitoxin unit, calling this quantity L0.
(L meaning Limes or threshold.)
It is clear that in judging of complete neutralization of a quan-
tity of toxin by antitoxin, there may be a strong subjective element,
since any very slight excess of toxin may cause unimportant local
reactions such as edema or small hemorrhages, which could escape
the attention of one observer while being noticed and recorded by
another. In order therefore to exclude definitely all subjective fea-
tures from such experimentation, Ehrlich now established another
toxin value — L+ dose ("Limes death" — now, for convenience, writ-
ten L + ) — which eliminated all possible variations of personal per-
ception. He designated by this symbol that quantity of toxin which
not only neutralized one antitoxin unit but included enough toxin,
in excess of this, to give the result of one free toxin unit, that is, to
cause death in 4 to 5 days in a guinea pig of 250 grams. Since
the three values just defined form the basis of Ehrlich's experiments
as well as that of all practical diphtheria serum standardizations we
will briefly restate them for the sake of clearness.
Thus:
M L D or "T" = the amount of toxin which, subcutaneously injected,
causes death in a 250-gram guinea pig in from 4 to 5 days.
L0 = the amount of toxin which is just neutralized by one antitoxin unit so
that no trace of reaction, local or otherwise, ensues
and
L+ = that amount of toxin which will cause death in 4 to 5 days in a guinea
pig of 250 grams if injected together with one antitoxin unit.
It will further clarify the meaning of these terms to examine
experimental protocols which show how these values are determined.
Thus in the following:
I. Injections of toxin
(1) .005 c. c. — guinea pig lives.
(2) .009 c. c. — guinea pig dies in 6 days.
(3) .01 c. c. — guinea pig dies in 4 days.
(4) .02 c. c. — guinea pig dies in 2 days.
.01 = M L D or T.
II. 1 Antitoxin unit + .19 toxin = late paralysis.
1 Antitoxin unit + .20 toxin = sometimes late paralysis and sometimes
acute death.
1 Antitoxin unit + .21 toxin = death fourth day.
1 Antitoxin unit + -22 toxin = death in 2 to 3 days.
.21 = L+ dose.
110 INFECTION AND RESISTANCE
III. Antitoxin unit -f .14 c. c. toxin = no reaction.
Antitoxin unit -j- .15 c. c. toxin = no reaction.
Antitoxin unit + .16 c. c. toxin = slight congestion about point of injec-
tion, scarcely visible.
Antitoxin unit + .17 c. c. toxin = apparent reaction at site.
Antitoxin unit -j- .18 c. c. toxin = edema at site.
Lo = .16."
In determining these values with a large number of toxins Ehr-
lich discovered the curious fact that, although there was a rapid and
extensive diminution of toxicity in every toxic filtrate in the course
of time, there was no corresponding alteration in the L0 amount.
In other words, although more and more of the toxic broth was
necessary to kill a guinea pig of standard weight in the required
time, the amount of the same broth which neutralized one antitoxin
unit remained approximately the same.12
In seeking an explanation for this apparent paradox, Ehrlich
concluded that we must assume that the toxin is complexly con-
structed, consisting of a toxophore and a haptophore group. Assum-
ing that chemical union between the toxin and the antitoxin (or, in
disease, the body cell) takes place, it is by means of the haptophore
group that such union is brought about. The toxophore group, how-
ever, is the element by which toxic action is exerted after union
by the haptophore group has been accomplished. It would be
conceivable, therefore, that in deteriorating in toxicity the toxin
might undergo alterations in the toxophore group only, its hapto-
phore group, and, therefore, its antitoxin neutralizing properties
remaining intact. Such modified toxins, modified only as to the
toxophore groups, Ehrlich now refers to as "toxoids."
In the production of diphtheria toxins for practical purposes it
has been found advisable to allow them to "season," that is to stand
for prolonged periods until they have reached a state of "equili-
"II and III are taken from the article by Rosenau, P. H. & M. H. S.,
Hyg. Lab. Bulletin, 21, 1905.
12 This statement plainly contradicts the definition of a toxin unit; i. e.,
the amount which neutralizes 100 toxin units and often leads to confusion
among students or others who are unfamiliar with this subject. It should
be borne in mind that, while the definition of an antitoxin unit is the one
accepted when Ehrlich first arbitrarily established it, the antitoxin unit,
as at present in use, is really an amount of antitoxin standardized against
L+ quantities of toxin, this last value again obtained by measurement
against the original unit. It represents a neutralization value equal to the
original one, but may protect the guinea pig against 85, 110, 130, etc.
(variable) toxin units, according to the constitution of the particular toxic
filtrate employed in the experiment. Indeed, if, in the following pages, the
reasoning of Ehrlich is consistently adhered to, our definition of an anti-
toxin unit should be: The amount of antitoxin which contains 200 binding
affinities for toxin. This will become clearer as the following paragraphs
are read.
TOXIN AND ANTITOXIN 111
brium" at which the conversion of toxin to toxoids has been reduced
to a minimum and the change of relationship between L0 and aT"
or M L D has practically ceased to go on. From the very begin-
ning of the growth of the culture in the incubator the process of
toxoid formation has probably occurred, and even freshly prepared
toxic filtrates therefore are not pure "toxins," especially since the
conversion of toxin to toxoid seems to diminish in velocity as time
goes on.
^sTow in spite of the presence of such alteration products, in com-
paring the values L0 and L+ of any given toxin preparation, one
would naturally suppose that L+ minus L0 should be equal to one
M L D, or the quantity just sufficient to kill a guinea pig of 250
grams in 4 to 5 days. For we have seen that L0 just neutralizes
one antitoxin unit while L+ is the quantity which, in addition to
such neutralizing power, has an excess of toxin equal in action to
one minimal lethal dose. This, however, is not the case. Let us
illustrate this by a concrete case. One of Ehrlich's toxins on meas-
urement showed a minimal lethal dose or M L D of 0.0025 c. c.
The L+ dose of this was . 25
while The L o dose of this was . 125
The difference was .125 or 50 M L D instead of 1 M L D as we
would suppose.
Stated in words, this measurement means that after neutralizing
completely one antitoxin unit with the toxic filtrates, in order to
obtain death in a guinea pig in 4 days with such a mixture, it was
necessary to add, beyond the neutralizing quantity, 50 M L D, or
again as much as was necessary for neutralization.
This last relation is merely coincidence, since it might have been
30 or 40 or 60 M L D just as well. The important point is the fact
that more than 1 M L D was necessary, and by this fact Ehrlich was
led to resort to an assumption which forms one of the basic princi-
ples of many of his explanations for serum phenomena, namely, the
assumption of differences in combining avidity or affinity.
As applied to the present problem he reasoned as follows:
It is conceivable that the toxoids resulting from deterioration of
toxin might possess three different degrees of affinity for the anti-
toxin. They might have a stronger, an equal, or a lesser affinity than
the toxin itself. If their affinity for antitoxin were equal to that of
toxin they would, of course, not influence the L+ dose itself; if
stronger than toxin their influence would be so exerted that toxin
would be forced out of combination with antitoxin, giving place to
the toxoid, and the effect would be the opposite from that experi-
mentally observed. If, however, their affinity for antitoxin were
weaker than that of toxin each additional toxin unit added to the
L0 dose would unite with antitoxin, replacing a corresponding quan-
INFECTION AND RESISTANCE
tity of the toxoid of weaker affinity. In consequence, as far as the
poisonous properties of the mixture are concerned, the addition of
toxin would not render the neutral mixture poisonous for guinea
pigs until the toxoids had been completely displaced from union with
antitoxin. Finally, after all the antitoxin had been bound to un-
changed toxin, further addition would then result in the presence of
free toxin and poisonous properties would again appear in the mix-
ture. Ehrlich at first spoke of the toxoids possessing weaker affinity
for antitoxin than the toxin itself as "epitoxoids." This conception
can be rendered clear by the following example :
In the case cited above we had
TorMLD = 0.0025 c. c.
L+ = 0.25 c. c.
Lo = 0.125 c. c.
The difference = 0.125 = 50 M L D.
Supposing that the toxoids (epitoxoids) present in the mixture
possessed an affinity for antitoxin less than that of toxin, the follow-
ing conditions might obtain :
151 toxin-antitoxin -|- 49 epitoxoid-antitoxin = L0.
Add 1 M L D or T and we have:
152 toxin-antitoxin + 48 epitoxoid-antitoxin + 1 epitoxoid free.
Add 2 M L D or T and we have:
153 toxin-antitoxin + 47 epitoxoid-antitoxin -f 2 epitoxoid free until,
finally, adding 50 T, we get:
200 toxin-antitoxin + 49 epitoxoid free + 1 toxin free = L+.13
Later experience led Ehrlich to abandon the opinion that the
epitoxoids were deterioration products of the toxin. He found that
the relation between L0 and L+ which we have just outlined, was
demonstrable in the same way, in freshly prepared toxin filtrates,
in which there had been little time for toxoid formation. He further
13 An example identical in significance with the one just given, but some-
what simpler in its arithmetical conditions, is here added for the sake of
permitting no possibility of unclearness. This example is taken from Ehr-
lich's own work.
T = .01 c. c. of the toxin bouillon.
L+ (neutral, of antitoxin unit yet killing 1 pig) = 2.01 c. c. or 201 T.
Lo (complete neutral, of 1 antitoxin unit) = 1 c. c. or 100 T.
Difference = 1.01 c. c. or 101 T
100 toxin-antitoxin + 100 epitoxoid antitoxin = L0;
Add 1 T, and we have:
101 toxin-antitoxin + 99 epitoxoid-antitoxin + 1 epitoxoid free;
Add 101 T, and we have:
200 toxin-antitoxin +100 epitoxoid free -f 1 T free = L+.
TOXIN AND ANTITOXIN 113
noticed that, even after deterioration had occurred to a considerable
extent, and the amount necessary to kill a guinea pig had been much
increased (the number of fatal doses in L0 constantly decreasing as
toxoids replaced toxin), the L+ nevertheless remained unchanged.
This, he held, could mean one thing only. The elements present in
toxic broth which possessed a weaker affinity for antitoxin than the
toxin itself, namely, the epitoxoids, were present from the very begin-
ning and were probably separate and primary secretion products of
the diphtheria bacilli, remaining relatively stable and constant as the
broth was preserved. In order to avoid confusion, therefore, he now
referred to the "epitoxoids" as "toxons" — to preclude their confu-
sion with the other toxoids or true toxin deterioration products.
These toxons possess, according to Ehrlich, a "haptophore" group
identical with that of the toxin, but have a different toxophore group.
For there is reason to believe that the toxon, lacking the power of
causing acute death, gives rise to slow emaciation and paralysis,
finally killing after a subacute or chronic course. Thus, in the tab-
ulation just preceding, we have seen that the toxic broth added to
neutral mixtures of toxin and antitoxin (containing the L0 dose),
does not give rise to the acutely toxic effect of one M L D or T until
an amount has been added which considerably exceeds one toxin
unit. This, we explained, by Ehrlich's reasoning, on the supposi-
tion that "epitoxoids" or "toxons" are displaced from their union
with antitoxin, giving place to toxin and becoming free. Such toxin-
antitoxin mixtures — in which the amount of toxin broth ranges be-
tween the L0 and the L+ doses — therefore, contain no free toxin
units but contain varying amounts of free toxon. An injection into
guinea pigs is not followed by acute death in these cases, but leads
with considerable regularity to emaciation, paralysis, and death
after a long incubation period.
It has been objected to this, as we shall see, that the slow poison-
ing produced by such mixtures is due, not to a qualitatively differ-
ent poison but to the presence of minute quantities of free toxin too
small to produce acute death, yet sufficient to produce this gradual
injury. This Dreyer and Madsen 14 tried to disprove by experi-
ments in which they prepared antitoxin-toxin mixtures so bal-
anced that they possessed the toxon effect, and of these mixtures
injected increasing multiples. In no case did they succeed in pro-
ducing acute death even when the amount injected had been multi-
plied to such an extent that free toxin, if present, must have asserted
itself. The same workers were able to show that the injection of
these mixtures, in which free toxons were assumed to be present,
produced immunity against toxin, thus indicating the similarity of
the haptophore group of toxin and toxon — a conception which will
14 Dreyer and Madsen. ZeitscJir. /. Hyg., Vol. 37, 1901.
INFECTION AND RESISTANCE
become more and more clear as we consider the "Side-Chain Theory"
which Ehrlich evolved as a result of his toxin analysis.
Ehrlich had thus elicited facts which seemed to him to indicate
the presence of three qualitatively different substances in toxic fil-
trates of diphtheria cultures. Two of these, the toxin and the toxon,
were present, he assumed, in freshly prepared filtrates, as indepen-
dent primary secretion products of the bacilli, the toxin an acute
poison, the toxon a substance with slower and qualitatively different
poisonous effects. Both of them, toxin and toxon, possessing similar
haptophore groups, could unite with antitoxin and neutralize it, but
the affinity of toxon for antitoxon was weaker than that of toxin. For
this reason toxin could displace toxon from its union with antitoxin,
this accounting for the discrepancy between the L+ and the L0
doses. The third class of substances, the toxoids, were deterioration
products of toxin, the deterioration implying an alteration in the
toxophore group only, the haptophore group remaining the same.
It is plain from this reasoning that Ehrlich's conception implies
complete analogy between chemical reactions in general and the
neutralization of toxin by antitoxin. Accordingly it is but another
step in the same direction to speculate concerning the actual rela-
tions of valency existing between the two substances. It seemed to
Ehrlich that there were many reasons for assuming that the union
between toxin and antitoxin occurred in proportions of 200 to 1 ;
that is, just as the formula for water is H2O, that of toxin-antitoxin
combinations would be "Toxin20oAntitoxin."
The considerations on which this opinion was based were as fol-
lows: In examining a large series of toxic filtrates, Ehrlich,15 as
well as Madsen, had found that the number of toxin units ("T" or
M L D) necessary to neutralize one antitoxin unit (that is, the
number of toxin units contained in the L0 dose) corresponded,
with great regularity, to multiples of 100. Values of 25, 33, 50,
100, etc., recurred again and again. This indicated that the de-
terioration of the toxin into toxoids followed a certain regularity o*f
progression and seemed to justify the assumption that the absolute
binding power possessed by antitoxin was represented by a valency
corresponding to a multiple of 100. Since the number of toxin units
contained in an Lo dose rarely fell below, and usually above 100, the
valency could not be less than 100. On the other hand, repeated
measurements of L0 and L+ doses never showed as many as 200
toxin units. Ehrlich's own highest value was 133, and the highest
ever obtained by any one was a measurement by Madsen of 160.
Now considering the fact that no toxin is "pure" but that, in every
case, it contains admixtures of toxoid and toxon, the values 133 or
160 cannot represent all the valencies of an antitoxin unit. They
represent only that part of the antitoxin unit which is neutralized
10 Ehrlich.. Deutsche med. Woch., No. 38, 1898, Vol. 24,
TOXIN AND ANTITOXIN 115
by the "toxin," as measurable upon guinea pigs, a certain fraction
of antitoxin being united to toxoid or toxon. It is likely, therefore,
as Ehrlich reasoned, that, being higher than 100, and in an ob-
viously impure condition approaching but never reaching 200, the
valency of antitoxin for toxin was just 200. The correctness of this
surmise seemed rendered more probable by Ehrlich's further studies,
since analysis of the ingredients of various toxic filtrates, that is, the
determination of the relative contents of toxin, toxoids, and toxon,
appeared to show constantly definite relations to the valency 200.
The method by which Ehrlich carried out these subsequent stud-
ies is spoken of as the method of "Partial Absorption." In prin-
ciple it represents a reversal of his earlier methods of measurement.
In these he had titrated various amounts of toxin broth against a
constant quantity (one unit) of antitoxin, establishing the L+ and
L0 values. In the method of Partial Absorption, on the other hand,
he measured varying fractions of an antitoxin unit against a con-
stant amount of toxin, employing for this a previously determined
L+ and L0 dose. A measurement carried out in this way is shown
in the following tabulation in which, at the same time, there is indi-
cated how many valencies each antitoxin fraction represents, on the
basis of an assumed total of 200 for each unit.16
0 antix. unit representing 0 valency + L+ = 85 free T units
.1 antix. unit representing 20 valencies -f L+ = 85 free T units
.25 antix. unit representing 50 valencies + L+ = 60 free T units
.8 antix. unit representing 160 valencies 4" L+ = 10 free T units
.9 antix. unit representing 180 valencies + L+ = 3.5 free T units
1.0 antix. unit representing 200 valencies + L+ = 1 free T unit
It is immediately evident in this table, as it would be evident in
any other citation of similar measurements, that there is no regu-
larity in the progress of neutralization ; or, in other words, that addi-
tion of a definite fraction of antitoxin does not remove the arithmet-
ically corresponding amount of toxic properties from the L+ dose.
The first 0.1 unit of antitoxin in this table has removed no free
toxin whatever. The addition of the next 0.15 of an antitoxin unit,
representing 30 valencies, has removed f-f or T\ of the total toxicity.
Throughout the scale there is not the regular progression of neutral-
ization, multiple by multiple, which would be expected if antitoxin
could be titrated against a pure toxin. This, according to Ehrlich, is
due to the presence of various toxoids which possess varying affinities
for the antitoxin molecule. The first 0.1 of a unit added, in this case,
does not diminish the toxicity of the mixture because it is bound by
"protoxoids" which possess a higher affinity for antitoxin than the
16 This measurement is taken from one cited by Ehrlich in Deutsche med.
Woch., No. 38, 1898, Vol. 24, and is taken literally except for the first value
of 1/10 antitoxin unit, which is inserted to illustrate the formation of pro-
toxoids.
9
116
INFECTION AND RESISTANCE
toxin itself. Next are bound the toxins themselves together with
varying amounts of "syntoxoicls" which possess the same affinity as
toxin. Finally there are left the toxons which possess a lesser affin-
ity than toxins or toxoids, and therefore again have the discrepancy
between the L0 and L+ dose. Ehrlich utilizes this method in order
to determine the composition of the constituents of any given toxic
filtrate and expresses the results in the so-called "toxin spectra."
The construction of these spectra and the principles underlying
the measurements on which they are based are very clearly illus-
trated by Madsen,17 from whose article the following type spectra
are taken : 18
9 » 20 K> 40 M 60 TO 00 so too no
TOXIN SPECTRUM AFTER MADSEN, Ann. de I'lnst. Past., Vol. 13, 1899, p. 57.
This figure represents the diphtheria filtrate composed of 50
valencies of protoxoid, 100 toxin and 50 toxon equivalents. The
measurements in this case may be represented by the following tab-
ulation :
LO -h 1 antitox. unit = 200 valencies = 0 lethal dose
Lo -j- .75 antitox. unit = 150 valencies = 0 lethal dose
L0 + .25 antitox. unit = 50 valencies = 100 lethal doses
Lo -j- 0 antitox. unit = 0 valency = 100 lethal doses
The following diagram, also from Madsen, represents the same
poison after it had deteriorated to ^ its toxic power. Lo, therefore,
would contain only 50 toxic doses.
AFTER MADSEN, Ibid., p. 578.
The measurements corresponding to this table are as follows:
L0 + I . antitox. unit = 200 valencies = 0 lethal dose
Lo 4" .75 antitox. unit = 150 valencies = 0 lethal dose
Lo + .745 antitox. unit = 149 valencies = 0 lethal dose
Lo 4- -74 antitox. unit = 148 valencies = 1 lethal dose
etc. until
17 Madsen. Ann Past., Vol. 13, 1899, p. 576.
18 We have chosen to illustrate the principles of the toxin spectrum from
the article of Madsen rather than from Ehrlich's original work, because the
former presents this difficult phase of the subject in a hypothetical toxin of
extremely simple structure. Some of Ehrlich's spectra constructed from
actual measurements may be found in Deutsche med. Woch., No. 38, 1898.
TOXIN AND ANTITOXIN
117
Lo + -25 antitox. unit = 50 valencies = 50 lethal
Lo -f- 0 antitox. unit = 0 valency = 50 lethal
The following spectrum, the third in Madsen's article, represents
the same toxin described in the preceding spectrum but at a later
period, at which considerable further deterioration had taken place.
The L0 dose now contained but 30 M L D or, in other words, the
amount of toxin contained in the L0 dose was sufficient to kill 30
guinea pigs only.
PKOTOTOXOID
ao 90100,10
AFTER MADSEN, Ibid,., p. 579.
Madsen's description of the method in which this spectrum is
constructed is the following :
LO + f o& of an antitoxin unit kills 0 guinea pig
LO + mi of an antitoxin unit kills 0 guinea pig
L0 +
H*of
an
antitoxin
unit
kills
0 guinea
Pig
L0 +
14D nf
3~OQ OI
an
antitoxin
unit
kills
1
guinea
Pig
L0 +
mof
an
antitoxin
unit
kills
2
guinea
pigs
L0 +
H$ Of
an
antitoxin
unit
kills
3
guinea
pigs
L0 +
M$ of
an
antitoxin
unit
kills
5
guinea
pigs
L0 +
^r of
an
antitoxin
unit
kills
5
guinea
pigs
L0 +
/o8^ of
an
antitoxin
unit
kills
6
guinea
pigs
L0 +
~/Wo of
an
antitoxin
unit
kills
6
guinea
pigs
L0 +
/A of
an
antitoxin
unit
kills
7
guinea
pigs
L0 +
A% of
an
antitoxin
unit
kills
10
guinea
pigs
L0 +
•2^V Of
an
antitoxin
unit
kills
30
guinea
pigs
L0 +
^^ of
an
antitoxin
unit
kills
30
guinea
pigs
The amount of toxon has remained the same in spite of deteriora-
tion. As less and less antitoxin is added, between the values of -J£$
and ^$J of an antitoxin unit, there are now liberated only 5 fatal
doses of the toxin. It is in this zone that deterioration has taken
place, since, in the preceding spectrum, the difference between the
addition of %$-% and -J-J^ of an antitoxin unit represented 25 fatal
doses for guinea pigs. When in this last spectrum the amount of
antitoxin is gradually reduced from 100 valencies to 50 valencies 25
fatal doses are liberated, a quantity corresponding to the similar
zone in the preceding spectrum. Thus in this particular zone of the
spectrum no change has taken place. The same is true of the pro-
toxoid zone.
It is unnecessary to cite a larger number of such measurements
in this place, since the ones given sufficiently illustrate the methods
118 INFECTION AND RESISTANCE
and the conclusions drawn from them. As a result of such experi-
ments Ehrlich concludes:
I. That the diphtheria bacillus produces primarily two kinds
of substances (a) toxin, (b) toxon, both of which bind the antibody.
II. The toxins (and perhaps also the toxons) may deteriorate
and be modified into secondary substances (toxoids) which may be
distinguished by their different degrees of affinity for antitoxin.
III. This classification does not exhaust all possible complica-
tions, since each subdivision of toxin consists apparently of equal
parts of two different modifications which are similar to each other
in their relation to antitoxin but differ in varying resistance to in-
fluences of deterioration. A more complete analysis of these condi-
tions may be found, together with a series of illustrative spectra, in
Ehrlich's article in the Deutsche med. Wochenschr., Sept., 1898,
which has been quoted above.
The complex deductions arrived at by Ehrlich are largely de-
pendent, as we have seen, upon strict adherence to the analogy be-
tween the toxin-antitoxin reactions and those occurring between
strong acids and strong bases. In such cases there is a complete
reaction, in which chemical change ceases only when there has been
a complete neutralization of one by the other. If, for instance, we
mix molecular equivalent amounts of H2SO4 and NaOH, an ap-
parently complete change into Na2SO and H2O occurs:
H2S04 + 2 NaOH = Na2S04 + 2 H20
The reverse process does not seem to take place, and if traces
of uncombined H2SO4 and NaOH are present, as may be theoret-
ically assumed, they are so slight in amount that they are not dem-
onstrable. There are, however, many chemical reactions in which
the process is not a complete one, in that the chemical change does
not proceed until the reagents are completely used up. Reaction in
these cases ceases when an equilibrium is reached at which there are
present definite amounts of the reaction products and of the original
substances at the same time.19
This occurs when a weak acid is added to a weak base. In such
cases the reaction is incomplete and reversible and, together with the
neutralization products, both free acid and free base may be present.
The conditions are best explained by citing an example of a reversi-
ble reaction which is commonly given in text-books of physical chem-
istry, namely, the reaction between ethyl-alcohol and acetic acid.
(Our citation is taken from Philip's "Physical Chemistry," London,
Arnold, 1910) : "When one gram mol. of ethyl alcohol is added to
one gram mol. of acetic acid, a reaction takes place which results in
19 See Cohn. "Vortrage f . Artze iiber Physik. Chem.," Engelman, Leip-
zig, 1901.
TOXIN AND ANTITOXIN 119
the formation of ethyl acetate and water; the reaction, however, is
incomplete and stops at an equilibrium point at which the reaction
mixture contains % gram mol. alcohol, % gram mol. acid, % gram
mol. ethyl acetate, and % gram mol. water. If, on the other hand,
1 gram mol. of ethyl acetate is mixed with 1 gram mol. of water, a
reaction sets in which results in the formation of ethyl alcohol and
acetic acid. This change likewise stops in equilibrium at a point at
which the composition of the reaction mixture is the same as that
already stated." The reaction is thus reversible and may be written:
C2H6OH + CHaCOOH ^±1 CH3COOC2H6 + H20
Another example somewhat simpler and more easily brought into
analogy with the toxin-antitoxin reaction is that of the dissociation
of phosphorus pentachlorid into phosphorus bichlorid and chlorin
(see Alexander Smith, "General Chemistry," Century Company, N.
Y., 1911, p. 181).
Here the reaction takes place:
PC15 T-** PCls + C12
Chemical equilibrium is reached when the reaction speed is the same
in both directions, and there will be present, at equilibrium, PC13,
C12, and PC15. Now the "Law of Mass Action" (Guldberg &
Waage) states that reaction goes on at a velocity proportionate to
the concentration of the reacting molecules. It is plain, therefore,
that at the point at which the reaction takes place with equal veloci-
ties in both directions, that is, at the equilibrium point, a very defi-
nite relation of molecular concentrations must obtain, and this rela-
tion can be expressed as a formula. For the example given above
this may be written as follows :
Cone. PC13 X Cone. Cl ^ , ,.
7^ s™ = K (constant)
Cone. PCI 5
This formula is expressed in words by Alexander Smith as follows:
"If we change the amount of the pentachlorid placed in the vessel,
or if we use amounts of chlorin and trichlorid which are not equiv-
alent, the numerical value at equilibrium of each concentration will,
of course, be different, but the product of the concentrations of tri-
chlorid and chlorin, divided by the concentration of the pentachlorid,
will always give the same numerical value for (K), the constant, at
the same temperature."
Now to return to the application of these facts to the neutraliza-
tion of toxin by antitoxin, if the reaction is one analogous to that of
a strong acid and alkali, as cited above in the case of H2SO4 and
120 INFECTION AND RESISTANCE
NaOH, we would expect a complete neutralization of one by the
other, multiple for multiple, and the explanation of Ehrlich based
on the assumption of different toxin constituents, of varying affin-
ities, and different pharmacological effects, is the only one which will
account for the experimental results. Assuming, however, that the
reaction is one analogous to that taking place between a weak acid
and a weak base — such as boric acid and ammonia — we have an en-
tirely different state of affairs. For here the reaction goes on to a
point of equilibrium, and in mixtures containing equivalent amounts
of the weak acid and the base there will be present the reaction prod-
ucts and also small amounts of unbound free acid and free base.
And according to the law of "Mass Action," the quantities of free
acid and base present will depend entirely on the masses of the
reagents put together. Thus, for each particular mixture of the two,
different quantities of the original substances will be found uncom-
bined, and, furthermore, the gradual addition of one to the other
will not have a neutralizing value in exact proportion to the amount
added. Were the toxin-antitoxin reaction analogous to such chemical
systems, then we could assume that every mixture of the two sub-
stances, whatever the relative amounts, would contain not only the
united toxin-antitoxin molecule, but also varying amounts of disso-
ciated free toxin and free antitoxin, the quantities of each depending,
according to the law of mass action, upon the molecular concentra-
tions obtaining in the individual mixture. This, indeed, is the con-
ception of toxin-antitoxin union formulated by Arrhenius and Mad-
sen.
Arrhenius and Madsen,20 21 bearing in mind these conditions,,
made comparative studies of the neutralization of tetanolysin by its
antilysin on the one hand, and that of ammonia by boric acid on the
other. Ammonia, like most bases, is a hemolytic agent, while boric
acid, unlike stronger acids, has no hemolyzing properties. For this
reason, in mixtures of the two, the toxicity is proportional to the
concentration of free ammonia (though, as Arrhenius states, "a cor-
rection must be made for the lowering action of the ammonium salt,
as indicated by experiments on this action"). Because the reaction
between boric acid and ammonia is reversible, that is, the salt formed
is dissociated by the hydrolytic effect of the water, there is always
present a slight amount of free ammonia even if the largest possible
quantities (to saturation) of boric acid are added. (See Arrhenius,
"Immunochem.," p. 174.) The curve of toxicity indeed descends
as more boric acid is added, but never reaches 0.
By a modification of the formula expressing the law of Mass
Action, Arrhenius and Madsen could calculate the amount of free
20 Arrhenius and Madsen. Zeitschr. f. physik. Chem., 44, 1903, and
Festschrift Kopenhagen Serum Instit., 1902.
21 Arrhenius. "Immunochemistry," Macmillan, N. Y., 1907.
TOXIN AND ANTITOXIN
ammonia present in a series of mixtures in which increasing quanti-
ties of boric acid were added to a constant quantity of ammonia, and
CURVE EEPRESENTING THE NEUTRALIZATION OF TETANOLYSIN BY DIFFERENT QUAN-
TITIES OF ANTITOXIN.
Taken from Arrhenius, "Immunochemistry," Macmillan, 1907, p. 175.
the values so obtained corresponded with much accuracy to those re-
sulting from measurements of toxicity upon red blood cells. The fol-
lowing table taken from Arrhenius and Madsen illustrates this :
TOXICITY (Q) OF 0.1 N. NH3 (1 EQUIVALENT) WITH N EQUIVALENTS OP BORIC
ACID. (Taken from Arrhenius, loc. cit. p. 176.)
n =
Equivalents of boric
acid added
Quantity of free
ammonia — i. e.,
toxicity — observed
q = Ammonia
toxicity calculated
from formula
Aq oba.
0
100
(100)
0.17
85
79
15
0.33
69
64
16
0.67
43
42
26:2 = 13
1
25
27
18:2 = 9
1.33
20
18
5:2]
1.67
13
13
7:2 }2.5
2
10
10
3:2 J
Here, in the last column, there is indicated the proportion of
toxicity which is neutralized by the successive addition of % of an
equivalent of boric acid. The first additions lower it to a degree
proportionate to the amount of acid added; the next additions neu-
tralize it to a much slighter degree, and, as further additions are
132 INFECTION AND RESISTANCE
made, each successive one possesses progressively less relative neu-
tralizing power than the preceding.
This, it is plain, is closely analogous to the phenomena observed
by Ehrlich in his "Partial Absorption" method, and Arrhenius con-
cludes that the two phenomena, toxin-antitoxin and boric acid-am-
monia neutralization, are closely analogous. His point of view is
further strengthened by his experiments with tetanolysin and its
specific antibody, in which he constructed a curve similar to that
given for boric acid, derived a formula and found that the observed
and the calculated values closely coincided for various mixtures of the
two. He claims, in consequence, that the phenomena observed by
Ehrlich should not be interpreted as due to "partial toxins" — toxoids
or toxons, but dependent rather upon the presence of varying quan-
tities of free toxin dissociated from union with antitoxin because of
the reversibility of the union.
The opinions of Arrhenius and Madsen are not generally ac-
cepted. It is in the first place doubtful whether substances like toxin
and antitoxin, which, as far as we know their chemical nature at all,
belong to the class of substances spoken of as colloids, can be re-
garded as subject to the laws of Mass Action in their reactions.
Nernst 22 has criticized Arrhenius' deductions chiefly on the
basis of their assumption of the reversibility of the union of toxin
and antitoxin, since reversible reactions between colloids, though not
at all inconceivable, have so far not been definitely shown. Further-
more, as Nernst states, if complete reversibility of such reactions
were possible it would be hard to understand how antitoxin can pro-
tect the animal against the actions of toxin.
Another point of view concerning the toxin-antitoxin union which
has been gaining ground especially through the work of Landsteiner
and his pupils, is that of Bordet.23 Bordet expresses his views in
the following way:
I. When one mixes with a certain quantity of toxin an amount
of antitoxin which is insufficient to produce a complete neutraliza-
tion, the molecules of antitoxin are not taken up by a definite frac-
tion of the toxin molecules, satisfying the affinities of these entirely
while other units remain intact ; on the contrary, the antitoxin mole-
cules distribute themselves equally upon all the toxin molecules pres-
ent, and these are therefore, all of them, partially saturated, and
lose proportionately a part of their initial toxicity. One could say
that there is an attenuation of the toxin since there is a formation of
a less poisonous complex.
II. The symptoms of poisoning produced by such a complex in-
jected into animals or placed in contact with sensitive cells cannot be
22 Cited from Landsteiner in "Kolle u. Wassermann Handbuch," 2d Ed.,
Vol. 5.
23 Bordet. Ann. de I'Inst. Past., Vol. 17, 1903.
TOXIN AND ANTITOXIN 123
identical with those which would be produced by a fully saturated
mixture of toxin and antitoxin, or by intact toxin.
III. Between these two extremes, free toxin and entirely neu-
tralized toxin, one can imagine many transitions, progressive stages
of attenuation. Every time that one mixes toxin and antitoxin in
the same way one attains the same degree of attenuation.
Briefly put, this means that Bordet estimates toxin-antitoxin
combinations of different degrees of toxicity as representing differ-
ent stages in the completeness of the saturation of the individual
toxin units. When 10 parts of toxin are added to 1 part of anti-
toxin, the result, according to him, would not be such that 1 part is
neutralized by 1 part of antitoxin, leaving 9 parts of toxin free. He
assumes rather that each unit of toxin is attenuated by the absorp-
tion of T^th of a part of antitoxin. He compares this process to the
action of iodin upon starch. Starch can absorb variable quantities
of iodin and, according to the amount taken up, is colored slightly
or deeply blue. This mode of action is common to most staining
processes. The substance that is stained fixes varying quantities of
coloring matter and the coloring matter does not limit itself to a
definite fraction of the substance stained but distributes itself equally
to the material, coloring it slightly or deeply, in its entirety, accord-
ing to the relative amount of color added. We will see later that
there are many reasons for regarding other antigen-antibody com-
binations as following similar laws of proportion.
Bordet and others speak of this point of view as the "Absorption
Theory," and Biltz, in studying this point of view by physical
methods, comes to the conclusion that the observed figures of the
quantitative relations between toxin and antitoxin in the process of
neutralization are fairly consistent with the values to be expected if
the process were actually an absorption phenomenon.
A curious occurrence which seems to bring the toxin-antitoxin
reactions close to colloidal reactions in general is that which is
known as the "Danysz 24 Effect" or as the "Bordet 25-Danysz Phe-
nomenon." Danysz discovered that when ricin or diphtheria toxin
were brought into contact with their Homologous antibodies the de-
gree of neutralization depended upon the manner in which the two
were put together. When the toxin was added to the antitoxin in
two fractions, a considerable time being allowed to elapse between
the additions, the final mixture was much more toxic than when the
total amount was added at once. In other words, although both
mixtures contained exactly the same quantities of the two reacting
substances, nevertheless the amount of toxin left free varied in the
two cases, according to the speed with which they had been put to-
24 Danysz. Ann. de I'Inst. Past., Vol. 16, 1902.
25 Bordet. Ann. de I'Inst. Past., Vol. 17, 1903.
INFECTION AND RESISTANCE
gether. This was confirmed in 1904 by von Dungern 26 for diph-
theria toxin, and Craw 27 was able to observe it in the case of mega-
theriolysin and its antilysins.
Von Dungern interpreted this in the sense of Ehrlich, by assum-
ing it to be due to what he calls "epitoxonoids." This epitox-
onoid he assumes to be a constituent of toxic broth, which has still
less affinity for antitoxin than the toxon. It can combine with diph-
theria antitoxin, but not until all the true toxin is bound. However,
when it is once united with diphtheria antitoxin it is not very easily
displaced from the union, especially when a considerable time has
elapsed since the union. Therefore, he thinks, when the toxin is
added to the antitoxin in two fractions, this epitoxonoid is bound and
keeps the toxin, which is added later, out of combination. Whereas
if the toxic broth is added as a whole, it is the epitoxonoid which is
left unbound. This explanation of von Dungern's may be looked
upon as an ingenious refinement of the reasoning introduced by Ehr-
lich into this field. As a matter of fact reactions similar to the
Danyz phenomena have been very commonly observed in the reac-
tions between various colloids.
THE SIDE-CHAIN THEORY
Mechanism of Antibody Formation
The discovery of antitoxins in the blood serum of toxin-immune
animals by Behring and his collaborators furnished a point of new
departure for the investigation of the phenomena of immunity, and
Ehrlich' s work upon the nature of the reaction between toxin and
antitoxin, both in vitro and in the animal body, firmly established
that the protective effect of the latter was one of direct neutraliza-
tion, and not, as at first supposed, one of toxin destruction or of
indirect influence through the mediation of the body cell. As we
have seen, moreover, it was quickly noted that these reactions Were
strictly specific in that an antitoxin produced with any one of the
known toxins reacted solely with this one to the exclusion of all
others. All these facts were of the utmost practical importance and
gave hope of ultimate extensive therapeutic application, a hope
which has, in part, been realized.
The physiological mechanism by which these phenomena were
brought about, however, was, and is, to a great extent still, a mys-
tery, and a most extensive and painstaking series of researches has
occupied itself with its elucidation.
26 Von Dungern. Deutsche med. Woch., 30, 1904.
27 Craw. Jour. Hyg., Vol. 7, 1907.
TOXIN AND ANTITOXIN 125
When we consider the invariable production of a specific anti-
toxin in response to the treatment of an animal with a toxin it is
but natural that Buchner and others should have at first assumed
that the antitoxin is, in each case, a product obtained by the action
of the body tissues from the toxin itself. While difficult to refute
at a time when little was known of the laws of antitoxin production
and of quantitative relationships, such an assumption is entirely un-
tenable in the light of more recent knowledge. We now know that
such a simple conversion of toxin into antitoxin cannot explain the
phenomenon because the amount of antitoxin incited in the immu-
nized animal is out of all proportion great in comparison with the
amount of toxin injected. Thus Knorr 28 has found that 100,000
units of antitoxin may be produced by the injection of the toxin
equivalent of one unit. Moreover the discovery by Salomonsen and
Madsen 29 that pilocarpin injections will increase the amount of
antitoxin produced by an animal distinctly pointed to the likelihood
of the participation of the general physiological activities of an
immunized subject in the production of antibodies. Unquestionable
proof of this was also brought by the experiments of Roux and Yail-
lard,30 in which antitoxin production in immunized animals con-
tinued even after the entire volume of blood had been removed by
repeated bleeding. This observation points distinctly to the direct
secretion of antibodies by the tissue cell, in the nature of what has
been termed by Roux 31 an "internal secretion.77 And it is this
activity of the body cell in the production of antibodies which forms
the fundamental premise from which the now classical "Side-Chain
Theory7' of Ehrlich takes its departure.
In order to approach this theory logically it will be of advantage
to consider briefly the general subject of the assimilation of food-
stuffs and other substances distributed by the circulation to the cells
of the animal body. For, as Ehrlich has expressed it, "The Reac-
tions of Immunity, after all, represent only a repetition of the
processes of normal metabolism, and their apparently wonderful
adjustment to new conditions is only another phase of 'Uralter
Protoplasma Weisheit.7 77 32 It is impossible to conceive the nutri-
tion of body cells without assuming that the assimilable nutritive
substances come into physical and, eventually, chemical relationship
with the protoplasm of the nourished cell. Considering the large
variety of substances which may thus be brought into contact with
cells in the course of normal and abnormal metabolism, the body cell,
28 Knorr. Munch, med. Woch., 1898, pp. 321, 362.
29 Salomonsen and Madsen. Ann. de I'Inst. Past., Vol. 12, 1898.
30 Roux and Vaillard. Ann. de I'Inst. Past., Vol. 7, 1893.
31 Roux. Ref . in Semaine Medicale, 1899.
32 Ehrlich. Introduction to "Gesammelten Arbeiten," Berlin, Hirsch-
wald, 1904.
126 INFECTION AND RESISTANCE
chemically considered as a complex of enormous molecules, must
possess a correspondingly great variety of atom groups, by means of
which it can unite with these substances to assimilate them and make
them a part of its own protoplasm. In order to enter into similar
relationship with toxins and other antigens, then, it is only logical
to suppose that the cell, in the same way, unites chemically with the
antigenic substance, and either assimilates it without sustaining
harm, as in the case of non-poisonous complexes, or is injured in the
process, as in the case of the poisons.
The living cell, from this point of view, is conceived as consist-
ing of a central chemical nucleus, the "Leistungskern," more or less
stable, in that the specialized tissue function is dependent upon it,
and a manifold variety of "side chains," or atom groups by means
of which it can enter into relationship with the nutritive and other
materials carried to it by the body fluids. The latter term, "side-
chains," is taken from the nomenclature of chemistry, and, although
the analogy is a loose one, it serves satisfactorily to elucidate Ehr-
lich's meaning. Thus we may conceive the "Leistungskern" as the
central carbon ring of any compound of the Benzol series, as, for
instance, in salicylic acid in which the hydrogen atoms, the hydroxyl,
OH OH
°
H— C C— COOH ( J C02CH,
H— C C— H
\ /
Methyl salicylate
Salicylic acid
|
and the acid radicles represent "side chains." By means of the lat-
ter the compound can enter into relation with other substances, as,
for instance, with CH3 in the formation of methyl salicylate.
Graphically, though this analogy formulates Ehrlich's fundamental
conception, it must not be taken as too literally representing the
existing conditions, since, in actual metabolic interchange, there is
an infinite variety of possible "side-chain" groups ; for we are deal-
ing with an enormous number of assimilable substances, most of
them of chemically unknown constitution. The cell, therefore, is
looked upon as an active chemical complex, retaining its own peculiar
functional characteristics by reason of the "Leistungskern," but
constantly getting rid of waste products and entering into new union
with extraneous materials by virtue of its "side chains." These side
chains, because of their "receiving" function, are spoken of by Ehr-
lich as cell "receptors."
11
J.1
li
f
9!
I
128 INFECTION AND RESISTANCE
That the chemical structure of certain bodies determines their
ability to enter into relation with cell derivatives such as enzymes
is, of course, a fact well established by experiment and explains the
specific action of bacterial and other ferments upon certain sub-
stances to the exclusion of others. Thus Pasteur noted the fact that
bacterial ferments could decompose dextrorotatory tartaric acid
while they did not affect the levorotatory variety, and Emil
Fischer 33 showed that only those carbohydrates possessing 6 and 9
carbon atoms were subject to fermentation by yeasts, and of these
only the ones belonging to the "d" series, observations which, by
demonstrating the relationship between these active agents of extra-
cellular digestion, and the stereochemical configuration of the mole-
cules acted upon, lend much support to the logic of Ehrlich's con-
tentions.
Moreover, the recent experiments upon the growth of tissues in
plasma outside of the animal body in which cartilage cells produce
cartilage, kidney cells, etc., have shown that, given the same nutri-
tive materials, the cells themselves must command a certain selective
power in the choice of these materials, which can only depend upon
a specific element in the structure of the cell receptors. As Fischer
has expressed it for fermentation, the ferment must possess an atom
group which fits into some group of the fermentable substance as a
"key does into a lock," an analogy which is equally applicable to
Ehrlich's conception of the relation of the "side chain" to a nutri-
tive molecule.
Now the toxins and other antigens are, all of them, so far as we
know, complex chemical substances, derivatives of animal and vege-
table cells, and, for this reason, should have much in common with:
the materials available for nutrition. It is not strange, therefore,;
that, coming into contact with the cells of the body during the acci-
dents of disease or other abnormal conditions, they should find re-%
ceptors by means of which they can combine with the cell. Under
the ordinary conditions of nutrition a suitable particle taken up by
the cell in this way would be assimilated and the receptor either
freed for further use or regenerated for the further absorption of j
similar substances, by virtue of a mechanism delicately coordinated
to the needs of cell-nutrition. In the case of the absorption of sub-
stances belonging to the class of antigens, however, foreign proteins
difficult of assimilation, or of toxins even directly harmful, the re-
ceptors occupied by these substances are rendered useless to the cell,
and, if the cell continues to live, must be regenerated. If the degree;
of poisoning or the amount of other antigen introduced has been
extremely slight, this regeneration may possibly take place, as in
the course of nutritive processes, without further disturbance. If,
however, the amounts of antigen are greater than this, or are repeat-
33 See Oppenheimer, "Die Fermente," Vol. 1.
TOXIN AND ANTITOXIN 129
edly thrust upon the cell, the process of regeneration may be not only
sufficient to compensate for the loss of the eliminated receptors, but
may follow the general law of overcompensation, formulated by Wei-
gert, and receptors of the variety occupied by the antigen are pro-
duced in excessive number.
Here again Ehrlich has called analogy to his aid, and has taken
his conception of "overcompensation" from the well-known phe-
nomena of pathological anatomy where, for instance, in the restora-
tion of cellular elements after injury, there is often an overpro-
duction of granulation tissue, far beyond the needs of simple healing.
Thus the restitution of cell receptors, if sufficiently stimulated
by large quantities or repeated administration of the antigen, far
exceeds the quantity normal to the cell, and may proceed to such a
degree that the cell, becoming as it were "top-heavy" with these
elements, sloughs them off into the surrounding lymph and blood,
where they circulate as free receptors. These free receptors then,
having specific affinity and combining power for the antigen which
incited their production, unite with subsequently introduced antigen
in the blood stream, diverting it from the cells themselves, and, in
the case of the variety of antigens spoken of as toxins, this union
with the free receptors in the blood stream would serve to protect
the cells from harm, exerting thereby an antitoxic action.
The antibodies appearing in the blood of immunized animals,
therefore, represent atom complexes, normally parts of the body cells
and concerned in the metabolic processes, but now produced in ex-
cess and extruded into the body fluids under the influence of the
stimulation of immunization. The very substances, as Behring has
put it, which make possible the poisoning of the cell by the toxins be-
come protective wThen, detached from the cell, they circulate in the
blood. Thus the theory, beside explaining the causes leading to anti-
body formation, offers a plausible reason for the relatively strict
specificity observed in antibody-antigen reactions.
Formulated in direct connection with the investigations upon
toxins and antitoxins, the side-chain theory has been extended by
Ehrlich and his associates to all known phases of antibody-antigen
reactions. The differences in the nature and complexity of various
antigens would naturally necessitate variation in the receptors capa-
ble of assimilating them, and these receptors, appearing subsequently
in the blood as antibodies, must, of necessity, differ from each other.
On this basis Ehrlich has conceived of three main varieties or "or-
ders" of receptors or "haptines," as he calls them. Of these the
simplest are those of the first order which attach to the toxins, and
by over-regeneration appear in the blood stream as antitoxins. Those
of the second order, adapted to the assimilation of more formidable
protein molecules, are, of necessity, of greater structural complex-
ity, appearing in immunized animals as the agglutinins and precip-
130 INFECTION AND RESISTANCE
it ins, while those of the third order, dependent upon the coopera-
tion of alexin or complement, for proper functionation, appear as
the cytotoxins or lysins. The detailed structure of these various
haptines will be discussed in connection with other considerations
dealing with their special reactions.
Limiting ourselves, for the present, to a broad consideration of
the theory as a whole, it may be briefly recapitulated as follows:
Toxins or other antigens, in order to exert any influence upon the
animal body, must enter into chemical relationship with the cells.
This they do by virtue of union with chemical units or atom groups
of the cells, spoken of as "side chains." These side chains or recep-
tors, thrown out of function by this union, and necessary for the
metabolic processes of the cell, are regenerated, and under the influ-
ence of repetition of this process are produced in excess, to such a
degree that they are eventually thrown off by the cells and enter the
circulation as antibodies. Thus far the theory, comparing the union
of antigen with cells to the processes of nutrition, is eminently log-
ical and likely, necessitating the assumption of over-regeneration as
the only criterion not directly amenable to experimental proof.
That the antigen can be bound by the body cells has been vari-
ously shown in a large number of investigations, some of which have
been reviewed in our section on the action of bacterial poisons. We
have there seen that Donitz demonstrated the rapid disappearance
of tetanus and diphtheria toxins from the circulation of susceptible
animals, and that conversely Metchnikoff showed that the poison may
persist unabsorbed and unchanged for weeks and months in the
blood of such insusceptible animals as the turtle and the lizard,
facts which furnish indirect evidence of the absorption of the toxins
by the body cells. More direct evidence has, of course, been possible
in the test tube experiments with hemolytic and other cell poisons
where a directly specific combination between antigen and antibody
has been easily demonstrable. Thus, in his earlier experiments with
spider poison, Sachs was able to show that rabbit erythrocytes, which
are sensitive to the poison, could absorb it out of solution, while dog
and other corpuscles, which were insusceptible to the poison, did not
bind or absorb it. This can be easily demonstrated for many anti-
gens and antibodies and may be accepted as a fact.
This point established, and repeatedly confirmed, and the origin
of antitoxins from the cells of the body having been rendered likely
by the experiments of Salomonsen and Madsen, and by those of Roux
and \7aillard just cited, it would follow, by the theory of Ehrlich,
that we should find the site of antibody production in the very cells
which possessed specific affinity (receptors) for the antigen. This
question has been variously investigated, chiefly in the case of the
toxins and antitoxins, since this phase of the subject is most easily
amenable to experiment. It will be remembered also that Wasser-
TOXIN AND ANTITOXIN 131
mann and Takaki discovered that emulsions of the tissue of the cen-
tral nervous system of rabbits and guinea pigs, shown by Meyer and
Ransom and others to be the special points of attack for tetanus toxin,
possessed the power of neutralizing this poison in vitro, while emul-
sions of spleen Jddney and other organs had no such effect. They
assumed from this that the poison was fixed by cell receptors, ante-
cedents of antitoxin in the sense of Ehrlich. Kempner 34 35 made
similar observations with botulinus toxin and further confirmation
has been derived from experiments like those of Blumenthal,36 who
found that the toxin was neutralized by the brain tissue of susceptible
animals but showed conversely that the brain substance of the
chicken, an animal but slightly susceptible to tetanus, possessed little
or no neutralizing power. Similar results were obtained by Metchni-
koff in the cases just cited.
The great importance of these experiments lies not only in show-
ing that body cells may absorb the toxins, but that there is direct
relationship between the susceptibility of tissue and the toxin-bind-
ing properties. Furthermore the facts demonstrated by Metchnikoff
that no antitoxin was produced by those animals (turtle, lizard) in
which the tissues had no power of fixing poison and which are con-
sequently insusceptible, furnish powerful evidence in favor of Ehr-
lich's view.
It becomes of great importance, therefore, to determine whether
in the case of the fixation of tetanus toxin by the brain cells the
union between cell and toxin is a specific and chemical one compar-
able in every way to the union of toxin with antitoxin.
Metchnikoff, in spite of his results in the experiments just cited,
objected to this interpretation on the ground that although the brain
emulsion of a guinea pig neutralized tetanus toxin in vitro, the in-
jection of the toxin into such an animal, subdurally, produced the
disease. This can hardly be regarded as a valid argument against
Wassermann's interpretation, since the very premises of the Ehrlich
theory require that these neutralizing elements, when still attached
to the living cell, as "sessile" receptors, are the cause of the poison-
ing, since they serve to "unlock" the cell to the entrance of the toxin.
Similar objections on the part of Metchnikoff37 were based on some
of his own experiments, as well as on those of Courmont 38 39 and
Doyen, in which it was found that the poison disappears but slowly
(in 2 to- 3 months) from the circulation of frogs, and the brain cells
show hardly any toxin neutralization in vitro, whereas these animals
34 Kempner and Pollak. Deutsche med. Woch., 1897, No. 23, p. 505.
35 Kempner and Shepilewsky. Zeitschr. f. Hyg., Vol. 36, 1901, p. 1.
36 Blumenthal. Deutsche med. Woch., 1898, No. 12, p. 185.
37 Metchnikoff. Ann. de I'lnst. Past., Vol. 12, 1898.
38 Courmont and Doyerf. Arch, de Physiol., 1893.
39 Courmont and Doyen. Compt. rend, de la soc. de biol, 1893.
132 INFECTION AND RESISTANCE
can be rendered tetanic if they are warmed to 25° to 30° C. Further
work, however, by these authors as well as by Morgenroth 40 has
satisfactorily cleared up this difficulty. As a matter of fact, tetanus
poison disappears more rapidly (that is, is bound by the cells more
rapidly) from the circulation of frogs, if the frogs are warmed to
30° C. or more. Furthermore, if the toxin is injected into these
animals, and they are kept at low temperatures, no disease results,
but if they are then warmed up to the temperature stated, they grad-
ually succumb to the disease. Morgenroth has shown that the ap-
parently anomalous behavior of frogs in this respect is actually a
question of temperature. At low temperatures the poison is bound,
though with extreme slowness, but the toxophore group of the toxin
does not functionate. When the animals are warmed, not only does
the binding proceed more rapidly, but the toxophore group becomes
active. He thus not only has answered Metchnikoff's objections to
Ehrlich's theory on this ground, but has furnished an additional in-
direct confirmation of the dual constitution of toxin, that is, its
constitution of a haptophore and a toxophore atom group, suggested
by Ehrlich in his diphtheria-toxin analysis.
There is apparently, then, a strong absorption of tetanus toxin
by the brain and nervous tissue of all animals which are susceptible
to the poison, an absorption which amounts, as we have seen, to neu-
tralization, the brain emulsion acting like antitoxin when mixed
with the toxin before injection, as in Wassermann's and Takaki's
experiments.
A serious objection has been brought, however, to the assumption
that this binding can be identified in its nature with the similar bind-
ing of toxin by antitoxin, and a number of authors have claimed that
the binding by the brain is not a binding by specific receptors, but
an accidental property due to the presence of some fortuitous fixing
substance in the central nervous system. Besredka 41 showed, for
instance, that the brain of susceptible animals could bind much more
toxin than it could actually neutralize, and that, if antitoxin was
added to a brain emulsion previously saturated with the toxin, the
toxin is removed from its combination with the brain cells and these
again regain their original absorbing property. These experiments
would seem to point to a difference, especially in regard to affinity
and firmness of union between the nature of the combination between
toxin and brain emulsion on the one hand, and toxin and antitoxin
•on the other. This, of course, would prove a serious obstacle to the
interpretation of the binding of toxin by susceptible cells in the sense
of Ehrlich, as depending as it were upon union with specific recep-
tors, or, as they might be termed, "sessile" antitoxin. Moreover, to
strengthen such objections to this point of view, the work of Land-
40 Morgenroth. Arch, internat. de Pharm., Vol. 7, 1900, pp. 265-272.
41 Besredka. Ann. de VInst. Past., Vol. 17, 1903, p. 138.
TOXIN AND ANTITOXIN 133
steiner and v. Eisler 42 has brought out the fact that extraction of
brain tissue with ether materially reduces its toxin-binding powers
by removing fatty or lipoidal substances, such as cholesterin and
lecithin. And it has indeed been confirmed that lipoids can possess,
in many instances, binding properties not only for toxins but for
other forms of antibodies. On the basis that at least a part of the
toxin absorption by brain emulsions depends upon such lipoidal fixa-
tion, the results of Besredka are readily explained, but were this the
sole cause of toxin fixation by these tissues it would indeed be diffi-
cult to interpret the phenomenon, with Wassermann and Takaki, in
support of Ehrlich's theory. For, without going into further refine-
ments, the fact of the probable proteid, certainly not lipoidal, nature
of the antitoxins, discussed in a previous section, would alone serve
to distinguish the two modes of toxin fixation.
However, a number of facts have been ascertained which show
that, although the lipoids play some part in the antitoxic action of
brain cells, they do not by any means account for the entire process.
In the first place it is found that the heating of brain emulsions al-
most completely removes their power to bind the toxin, while no
such reduction of the fixative property follows the heating of lipoids
like cholesterin or lecithin. The experiments of Marie and Tif-
feneau 43 have done much to clear up the confusion regarding this
point. They determined that the "lipoidal binding" constituted only
about one-tenth of the total binding power of the brain emulsions,
by showing in the first place that only one-tenth of the total was left
after heating, and that all but one-tenth could be destroyed by sub-
jecting the tissue to the action of proteolytic enzymes. It appears
from this that a large part, at any rate, of the toxin fixation of the
brain tissues is dependent upon substances of an albuminous nature,
a smaller but definite part being dependent upon fixation by lipoids,
a phenomenon entirely apart from the former in underlying princi-
ples. This would, it seems, both justify the original interpretation
of Wassermann and still explain the apparently contradictory results
of Besredka and others.
42 Landsteiner and v. Eisler. Centralbl. f. Bakt., Vol. 39, 1905.
43 Marie and Tiiteneau. Ann. de I'Inst. Past., Vol. 22, pp. 289 and 644,
1908.
10
CHAPTER VI
THE BACTEEICIDAL PKOPERTIES OF BLOOD SEBUM,
CYTOLYSIS, AND SENSITIZATION
IN spite of the profound physiological alteration of the animal
body which is implied by the acquisition of immunity against any
particular infection, we have seen that no anatomical or histological
changes in the organs and tissues accompany such alteration. The
same is true of the difference between animals of different species,
in which the most marked variation in resistance against any given
infection is inexplicable on the basis of structural or microscopic
characteristics in the organs. We have mentioned briefly the at-
tempts that have been made to discover chemical and physical changes
or differences to account for such conditions and have seen that the
attention of investigators was soon attracted to the blood.
A possible relationship between the blood and the defence of the
body against infection had been foreshadowed by observations made
long before the days of bacteriological knowledge. As early as 1792,
John Hunter, in his " Treatise on the Blood, Inflammation and Gun-
shot Wounds," had noted that the blood did not decompose as readily
as other putrescible material, and a century later, during the period
of great interest in the living nature of fermentation and putrefac-
tion, Traube (1874) expressed the opinion that blood could destroy
bacteria. Similar observations were made by Lister and by Groh-
man 1 but no experimental work aimed at this point was carried on
until 1886, when the subject was taken up by Nuttall,2 von Fodor,3
and Fliigge, and a little later by Buchner.4 These authors, working
with defibrinated blood, peptone blood, and blood serum, showed that
such substances all exerted a definitely measurable destructive influ-
ence upon bacteria, and Nuttall, later confirmed by Buchner, further
found that this bactericidal power was weakened on standing, and
could be rapidly destroyed by heating to 60° C.
Their method of procedure consisted in the planting of controlled
amounts of various bacteria in measured quantities of blood and,
1 Grohman. Quoted from Adami, "Principles of Pathology," Vol. 1, p.
497.
2 Nuttall. Zeitschr. f. Hyg., 4, 1888.
3 Von Fodor. Deutsche med. Woch., 1887.
4 Buchner. Centralbl f. Bakt., Vol. 5, 1889.
134
BACTERICIDAL PROPERTIES OF BLOOD SERUM 135
after several hours at 37° C., pouring plates and thus determining
the numbers of surviving organisms. The fact of bactericidal power
established, there was, of course, much early difference of opinion
as to the mechanism responsible for the destruction of the bacteria,
and a number of simple explanations were suggested which, though
entirely refuted at the present time, still possess considerable inter-
est in showing the stages of development through which the concep-
tions of the mechanism of immunity have progressed.
These early theories were formulated chiefly upon the under-
lying thought that the animal body was primarily passive in its rela-
tion to the invading micro-organisms, and that the disappearance of
bacteria in the body fluids was due to the existence of a chemically
or physically unfavorable environment which prevented their multi-
plication and therefore induced gradual mortality among them.
Thus Billroth 5 believed that bacteria could thrive in the body only
after a preceding putrefactive change had prepared a favorable pab-
ulum. Others attempted to discover a relation between the degree
of alkalinity of the blood serum and the destruction of bacteria.-
This argument was soon refuted by the experiments of Buchner, who
showed conclusively that the bactericidal power of serum was not
reduced by the neutralization of its natural alkalinity with weak
acetic acid.
Another theory which has been kept alive until the present day
by Baumgarten,6 and in favor of which much has been written by
Fischer, is the so-called "Osmotic" explanation. The basis of this
conception is the observation that vegetable and other cells, which
are in themselves delicate osmotic systems, undergo changes when
they are placed into fluids of different osmotic tension.7 Thus, of
course, cells of all kinds may be destroyed by being placed in dis-
tilled water on the one hand, or in hypertonic salt solution on the
other. The point of view of Baumgarten, as explained in a recent
edition of his "Text-book of Bacteriology," is the following: The bac-
terial (or blood) cell, like all cells, is surrounded by a semi-per-
meable membrane. Under ordinary conditions, this membrane per-
mits the passage of certain substances which must enter and leave
the cell in the course of normal metabolism. When the bacteria are
placed in a specific bacteriolytic serum there is a chemical union
between the antibody and the cell membrane, and the latter is, in
consequence, injured. The result of the injury is that now the cell
becomes permeable for salts and other substances to which it was
impermeable before, and there are consequent swelling and in-
5 Billroth. Quoted from Sauerbeck, "Die Krise in der Immunitatsforsch.,"
Klinkhardt, Leipzig-, 1909.
6 Baumgarten. "Lehrbuch der pathogenen Mikroorg.," Hirzel, Leipzig,
1911.
7 See also Pfeiffer's "Pflanzen Physiologic."
136 INFECTION AND RESISTANCE
creased intracellular pressure. This, in turn, brings about the ex-
trusion from the cell of proteins and other ordinarily non-diffusible
substances, and destruction of the cell results. This explanation is
practically an adaptation of the. earlier more primitive osmotic the-
ories to the facts subsequently discovered. It stands in direct con-
tradiction to the prevailing opinion that the process of bacteriolysis
and cytolysis in general is an enzymotic process, brought about by
the injury of the cell by specific substances comparable to digestive
ferments.. Interesting' though the suggestion of Baumgarten is, it
can hardly receive more than casual attention given it for the sake
of completeness, since careful experimental work by von Lingel-
sheim8 has shown definitely that altered salt contents of serum
do not exercise the effect upon bacteriolysis which we would be en-
titled to expect from Baumgarten's reasoning.
In explanation of the natural immunity possessed by many ani-
mals against various infections, Baumgarten has offered another
explanation which, like the preceding, we may classify, in agreement
with Sauerbeck,9 with the "passive" theories. This theory, which
he calls his "Assimilation Theory," assumes that the bacteria do not
find suitable food material in the tissues and fluids of certain ani-
mals, and, since bacteria do not have to be killed to be eliminated,
but may be checked merely by their inability to grow and multiply,
they must soon succumb in surroundings in which they find no suit-
able foodstuffs. This point of view approaches somewhat the earlier
exhaustion theory of Pasteur, which has been mentioned in another
place.10
In contrast to these "Passive" theories of immunity are the now
prevailing and well-founded opinions that the resistance of the ani-
mal body against bacterial invasion is not a mere fortuitous result
of chemical and physical conditions encountered by the infectious
agents, but is rather the result of the struggle against the invasion
by active forces of the body cells and fluids. The part played by the
cells had already been emphasized by Metchnikoff and his school
when the discovery of the bactericidal power of the normal blood
was made. The study of the antibacterial powers of the blood now
introduced a new element which became the basis of the so-called
"humoral" theories. In the prolonged controversies waged, with
great astuteness and experimental skill, between the adherents of
these two schools, most of the facts which we possess regarding im-
munity were discovered, and it is only within recent years that we
8 Von Lingelsheim. Zeitsclir. f. Hyg., Vol. 37, 1901.
9 Sauerbeck. "Die Krise in der Immunitatsf orschung," Klinkhardt,
Leipzie, 1909.
10 The influence of foodstuffs, temperature, and other environmental con-
ditions upon natural immunity has been discussed in an earlier section.
BACTERICIDAL PROPERTIES OF BLOOD SERUM 137
have obtained information which has made possible a correlation
between these two main paths of thought.
The humoral theory was conceived by Buchner, as the first
important theoretical result of NuttalFa discovery. Buchner, as
we have seen, confirmed the observations of Nuttall both as
to the primary fact of the bactericidal power of the fresh normal
blood and as to the unstable nature of this bactericidal property.
He looked upon the antibacterial power as depending upon a con-
stituent of the fresh blood plasma, which he named <f alexin' (pro-
tective substance), and which he believed to be comparable to a
proteolytic enzyme. The action of this alexin was conceived as
potent against all bacteria equally, without showing specific selection
of various species to any great extent. The analogy to ferment
action was formulated by Buchner because of the heat sensitiveness
and the instability of the bactericidal substance on standing ; and he
suggested that this alexin might possibly be a product of the tissue or
blood cells, possibly leukocytic in origin.
Buchner found that the action of the ferment-like alexin upon
bacteria was most marked at the temperature of the body, and that
it was capable of destroying bacteria in the subcutaneous tissues and
the serous cavities of the animal body, without the aid or coopera-
tion of cellular elements. He inferred that there was a direct rela-
tion between the potency of the alexin and resistance against infec-
tion.
The next great step in the understanding of the bactericidal
processes was now made by Pfeiffer as a consequence of studies upon
the nature of cholera immunity. Pfeiffer n 12 found that the injec-
tion of cholera spirilla into the peritoneal cavity of a guinea pig
which had recovered from a previous cholera infection was fol-
lowed by a rapid destruction of the bacteria. If small quantities of
exudate were taken out of the peritoneum at varying intervals after
the injection, a granular change and swelling of the bacteria were
noticed, followed, soon after, by complete dissolution and disappear-
ance. Such animals would recover from doses of bacteria which, in
control animals of the same weight, resulted in death. He further
found that the phenomenon was specific, in that the dissolution of
cholera organisms only occurred in the cholera-immune animals,
other bacteria being unaffected. In other words, the guinea pig had
acquired a specific antibacterial power, expressed by the process of
"bacteriolysis," a property possessed to only a very slight extent by
the peritoneal exudate of a normal animal. It was the next logical
step to determine whether the bacteriolytic power could be trans-
ferred to the peritoneal cavity of a normal animal by injecting, to-
gether with the bacteria, a small amount of the serum of such an
"Pfeiffer. Zeitschr. f. Hyg., Vol. 18, 1894; also Vols. 19 and 20.
12 Pfeiffer & Isaeff. Deutsche med. Woch., No. 18, 1894.
138 INFECTION AND RESISTANCE
immune animal. This was indeed found to be the case and, al-
though such immune serum, like normal serum, is deprived of its
in vitro bactericidal power on heating, Pfeiffer found, in his intra-
peritoneal experiments, that heated serum is quite as effectual as
fresh immune serum in transferring passive immunity to a normal
guinea pig. We may summarize the important harvest of facts ob-
tained from these experiments of Pfeiffer in the following state-
ments :
1. Rapid dissolution of cholera spirilla takes place in the
peritoneal cavity of a cholera-immune guinea pig. Similar lysis
takes place not at all, or only to a slight extent, in the peritoneum
of a normal pig. In consequence of the lysis the immune pig will
survive the injection of quantities of bacteria which invariably kill
normal animals of the same weight.
2. The protection obtained in this way is specific.
3. The protection may be transferred from an immune to a
normal guinea pig, by injecting a little immune serum together with
the bacteria into the peritoneum of the normal animal. In a normal
animal so treated lysis is in every way similar to that observed in
the immune pig.
4. The transfer of the lytic power and consequent immunity
can be brought about not only by means of fresh immune serum but
by heated serum as well, although the latter has lost all its alexic
power because of the heating.
Of the phases of this "Pfeiffer phenomenon" the one most diffi-
cult to understand, in the light of the knowledge of that time, was
the transference of the lytic property with the heated serum. Pfeiffer
very naturally took his experiments to signify that the actual de-
struction of bacteria in the animal body could take place entirely
without the phagocytic participation of the body cells, a view in
sharp contrast to that of the Metchnikoff school, and based upon his
observation of the complete extracellular disintegration of the spir-
illa in the peritoneal exudate. He assumed, however, that there was
an indirect participation on the part of the cells. The observation
that heated serum, inactive outside the body, was efficient when in-
troduced into the peritoneum, persuaded him that the cooperation
of the living tissues was a necessary factor, and he assumed a pos-
sible activation by substances derived from the endothelial cells lin-
ing the peritoneal cavity. In the same way he explained his failure
to observe actual bacterial dissolution in hang-drop preparations,
even when fresh serum was used in the experiment.
It will be interesting to examine a protocol of an experiment such
as those carried out in the performance of the Pfeiffer phenomenon
in order to make the actual occurrences entirely clear. In such
experiments the quantity of bacteria used must be chosen with some
regard to the virulence and toxicity of the particular culture em-
BACTERICIDAL PROPERTIES OF BLOOD SERUM 139
ployed, since, as we shall see, protection of animals by bactericidal
or bacteriolytic sera does not follow the law of multiple proportions
as in the case of the protection against toxins by antitoxins. While
the dose of bacteria chosen should be considerably above the minimal
lethal dose for an animal of the weight used, it should nevertheless
be remembered that the bactericidal serum does not possess antitoxic
properties against the poisons liberated or produced as the bacteria
undergo dissolution, and at best the protection by bacteriolysis is
limited to a very definite maximum of bacteria, beyond which no
further increase of serum quantity will avail. The following table
will illustrate an experiment of this kind in which, in a series of
guinea pigs, the bacteriolytic protective power (titre) is determined
by comparative tests.13
PFEIFFER PHENOMENON
Weight
of
guinea pig
Dose of bacteria*
cholera spirilla
Amount of
inactivated
immune serum
Result
(1) 215 gm.
2mg.
0.1 c. c. in 1 c. c.
Complete dissolution in less than
salt solution.
1 hour. Lives.
(2) 230 gm.
2mg.
0.05 c. c.
About the same as first.
(3) 200 gm.
2 ing.
0.01 c. c.
Somewhat slower than in other
,
two; a few unchanged spirilla
after 1 hr. Final dissolution.
(4) 245 gm.
2mg.
0.005 c. c.
Pig lives.
Similar to (3) but complete dis-
solution in 2 hrs. Pig lives.
<5) 220 gm.
2mg.
0.001 c. c.
After 30 min. the spirilla seem
to have begun to multiply.
Dies with innumerable active
spirilla in peritoneum.
Normal
control
(6) 210 gm. !
2mg.
0.1 c. c. normal
Very slight lysis at the beginning.
inactive rab-
Soon rapid multiplication.
bit serum.
Dies.
"The bacteria may be measured for such an experiment by standard loopfuls ( 1 loop be-
ing equal to 2 milligrams), or by volume in emulsion with salt solution.
Pfeiffer has established a system of standardization for the meas-
urement of sera by this technique. He speaks of one immunity unit
as the smallest amount of such a serum which is capable of causing
13 For extensive discussion of the technique of such tests see Boehme in
Kraus u. Levaditi Handbuch, etc., Vol. 2, p. 366. The scheme of presenta-
tion of our example is taken from that used by him. See also Pfeiffer,
Zeitschr. /. Hyg., Vol. 19, 1895, p. 77.
140 INFECTION AND RESISTANCE
complete dissolution of 2 milligrams of culture material 14 (of a
standard culture) and saving the life of the animal. The unit of
the serum in the preceding test would accordingly be 0.005 c. c., and
the ticre of the serum, expressed in Pfeiffer's language, would be
200 units to the cubic centimeter. Owing to the great variation in
the virulence and toxicity of different strains of the same organism,
and also because of the difficulties opposed to the visible dissolution
of many bacteria, which may be killed by the serum without show-
ing much evidence of solution, the practical application of Pfeiffer's
standardization is not universally possible. In doing experiments
by this technique, whatever their purpose may be, accurate adjust-
ment of bacterial amounts and preliminary studies of virulence must
be made in order that the tests may be of real value and, failing
visible lysis, the death of the animals must be taken as the indicator
of the titration. Comparisons of results obtained with two different
cultures of the same species are consequently of value only when the
minimal lethal dose of each and its toxicity have been studied before
the final tests are made.
The cardinal points of Pfeiffer's phenomenon were rapidly con-
firmed, but his assumption that the process could take place only
within the animal body was soon corrected by both Metchnikoff 15
and Bordet.16 Both of these investigators succeeded in producing
extracellular lysis of cholera spirilla in hang-drop preparations. The
former produced the phenomenon by adding to the hang-drop prep-
arations small quantities of extracts of leukocytes, and thus at-
tempted to correlate Pfeiffer's observations with his own opinions
regarding the importance of the leukocytes in bacterial destruction.
The latter, however, subjected the phenomenon of bacteriolysis, both
in vivo and in vitro, to a careful analysis and obtained results which
definitely disproved the necessity of cellular intervention in this
phenomenon, and furnished facts regarding the process which stand
uncontradicted to the present day. Upon the basis of these our
modern views of the mechanism of cytolysis in general are founded.
Bordet showed that the bacteriolytic properties of immune serum
are indeed destroyed by heating to from 50° to 60° C. If, however,,
to such a heated immune serum there is added a small quantity of
fresh normal serum, the bacteriolytic power is restored with undi-
minished vigor. He recognized in consequence that there were two
distinct serum elements necessary for the process. Fresh normal
serum by itself had very slight or no bacteriolytic power. Fresh
immune serum had powerful and rapid effects. Heated immune-
14 The standard "loop" used in many laboratories for the rough meas-
urement of quantities of bacteria from agar cultures takes up approximately
2 milligrams of the material.
15 Metchnikoff. Ann. de I'Inst. Past., Vol. 9, 1895.
16 Bordet. Ann. de I'Inst. Past., Vol. 13, 1899.
BACTERICIDAL PROPERTIES OF BLOOD SERUM 141
serum had lost its power completely, but this was restored to it
the addition of the fresh normal serum. He noted, furthermore,
that the specific nature of the bacteriolysis by the immune serum was
unchanged after it had been inactivated by heat and reactivated sub-
sequently by the normal serum. The inference was plain. Immu-
nization of an animal incites the production, in the blood of this
animal, of a "preventive" substance, which is moderately resistant
to heat, and which is specific for the bacteria employed in the im-
munization. This substance cannot act upon the bacteria alone, how-
ever, but depends for its effective functionation upon the coopera-
tion of another substance present universally in normal serum, the
"bactericidal" substance, which is non-specific, corresponds to Buch-
ner's alexin, and is apparently not increased by the process of im-
munization. These are the fundamental facts revealed by the early
studies of Bordet, and they are stated in the present connection
merely as experimental facts, without further elaboration of the
later theoretical interpretation placed upon them by Bordet himself
and by Ehrlich and his followers.
In the course of these studies Bordet 17 had used the immune
serum produced in a goat by injection of cholera spirilla. As normal
serum he had used guinea-pig serum, and the latter frequently con-
tained a few blood corpuscles. He noticed that these corpuscle? were
frequently clumped in the goat serum and correlated this with the
similar clumping (agglutination) of cholera organisms which he
had noticed in this and other sera. In his incidental observation of
the phenomenon of agglutination he had concluded that the living
nature of the bacteria had no importance as far as their agglutina-
tion was concerned, dead organisms being as readily agglutinated as
living.
Reasoning from this similarity between blood cells and bacteria
in their behavior in serum, it occurred to him that the phenomena
both of agglutination and of lysis might be expressions of general
biological laws, not limited to bacteria. Accordingly he injected
rabbit blood into guinea pigs, and examined the serum of animals
so treated for its action upon rabbit corpuscles, in vitro. He found
that the sera of "blood-immune" animals had acquired not only in-
creased agglutinative power against the corpuscles injected, but had
also acquired specific "hemolytic" powers, that is, the property of
causing a solution of hemoglobin out of the red cells. (Eor the
process of serum hemolysis does not consist of a complete dissolution
of the red corpuscles, but rather in the liberation of the hemoglobin
from the cell stromata.) The latter (shadow forms) can be recovered
undisintegrated by the centrifugation of hemolyzed blood. The
17 See Bordet's own account in a "Resume of Immunity"; "Studies in
Immunity," Bordet, collected and translated by Gay, Wiley & Son, N. Y.,
1909.
INFECTION AND RESISTANCE
process, like that of bacteriolysis, was specific in that the hemolytic
power was lost if the serum was heated to from 50° to 60° C., but
could be restored undiminished by the addition of a little fresh nor-
mal serum, in itself possessing no hemolytic properties for the given
species of cell. The specificity of the phenomenon again was seen
to reside entirely in the heat-stable factor, the heat-sensitive or
"alexm" factor being non-specific, and not increased during the
process of immunization.
Observations related to those of Bordet concerning hemolysis
were made independently, in the same year, by Belfanti and Car-
bone, who had observed that the serum of animals treated with blood
cells of another species became toxic for this species, and extensive
confirmation of the phenomenon of hemolysis was obtained, in the
year following, by the work of von Dungern, and by that of Land-
steiner.
After Bordet had thus established the important fact that hemol-
ysis was in every way analogous to bacteriolysis in that, like bac-
teriolytic sera, hemolytic sera could be inactivated by heat, but re-
activated by the addition of small quantities of fresh normal serum,
Ehrlich and Morgenroth 18 undertook an elaborate study of the
mechanism of hemolytic phenomena, hoping thereby to elucidate the
mechanism of lysis in general. For it is obvious that hemolysis
lends itself far more easily to experimentation than does bacteriol-
ysis, and, as we shall see, experiments on hemolysis can be made
with a considerable degree of accuracy. Ehrlich and Morgenroth
approached the investigation of the hemolysins from the point of
view of the side-chain theory, formulated by Ehrlich in connection
with his work on the toxins. According to this theory, it will be
remembered, the hemolytic substances in the sera of animals treated
with blood corpuscles represent the receptors or side chains of tissue
cells. These receptors were originally integral chemical elements of
the body cells, by means of which the cell became united to the
injected erythrocyte (or bacterial) protein. Since union with the
foreign substance blocked these receptors or side chains, thereby
rendering them useless, they had been regenerated and, under the
influence of immunization, regenerated in excess, cast off by the cell,
and were now free in the blood stream as hemolysins (or bacteriol-
ysins).
If this conception of the process was the correct one, Ehrlich
and Morgenroth argued, the hemolytic substances of any immune
hemolytic serum should possess specific chemical affinity, "hapto-
phore groups/' as they expressed it, for the blood cells which had
been used in the immunization.
In order to show this, they inactivated at 5fi° C., by the method
of Bordet, a goat serurn which was hemolytic for beef blood, left it
18 Ehrlich and Morgenroth. Berl. klin. Woch., Nos. 1, 21, and 22, 1900.
BACTERICIDAL PROPERTIES OF BLOOD SERUM 145
in contact with beef blood corpuscles for 15 minutes at 40° C., and
then separated the cells from the supernatant fluid by centrifuga-
tion. To the blood cells they then added a little normal goat serum
(by itself not hemolytic for beef blood) and found that complete
hemolysis occurred. The addition of normal goat serum and beef
blood cells to the supernatant fluid, however, resulted in no change.
In the following diagram we have tried to represent this basic
experiment, giving the facts only of the experiment without using
any of the usual symbols which imply agreement with a theory.
EXPERIMENT TO SHOW THAT THE ANTIGEN (iN THIS CASE RED BLOOD CELLS)
ABSORBS THE SPECIFIC HEAT STABLE ANTIBODY OUT OF THE IMMUNE SERUM.
f 4 c. c. of 5 per cent, emulsion of washed beef blood.
In a test tube \ 1 c. c. of inactivated blood serum of a goat treated with beef
i blood.
These substances are left together at 37.5° C. for one hour and
then centrifugalized into:
I II
Sediment of Corpuscles. — To this are Supernatant Fluid Containing the Serum
added 4 c. c. salt solution and 0.8 c. c. and Salt Solution. — To this are added
fresh normal goat serum, by itself washed beef corpuscles and 0.8 c. c.
not hemolytic for beef corpuscles. fresh normal goat serum.
Result = Complete hemolysis. Result = No hemolysis.
Summarized, together with the facts we have already outlined,
this basic experiment has the following significance : the fresh serum
of the goat, previously injected ("immunized") with the beef blood,
possessed the property of dissolving the hemoglobin out of beef cor-
puscles, viz., hemolyzing them. Heating this serum to 56° C. for
20 minutes, as Bordet has shown, deprives the serum of all hemolytic
power, i. e., inactivates it. The addition of a little fresh goat serum,
in itself inactive, completely reactivates the hemolytic properties of
the heated immune serum. So far, as we have already seen, this
shows that hemolysis is a dual process in which a heat-sensitive
and a heat-stable substance co-operate, neither of them capable of
producing lysis by itself. The heat-sensitive ingredient, correspond-
ing to Buchner's "alexin," is present in normal serum, and, as Bor-
det 19 and von Dungern 20 had shown, is not increased in the process
of immunization, and is apparently not specific. The heat-stable
substance, therefore specific and increased in immunization, must
represent the receptors, overproduced and cast off into the circula-
tion. And, as Ehrlich and Morgenroth have now shown in the ex-
periment just described, this heat-stable element is actually bound
to the red corpuscles, and renders them susceptible to the action of
19 Bordet. Ann. de I'Inst. Past., Vol. 12, 1898.
20 v. Dungern. Mimch. med. Woch., No. 20, 1900, p. 677.
144 INFECTION AND RESISTANCE
the heat-sensitive substance in the normal goat serum. And further-
more, in attaching to this heat-stable element, the blood cells have
removed it from the solution. For we have seen, in the experiment,
that addition of corpuscles and normal serum to the supernatant fluid
resulted in no hemolysis, showing that the third necessary element,
originally in the mixture, had been carried down with the red cells.
In these and other experiments then, it was shown that only the
heat stable substances could be fixed by the red cells, and this even
at temperatures at or about 0° C. (a fact which indicates the strong
affinity between the two substances), while the heat-sensitive
"alexin," which Ehrlich now called "complement ," could not attach
directly to the red cells. For if such complement, in the form of
fresh serum, was added to washed red blood cells, and the mixture
after standing at 40° C. for some time was centrifugalized, the com-
plement remained in the supernatant fluid, as could be easily shown
by an experiment such as the one represented in the following proto-
col.
EXPERIMENT TO SHOW THAT COMPLEMENT OR ALEXIN IS NOT ABSORBED BY
UNSENSITIZED CELLS
(4 c. c. of 5 per cent, emulsion of washed beef Llood.
0.8 c. c. of fresh normal goat serum (alexin or comple-
ment), not, by itself, hemolytic for beef blood.
These substances are left together at 37.5° C. for one hour, then
centrifugalized into:
I II
Sediment of Cells. — To this is added Supernatant Fluid (salt solution and
inactivated serum of immune goat serum). — To this is added washed
which would cause hemolysis if beef blood and inactivated serum of
alexin were present. immune goat containing heat stable
element.
Result = No hemolysis. Result = Complete hemolysis.
Although, therefore, the red cells bind the thermostable specific
antibody of the immune serum and not the complement or alexin,
it was shown both by Bordet and by Ehrlich and his collaborators
that the red cells, after absorption of the thermostable substance,
when exposed to the action of the complement, were not only disin-
tegrated by hemolysis but, in the process, fixed or attached the com-
plement, so that this was no longer available for further activation
of other sensitized cells.
The fact that the alexin or complement is used up during proc-
esses of lysis, as first described by Bordet, Ehrlich, and others, has
recently been made the subject of repeated investigation, srnce this
is out of keeping with the general enzyme or fermentlike nature of
complement indicated by many of its other properties.
BACTERICIDAL PROPERTIES OF BLOOD SERUM 145
Muir,21 who studied the conditions thoroughly , comes to the con-
clusion that the complement is in truth used up in hemolysis, but
that it does not always disappear completely, this depending upon
the relative amount of sensitizer or amboceptor present. (He con-
firms the quantitative ratios between the two substances found by
Morgenroth and Sachs in hemolytic reactions, a subject discussed
by us in another place.)
Liefmann and Cohn,22 in a more recent publication, have come
to different conclusions. They believe that the disappearance of free
complement from hemolytic complexes is not due to its chemical
union with the sensitized cells in the process of hemolysis, but is due
rather
(1) to a fixation by the products of hemolysis (stromata, etc.)
after the reaction is accomplished,
(2) to dilution, and
(3) to weakening because of prolonged preservation in dilute
solution at 37° C.23
Theoretically this is of considerable importance if confirmed,
since it would bear out strongly the conception of complement as a
true enzyme or ferment. From the point of view of the practical
utilization of complement fixation for various purposes it makes
little difference, since here the disappearance of complement is the
essential thing, irrespective of whether this occurs in the course of
its activity or because of fixation by the products of its own action.
We now have the basic principle of hemolysis ; facts which can
easily *be shown to hold good for bacteriolysis and for the bacteri-
cidal processes even when no actual solution takes place. Briefly
reviewed, these facts are as follows: The antigen (blood cells, bac-
terial cells, etc.) undergoes hemolysis or bacteriolysis when acted
upon by two factors, one a thermostable substance, specific and in-
creased during immunization, the other a thermosensitive substance
present in fresh serum, not increased 24 by immunization of the ani-
mal with the antigen and not specific. The specific thermostable
substance becomes united with or fixed to the antigen regardless of
the presence or absence of the thermosensitive alexin or comple-
ment, and with such avidity that the union takes place even at 0° C.
The alexin or complement, however, cannot enter into relation with
the antigen unless this has been rendered susceptible to it by attach-
ment to the thermostable specific substance. When this has taken
21 Muir. Lancet, Vol. 2, 1903, p. 446.
22 Liefmann and Cohn. Zeitsch. f. Immunitatsforsch. Or., Vol. 8, p. 58,
1911.
23 In tlie ordinary dilution used in Wassermann tests, the unit of comple-
ment employed may deteriorate entirely within several hours at 40° C.
24Bordet. Ann. de I'Inst. Past., Vol. 12, 1898. Confirmed by v. Dun-
gern, Munch, med. Woch., No. 20, 1900.
146 INFECTION AND RESISTANCE
place, union with complement occurs, but only at temperatures above
0° C. (the speed and completeness of the union increasing as the
temperature approaches 40° C.), and the result of the union is lysis
or, in the case of bacteria not easily soluble, the bactericidal effect.
Early in their researches, Ehrlich and Morgenroth were led to
speculate upon the possibility of the formation of lytic antibodies
within the animal against its own tissue cells. It would be of
the greatest importance to pathology, as they point out, if it could
be shown that an animal could produce hemolysins, for instance,
against its own blood cells. Thus, if an extensive internal hemor-
rhage occurred from trauma or other cause, in the course of which
considerable quantities of erythrocytes are subjected to disintegra-
tion and absorption, it is at least conceivable that specific "auto-
hemolysins" might appear which would lead to a chronic destruction
of the red cells, with consequent anemia. This form of reasoning,
as we shall see, has been extensively applied in the case of the cyto-
toxins for the explanation of a variety of pathological conditions.
Ehrlich and Morgenroth approached the question experimentally in
their further work on the hemolysins in goat blood. They found that
it was comparatively easy to produce hemolysins in one goat by
treatment with the erythrocytes of other goats, isohemolysins, as
they called them.
Although, however, the blood serum of such an immunized goat
was strongly hemolytic, not only for the blood cells of the goats
whose blood had been injected, but also for the erythrocytes of cer-
tain other goats (though not, as we shall see, for goats in general),
it was never in any case active against this goat's own cells. More-
over, while the other sensitive erythrocytes could absorb the hemo-
lytic antibody out of the inactivated serum, the insensitive corpuscles
of the goat himself seemed to possess no affinity whatever for the
lysin of his own serum ; mixed with the serum they failed to absorb
out the hemolysin. This was in no sense, therefore, an autolysin.
These experiments show a remarkable individual variation be-
tween the similar tissues of animals of the same species, since Ehr-
lich and Morgenroth were indeed able to show that the insensibility
of the goat's own corpuscles depended upon a complete absence of
receptors for the isolysin. For, to explain the lack of "autolytic"
action of such a serum, two possibilities could be assumed. One, as
above, that the corpuscles of the goat possessed no receptors by means
of which the isolysin could be "anchored" or, second, that, although
such receptors were present, they were already satisfied, or saturated
with the lysin in the blood stream. In the latter case it would be
hard to understand why hemolysis had not taken place.
In order to completely disprove the latter possibility, Ehrlich
and Morgenroth did not allow the matter to rest upon conjecture, but
BACTERICIDAL PROPERTIES OF BLOOD SERUM 147
resorted to an ingenious method of experimentation which yielded a
further important result, namely, the discovery that the injection of
antibodies into animals may give rise to "anti-antibodies." They
injected inactivated hemolytic serum into goats whose corpuscles
were sensitive to its action, and found that an "anti-isolysin" was
formed, which, mixed with hemolysin and sensitive corpuscles, pre-
vented hemolysis. Injection of such an isolysin into the goat from
which it had been obtained, however, did not yield anti-isolysin, and
it was therefore reasonable to suppose that its tissue cells possessed
no suitable receptors. This failure of the production of antibodies
by an animal against its own tissue cell has been spoken of by Ehr-
lich as "Horror Autotoxicus."
These rather involved experimental data will be shown to have- a
more than purely academic value when we come to speak of the
problems of cytotoxin formation, and although they seem to show
that auto-antibodies do not form, as a rule, exceptions to this gener-
alization have been observed. The most notable of these is the ob-
servation of Landsteiner and Donath 25 made in connection with the
condition of paroxysmal hemoglobinuria. It was found that in such
cases, in which hemoglobinuria follows exposure to cold, the blood
serum of the patient contains an "autohemolysin." If the blood of
such a case is taken into oxalate or citrate solution, and allowed to
stand at ordinary or incubator temperature, nothing occurs. If,
however, such blood is cooled to 0° to 10° C. and then warmed grad-
ually to the temperature of the body, rapid hemolysis occurs. In
this case the "amboceptor" of the serum is apparently fixed or an-
chored by the blood cells only at a low temperature, the complement
becoming active as the blood is warmed. Although Landsteiner' s
observations are undoubtedly accurate, it is likely that this mechan-
ism does not explain all such cases. The writer has had occasion to
examine carefully a number of clinically diagnosed cases of this
sort with a partially successful "Landsteiner" phenomenon in one
of them only. Other observers have, however, confirmed Land-
steiner's observation in well-established cases of the condition.
Before we leave the subject of iso-antibodies it will be interest-
ing to discuss for a moment the existence of isolysins in animals
other than goats and more especially those occurring in human
beings, phenomena which have recently assumed considerable impor-
tance in view of the frequent therapeutic performance of blood
transfusion.
The peculiar facts unearthed by Ehrlich and Morgenroth 26 indi-
cated specific differences between red blood cells of individuals in
the same species (goats), which could only be recognized by the
25 Landsteiner and Donath. Munch, med. Woch., 1904, p. 1590.
26 Ehrlich and Morgenroth. "tiber Hamolysine," Berl. kl. Woch., 1900,
No. 21.
148 INFECTION AND RESISTANCE
development of immune-isolysins. Work on other species of animals
has indicated that this fact has a broad significance and that similar
differences between individuals of the same species occur in many,
if not all, species of animals. Isolysins similar in principle to those
of Ehrlich and Morgenroth were produced by Ascoli 27 in rab-
bits; by Todd and White,28 in oxen; by Ottenberg, Kaliski, and
Friedmann 29 in dogs ; by Ottenberg and Thalhimer 30 in cats, and by
Hada and Rosenthal 31 in chickens. In all these instances the iso-
lysins developed showed the same peculiarities, namely, that they
attacked the cells of certain individuals and left the cells of other
individuals of the same species unharmed. Recent work on the
isolysins occurring naturally in the human blood has thrown con-
siderable light on the nature of immune isolysins.
The occurrence of isolysins in human blood was first noted by
Maragliano 32 in 1892, and a large amount of work had been done
before it was clear that the occurrence of isolysins is not a charac-
teristic of disease. The work of Moss 33 and of Grafe and Graham 34
has shown that the occurrence of isolysins is parallel with that of
iso-agglutinins (see chapter on agglutination), and that there are in
human bloods two isohemolysinogens, A and B (corresponding to the
two agglutinogens, A and B), and two isohemolysins, a and/3. The
hemolysinogens occur regularly according to the same rule as the
agglutinogens, but the hemolysins, while they always follow the
same rule when present, may be present or latent. Thus a person
whose red cells contain A may or may not have fi , and never has a;
a person whose red cells contain B may or may not have a, but can
never have ft; a person whose red cells are susceptible to both a and
/? never has any hemolysin in the serum. It seems likely that the
substances A and B, which cause the susceptibility of red cells to
the corresponding hemolysins, are definite biochemical structures
which possibly may be inherited in a similar way to the iso-agglu-
tinogens and that similar substances (probably a larger number of
them) are present in the blood cells of various species of lower ani-
mals. This readily explains the apparent irregularity attending the
development of isolysins in the lower animals. The reason for the
natural occurrence of such isolysins in human sera and occasionally
in the sera of lower animals, however, is a complete mystery. From
27 Ascoli. Munch, med. Woch., 1901.
28 Todd and White. Nature, June 23, 1910.
29 Ottenberg, Kaliski, and Friedmann. Jour. Med. Bes., Vol. 28, 1913.
80 Unpublished personal communication.
81 Hada and Rosenthal. Zts. f. Imm., 1913, 16, p. 524.
32 Maragliano. "IX Kongr. f. Innere Med.," 1892.
33 Moss. Johns Hop. Hosp. Med. Bull, March, 1910.
34 Grafe and Graham. Munch, med. Woch., 1911, pp. 2257, 2338.
BACTERICIDAL PROPERTIES OF BLOOD SERUM 149
the work of Matsuo35 it seems likely that the autolysins of par-
oxysmal hemoglobinuria are not identical with the isolysins a and ft .
Since the reintroduction of blood transfusion as a therapeutic
measure the occurrence of hemolysis between the blood of two human
beings has become of great practical importance. A number of seri-
ous or fatal accidents following the transfusion of hemolytic blood
have been reported. Ottenberg and Kaliski36 have shown that it
is possible regularly to avoid such accidents by preliminary blood
tests.
These tests are easily carried out by obtaining serum and washed
blood cells from both prospective recipient and donor, and testing
them one against the other for hemagglutination and hemolysis, as
follows :
1. Active Serum Donor, 0.5 c. c. + Red Cells Recipient, 0.5 c. c.
(5% emulsion in NaCl)
2. Active Serum Recipient 0.5 c. c. + Red Cells Donor 0.5 c. c.
Controls of both varieties of cells in salt solution.
Such tests should be observed for at least two hours before final
readings are taken.
Although we have by no means covered in detail the entire ex-
perimental plan followed by Ehrlich and his collaborators during
their early work, we are now ready to consider the basic views on
the structure of the lytic antibodies which they deduced.
It appears from the preceding that the thermostable hemolytic
antibody must, of necessity, unite with the red cell before the com-
plement or alexin can exert its action upon it. Ehrlich conceives
this process as a mediation on the part of the heat-stable substance
between the antigen and the alexin or complement. The heat-stable
body, which he calls "amboceptor," because of its assumed mode of
action, possesses two combining groups — one the "cytophile," by
means of which it is anchored to the sensitive cell, the other the
"complementophile," by means of which it exerts affinity for the
complement. The original cell receptor, from which such an "ambo-
ceptor" takes its origin, is one which not only can combine with the
antigenic substance offered for assimilation, but which also possesses.
another atom group by means of which it can enlist the aid of the
digestive ferment of the' blood, the alexin or complement. Cast off
into the blood stream, as a result of overregeneration, it now appears
as a "double" receptor, which can form a link between antigen and
complement.
35 Matsuo. D. Arc. f. kl Med., Bd. 107 H4, p. 335.
, 3G Ottenberg and Kaliski. J. A. M. A., Vol. 61, 1913.
150
INFECTION AND RESISTANCE
0NTIGEN
BACTERIflLO
OTHER CELL
flHBOCEPTOZ
ANTIGEN
&
SCHEMATIC EEPRESENTATIONS OF A EECEPTOR OF THE THIRD ORDER.
Ehrlich ?s conception of the relationship of antigen, amboceptor, and complement
in the bactericidal and hemolytic process. In A the receptor is still a part
of the body cell, in B it has been overproduced, and is free in the circulating
blood.
In his general scheme of diagrammatic representation of these
processes Ehrlich refers to the aamboceptors" as "haptines" of the
third order.
Now it is quite plain, from the extreme specificity which results
when an animal is immunized with any given variety of blood cells
or bacteria, that there must be as great a variety of such amboceptors
as there are different antigens, and indeed an animal immunized
with two or more antigens may simultaneously contain in its blood
serum a corresponding number of different amboceptors.
This assumption of the multiplicity of "amboceptors" in the
same serum is, of course, forced upon us by the fact of specificity,
and the frequently repeated observation that the same serum may
contain heat-stable lytic antibodies against a variety of antigens, each
antigen absorbing out of such a serum that antibody only which
specifically reacts with it. This fact has, of course, never been
denied, and it is a frequent misunderstanding of the views of Bor-
det, which will be discussed directly, to assume that he has combated
the "multiplicity of amboceptor" in the sense just outlined. Ehrlich
.and Morgenroth, however, have expressed themselves in favor of the
conception of a multiplicity of "amboceptor" not only in this sense,
but as occurring in response to immunization with one and the
;same antigen.
Ehrlich and Morgenroth37 assume that any cellular antigen,
blood or bacterial cell, substances of great complexity of chemical
structure, must necessarily be possessed of a large number of differ-
ent side chains or receptors. When immunization is practiced with
such cells a correspondingly varying number of different ambocep-
tors must result. They found, for instance, that when rabbits are
37 Ehrlich and Morgenroth. Berl. klin. Woch., Nos. 21 and 22, 1901.
BACTERICIDAL PROPERTIES OF BLOOD SERUM 151
immunized with ox blood, the resulting antiserum is capable of pro-
ducing hemolysis not only of ox blood but of goat's blood as well,
though to a lesser degree. They conclude from this that the hemo-
lytic action of the serum must be referred to the presence of at least
two kinds of amboceptor, especially since repeated experiments with
different anti-ox-blood sera showed that there was no regularity in
the proportions of hemolysins for ox and goat blood, respectively.
This opinion they further fortify by showing that exposure of the
serum to ox blood deprives it of all its hemolysins, both those for
ox and those for goat's blood, whereas absorption with goat's blood
alone removes the specific goat's blood hemolysins only. They trans-
late their understanding of the conditions to graphic form by the
following diagram:38
If ox blood is in-
jected, « and ft receptors
being present, <* and ft
amboceptors are formed,
and ox blood can conse-
quently anchor both am-
boceptors. The presence
of ft receptors in goat's
blood also explains the
moderate hemolysis of
this blood by the anti-
serum, but lacking the <*
receptors which, in this
case, represent the larger
proportion, these blood
cells cannot remove all
the amboceptor for ox
blood out of the serum.
The example given, of
course, represents the simplest assumed case, and Ehrlich and Mor-
genroth believe that the same blood or bacterial cells may possess an
entire series of such receptors, some of them being dominant for the
given species, others being merely secondary or "partial," in varying
proportions.
If we grant the fundamental premises of Ehrlich respecting the
"double receptor" or "amboceptor" nature of the specific antibody
and its mediation between antigen and complement by means of a
cytophile and a complementophile receptor, certain logical conse-
quences of this conception suggest themselves, which, in their many
ramifications, have been the subject of much investigation. And
although many phases of these researches are no longer commonly
accepted, some, indeed, being untenable in the light of more recent
38 Ehrlich. "Gesammelte Arbeiten," p. 147.
OX
GO/77
SCHEMATIC KEPRESENTATION OF EHRLICH AND
MORGENROTH'S CONCEPTION OF THE COMPLEX
STRUCTURE OF AN ANTIGEN.
(After Ehrlich and Morgenroth. Berl. Iclin.
Woch., Vol. 38, 1901.)
152 INFECTION AND RESISTANCE
discoveries, the influence of this work upon the development of
immunology has been so important that it must be briefly reviewed
in order that controversial questions may be justly considered.
The comparison of the action of hemolytic sera with that of fer-
ments, and the possibility of producing antiferments by the injection
of the ferments into animals, obviously suggests a similar induction
of antihemolysins by the treatment of animals with lysins. This,
we have seen, was the method employed by Ehrlich and Morgenroth
in their studies of the causes of the failure of autolysin formation
in goats. They extended this work with the purpose of ascertaining
whether or not there were differences in the structure of the cyto-
phile groups of the various amboceptors formed when various ani-
mals were injected with any given species of red blood cells. After
obtaining a strong hemolytic serum by injecting ox blood into a
rabbit, they treated a goat with the inactivated serum of this rabbit.
The result was that the serum of the goat so treated, when mixed
with ox blood cells and the hemolytic serum, prevented the sensiti-
zation of the cells by the hemolysin. They then measured the neu-
tralizing power of such an "anti-amboceptor" or "anti-sensitizer"
against a variety of hemolytic sera produced with ox blood in differ-
ent animals and found that, while this "anti-amboceptor" neutralized
the hemolytic action of an antiserum produced in rabbits, it had but
an indifferent or entirely ineffective neutralizing power upon sim-
ilar ox blood hemolysins derived from goats, geese, dogs, rats, or
guinea pigs. They concluded from this that, although these various
lysins had been produced in the different animals by the injection of
the same antigen, viz., ox blood, and possessed affinity for the ox
blood in consequence, they must necessarily differ from each other in
some way, since they were not equally neutralized by the same anti-
lysin. It seemed to them that the difference in such cases must
depend upon variations in the structure of the cytophile group of
the amboceptor, a conclusion which they based upon the foregoing
experiments and sought to support by the following reasoning:
When an animal is treated with sensitizers or amboceptors, they
reasoned, these bodies react with the tissue cells by means of the
cell-receptors. These receptors are then overproduced and extended
into the circulation as free atom-groups.
They now act as "anti-amboceptor," free in the serum, but are
in structure merely overproduced cell receptors, identical with those
which originally united on the cell with the injected amboceptor.
Ehrlich and Morgenroth,39 therefore, believed that the neutral-
ization of the amboceptor by the antilysin depended upon a union of
the latter with the "cytophile" group of the former, preventing its
subsequent union, with the red cells. And since one and the same
antilysin did not thus invalidate the action of all the amboceptors
39 Ehrlich and Morgenroth. Berl. kl Woch., No. 22, 1901, p. 600.
BACTERICIDAL PROPERTIES OF BLOOD SERUM 153
for ox blood (derived from different animals), they concluded that
these "amboceptor" must possess different "cytophile groups."
That this conclusion of Ehrlich and Morgenroth is not correct
seems to follow the subsequent work of Bordet.40 He demonstrated
that it is not necessary to inject animals with specific hemolytic sera
in order to obtain antilytic sera, but that the same object may be
attained by injecting animals with the normal serum of an un-
treated animal. Moreover, if an "antisensitizing" serum so pro-
duced was added to corpuscles which had al-
ready absorbed "amboceptor," it prevented the
subsequent union of these sensitized cells with
alexin or complement. From this it becomes
clear that, in the first place, the antisensitizer
or anti-amboceptor cannot be identical with the
cell receptors of the corpuscles, and, further,
that the inhibition of the hemolysis which such
an antisensitizer exerts, cannot be due to union
with the "cytophile" group. This both contra-
dicts the Ehrlich conception of the mechanism
of "anti-amboceptors" and invalidates his ar-
gument, in this instance, in favor of the plural-
ity of the amboceptors produced by the injec-
tion.
Bordet's experiments were later confirmed
by Ehrlich and Sachs,41 who admit the error
of the former "anticytophile" interpretation
of Ehrlich and Morgenroth's experiments, but
they still maintain that Bordet' s experiments
do not disprove the conception of an "ambo-
ceptor" or "Zwischenkorper" of Ehrlich. They
claim that Bordet's results merely prove that
the anti-amboceptor or anti-sensitizer is "anticomplementophile" in-
stead of "anticytophile."
The principles involved we will discuss in another place in con-
nection with Moreschi's analysis of the "anticomplements." How-
ever this may be, we may conclude that Ehrlich and Morgenroth's
differentiation of amboceptors or sensitizers by the cytophile group
is no longer valid.
The studies of Bordet on the antisensitizers (anti-amboceptor)
had important results apart from their refutation of Ehrlich and
Morgenroth's opinion. In addition to showing that such antisensi-
tizer did not represent cell receptors identical with those that an-
chored the sensitizer (amboceptor) to the red blood cells, his experi-
ments revealed the fact that such an antisensitizer neutralizes un-
40 Bordet. Ann. tie I'Inst. Past., Vol. 18, 1904, p. 593.
41 Ehrlich and Sachs. Berl. klin. Woch., No. 19, 1905.
COMPLEMENT
flVBOCEPTOR
flNTMHBOCEPTOR
SCHEMATIC KEPRESEN-
TATION OF EHRLICH
AND MORGENROTH 's
CONCEPTION OF THE
NEUTRALIZATION OF
A HEMOLYTIC
SERUM BY ANTILY-
SIN OR ANTIAMBO-
CEPTOR, EEACTING
WITH THE CYTO-
PHILE GROUP. (Ehr-
lich and Morgenroth,
loc. cit.)
This conception, as we
shall see, has be-
come untenable.
154 INFECTION AND RESISTANCE
specifically various specific sensitizers as well as normal antibodies
in the serum of the same animal; and this showed that there is no
necessity of assuming a variety of specific antisensitizers, as had
been done by the German workers.
As regards the multiplicity of amboceptor or sensitizer, however,
though the proof of this, by means of anti-amboceptors, has had to
be abandoned, as we have seen, there is still a great deal of evidence
advanced in favor of such an assumption. The chief support for
such an opinion is found in the "group reactions" among bacteria,
similar to those observed for blood cells by Ehrlich and Morgenroth,
and described above (see page 151). For it is frequently observed
that the antibodies produced by immunization with one species of
bacteria may have a certain though lesser degree of action upon
other related forms, these in turn absorbing only a part of the ambo-
ceptor out of the serum, while the species originally used for im-
munization takes out all the amboceptor present. Considering the
great chemical complexity of the bacterial or tissue cells, moreover,
we may well expect such multiplicity. And it is, indeed, entirely
reasonable to suppose that a structure as complex as the bacterial
cell may contain a number of antigens and consequently give rise
to a number of sensitizers which differ in that each is specific for its
particular antigen only. This is merely a restatement of the phe-
nomenon of specificity and has, as a matter of fact, no modifying
influence on the general principles involved.
From the point of view of a general understanding of the proc-
esses of immunity, however, the question of multiplicity of sensitizer
is not so fundamentally important as is the similar controversy which
has been waged regarding the unity or multiplicity of alexin or com-
plement. Here again there has been some misconception as to the
meaning of those who maintain the unity of alexin. Neither Bordet,
nor anyone else familiar with experimental conditions, has ever main-
tained that the alexins of different animals were functionally iden-
tical. It is a well-known fact that the fresh blood sera of various
animal species differ from each other considerably in their power to
activate bactericidal or hemolytic systems. In regard to hemolysis,
fresh guinea-pig serum is very powerful in activating many sensi-
tized blood-cell complexes, but weak in activating sensitized guinea-
pig corpuscles. Often one finds that the alexin of an animal is en-
tirely impotent or but weakly capable of producing hemolysis of the
sensitized cells of its own species, though this is not a general rule.
Again, even without such species relationship, a given alexin may
be very weak for certain complexes and strong for others. The
alexin of horse blood can even be fixed to sensitized cells 42 without
42 For the sake of clearness it may be repeated here that by sensitized
cells we mean cells which have absorbed specific "amboceptor" or "sensitizer,"
and have thereby become amenable to the action of alexin or complement.
BACTERICIDAL PROPERTIES OF BLOOD SERUM 155
producing much, if any, hemolysis.43 An alexin which may be strong
for a given hemolytic complex may be weak for certain bactericidal
complexes, or vice versa. Thus there is a large mass of evidence
which shows that no two alexins are exactly alike, though the
difference between them can, of course, be defined functionally
only.
The difference between the opinions of Ehrlich and his school
on the one hand, and the followers of Bordet, on the other, revolves
not about this point, upon which all agree, but about the question of
whether one and the same serum may contain more than one alexin
or complement. Ehrlich and Morgenroth44 and Ehrlich and
Sachs 45 have brought forward evidence from which they deduce the
existence of a number of different alexins or complements for hemo-
lytic complexes in the same serum. The earlier experiments of
Ehrlich and Morgenroth on this question were carried out by means
of the filtration of normal goat serum through Pukall filters ; 46 in
these it appeared that the serum which passed through the filters was
complementary for sensitized guinea-pig cells, while that part which
had, in the original serum, activated sensitized rabbit cells was left
behind. Similar differentiation of complement they later based
upon experiments with anticomplementary sera which, they showed,
did not equally neutralize all the complementary functions of a
serum.
In support of their contention Neisser47 described two comple-
mentary substances in rabbit serum, the one active for bactericidal
complexes, the other for hemolytic, and similar experimental evi-
dence has been brought forward by Wassermann48 for guinea-pig
and by Wechsberg 49 for goat serum.
The evidence advanced by these writers is based chiefly on ex-
periments in which it was found that a normal serum which pos-
sessed both bactericidal and hemolytic powers could be deprived of
the complement for one or the other of these activities only, by ab-
sorption with the respective cells. In addition to this, Ehrlich and
Morgenroth, Ehrlich and Sachs,50 Wendelstadt,51 and others, claimed
to have differentiated various complements in the same serum by
careful heating, by the action of weak acids or alkalis, or such
methods as the digestion of sera by papain.
43 Browning. Wien. klin. Woch., No. 15, 1906.
44 Ehrlich and Morgenroth. Berl. kl. Woch., No. 31, 1900.
45 Ehrlich and Sachs. Berl. kl Woch., No. 21, 1902.
46 Sachs. Berl. kl. Woch., Nos. 9 and 10, 1902.
47 Neisser. Deutsche med. Woch., 1900, p. 790.
48 Wassermann. Zeitschr. f. Hyg., 37, 1901.
49 Wechsberg. Zeitschr. f. Hyg., Vol. 39, 1902.
60 Ehrlich and Sachs. Berl. kl. Woch., Nos. 14 and 15, 1902.
51 Wendelstadt. Centralbl. f. Bakt., I, Vol. 31, 1902.
156 INFECTION AND RESISTANCE
As a rule, these experiments have been carried out with normally
hemolytic serum and unsensitized cells, though in certain cases
Ehrlich has employed sensitized cells; but whenever this was done
exposure to complement for purposes of absorption has been for
much briefer periods than when normal serum was used. This point
is significant when we come to consider the objections to the inter-
pretation of the preceding experiment in favor of a plurality of com-
plement, objections raised chiefly by Wilde 52 and by Bordet.
Wilde refuted particularly the experiments of Neisser, who
claimed that the absorption of fresh rabbit serum with anthrax
bacilli deprived this serum only of its bactericidal but not of its
hemolytic complement. Wilde showed that, if a sufficient excess of
anthrax bacilli (or in given cases of typhoid bacilli or cholera spir-
illa) were added, both bactericidal and hemolytic complement could
be absorbed from normal serum. He concludes that there is actually
only one alexin present, but that the red cells and anthrax bacilli
differ in their susceptibility to this alexin (or, in other words, that
the sensitization of these cells by the normal serum is unequal, a
conclusion which seems rational in view of the fact, now well known,
that one and the same complement may differ greatly in the degree
of its activity upon different sensitized complexes.
Bordet has analyzed the conditions in a similar way. He found
that absorption of normal serum with unsensitized cells rarely de-
prived this serum of all of its alexin, even when these cells were used
in considerable amounts. This he attributed to the feeble sensitiza-
tion of the cells. If, however, strongly sensitized cells were added
to such a normal serum, all the alexin would be taken up. He refers
the phenomenon of specific alexin absorption, observed by previous
workers, to insufficiency in the perfection of sensitization on the
part of the cells used in the preliminary exposure ; and subsequent
work with complement fixation seems to bear him out.
Most of these arguments, though they seem to us perfectly valid
in the light of the experimental facts, have been answered by Ehrlich
and his school by the assumption of the existence of so-called "poly-
ceptors." Ehrlich now admits that the amboceptors cannot be shown
to differ from each other. However, he does not believe that differ-
ences in the intensity of sensitization explain variation in the
functional efficiency of different complements upon sensitized cell
complexes, nor does he accept, for proof of this, the fact that comple-
ment may be entirely absorbed out of a serum by a complex, even
though the complement may be comparatively inefficient as an acti-
vator in the given case. He assumes that the sensitizer or "ambo-
ceptor" may possess a number of complementophile groups (poly-
ceptors), by means of which a number of different complements may
52 Wilde. Habilitations Schrifft, Munich, 1901. Also Berl kl. WocJi.,
No. 34, 1901.
BACTERICIDAL PROPERTIES OF BLOOD SERUM 157
become active in the given case. Thus, although such a polyceptor,
of course, is capable of uniting with the complement which activates
the dominant complement, it is capable also
of union with a number of other comple-
ments which have slight or no functional
action whatever — the non-dominant com-
plements. This opinion is rendered dia-
grammatic by Ehrlich and Marshall 53 in
the following way :
If one carefully considers the reasons
advanced for the assumption of the exis-
tence of such polyceptors it does not seem
that they are sufficiently forcible to lead
one to desert the much simpler explanation
of Bordet.
Related to the problems discussed in
connection with
the production
of "anti-ambo-
ceptors" or "an-
t i sensitizers"
are those which
have arisen re-
garding the ex-
istence of " anti-
complement" or
"anti - alexins."
Ehrlich and Morgenroth claimed that, by
the injection of active horse serumvinto a
goat, they had obtained substances in the
goat serum which neutralized horse com-
plement. They believed that the "anti-
complements" thus produced neutralized
the complement by uniting with its hapto-
phore group, thus preventing its combina-
tion with the acomplementophile group"
of the amboceptor. This was their con-
clusion because they found that the "anti-
complementary" serum exerted no protec-
tive influence upon sensitized cells, when
these were exposed to the serum and then
removed, but that it protected against
hemolysis when added to the cells together
with the complement. There was apparently no union of the pro-
tective substance with the "complementophile" group of the ambo-
53 Ehrlich and Marshall. Berl kl. Woch., No. 25, 1902.
POLYCEPTOR ACCORDING TO
EHRLICH AND MARSHALL.
(a) Eeceptor of the Cell.
(b) Haptophore Group of
the Amboceptor.
(c) Dominant Complement.
(d) Secondary Complements.
Complementophile Groups of
the Amboceptor:
(1) for the Dominant Com-
plement.
(2) for the Secondary Com-
plement.
(After Ehrlich and Marshall,
Berl. Uin. Woch., No.
25, 1902.)
EHRLICH AND MORGEN-
ROTH's CONCEPTION OF
THE ACTION OF ANTI-
COMPLEMENT.
A. Scheme of Hemolysis.
B. Action of Anticomple-
ment upon Hemolysin.
b. — blood cell, c. = com-
plement, i. = immune
body, a. = anticomple-
ment.
The complementoids are
not included in the
scheme, since in this
case they are without
influence.
158 INFECTION AND RESISTANCE
ceptor, but the protecting substance did act in direct antagonism to
the complement itself.
From the fact that similar anticomplements could be produced
when inactivated serum was injected into animals, they concluded
that, on inactivation, there was not a complete destruction of the
complement, but that during the process of heating the zymophore
group of the complement only was- injured, the' "haptophore group,"
by means of which union to the tissue elements would take place,
and through which, therefore, specific antibody production would be
incited, remaining intact. Such altered complement they speak of
as "complementoid."
Bordet has made similar observations upon the production of anti-
alexins by the injection into animals both of active and of inactive
serum, but in the light of further researches, which will be discussed
in connection with the problems of alexin-fixation, chiefly those of Mo-
reschi and of Gay, we are forced to the conclusion that the existence of
true anticomplements- is by no means certain, and that the older evi-
dence in their favor is found to be. unconvincing at the present time.
In the preceding paragraphs we have emphasized the. conceptions
of the cytolytic phenomena formulated* by Ehrlich and his followers,
and although we have brought out, whenever possible, »the objections
of other investigators to many of these opinions, we have not yet
followed out in a systematic manner the reasoning of any of Ehrlich's
opponents. In opposition to the views of his school the leading
position has been taken by Bordet, who, after all, furnished in his
investigations the fundamental facts which have led to a comprehen-
sion of the cytolytic processes. In explaining Bordet's views we can
do no better than to follow out his own exposition as set forth in his
article, "A General Resume of Immunity/' 54 published with a collec-
tion of his papers. He expresses himself, in substance, as follows :
That the antigen, in the form of bacteria, blood cells, or cells of
any other nature, meets in the body of the treated animal a "recep-
tor" complex with which it unites is, of course, plain and agreed to
by everyone. That the antibody produced by the tissues in response
to such union of antigen with receptor is a direct product of the cells
containing the receptors is likely. It is by no means certain, how-
ever, or, at any rate, it has never been experimentally demonstrated,
that, as Ehrlich maintains, the antibody is identical with the original
receptor by which the antigen was fixed or anchored to the tissue
cell. It might be assumed with equal justice that the cells of the
immunized animal could build up a new substance, not identical with
the receptors, in consequence of stimulation by the antigen. It is
also by no means certain whether the injected antigen reacts with
the body cells themselves or with the normal antibodies which we
54 "Studies in Immunity" oy Bordet and collaborators. Gay, Wiley &
Sons, N. Y., 1909.
BACTERICIDAL PROPERTIES OF BLOOD SERUM 159
know to exist in many cases. Thus the blood serum of goats may
normally often contain hemolysins against rabbit corpuscles. Is it
not reasonable to suppose that possibly these may furnish the point
of attachment and the source of further antibody production when
rabbit cells are injected into goats? In criticism of Ehrlich's as-
sumption of the mode of action of heat-stable lytic antibody, Bordet
very justly maintains that no proof whatever exists of the "ambo-
ceptor" nature of this substance. All that is certain is that the
stable substance must unite with the antigen before the alexin or
complement can exert its action upon it or be fixed by it. There is
no entirely valid proof of
the existence in this anti-
body of a "complemento- co
phile" and a "cytophile"
1 f« f SENSITIZE*
group, and no satistactory
instance has been observed
in which alexin has united
with a heat-stable anti-
body which has not previ- *********<
ously been united with an
antigen.55 All that has
been shown is that the an- SCHEMATIC EEPRESENTATION OF BORDET 's VIEW
,. ,1 .., . , CONCERNING THE INABILITY OF COMPLEMENT
tigen, together with its T0 UNITE WITH ElTHER ANTIGEN OR SEN-
specific antibody, forms a SITIZER ALONE AND ITS ABILITY TO BE
pornr>W w h i P h ha<* an FIXED BY THE COMPLEX FORMED WHEN
.ompiex w hi c h THE ANTIGEN Is SENSITIZED.
avidity for alexin, a com- Compare this figure with that representing
plex which IS "endowed Ehrlich's conception of the same process,
with properties of absorp-
tion for complement which neither of its constituents alone possesses."
Bordet speaks of the "amboceptors," therefore, as "sensitizers,"
meaning by this that the antigen, by union with its antibody, is sensi-
tized to the action of the alexin. The term "sensitizers" in no way,
therefore, implies a preconceived notion, experimentally unproved,
of the mode of action or structure of the sensitizer. Since we have
graphically explained Ehrlich's opinions, a similar diagrammatic
representation may be permitted of Borders opinion of the same
process of union of antigen and heat-stable antibody with the conse-
quent development of alexin-fixing property.
In this diagram the ability to absorb or unite with complement
becomes evident only after a complex has been formed by the union
of the two elements, antigen and antibody. The diagram must not
be assumed to mean that the notch into which the complement fits
symbolized necessarily an "atom group," but merely expresses the
idea of "ability to absorb alexin," not assuming that this ability is
55 Refer also to the discussion of the conglutinins at the end of this
chapter.
160
INFECTION AND RESISTANCE
either chemical affinity by means of a definite atom group or a mere
physical change of molecular equilibrium permitting a specific com-
plement absorption.56
It will be seen from the preceding that the controversy between
Ehrlich's "amboceptor" conception and the "sensitization" idea of
Bordet turns largely upon the existence of a so-called complemen-
tophile group of the thermostable antibody. For if it were the case
that this antibody possessed an atom group which permitted it to
unite with alexin, independent of previous union with antigen, it
would go far to support Ehrlich's view. One of the strongest argu-
ments brought into the field in favor of such an occurrence by Ehr-
lich's followers is the phenomenon of Neisser and Wechsberg, which
is usually spoken of as "complement deviation" (Komplement Ablen-
kung).
In order to make the conditions underlying this phenomenon
clear, it will be of advantage to consider for a moment the methods
of determining quantitatively the amount of bactericidal antibody
(sensitizer amboceptor) in any given immune serum, since it was in
working with such titrations that Neisser and Wechsberg made their
observations.
In carrying out such measurements, -it is customary to add in
series, to constant amounts of bacteria, varying amounts of inacti-
vated antiserum and constant amounts of complement or alexin.
These mixtures are set away in the thermostat for 3 to 4 hours, are
then mixed with agar and plates are poured. The colonies which
develop will give an indication of the number of bacteria killed in
each mixture when compared with similar plates poured from tubes
in which the same original amounts of bacteria had been mixed with
alexin alone. The following table will exemplify such a test :
Typhoid bacilli
Typhoid
antiserum
inactive
Alexin
Result in colonies
after 3 hours,
incubation
Constant quantity
1 C C.
.07 c. c.
Many thousand
Constant quantity
.01 C. C.
.07 c. c.
Many thousand
Constant quantity
.005 C. C.
.07 c. c.
150 colonies
Constant quantity
001 c. c.
.07 c. c.
200 colonies
Constant quantity
.0005 c. c.
.07 c. c.
800 colonies
Control I, constant quantity. .
.07 c. c.
Many thousand
Control II, constant quantity
Many thousand
56 The diagram on page 159, though possibly not expressing with absolute
accuracy the idea of sensitization, was devised because it will remove what
seem to the writer frequent misconceptions of Bordet's views. Statements are
found in the literature which imply (Ehrlich, "Kraus und Levaditi Hand-
buch," Vol. 1, p. 8) that Bordet assumes "dass das Komplement direkt an
die Zelle angreift," and deny that there is experimental evidence to support
this. It is perfectly true that there is no evidence to show such "direktes
BACTERICIDAL PROPERTIES OF BLOOD SERUM 161
MM
Ufl
In this table it is noticeable that, although there has been con-
siderable bactericidal action in the mixtures in which 0.005, 0.001,
and 0.0005 c. c. of antiserum were used, the mixtures in which as
much as 0.1 and 0.01 c. c. were present, and in which one would nat-
urally expect a still greater antibacterial action, the contrary occurred.
This surprising and curious phenomenon, showing that an excess of
antibody could ac-
tually be harmful
to the functiona-
tion of the bacteri-
cidal complex, was
explained by Neis-
ser and Wechsberg
by t h e following
reasoning. In tests
like the one given
above a limited
amount of bacteria
and a 1 e x i n has
been mixed with
the enormous
amount of anti-
body represented
in the immune
serum. Although
bacteria can absorb more of this antibody than is necessary for their
solution or destruction, nevertheless the higher concentration given
in the table will contain quantities of "amboceptor" so far in excess
of the amount that can be absorbed that much of it must remain
free in the fluid. Now this amboceptor, possessing a complemen-
tophile group, is able to anchor complement or alexin as well as that
which has become united with the bacteria. In consequence, there
being only a limited amount of complement, some of this is deviated
from the amboceptor-antigen complexes by the free amboceptor, and
is, in consequence, ineffective so far as bactericidal action is con-
cerned. In the higher dilutions of the antiserum, in which no such
excess is present, the complement will be concentrated upon the "at-
tached" or "anchored" amboceptor, and greater efficiency will result.
Graphically Neisser and Wechsberg express their idea in the figure
which we reproduce.
As to the accuracy of the observations of Neisser and Wechsberg
there can be no question, and everyone who has occasion to carry out
Angreifen" upon the unaltered cell, but there is evidence that this union
takes place after the cell has absorbed the antibody, and no satisfactory
evidence to show that the thermostable body is an intermediary, that is, forms
a link as conceived in the amboceptor idea.
COMPLEMENT DEVIATION AS CONCEIVED BY NEISSER AND
WECHSBERG.
The complement being united to the unbound ambo-
ceptor is thereby deviated from the amboceptor,
which has gone into relation with the antigen.
(After Neisser and Wechsberg, Munch, med. Woch.,
1901, p. 697.)
162 INFECTION AND RESISTANCE
bactericidal tests with any frequency is sure to meet with the phe-
nomenon again and again. But their explanation, which involves
the assumption of union between free sensitizer or amboceptor, and
alexin or complement, without the participation of antigen, cannot
be accepted since, search as we may, through the extensive experi-
mentation that this problem has inspired, there is no instance on
record in which indisputable evidence of such an occurrence has
been advanced. On the contrary, there is a mass of satisfactory
evidence available which indicates clearly that amboceptor or sensi-
tizer alone cannot absorb alexin, and the Neisser-Wechsberg explan-
ation seems consequently to be merely an interesting and cleverly
conceived but improbable possibility.
What, then, is the explanation of the diminution of bactericidal
effect in the presence of an excess of sensitizer ? We will see that, in
the study of agglutinin and precipitin reactions, phenomena exactly
analogous to the Neisser-Wechsberg effect have been noticed, in the
case of the agglutinins, the so-called "pro-agglutinoid" zone being a
case in point. For these phenomena, as well as for that of
Neisser and Wechsberg, explanations have been advanced by
the Ehrlich school, similar in principle in that they all depend upon
more or less arbitrary assumptions regarding affinity between the
reacting bodies. Such explanations, though not outside the realm of
possibility, have, however, lost much force since it has been recog-
nized that the reactions between serum antibodies and their antigens,
in general, take place according to laws far more closely analogous
to those governing reactions between colloids than to those governing
chemical reactions in which the laws of definite proportions can be
applied. And, indeed, the reacting substances in antigen-antibody
complexes are, beyond doubt, of the nature of colloids. Now, in
many precipitations resulting when two colloids are mixed, an ex-
cess of one or the other factor will completely inhibit the occurrence
of the precipitation; the reaction taking place only when definite
proportions between the reacting bodies are present. The occurrence
of such inhibition zones, due to an excessive concentration of one
reagent, can be shown for agglutination and precipitation, exactly
as it can in ordinary colloidal reactions, and it is more than likely
that the Neisser- Wechsberg phenomenon is merely an example of a
similar phenomenon.
Looked at from this point of view, far from supporting the sup-
position of a separate complementophile group and therefore of the
"amboceptor" nature of the heat-stable lytic antibody, the Neisser-
Wechsberg phenomenon indeed becomes rather a strong argument
in favor of Bordet's views, and against those of Ehrlich. For, by
introducing the analogy between the lytic and bactericidal processes
with colloidal reactions, it takes away much force from the supposi-
tion that antigen-sensitizer alexin reactions take place according to
BACTERICIDAL PROPERTIES OF BLOOD SERUM 163
laws of definite proportion, an idea which still underlies, though
somewhat loosely, many of the more important views of antigen-
antibody reactions as conceived on the basis of the amboceptor theory.
Gay has suggested also that the ^sTeisser-Wechsberg phenomenon
may well be explicable on the basis of the fixation of complement by
precipitates. In a succeeding section we will discuss the fixation of
alexin, which occurs when a dissolved protein is brought together
with its specific antiserum. It is not impossible that this may occur
when bacterial emulsions, from which a small amount of bacterial
protein may well go into solution, are brought together with anti-
serum in concentration. Under such conditions a reaction might
readily occur which would lead to the fixation of alexin and its con-
sequent deviation from the sensitized bacteria.
Of all explanations considered, therefore, that of Neisser and
Wechsberg seems to be the least likely. It would seem to us that
Bordet's interpretation of these facts is borne out indirectly by cer-
tain experiments of Morgenroth and Sachs57 themselves, in which
the mutual quantitative relations between complement and "ambo-
ceptor" were studied. In these experiments it was shown th$,t the
more highly cells were sensitized, the smaller was the quantity of
complement which was needed for their hemolysis, and vice versa,
the less the sensitization (the smaller the quantity of amboceptor)
the more complement was necessary to produce the same result.
The following extract from one of their protocols will illustrate this :
BEEF BLOOD CELLS 5%, 1 C. C., ANTIBEEF GOAT SERUM, GUINEA-PIG COMPLEMENT
Amount of
amboceptor
Relative amount of
amboceptor
Amount of
complement for
complete hemolysis
.05
1
.008
.2
4
.0025
.4
8
.0014
A similar relation may be observed by all who have occasion to
work with hemolytic reactions. In the present connection this seems
to bear out Bordet's interpretation, since, knowing the differences in
functional efficiency of various complements for different hemolytic
and bactericidal complexes, we could well expect that insufficient
sensitization of a red cell or bacterial antigen, not particularly amen-
able to the complement employed, might fail to absorb it completely
out of the serum, thus giving a negative result which would simulate
complete lack of affinity.
This research of Morgenroth and Sachs seems further of funda-
57 Morgenroth and Sachs. Berl kl Woch., No. 35, 1902.
164
INFECTION AND RESISTANCE
mental importance in its contradiction of the regularly progressive
quantitative relations which strict adherence to the "amboceptor"
idea would seem to impose.
The quantitative relations here outlined have been diagrammat-
ically represented by Noguchi as follows:
- Complement
Purpto - Amboceptor
Mtf - foemo/ysu
I unit of Amboceptor
used in eac/i tv/tn various
fractions of- a comp/emerrf
unft
*
NOGUCHI 's DIAGRAM ILLUSTRATING THE QUANTITATIVE EELATIONS BETWEEN ANTI-
GEN, AMBOCEPTOR AND COMPLEMENT.
< Taken from Noguchi, "Serum Diagnosis of Syphilis," Lippincott, Philadelphia,
1910.)
The essential point of difference between the opinions of Ehrlich
and Bordet concerning the processes of hemolysis and bacteriolysis
lies, as we have seen, in the conception of the union of alexin or com-
plement with amboceptor or sensitizer. Although Ehrlich and his
followers admit that the union of complement with amboceptor does
not usually occur unless the amboceptor has previously united with
the antigen, they still maintain that this may occasionally take place
BACTERICIDAL PROPERTIES OF BLOOD SERUM 165
in the case of special complexes in which the complement may di-
rectly unite with free amhoceptor. This, we have seen, is the hasis
of the Neisser-Wechsberg conception of complement — "Ablenkimg"
or deviation, and of other ramifications of this theory. Bordet, on
the other hand, consistently holds that alexin or complement is at-
tached only by the complex antigen-sensitizer (antigen-amboceptor).
In the controversy which this difference aroused, an observation was
reported by Ehrlich and Sachs,58 which seemed to represent, as they
themselves express it, an "Experimentum Crucis" proving Ehrlich's
contention of the intermediary function of the amboceptor in con-
trast to Bordet's asensitization" idea. The facts, as they record
them, are as follows: When fresh horse serum is added to guinea
pig corpuscles, slight hemolysis results. When inactivated ox serum
alone is added to such corpuscles, of course no hemolysis results. If
the corpuscles are, on the other hand, exposed to the action of the
inactive ox serum, together with fresh horse serum, very active hem-
olysis is brought about. Apparently the ox serum sensitizes (or
furnishes amboceptor to) the guinea pig corpuscles, rendering them
amenable to the action of the complement in the fresh horse serum.
In other words, inactivated ox serum can be reactivated by the addi-
tion of fresh horse serum. From this one would expect that if the
guinea pig cells were exposed to inactive ox serum, then separated
from the serum by centrifugalization and fresh horse serum subse-
quently added, hemolysis would ensue. However, this was not the
case. When the cells were so treated it was found that they had
not been sensitized, and, what is more, it could be shown that the
ox serum so employed had lost none of its ability to produce strong
hemolysis when added to another complex of cells and fresh horse
serum. Ehrlich and Sachs concluded that this experiment defi-
nitely showed the ability of the amboceptor in the ox serum to unite
with alexin independently. The relation to the cell occurred
only after the union of the amboceptor in the ox serum and the
complement in the horse serum had been established, and if
their interpretation is correct, of course, it constitutes strong-
evidence against the general principle of "sensitization" as conceived
by Bordet.
This apparent inability of the corpuscles to absorb amboceptor
independently out of the inactivated ox serum, and the fact that
hemolysis results only if the corpuscles, ox serum, and fresh horse
serum are all simultaneously present, are extraordinary and not at
all in keeping with the preceding work of Ehrlich and Morgenroth,
and indeed with experience of these phenomena in general. It is log-
ical therefore to examine more closely the peculiar conditions main-
tained in these experiments before applying the reasoning deduced
from obviously different phenomena to their explanation.
58 Ehrlich and Sachs. Berl klin. Woch., No. 21, 1902.
166 INFECTION AND RESISTANCE
Bordet and Gay59 accordingly studied the Ehrlich-Sachs phe-
nomenon carefully and obtained results which confirmed the experi-
mental data of these writers but cast much doubt upon the validity
of their conclusions.
In going over the experiments of Ehrlich and Sachs, Bordet and
Gay made an observation which had apparently escaped the atten-
tion of the former investigators. Heated bovine serum has but a
slight agglutinating power for guinea pig corpuscles. Fresh horse
serum agglutinates them only slightly and slowly. On the other
hand a mixture of the two sera agglutinates them very rapidly and
completely. The bovine serum apparently possessed an accelerating
or fortifying influence both upon the weakly active normal hemoly-
sins and agglutinins in the horse serum. Bordet and Gay conse-
quently suspected that this property might be due to an undescribed
substance, peculiar to the bovine serum. To eliminate the uncertain
elements obtaining in experiments in which normal sensitizer is
used they now experimented with guinea pig corpuscles, anti-guinea
pig sensitizer (from a rabbit immunized with guinea pig blood cells)
and guinea pig alexin.
They found that sensitized guinea pig cells are hemolyzed by
guinea pig alexin very slowly and imperfectly, as is often the case
when the alexin comes from the same animal species as the cells.
When heated bovine serum was added to the complex of sensitized
cells and alexin, rapid agglutination and hemolysis resulted. Their
experiments may be tabulated as follows :
1. Cells + guinea pig alexin + heated bovine serum = no agglutination;
very slight hemolysis on next day.
2. Cells H- sensitizer + heated bovine serum = slight agglutination; no
hemolysis.
3. Cells + sensitizer + alexin -f bovine serum = powerful agglutination
and complete hemolysis in 10 minutes.
4. Cells + sensitizer -f- alexin = very slight agglutination and incomplete
hemolysis in 30 minutes.
5. Cells + sensitizer = slight agglutination; no hemolysis.
In tube (1) the slight hemolysis was due to the small amount of
normal sensitizer present in the bovine serum, and the slight agglu-
tination in tube (5) is referable to the agglutinating power of the
sensitizer. In tube (3) we see the powerfully accelerating effects
exerted both upon agglutination and hemolysis when bovine serum
acts upon sensitized corpuscles in the presence of alexin.
Bordet and Gay's interpretation of the Ehrlich-Sachs phenom-
enon, in the light of these new experiments then, is, in their own
words, as follows: "When guinea pig corpuscles are added to a
mixture of the two sera they are affected by the sensitizer of the
horse serum and, to a certain extent, by the sensitizer in the heated
69 Bordet and Gay. Ann. de I'Inst. Past., Vol. 20, 1906, p. 467.
BACTERICIDAL PROPERTIES OF BLOOD SERUM 167
bovine serum. This second sensitizer is, however, superfluous. Its
presence is by no means necessary for the experiment. When this
sensitization is effected the corpuscles are then in condition to fix
the horse alexin. This alexin, however, has only slight hemolytic?
power. But once the corpuscles have become sensitized and laden
with alexin they are modified in their properties of molecular ad-
hesion to such an extent that they become able to attract a colloidal
substance of bovine serum, which unites with them. The adhesion
of this new substance produces two results : it causes the blood
corpuscles to be more easily destroyed by alexin and also agglutinates
them energetically. Consequently, a powerful clumping, followed
by hemolysis, is observed."
Bordet and Gray, therefore, assume that the action of the bovine
serum is due to a new substance which they speak of as "bovine
colloid." This substance resists heating to 56° C., is probably al-
buminous, and has the property of uniting with cells that are laden
with sensitizer and alexin, but remains free in the presence of nor-
mal or merely sensitized cells.
They fortify this opinion by showing experimentally that the
"colloid" is removed from bovine serum by absorption with sensi-
tized bovine corpuscles which have been treated with horse alexin.60
Bordet and Streng61 later studied this "colloid" more thor-
oughly and have suggested for it the name " congluiimn." Streng 62
later showed that the agglutinating action of this substance could be
shown not only for sensitized and "alexinized" red blood cells, but
also for similarly treated bacteria, and that conglutinins were pres-
ent not only in bovine serum, but in that of goats, sheep, antelopes,
and a number of other herbivores, but apparently absent in cats,
dogs, guinea pigs, and birds.
The body described by these workers as conglutinin is probably
identical with a similar heat-stable serum component reported by
Manwaring63 and called by him "auxilysin."
60 Browning (Wien kl. Wochenschr., 1906) had shown that horse alexin
may be absorbed by sensitized beef cells without causing hemolysis.
61 Bordet and Streng. Centralbl f. Bakt., Orig. Vol. 49, 1909.
62 Streng. Zeitschr. f. Immunitatsforsch., Orig. Vol. 2, 1909, p. 415.
63 Manwaring. Centralbl. f. Bakt., 1906; Orig. Vol. 42.
CHAPTER VII
FUKTHEK DEVELOPMENT OF OUK KNOWLEDGE CON-
CEKNING COMPLEMENT OE ALEXIN. COMPLE-
MENT FIXATION
IT will be remembered that Buchner in his first studies upon the
"alexin" compared its action to that of an enzyme or ferment, and
suggested that the source of this substance might possibly be found
in the white blood cells. This thought was very obviously suggested
by the observation that bacteria were destroyed within the white
blood cells, after phagocytosis, by a process analogous in many ways
to that by which they were destroyed by the serum constituents.
Hankin,1 in an elaborate study dealing with the problem, maintained
the leukocytic origin of alexin on the basis of the observation that
increased bactericidal properties closely followed upon the heels of
periods of leukocytosis. He assigned the particular property of
alexin production to the eosinophile cells, proposing for them the
designation "alexocytes." Further study, however, has not justified
such an association with the eosinophiles, and Hankin' s opinion has
not been experimentally upheld.
After Hankin the problem occupied the attention of a number
of other investigators, and many of them succeeded in showing that
there was, indeed, an increased bactericidal power in exudates rich
in leukocytes, and further that bactericidal substances could be
directly extracted from leukocytic emulsions. We refer particularly
to the early work of Denys and Havet,2 of Hahn,3 of Van de Velde,4
and others, studies which will be described in our chapter on
phagocytosis. This work was done before the complex nature of the
bactericidal constituents of serum had been demonstrated and before
the work of Schattenfroh and others had shown that the bactericidal
substances extracted from leukocytes were of a nature quite distinct
from the active elements of the serum, and were independent of the
participation of alexin. Although these earlier investigations cannot
properly be regarded, therefore, as proving the leukocytic origin of
1 Hankin. Centralbl. f. Bakt., Vol. 12, 1892.
2 Denys and Havet. La Cellule, Vol. 10, 1894.
3 Hahn. Archiv f. Hyg., Vol. 25, 1895.
4 Van de Velde. La Cellule, Vol. 10, 1894.
168
FURTHER DEVELOPMENT OF KNOWLEDGE 169
alexin, Metchnikoff and his school have nevertheless adhered to this
conception for various additional reasons.
Metchnikoff distinguishes between two kinds of alexin — the
microcytase, which is the bactericidal complement or alexin, and
is supposed to originate from the microphages or polynuclear leu-
kocytes, and the macrocytase, which represents the hemolytic and
cytolytic alexin or complement, and originates from the mononuclear
cells or macrophages. As in the case of the bactericidal alexin, ex-
traction methods have been employed to demonstrate that the hemo-
lytic alexin took its origin in the macrophages, and at Metchnikoff' s
suggestion, Tarassewitch 5 prepared hemolytic substances by extract-
ing spleen tissue and other amacrophagic organs" in various ways.
Here again the identity of the hemolytic extracts with serum hemoly-
sins has been placed in doubt. Korschun and Morgenroth 6 have
shown that the hemolytic organ extracts were heat stable and alcohol
soluble ; Donath and Landsteiner,7 and others, have obtained similar
results. It would be quite thankless to review the extensive literature
which has accumulated upon this point. It would seem, in summariz-
ing it, that no definite proof of the presence of true, active alexin,
either hemolytic or bactericidal, within the leukocyte or mononuclear
cells has been brought by methods of extraction, and the apparently
positive results reported by earlier observers are adequately explained
by the discovery of the heat-stable and non-reactivable bactericidal
and hemolytic substances in extracts of such cells by Schattenfroh,
Korschun and Morgenroth, and many others. It appears, moreover,
from these investigations that probably the intracellular substances
by which the digestion of ingested bacteria or blood cells is brought
about are of a nature entirely distinct from that of the serum anti-
bodies and alexins. A very ingenious demonstration of this is found
in an experiment first made by RTeufeld. Neufeld 8 allowed leu-
kocytes to take up highly sensitized red cells. Instead of undergoing
prompt hemolysis, as they would if small amounts of alexin had been
added, they were slowly broken up without hemolysis, fragments of
hemoglobin remaining after complete morphological disintegration
of the erythrocytes. At no time were intraphagocytic "shadow'*
forms observed.
The failure to extract alexins from dead leukocytes does not,
however, preclude the possibility of the secretion of alexins by living
leukocytes. This point is one which is, of course, much more diffi-
cult to investigate directly. Indirectly the increased bactericidal
properties of exudates rich in leukocytes, as found by Denys and
Havet, would point in this direction. However, even this is not
5 Tarassewitch. Cited from Metchnikoff.
6 Korschun and Morgenroth. Berl kl Woch., No. 37, 1902.
7 Donath and Landsteiner. Wien. kl. Rundschau, Vol. 40, 1902.
8 Neufeld. Arb. a. d. kais. Gesundlieitsamt., Vol. 28, 1908, p. 125.
170 INFECTION AND RESISTANCE
•conclusive, since at the time when these investigations were carried
out no discrimination was made between the bactericidal serum sub-
stances and those other "endolysins" which might well have been
extracted from the accumulated white blood cells. The writer some
years ago attempted to approach this problem directly by keeping
leukocytes alive in inactivated serum and in Ringer's solution at
37.5° C. for several days in the hope that, after 48 hours, alexin,
hemolytic or bactericidal, might appear in these fluids. The experi-
ments were entirely negative, but were regarded as inconclusive, since
it was impossible to determine accurately how long, or in what pro-
portion, the leukocytes had remained alive.
One of the basic premises of MetchnikofFs theory on the nature
of alexin consists in the conception that alexin is not found in the
circulating blood plasma, but appears only when there has been leu-
kocytic injury, as in the clotting of blood or in the "phagolysis"
which, as we have seen in the chapter on phagocytosis, usually occurs
after foreign substances have been injected into the peritoneum, pre-
ceding a local accumulation of leukocytes. This point of view seems
to be rendered improbable because of the rapid hemolysis which
occurs when we inject sensitized red blood cells into the circulation
of an animal, but we might here, too, assume a preliminary injury
to white blood cells resulting from the intravenous injection.
Much less likely to be accompanied by cell injury is the method
of obtaining blood serum by creating an area of artificial edema by
ligating a limb — or, as in MetchnikofFs 9 10 experiments, the ear of
a rabbit. And, indeed, in edema fluids so obtained little or no alexin
is ordinarily found. This fact has been interpreted in favor of
MetchnikoiPs views, as has also the curious absence of alexin in the
aqueous humor of the anterior chamber of the eye.11 12 In this fluid
no alexin is present under normal conditions, but if puncture is prac-
ticed, and the fluid again taken after a period of three or four hours,
alexin is now found, probably, according to Metchnikoff s school, be^
cause of the coincident entrance of leukocytes into this space. It is
conceivable, however, that the aqueous humor may be free from
alexin for other reasons than the absence of leukocytes ; and an injury
which is followed by the invasion of leukocytes is pretty sure to be
followed also by the entrance of the fluid elements of the blood ; i. e.,
alexin.
Much experimental work has been done in which it has been
attempted to demonstrate directly that the blood plasma contains no
complement or alexin. The most important investigation of this
9 Metchnikoff. Ann. Past., Vol. 9, 1895.
10 Bordet. Ann. Past., Vol. 9, 1895.
11 Metchnikoff. Loc. cit.
12 Mesnil. Ann. Past., Vol. 10.
FURTHER DEVELOPMENT OF KNOWLEDGE 171
kind is that carried out by Gengou13 in 1901. It was Gengou's
primary purpose to obtain the plasma of mammals in such a way
that no cell injury would occur. This he accomplished by special
methods in which coagulation was avoided without the addition of
foreign anticoagulants like hirudin, etc. His technique was, in
essence, as follows : He took the blood directly through a paraffined
cannula into tubes that had been coated with paraffin, and centrifu-
gal ized it" at low temperatures until cell free. This plasma, taken
from the paraffin tubes, quickly clotted, and the material with which
the experiments were done actually consisted of blood serum. Upon
examining the serum so obtained, he found that it exerted practically
no bactericidal action. As a result of this investigation he claims
to have demonstrated the truth of MetchnikofFs contention that the
circulating blood plasma contains no alexin.
If borne out, it is true that Gengou's results would very power-
fully support this theory, and for this reason a large number of ex-
periments have been made since then, with the same end in view.
In all such investigations the technical procedures are extremely
difficult and, as Addis 14 has recently said, in our opinion quite cor-
rectly, it would be impossible to carry out bacteriolytic or hemolytic
experiments with mammalian paraffin plasma without obtaining
coagulation, and for this reason most of the writers who have re-
peated Gengou's experiments have worked, as did he, not with
plasma, but with serum. Falloise,15 following Gengou's method
exactly, obtained results diametrically opposed to those of Gengou;
Schneider/6 also with the same technique, failed to confirm Gengou's
results; Herman,17 on the other hand, confirms Gengou.
In order to overcome the technical difficulties encountered in
working with mammalian plasma a number of writers have more
recently experimented with bird blood, which, as is well known,
coagulates much more easily than does mammalian blood. Hewlett,18
who worked with goose plasma and peptone plasma, could not con-
firm Gengou' s results. Lambotte,19 examining the plasma of chick-
ens, found no difference between the serum and plasma in their con-
tents of bactericidal alexin, as measured against cholera spirilla.
Yon Dungern, working with fish plasma, obtained similarly negative
results, and recently Addis, in a careful comparative study of
chicken plasma, found no evidence of differences between plasma
and serum in either the bactericidal or the hemolytic alexin. As far
13 Geng-ou. Ann. de I'Inst. Past., Vol. 15, 1901.
14 Addis. Journ. of Inf. Dis., Vol. 10, 1912.
15 Falloise. Bull, de I'Acad. Roy. de Med,, 1905, p. 230.
16 Schneider. Archiv /. Hyg., 1908, Vol. 65, p. 305.
17 Herman. Bull, de I'Acad. Eoy. de Med., 1904, p. 157.
18 Hewlett. Archiv f. exp. Path. u. Pharmk., 1903, Vol. 49, p. 307.
19 Lambotte. Centralbl. f. Bakt., I, Orig., 1903, Vol. 34, p. 453.
INFECTION AND RESISTANCE
as we can tell at present, therefore, we cannot accept, as conclusively
proven, the contention that the circulating plasma contains no alexin.
Nevertheless the Metchnikoff school have not been discouraged by the
various contradictions of Gengou's work, found in the experiments we
have enumerated, because they are not satisfied that the technique of
other workers has conclusively excluded cell injury. Owing to the
great difficulties of investigations of this kind, when carried out with
mammalian blood, it is not impossible that they are justified in this,
but nevertheless the assumption of the absence of alexin in the
plasma finds so many objections in other observations that the bur-
den of proof would certainly rest with Gengou and his supporters.
Not the least important of these objections, it seems to us, is based
on the very simple experiment of injecting bacteria into the veins of
a living animal and finding a very rapid and active phagocytosis.
And considering the very probable participation of alexin in the
opsonic functions this would seem to point strongly toward the pres-
ence of these substances in the circulating blood. The evidence also
furnished by the recent developments of our understanding of ana-
phylaxis would further tend to strengthen our belief in the presence
of alexin or complement in the normal circulation. For, in the
process, as we shall see in a later chapter, complement plays
an important role. When 3 per cent, salt solution is administered (as
in Friedberger's experiments), and the action of complement is
thereby inhibited, anaphylactic shock may be greatly diminished.
It has also been claimed, chiefly by Walker 20 and by Henderson
Smith,21 that, as serum stands upon the clot it at first gains in alexin
or complement contents, an occurrence which they attribute to the
liberation of alexin from the leukocytes. This observation has not
been universally borne out and, even were it unquestionable, it might
be dependent upon any one of the numerous factors involved in the
complicated process of coagulation rather than upon leukocytic
changes only.
The failure to obtain definite proof of the origin of alexin from
the white blood cells has led to search for the source of these sub-
stances in various organs. An interesting series of investigations
on this subject are those of Mile. Louise Fassin,22 who believes that
she has found reasons for definitely associating the thyroid gland
with alexin production. She found that the subcutaneous injection
of thyroid extract into dogs and rabbits was followed by a rapid
increase of alexin, both hemolytic and bactericidal, and that the
same thing was true when thyroid substance was administered by
mouth. When the thyroid gland was removed from rabbits a reduc-
tion of alexin resulted. Although important, these researches do not
20 Walker. Journ. of Hyg., Vol. 3, 1903.
21 Smith. Proc. Roy. Soc., Series B, Vol. 79, 1906.
22 Louise Fassin. C. E. de Soc. Biol., Vol. 62, 1907.
FURTHER DEVELOPMENT OF KNOWLEDGE 173
necessarily prove that the thyroid can be looked upon as a source
of alexin, and, indeed, Fassin gives experimental results without
drawing any very sweeping conclusions. It might well be that the
thyroid secretion is simply concerned in stimulating the production
of alexin from another source. Marbe 2B has similarly associated
the thyroid gland with the production of opsonins, which, when we
consider the probable identity of alexin and normal opsonin, may be
taken as a confirmation of Fassin's work.
Of great interest also are the series of investigations which asso-
ciate the liver with the production of alexin. The basis of such
investigations is found in the observations made by Morgenroth and
Ehrlich 24 that there is a diminished production of complement or
alexin in dogs subjected to phosphorus poisoning, with consequent
degeneration of the liver. The first investigator to study this ques-
tion experimentally was Nolf.25 Nolf tried to approach it by extir-
pating the liver in dogs, and found that his results were unreliable
by this method. He then experimented with rabbits and found that
when the liver was extirpated in these animals and the vena cava
anastomosed with the portal vein (Eck fistula) the animals would
survive for three or four hours. This period, though short, was suffi-
cient to show definite changes in the blood. Taken just before death
it differed from that taken just before the operation in a number of
important respects. There was relative incoagulability, there was
autohemolysis, and with these there occurred an extreme fall of
alexin or complement. Serious objections may be brought against
Golf's experiments. In the first place the operation as performed by
him results in shock and injury so profound that rapid death ensues,
conditions under which not only the complement-producing functions
but all functions, secretory and otherwise, are reduced. Miiller 26 ob-
jects to Nolf 's experiments chiefly for the reason that he did not pre-
vent the absorption of toxic substances from the intestine, materials
which could now enter the general circulation without any longer be-
ing neutralized by the liver functions. M tiller, for this reason, re-
peated ~Nolf 7s work but, by a complicated technique, temporarily shut
off the intestinal circulation in addition to extirpation of the liver.
He found, in agreement with Nolf, that exclusion of the liver from
the circulation resulted in the prompt diminution of complement or
alexin. In all such experiments, however, the very profound shock
which necessarily occurs in the animals would seem to us to vitiate
the results. Moreover, Liefmann 27 has repeated Miiller's experi-
ments without being able to obtain the same results. Not satisfied
23 Marbe. C. E. de la Soc. Biol, Vols. 64, et seq., 1908-1909.
24 Morgenroth and Ehrlich. In Ehrlich's "Gesammelte Arb.," etc.
25 Nolf . Bull, de I'Acad. de Science de Belg., 1908.
26 Miiller. Centralbl. f. Eakt., Vol. 57, 1911.
27 Liefmann. Weichhart' s Jahresbericht, Vol. 8, 1912, p. 155.
174 INFECTION AND RESISTANCE
with these experiments, however, Liefmann experimented on frogs ;
in whom, as Friedberger has shown, extirpation of the liver is not
so rapidly fatal as in warm-blooded animals. He removed the livers
of frogs in a number of cases and, although his animals lived about
a week, there was no definite diminution of the hemolytic properties
of the serum. It seems, therefore, that the origin of alexin in the
body is by no means settled and requires further investigation.
Equally unsatisfactory have been the attempts to define the chem-
ical nature of the complement or alexin. In the investigations deal-
ing with the hemolytic action of cobra venom it seemed at first as
though a clue to this problem had been found. Flexner and JSTogu-
chi 28 made the interesting observation that cobra poison alone does
not hemolyze the blood cells of certain animals, namely those of
cattle, goats, or sheep, if these cells are washed entirely free of
serum. This seemed to suggest that the serum of these animals con-
tained some activating substance. It also seemed to indicate that the
cells of other animals, which were easily hemolyzed, even when en-
tirely freed of serum, might contain such an activating substance
within themselves. The behavior of this activating substance toward
snake- venom hemolysis was therefore very similar to the action of
complement, except in one important respect, namely, as Calmette 29
showed, almost all sera were rendered more efficient for the activa-
tion of snake venom when heated to 65° C., whereas complementary
properties of sera for other hemolyzing complexes are, of course,
destroyed at 56° C. Kyes,30 31 on further studying these phenomena,
extracted the red blood cells of rabbits and other animals whose cells
were hemolyzed by snake venom alone, by shaking them up with
distilled water, and showed that, with these extracts, he could activate
the venom against ox, goat, and sheep corpuscles, cells which were
not ordinarily hemolyzed by the venom without the addition of
serum. Similar activation of the venom with extracts of the ox,
goat, or sheep corpuscles was not possible. He concluded from this
that the blood cells of the rabbit, dog, guinea pig, and man pos-
sessed an "endocomplement" for the snake venom; that is, a com-
plementary substance contained within the cells, while in the other
species it was found in the activating serum only.
The thermostability of such venom "complements" encouraged
him to attempt their isolation, and he found that they were ether-
soluble, indicating their lipoidal nature; and, finally, after several
negative attempts with activation by other lipoids, he determined
that lecithin, added to the corpuscles and the snake venom, brought
28 Flexner and Noguchi. Journ. Exp. Med., Vol. 6, 1902 ; Univ. Pa. Med.
Bull., 1902 and 1903.
29 Calmette. C. R. de I'Acad. des Sciences, p. 134, 1902.
30 Kyes. Berl klin. Woch., Nos. 38 and 39, 1902.
31 Kyes and Sachs. Berl. klin. Woch., Nos. 2-4, 1903.
FURTHER DEVELOPMENT OF KNOWLEDGE 175
about a rapid hemolysis. This seemed to explain both why heated
serum could activate the venom in some cases, and why some varie-
ties of blood cells could be hemolyzed without serum, since lecithin
is a substance widely distributed both in the fluids and cells of the
animal body. His further studies seemed to show that, by proper
chemical manipulation (bringing together cobra poison with lecithin
in chloroform solution), he could produce a combination of the two
which he called "cobra lecithid," a substance which apparently "acti-
vated" cobra venom. He conceived it as the "amboceptor-comple-
ment" complex of the cobra hemolysin, which acted hemolytically
upon all varieties of blood cells.
These researches of Kyes aroused much interest, chiefly because
they seemed to furnish an example of a chemically definable com-
plement, lipoidal in its constitution. Recent researches by Von
Dungern and Coca,32 however, seem to prove that, while Kyes' ex-
perimental facts were perfectly accurate, his conclusions do not seem
to have been warranted. Von Dungern and Coca showed that the
cobra venom contains a lipoid-splitting ferment which acts upon the
lecithin, liberating substances from it which hemolyze in the same
way as do many other non-specific substances. The cobra-lecithid,
according to this, would represent merely a lecithin derivative which
happens to have hemolytic action without any specific -relationship to
the hemolytic properties of the venom itself. Thus, even in this case,
unfortunately, we are not in possession of facts which bring 'us
nearer to a chemical understanding of the complementary substances
or alexins.
In the further development of attempts to define alexin or com-
plement chemically, two further researches are of importance,
namely, those of Von Liebermann 33 and of Noguchi. In both in-
vestigations it is suggested that the alexin may consist of a combina-
tion of soaps and proteins. Noguchi 34 showed that the hemolytic
organ-extracts described by various observers were soaps, a possi-
bility which had been previously considered by Sachs and Kyes.35
Noguchi further established analogies between his soaps and com-
plement as follows: Sensitized blood cells are hemolyzed by mix-
tures of soaps and inactivated guinea pig serum, while normal ery-
throcytes are not hemolyzed by similar mixtures; furthermore, like
normal complement, such serum-soap mixtures are inactivated by
prolonged preservation and by heating at 56° C. Objections were
soon made to the findings of both Noguchi and Von Liebermann by
Hecker,36 whose experiments seemed to show that when sensitize^
32 Von Dungern and Coca. Munch, med. Woch., 1907, p. 2317.
33 Von Liebermann. Biochem. Zeitschr., Vol. 4, 1907.
34Noguchi. Biochem. Zeitschr., Vol. 6, 1907.
35 Sachs and Kyes. Berl. kl Woch., 2-4, 1903.
36 Hecker. Arb. auf dem konig. Inst. f. exp. Ther., Heft 3, 1907.
176 INFECTION AND RESISTANCE
blood cells were thoroughly washed free of serum soaps did not have
this hemolyzing action, and Friedemann and Sachs37 claimed that
they were unable in any case to inactivate the hemolytic serum soap
mixtures by heating to 56° C. These writers, as well as others,
attribute Noguchi's results to the fact that the sera which he used to
produce his "artificial complement," i. e., his serum soap mixtures,
were heated to 50-51° C. only, a fact which would justify doubt of
complete inactivation. Knaffl-Lenz38 has more recently carried out
experiments on the same question. His results seem to show that the
hemolytic action exerted by fatty acids or soaps is a phenomenon
quite incomparable to true complement action, and that these hemoly-
sins are heat stable, remaining unchanged by heating at 56° C. We
have referred in a number of places to the analogy between alexins
and ferments or enzymes. The chief objection to this conception
formerly brought forward was based upon the fact that the comple-
ment or alexin, unlike an enzyme, was used up during its reactions,
and that a definite quantitative relationship existed between the
alexin and the amount of cells or bacteria upon which it could act.
Recent experiments by Kiss 39 seem to show that this quantitative re-
lationship is not as strict and regular as was formerly supposed. He
showed that the action of complement depends very largely upon its
concentration. For instance, to cite his work directly: "0.05 com-
plement is sufficient to hemolyze completely a definite quantity of
sensitized blood cells if the experiment is done in a total volume of
5 c. c. 0.02 c. c. of complement gives absolutely no hemolysis in a
similar volume. When, however, the total volume is reduced to 2.5
c. c., then 0.02 c. c. of the complement begins to act, and it produces
complete hemolysis if the total volume is reduced to 1.25." In fur-
ther developing this observation he showed that, if sufficiently con-
centrated, a very small amount of complement can act upon an ex-
tremely large amount of red blood cells, an amount incomparably
larger than those acted upon in more dilute solutions. These observa-
tions would tend to strengthen considerably the conception of the fer-
ment nature of alexins in general.
Kiss' observations are furthermore in agreement with the inves-
tigations of Liefmann and Cohn,40 whose work we have mentioned
in the preceding chapter on Cytolysis. These writers assert that the
fixation of complement during hemolysis is not due to its chemical
union with the sensitized cells, but is due to fixation by the end
products of the reaction ; in other words, by the stromata of the red
cells and possibly by other substances given up by these cells. A
further factor contributing to the disappearance of complement in
37 Friedemann and Sachs. Biochem. Zeitschr., Vol. 12, 1908.
38 Knaffl-Lenz. Biochem. Zeitschr., Vol. 20, 1909.
39 Kiss. Zeitschr. f. Imm., Vol. 3, 1909.
40 Liefmann and Cohn. Zeitschr. f. Imm., Vol. 8, 1911.
FURTHER DEVELOPMENT OF KNOWLEDGE 177
such reactions is, they claim, its rapid deterioration at 37° to 40° C.,
when diluted. If they are right, these considerations also remove
important objections to the conception of complement as a ferment.
It is clear, therefore, that although we have gained much detailed
information regarding the 'functional activity of the complement or
alexin, and may assume, in a general way, that its action is similar
to, if not identical with, that of an enzyme, we are nevertheless still
very much in the dark concerning its chemical nature. The same
thing may be said in regard to its physical characteristics. One
method of investigating the physical properties of complement has
been that of filtration. It may be remembered that one of Ehrlich
and Morgenroth's 41 arguments in favor of the multiplicity of com-
plement was the fact that, when goat serum was filtered through a
Pukal candle, the complement which was active upon rabbit corpus-
cles was retained, while that which acted upon guinea-pig cells
passed through. Immune bodies or amboceptor always passed
through.
Vedder,42 in similar experiments upon bactericidal complements,
claims to have been able, in the same way, to separate the comple-
ments acting upon different bacteria. The problem has been more
recently investigated by Muir and Browning.43 Their conclusions
are briefly as follows: In the early stages of filtration through a
Berkefeldt filter complement is often completely held back. After
continued filtration it begins to pass through. If the complement is
inactivated by the addition of hypertonic salt solution (5 per cent.),
it passes through, and the filtrate can be reactivated by dilution to
isotonicity. Sensitizer or amboceptor always passes through. Just
how these experiments are to be- interpreted is a little obscure. The
fact that the addition of salt renders the complement capable of
passing through the filter would seem to indicate that its original
inability to permeate did not depend upon the size of the molecule.
On the other hand, "it is also possible that the addition of salt to the
complement may increase its dispersion in such a way that the indi-
vidual particles are rendered smaller. This, however, is purely
speculative, and we are at a loss for a fully satisfactory explanation of
the results of Muir and Browning. We have repeated some of the
experiments of Muir and Browning and, in substance, 'confirmed
their results. It is our opinion that new filters remove complement
by adsorption, just as this is accomplished when complement is
shaken up with kaolin or other finely suspended material.
That the addition of salts of various kinds in quantities greater
than isotonicity (or more than the equivalent of 0.85=0.9 per cent.
41 Morgenroth and Ehrlich. Ehrlich's "Gesammelte Arbeiten," etc.
42 Vedder. Journ. Med. Res., Vol. 9, 1903.
43 Muir and Browning. Journ. of Path, and Bact., Vol. 13, 1909.
178 INFECTION AND RESISTANCE
!N"aCl)44 exerts a profound action upon the activity of complement
is well known. ISTolf 45 noted this in 1900, and the problem has heen
studied since that time by many investigators. Von Lingelsheim,46
who studied it in connection with his work on the refutation of the
"osmotic" theories of immunity, showed that increasing the salt con-
tents of serum (KNO3, NaCl, K2HPO3, etc.) progressively dimin-
ished its bactericidal power. Hektoen and Ruediger 47 also, after a
very thorough study of this phenomenon, conclude that the action of
the salts in such cases is exerted upon the alexin or complement and
not upon the heat-stable sensitizers, and that it probably depends
upon "physicochemical" causes. However, the manner in which
such salt-inactivation is brought about is, to a great extent, obscure.
There is no visible precipitation from serum after the addition of
salts sufficient in quantity to weaken its action. Nothing is, as far
as we can tell, removed from solution, and yet there is temporary
inactivation which, at the same time, renders the complement fil-
trable, facts from which we can only surmise some physical altera-
tion.
Inactivation of the complement also follows the removal of salts,
but here the process is accompanied by a definite chemical change in
that the serum globulins are precipitated.
Studies of this process have led to important modifications in our
conception of the nature of alexin, since they have shown that this
body, formerly assumed to be single and homogeneous, may be sub-
divided into at least two component parts by a number of experi-
mental procedures. Ferrata was the first one to point this out as a
consequence of investigations undertaken by him primarily with the
purpose of determining the nature of the influence of salts upon
hemolytic processes. Older studies of Buchner and Orthenberger 48
had shown that bactericidal action was inhibited when salts were
removed from the medium, but the causes underlying such inhibi-
tion had not been made clear. Ferrata 49 found, in the first place,
that the absence of salts exerted no effect upon the mechanism of
sensitization, but that amboceptor or sensitizer became attached to
the cellular elements as readily when salts were absent as when the
reagents were suspended in normal salt solution. It was a natural
inference, therefore, that the failure of hemolysis, which he observed
in 'salt-free media (analogous to the similar experiences of Buchner
44 Alexin can be preserved in the refrigerator for long periods if hyper-
tonic salt solution (15 to 25%) is added. It will again become active if iso-
tonicity is restored with distilled water.
45 Nolf . Ann. Past., Vol. 14, 1900.
46 V. Lingelsheim. Zeitschr. f. Hyg., Vol. 37, 1901.
47 Hektoen and Ruediger. Journ. of Inf. Dis., Vol. 1, 1904.
48 Buchner and Orthenberger. Archiv f. Hyg., Vol. 10, 1C90.
49 Ferrata. Berl. kl Woch., 1907, No. 13.
FURTHER DEVELOPMENT OF KNOWLEDGE 179
in the case of bacteriolysis), must be attributed to failure of func-
tionation on the part of the complement. On further investigation
he obtained a very simple explanation. Ferrata removed the salts
from his sera by dialyzing for twenty-four hours against distilled
water. In this process, of course, there is a precipitation of the
globulins while the water-soluble albumins remain in solution. The
former may be redissolved in normal salt solution and the latter
rendered is'otonic by the addition of calculated amounts of concen-
trated salt. In this way the original serum components are divided
into two parts, neither of which, as Ferrata found, is alone capable
of producing hemolysis of sensitized cells. In order to obtain the
complementary action possessed by the original
serum it is necessary to combine the two. This
principle discovered by Ferrata is probably re-
sponsible also for the results obtained by Sachs
and Teruuchi,50 who likewise noted the destruc-
tion of the complementary function in sera
diluted with distilled water, but attributed this,
in their publication, to the action of a comple-
ment-destroying ferment, which is assumed to
be active in salt-free media only.
In his first experiments Ferrata reported
that the precipitated globulin fraction was
thermostable, the thermolability of complement CONCEPTION OF COM-
being due entirely to the unprecipitated albu- TING^AS
t^Mjiii/ji-j
• END' PIKE
• (ALBUMEN FIMCT/OK)
SB H, DP, tec
^•^B (GLOBULIN FftACTIOH)
^Z
min fraction. The work of Ferrata was soon SUGGESTED BY
continued, however, by a number of other work- BRAND.
ers, who confirmed the essential fact of the par-
tition of the complement but modified and considerably extended the
original observations. Brand 51 found that both fractions were equally
thermolabile, and that the globulin sediment, after being redissolved
in salt solution, could not be preserved in an active condition for more
than a few hours. Preserved in distilled water or as sediment, it
may retain its activity for several days, but dissolved in salt solution
it becomes inactive within 3 to 4 hours, at room temperature.
Michaelis and Skwirsky 52 have since shown that the globulin frac-
tion, thermolabile when free, is unaffected by a temperature of 56°
C. after it has become attached to sensitized cells. Brand further
studied the relationship of the two fractions to the sensitized cells
and found that the globulin fraction may attach directly to such
antigen-antibody complexes, but that the albumin fraction cannot be
bound in this way unless the globulin fraction has been previously
attached. For this reason he has referred to the former as the "end-
50 Sachs and Teruuchi. Berl kl Woch., 1907, Nos. 16, 17, and 19.
51 Brand. Berl. kl. Woch., 1907, No. 34.
52 Michaelis and Skwirsky. Zeitschr. f. Imm., Vol. 4, 1910.
180 INFECTION AND RESISTANCE
piece" and the latter globulin sediment as the "mid-piece," assum-
ing, on the basis of the conception of Ehrlich, that the globulin frac-
tion serves to establish a link between the sensitized cell and the
end-piece analogous to that formed by the "amboceptor" between the
cell and the whole complement. It is possible, therefore, to treat
sensitized cells with mid-piece in such a way that they are thereafter
susceptible to hemolysis by the end-piece alone. Such cell-sensitizer-
mid-piece combinations have been spoken of by Michaelis as "per-
sensitized" cells.
Tsurusaki53 confirmed the findings of Brand as to the thermo-
lability of both "mid-piece" and "end-piece," but was unable to sep-
arate the complement of normal hemolysins into the two components
in the same way, since he found that hemolytic power was, in such
cases, completely destroyed after twenty-four hours of dialysis. It
seems to us not impossible that the natural deterioration of alexic
power which takes place during such periods of time, at temperatures
of from 16° to 20° C., may easily be held accountable for this, since
the very feeble sensitization of cells in normal hemolysin complexes
requires a correspondingly larger amount of alexin for activation.
We have mentioned that the so-called "mid-piece" undergoes a
rapid change when dissolved in salt solution and, after 3 or 4 hours,
may lose its ability to induce hemolysis when added to sensitized
cells together with end-piece. Although Hecker54 was able to con-
firm this, he nevertheless showed that this fact does not imply a
destruction of the mid-piece. For when such apparently inactive
"mid-piece" was added separately to sensitized cells, and end-piece
wTas subsequently allowed to act upon the complex, hemolysis re-
sulted. This seems to show that the "mid-piece" undergoes a change
on standing in salt solution which does not alter its ability to com-
bine with the sensitized cells, but which subjects it to inhibition of
such union when end-piece is present. It is also a peculiar fact, evi-
dent in many of our own experiments, that when "mid-piece" and
"end-piece" are first mixed and then added to sensitized cells the ef-
fect in hemolysis is less powerful than when the "mid-piece" is added
to the cells first, and the "end-piece" later. This effect is so instan-
taneous that, if, in a series of experiments in which combinations of
mid- and end-piece are used, the end-piece is run into the tubes con-
taining the cells just before the mid-piece is added instead of the
other way round, hemolysis is inhibited.
In working with dialysis, also, we have regularly had an experi-
ence which may explain the difficulties which many other investiga-
tors have had in such experiments. The globulin precipitate, which
53 Tsurusaki. Biockem. Zeitschr., Vol. 10, 1908.
54 Hecker. Arb. a. d. konig. Inst. f. exp. Ther., Frankfurt a/M., Heft
3, 1907. See also Guggenheimer. Zeitschr. f. Imm., Vol. 8, 1911.
FURTHER DEVELOPMENT OF KNOWLEDGE 181
fell out after dialysis of 24 hours or more, almost without exception,
retained moderate or slight hemolytic properties, which could not be
removed until the precipitate had been dissolved in salt solution and
reprecipitated with distilled water two or three times. This would
imply that a minute amount of the end-piece, carried down in pre-
cipitation, must suffice to activate the mid-piece and would seem
to point to the fact that in whole serum the two fractions are present
as a complex and not separately. This question has been much dis-
cussed and many facts have been brought out on both sides. Hecker
showed that the combination of mid-piece with the sensitized cells
can take place at a temperature of 0° C., while that of end-piece
with the "persensitized" cells requires a considerably higher tem-
perature. The bearing this fact may have upon similar earlier ex-
periments of Ehrlich and Morgenroth upon the thermal conditions
governing the union of amboceptor and complement with antigen is
self-evident. In the present connection, however, the fact that the
two fractions may be separately absorbed out of the serum by sensi-
tized cells at 0° C. would suggest the probability of their being sep-
arate in the whole blood. ~No crucial experiment has so far been
possible, and there is not enough evidence on either side as yet to
justify a definite opinion. However, the experiments of Michaelis
and Skwirsky and later ones of Skwirsky alone have much indi-
rect bearing on this question, though final interpretation is as yet
impossible. Michaelis and Skwirsky,55 after determining that an
acid reaction inhibits the hemolysis of sensitized blood cells, found
that under such conditions "mid-piece" alone is bound, but that
"end-piece" or the albumin fraction is left unbound. They recom-
mend the use of strongly sensitized cells in an acid medium as a
method of obtaining free "end-piece" from serum.
Skwirsky 56 subsequently found that during the ordinary Wasser-
mann reaction the complex of syphilitic serum and antigen binds
the mid-piece only. If the Wassermann reaction has been strongly
positive; that is if there has been absolutely no hemolysis, and we
remove the supernatant fluid by centrifugation, active end-piece can
be demonstrated in it by the addition of persensitized cells. Bron-
fenbrenner and Noguchi have also studied this phenomenon, but da
not believe that Skwirsky' s experiments prove that end-piece is free
in such "fixation" supernatant fluids. These supernatant fluids, ac-
cording to them, differ from all other "end-pieces" in that they are
active upon persensitized sheep corpuscles only, but not upon other
cells. An explanation for this is lacking.
There is much that is confusing in the facts so far revealed about
the two component parts of the alexin. The most difficult fact to
explain is the peculiar inactivation of the mid-piece in salt solution,
55 Michaelis and Skwirsky. Zeitschr. f. Imm., Vol. 4, 1910.
56 Skwirsky. Zeitschr. f. Imm., Vol. 5, 1910.
182 INFECTION AND RESISTANCE
which prevents its functionation in the simultaneous presence of
end-piece, but does not seem to interfere with its ability to combine
with the sensitized cells. As was to be expected, explanation for
this has been sought by the Ehrlich school in changes of affinity.
Sachs suggests that the mid-piece, by its preservation in salt solution,
has lost its avidity for the sensitized cells and has gained in avidity
for the end-piece, an alteration which therefore prevents its union
with the cells. The same idea was suggested by Hecker himself. It is
a little difficult to reconcile this explanation, however, with the fact
that whole serum can be preserved and remain active in its comple-
mentary function for a number of days, mid-piece and end-piece
being present together, in a medium which, as far as salt contents are
concerned, is isotonic with the salt solution in which mid-piece de-
teriorates so rapidly when alone.
That there is, after all, much similarity between the alexins of
different animals is evident from the fact that, as Marks and others
have shown, the end-piece of one animal may activate the mid-piece
of another species. It appears also from experiments like those of
Ritz and Sachs 5T that an animal may possess a mid-piece for certain
sensitized cell complexes without possessing a corresponding end-
piece. Thus they found that the serum of mice contained a mid-
piece but not an end-piece for sensitized guinea pig corpuscles.
Much that has been found out about the so-called globulin portion,
moreover, tends to engender doubt as to the wisdom of applying to
these complement fractions the terms "mid-piece" and "end-piece,"
an objection which is based upon reasons similar to those which pre-
vent iBordet from accepting the term amboceptor. For so little is
actually known concerning the mechanism of complement functiona-
tion, that it seems unwise to establish on a firm basis a preconceived
idea of the mechanism by adapting the terminology to a theory. The
most confusing feature of the problem lies in the surprising quantita-
tive relations which seem to exist in the reactions of the two frac-
tions. Thus Liefmann and Cohn58 claim that in the presence of
moderately sensitized cells no measurable amount of the so-called
mid-piece or globulin fraction is bound, that is, removed from solu-
tion ; and yet, when both fractions are added to such cells, rapid and
complete hemolysis results. In the presence of heavily sensitized
•cells (20 to 50 units) a small quantity only is removed. Nevertheless
this fraction has had a demonstrable effect on the cells, since it has
Tendered them amenable to the action of the albumin fraction. In all
such experiments, therefore, as Liefmann justly points out, the de-
gree of sensitization must be taken into consideration before conclu-
sions are formulated. It is curious also that a slight excess of the
globulin fraction may prevent complement action completely. In
57 Ritz and Sachs. Zeitschr. f. Imm., Vol. 14, 1912.
58 Liefmann and Cohn. Zeitschr. f. Imm., Vol. 7, 1910.
FURTHER DEVELOPMENT OF KNOWLEDGE 183
experiments cited by Marks 59 it appears' that the most ineffective
complement is obtained when "mid-piece" and "end-piece" are added
to the sensitized cells in proportions of 1 to 1. If the proportion of
"mid-piece" is increased two or threefold over that of "end-piece,"
hemolysis is inhibited. This, however, is true only when the two
fractions are simultaneously added to the sensitized cells. When the
sensitized cells are exposed to the excessive quantity of the "mid-
piece" separately, and "end-piece" added later, the effect is one of
stronger hemolysis than when smaller amounts are used. It is thus
seen that the relations between the complement fractions in hemolysis
are very involved. All that we can be sure of is that there are at least
two separable parts, that one of these acts directly upon the sensitized
cells, forming a so-called persensitized complex and rendering them
amenable to the subsequent action of the unprecipitated albumin
fraction.
The many difficulties encountered in the interpretation of the
confusing phenomena observed in connection with this problem have,
very naturally, led to a corresponding multiplicity of opinion. Most
observers at present incline to the opinion that the globulin and
albumin portions of fresh serum, separated by Ferrata's or any other
of several common methods, represent actually two complement frac-
tions. This is not, however, accepted by all workers. Bronfenbren-
ner and Noguchi 60 believe that the entire active complement is con-
tained in the albumin fraction or so-called "end-piece." They hold
that "complement-splitting" by dialysis or other methods is an inacti-
vation of end-piece by change of reaction. In their experiments they
were able to restore the functional activity of end-piece by the adjust-
ment of reaction, either with acid or alkali, respectively, or by the
addition of amphoteric substances. The mid-piece activates, they be-
lieve, by reason of its amphoteric nature and consequently adjusts
any excessive acidity or alkalinity of the medium. They were able
to substitute for mid-piece indifferent amphoteric substances such as
alanin. Liefmann 61 has been unable to confirm the experiments of
Bronfenbrenner and Noguchi, and believes that their results were
caused by incomplete splitting of the complement. Incidental to a
study of normal opsonins the writer has also repeated the experi-
ments of ;Bronfenbrenner without being able to confirm them.62
The method of Ferrata for the separation of the two parts of the
complement is successful only if dialysis is very thorough and suffi-
ciently prolonged to lead to complete precipitation of the globulins.
iNeufeld and Haendel 63 have had difficulty in thus separating the
89 Marks. Zeitschr. f. Imm., Vols. 8 and 11, 1911.
60 Bronfenbrenner and Noguchi. Jour, of Exp. Med., Vol. 15, 1912.
61 Liefmann. Weichhardt's Jahresbericht, Vol. 8, 1912.
62 Zinsser and Gary. Journ. of Exp. Med., Vol. 19, 1914.
63 Neufeld and Haendel. Arb. a. d. kais. Gesund., 1908.
184
INFECTION AND RESISTANCE
fractions, and the writer has noticed similar failures but has always
been able to obtain eventual separation by sufficient prolongation of
the dialysis. Because of the occasional difficulties and because of
the time-consuming and inconvenient nature of the method other
means of separation have been devised. The one used with success
by many workers has been that introduced by Sachs and Altmann,64
namely, precipitation of the sera with weak hydrochloric acid -g^
to -j^. Liefmann has separated the components by precipitation of
the globulins by the introduction of CO2. In carrying out this
method, Fraenkel 65 has found it advantageous to dilute the serum
ten times with distilled water, then allowing the CO2 to flow in at
low temperatures. It is likely that any of the usual methods of
globulin separation will serve for complement partition. The salt-
ing out methods are, however, extremely inconvenient because of the
prolonged dialysis subsequently necessary to remove the salts.
The inactivation of complement or alexin by the addition of
salts or by splitting is, very apparently, a temporary inactivation in
which prompt restitution can be practiced by bringing back original
conditions either by dilution to isotonicity or by reconstruction of
the divided substance, respectively. Heating to 56° C., the simplest
and most commonly employed method of inactivations was, until of
late, regarded as an irreversible process, the complement being irre-
trievably destroyed in the procedure. Gramenitski 66 has recently
carried out experiments which seem to show that this opinion is
erroneous. His experiments were suggested by the fact, observed by
Bach and Chodat,67 that certain oxydases and diastases may spon-
taneously regain some of their activity after inactivation by heat.
His work with complement indicated a similar gradual return to an
active condition after moderate heating. The great theoretical im-
portance of this observation will justify our insertion of one of
Gramenitski's protocols.
Experiment 1. Complement 10 times diluted was heated to 56°
C. for 7 minutes. It was then tested against sensitized beef blood
at varying intervals as follows:
Time after heating at which test
was made
Quantity of hemoglobin gone into solution after
% 10 min.
% 20 min.
% 30 min.
% 40 min.
Immediately after heating
1]/2 hour
0
0
20
10
20
30
70
40
40
60
80
70
70
80
100
24 hours
48 hours
64 Sachs and Altmann. Cited from Sachs in "Kolle u. Wassermann Hand-
buch," Vol. 2, p. 877.
65 Fraenkel. Zeitschr. /. 7mm., I, Vol. 8, 1911.
66 Gramenitski. Biochem. Zeits., Vol. 38, 1912.
67 Bach and Chodat. Cited from Gramenitski, loc. cit., p. 511.
FURTHER DEVELOPMENT OF KNOWLEDGE 185
In other experiments in which heating was more prolonged a
similar regeneration was observed, though not as pronounced as in
the one cited above. The largest amount of restored complement
seemed to be present after about 24 hours. After this gradual de-
terioration again ensued. It is quite impossible to offer an adequate
explanation for this at the present time. Gramenitski 68 acknowl-
edges this, but permits himself certain speculations which we repeat
in nearly his own language, since there is much in them which seems
to us reasonable. The complement, as indeed all other active serum
constituents, must be looked upon as colloidal in nature. When heat
is applied to such substances alterations occur which gradually lead
to coagulation. As this occurs there is an aggregation of particles
and a consequent diminution of surface tension. This last point has
been experimentally demonstrated by Traube,69 who has regularly
found a fall of surface tension as serum was heated to 56° C. And
of greatest interest in this connection is the further determination by
Traube that a gradual restoration of the surface tension takes place
as the serum is allowed to stand. It is not inconceivable, therefore,
that the inactivation of complement by heat may depend upon an
alteration of its colloidal state, i. e., an aggregation of the particles,
which, if not carried too far, may be reversible and followed by a
gradual dispersion as the serum is kept 24 hours. On the same
grounds the gradual deterioration of complement on standing may be
compared to the slow settling out of colloidal suspensions which
eventually results in spontaneous precipitation, a process which
occurs not only in chemically well-defined colloids, but is often ob-
served in sera. Bechold has referred to this as "das Altern Kol-
loidaler Losungen."
Of great interest, furthermore, in connection with the physical
properties of complement is the discovery made by Jacoby and
Schiitze 70 that complement can be inactivated by shaking. This
astonishing observation has been confirmed by Zeissler,71 Noguchi
and Bronfenbrenner,72 Ritz,73 and others. It appears, according to
these observers, that guinea pig serum, when subjected to active
shaking, can eventually be robbed thereby of its activating prop-
erties. The success of such experiments depends somewhat upon the
concentration of the serum, and is best observed in a dilution of 1
part to 10 parts of salt solution. Under such conditions complete
inactivation may be observed within 20 to 25 minutes. Between the
inactivation of complement by heat and that which results from
68 Gramenitski. Loc. cit., p. 504.
69 Traube. Zeitschr. f. 1mm., Vol. 9, 1911, and Biochem. Zeitschr., 1908.
70 Jacoby and Schiitze. Zeitschr. /. 1mm., Vol. 4, 1910.
71 Zeissler. Berl kl. Woch., No. 52, 1909.
72 Noguchi and Bronfenbrenner. Journ. of Exp. Med., Vol. 13, 1911.
73 Ritz. Zeitschr. f. 1mm., Vol. 15, 1912.
186 INFECTION AND RESISTANCE
shaking, there are certain similarities which seem to strengthen the
opinion regarding the nature of heat inactivation which we have
'cited above. For it has been variously shown that prolonged shaking
of protein solutions, like heating, gradually leads to coagulation. It
would be important to determine whether or not the inactivation by
shaking, like that produced by heat, is accompanied by a fall of sur-
face tension.
ALEXIN OR COMPLEMENT FIXATION
The controversy regarding the multiplicity of alexin and the
existence of a "complementophile group" cannot, of course, be re-
garded as closed, however much we may lean toward the acceptation
of Bordet's point of view, since German experimenters of eminence
still adhere to the Ehrlich interpretations. Moreover, it is, of course,
extremely difficult to disprove such an assumption as that of the
"polyceptor" conception of the complementophile group. However,
we may safely assert that the functional unity of complement (and,
after all, that is all that Bordet has maintained) is being upheld by
the constantly increasing evidence in its favor which is being fur-
nished by the practical and experimental application of the phe-
nomenon of "alexin fixation" described, in 1901, by Bordet and
Gengou.74 It will be well to bear in mind that this phenomenon
should be strictly distinguished from the so-called "complement de-
viation" ("Ablenkung"), described by Neisser and Wechsberg. The
latter was advanced as an explanation of the inactivity of bacteri-
cidal sera when used in too great concentration, as described in an-
other place {p. 160) (Neisser and Wechsberg phenomenon), and has
been variously utilized as support for the assertion that alexin can
unite with unattached sensitizer. It is regarded by most observers,
moreover, as untenable in the light of later investigation. In spite of
this, the term "Komplement-Ablenkung" has been employed by a
number of German writers (see Citron, Vol. 2, "Kraus und Levaditi
Handbuch") as synonymous with "fixation" in the sense of Bordet
and Gengou.
The phenomenon of Bordet and Gengou, briefly described, is
nothing more than an experimental utilization of the fact which we
have discussed at length, that alexin is fixed by antigen and antibody
after union, but by neither alone.
The condition, as observed by them, may be best described by
submitting the protocol of the first experiment detailed in their
communication :
, An .emulsion of a 24-hour slant of plague bacilli was used as
74 Bordet and Gengou. Ann. de VInst. Past,, 1901, Vol. 15, p. 289.
FURTHER DEVELOPMENT OF KNOWLEDGE 187
antigen, heated antiplague horse serum represented the antibody,
and fresh guinea pig serum was used as alexin. A series of tubes
was then prepared as follows :
1.
2.
3.
4.
5.
6.
v
Alexin + plague bacilli + inactivated antiplague serum.
Alexin + plague bacilli -j- inactivated normal horse serum. '
Alexin -- inactivated antiplague serum.
Alexin -j- inactivated normal horse serum.
Plague bacilli + inactivated antiplague serum.
Plague bacilli and normal horse serum.
These mixtures were left together for 5 hours and, at the end of
this time, sensitized rabbit corpuscles were added to each tube. The
result showed hemolysis in all the tubes except "I,77 in which there
were plague bacilli, antiplague serum, and alexin, and in tubes 5 and
6, which had contained no alexin from the beginning.75
It was plain, therefore, that the bacilli when specifically sensi-
tized had become capable of absorbing alexin and preventing its sub-
sequent action upon the sensitized erythrocytes. That the occurrence
was not exceptional was stiown by the fact that, in the same series,
similar results were obtained with anthrax, typhoid, and proteus
bacilli, and their respective antisera.
Schematized in accordance with the conceptions of Ehrlich our
diagram would be as follows :
*- COMPLEMENT otALEXfN
SPECIFIC
ANTIBODY — »
(PRESENT Of? NOT?}
HAEMOLYTIC
ANTIBODY
REDBLOOD.
0NTIGEM _^
(BACTERIA ETC)
COMPLEMENT FIXATION SCHEMATIZED ACCORDING TO EHRLICH'S VIEWS.
If the antibody in I is present then complement is fixed by the antigen-antibody
complex, and is no longer free, to act upon the hemolytic complex II. In the
same way antigen I could be determined if a known antibody I were used.
For, in the absence of either of these parts of the complex I complement
would remain unfixed and free to act on complex II.
We represent the phenomenon graphically in the symbols of Ehr-
lich merely because they facilitate clearness of exposition.
In the presence of both parts of Complex I the alexin is held and
75 We will see later that imsensitized bacteria in emulsion will non-specifi-
cally fix small amounts of complement.
188 INFECTION AND RESISTANCE
is no longer available for Complex II. If either of the reacting
parts, antigen or sensitizer, of Complex I are lacking the alexin is
left unfixed and free to react with Complex II.
With this technique Bordet and Gengou were able to demonstrate,
by indirect experiment, the presence of specific sensitizers in the
sera of animals immunized with various bacteria, a fact which was,
of course, surmised but had been amenable to proof heretofore only
in the case of bacteria like the spirillum of cholera in which lysis
under the influence of immune serum and alexin could be directly
observed under the microscope. The practical possibilities of their
method were, of course, immediately apparent. By the use of a
known antigen specific sensitizers can be demonstrated in this way,
and, vice versa, in the presence of a known antibody, the method
will serve to identify the nature of a doubtful antigen. Thus bac-
terial differentiation can be carried out by adding to the suspected
bacteria, in emulsion, a small quantity of a known antiserum and
alexin, and determining whether or not the alexin has become fixed.
And, conversely, Bordet and Gengou 76 have more recently utilized
the method in support of their claim of the specific etiological im-
portance of the bacillus isolated by them from whooping cough, by
showing that the serum of children suffering from this disease
formed a specific alexin-fixing complex when treated with the ba-
cillus.
The phenomenon of Bordet and Gengou thus found rapid prac-
tical application in the diagnosis of a number of infectious diseases,
and has, of course, attained great clinical importance in the diag-
nosis of syphilis in the form of the "Wassermann" reaction and its
many modifications. Before discussing these practical features in
greater detail, however, it will be useful to discuss more particularly
the many important theoretical considerations which have followed
in the train of the complement-fixation phenomena.
A year after the publication of Bordet and Gengou's paper
Gengou 77 made another fundamentally important observation by
showing that complement or alexin fixation was not limited to the
complexes of cellular antigens and their antibodies, but that the sera
of animals immunized with dissolved proteins (animal sera, etc.),
when brought together with their specific antigens, likewise formed
combinations which fixed alexin. Thus egg-white or dog serum,
brought together with "anti-egg-white" or "anti-dog" rabbit serum,
respectively, strongly fixed alexin, whereas neither the antigenic sub-
stances nor the antisera exerted such fixation alone. The interpre-
tation put upon this by Gengou was the following : "In sera obtained
by injecting rabbits with large doses of cow's milk, etc., there are,
in addition to the precipitins of Bordet and Tschistovitch, substances
76 Bordet and Gengou. Ann. de I'Inst. Past., 1906.
77 Gengou. Ann de I'Inst. Past., 16, 1902.
FURTHER DEVELOPMENT OF KNOWLEDGE 189
•analogous to the sensitizers described by Bordet in bacteriolytic and
hemolytic sera, and later found in the majority of antimicrobial
sera." The important point in this interpretation is that Gengou
conceived the existence of antiprotein sensitizers, in addition to the
precipitins, formed as a response to immunization with amorphous
protein. Moreschi 78 soon confirmed Gengou's experimental deter-
minations, and Neisser and Sachs79 took the further logical step of
applying this knowledge to the determination of proteins for foren-
sic purposes. This, too, we will further discuss when we speak of
the practical features of these phenomena. It thus appears that the
fixation of alexin is a generalized property of all mixtures in which
an antigen is brought into contact with its specific antibody, whether
the antigen is in the form of the whole bacterial or other cell, or in
that of a dissolved protein, animal serum, or egg-white, etc.
The observation of Gengou, though for a time insufficiently val-
ued, has had a profound influence upon the subsequent under-
standing of serum reactions. The fundamental importance of
this work was not fully recognized until his studies had found
logical continuation in the investigations of Gay80 and in those of
Moreschi.
Moreschi 81 studied the antihemolytic properties possessed by the
serum of a rabbit which had been treated with normal goat serum.
He found that such a serum had distinct anticomplementary powers
when it was added to a hemolytic system of ox blood sensitizer (ob-
tained against ox blood from rabbits), and goat complement. With
such a hemolytic system, however, there was anticomplementary ac-
tion only against goat complement and not against rabbit or guinea
pig complement. If, however, he used a hemolytic system in which
the amboceptor or hemolytic sensitizer employed was one obtained
from a goat, the serum was anticomplementary for all complements
which were used. Moreschi concluded from this that the apparent
anticomplementary action of the serum could not be interpreted as
the action of a specific anticomplement in the sense of Ehrlich, but
that it resulted from the reaction which took place as the consequence
of union of the antibody in the anti-goat rabbit serum and goat pro-
tein, which was introduced into the tubes, in the first case with the
complement, and in the second with the amboceptor. He proved his
contention by obtaining similar universal anticomplementary action
when he added a little normal goat serum to the tubes set up as above
described. It is plain, therefore, that anticomplementary action can
be explained in observed cases by the simple consideration of the phe-
nomenon of Gengou. Similar findings were later recorded by Muir
78 Moreschi. Berl. kl Woch., No. 37, 1905.
79 Neisser and Sachs. Berl kl. Woch., No. 44, 1905.
80 Gay. Centralbl. f. Bakt. I Orior. Vol. 93, 1905, p. 603.
P1 Moreschi. Berl. kl. Woch., 1905, No. 37, ibid., No. 4, 1906.
190 INFECTION AND RESISTANCE
and Martin,82 and it may well be doubted, as a result of these and
other researches, whether we are at all justified in assuming the ex-
istence of anticomplements.
The work of Gay, published independently in the same year as
that of Moreschi, has, in a general way, the same significance, out
Gay recognized the relation of the conditions observed by him to the
precipitin reaction, a feature absent from both the original study of
Gengou and the work of Moreschi. Gay noticed that an inactivated
hemolytic immune serum, left for some time in contact with its spe-
cific cells, and then separated from them by centrifugation, would
often possess anticomplementary or anti-alexic properties. He fur-
ther noted that after such a serum had been freed from the cells by a
short centrifugation, if it was again vigorously centrifugalized, a
slight, cloudy sediment would appear at the bottom of the tubes. If
this sediment was removed the serum lost its alexin-fixing properties.
He recognized that the precipitate formed in these tubes was a spe-
cific precipitate resulting from the union of a precipitinogen and its
antibody. The reaction was due entirely to the fact that insufficient
washing of the cells used in producing the hemolysin, gave rise to
the formation of precipitin against the serum of the animal from
which the cells had been taken, and subsequently insufficient washing
of the cells of this same species employed in the tests furnished
enough antigen to give a precipitin reaction in the tubes in which
the inactivated hemolytic (and precipitating) serum was 'mixed with
the 'cells. Subsequently, numerous investigations 83 have shown
Gay's interpretation to be correct, and we may now accept it as a
fact that precipitates formed by the union of specific antigen with
its antibody possess the power of fixing alexin and that, in a general
way, this fixation is proportionate in energy to the amount of pre-
cipitate which is formed.
Gay utilized his results primarily to contradict certain assertions
of Pfeiffer and Friedberger concerning antibacteriolytic substances
supposed to occur in normal sera. These authors had found that, if
normal sera possessing no "antagonistic" properties in the first place
were left in contact with certain bacteria, they acquired antibac-
teriolytic properties for these particular bacteria. Thus normal
inactive rabbit serum, left in contact with typhoid bacilli, and again
separated from the bacteria, now prevented the lysis of sensitized
typhoid bacilli if tested by the intraperitoneal method spoken of as
the Pfeiffer reaction. Sachs 84 applied these observations to analo-
gous hemolytic reactions and obtained similar results. He found
that, if normal, inactive rabbit serum was left in contact with sheep
82Muir and Martin. Journ. of Hyg., Vol. 6, 1906. See also Muir's
"Studies on Immunity," Froude, London, 1909.
83 Dean. Zeitschr. f. Imm., I, Vol. 13, 1912.
84 Sachs. Deut. med. Woch., 1905, No. 18.
FURTHER DEVELOPMENT OF KNOWLEDGE 191
or guinea pig corpuscles, it acquired the property of preventing the
hemolysis of these corpuscles if, later, it was brought together with
them in the presence of specific hemolysin and alexin. Gray now
showed hy experiment that Sachs' method was referable to insuffi-
cient washing of the corpuscles. When, in the first contact, the rab-
bit serum was exposed to the sheep corpuscles, a certain amount of
sheep serum adherent to the cells was carried over into the rabbit
serum. This sheep antigen later reacted with the antisheep pre-
cipitin present in the hemolytic immune serum and, in this way,
fixed alexin and prevented hemolysis.
It seems that the analysis of Gay is correct, and that Sachs'
conclusion as well as thpse of Pfeiffer and Friedberger, by analogy,
cannot be taken as demonstrating the existence of specific anticom-
plements or anti-amboceptors. Gay has further offered the same
mechanism as an explanation of the Neisser-Wechsberg phenomenon,
which has been discussed in another place.
To summarize, then, we have learned that there are a number of
varieties of specific alexin absorption or fixation processes, one that
is exerted by cells treated with specific sensitizer, be they blood or
bacterial, the other that which occurs when unformed protein is
brought into contact with its specific antiserum. The latter has been
correlated with the precipitin reaction, in that it has been found that,
whenever a specific precipitate is formed in such reactions, it is this
precipitate on which the fixation depends. On the other hand, it is
necessary to note that the formation of a precipitate is by no means
necessary for the fixation, for, as is well known, if a series of pre-
cipitin tubes are set up, in each successive one of which the amount
of antigen is diminished, a degree of dilution will soon be reached
at which no visible precipitate will occur, but which nevertheless
will show alexin fixation. The following is an illustration of such
an experiment:
Sheep serum -f antisheep serum
Precipitate
Fixation of 0.5 c. c.
guinea pig complement
05 c c. (1-20) 4- 0.5 c. c . ..
4-
Complete
0.5 c. c. (1:50) 0.5 c. c
+4-
Complete
05 c c (1-100) 05 c c
+++
Complete
0.5 c. c. (1 :200) 0.5 c. c
4-4-4-
Complete
0 5 c c (1 -500) 0 5 c c
+ + +
Complete
0.5 c. c (1-1000) 0.5 c. c
4-
Complete
0.5 c. c. (1:2,000) 0.5 c. c
4-
Complete
05 c c. (1-5000) 05 c. c
Partial
0.5 c. c. (1:10,000) 0.5 c. c
Partial
0 5 c. c. (1 '20 000) 0 5 c. c
None
From such experiments it follows moreover that the fixation of
alexin, carefully titrated, is a more delicate method of determining
192 INFECTION AND RESISTANCE
the presence of an antigen or, vice versa, of an antibody than is the
observation of a visible precipitate, a fact which has been made use
of, as we have mentioned, by Neisser and Sachs and others for for-
ensic antigen determinations.
It should also be remembered that, if to such a precipitate there
is added an excess of the antigen, the precipitate may be partially
dissolved, and this dissolved precipitate, as Gay 85 has shown, may
possess fixation properties. This, too, accounts for the fact, observed
by a number of workers, that if, in a series of precipitin tests the
supernatant fluids and the washed precipitates are separately exam-
ined for alexin fixation, the fixation properties reside entirely in the
precipitates except in those tubes in which a considerable excess of
antigen was used and in which, as in tubes 1 and 2 of the preceding
protocol, the precipitates were relatively slight. The subject, though
involved, is worthy of detailed consideration in this place since it
seems to us to have an important bearing on certain theoretical con-
ceptions which will be taken up below.
The important question now arises: what is the nature of the
alexin fixation by the complexes formed by unformed proteins with
their antibodies and, more especially, what is the nature of the
alexin fixation exerted by specific precipitates? There have been
much experimentation and speculation concerning this, and a number
of different views are held. Gengou assumed, as we have seen, that
this fixation, as studied by him, was entirely analogous to the fixation
by sensitized bacterial or blood cells. He expressed the belief that
treatment of an animal with an unformed protein produced not only
specific precipitins but also specific sensitizers, analogous to those
produced in response to treatment with bacterial or other cells. He
noticed the parallelism between the quantity of the precipitate
formed and the alexin fixation, but did not associate the two proc-
esses.
His conception of specific antiprotein sensitizers was accepted by
a number of workers, and Wassermann and Bruck,86 Friedberger 87
and several others brought out the facts that actual precipitate forma-
tion is not a necessary criterion of fixation. Thus the last-named
writer showed that the precipitating power of a serum may be de-
stroyed by moderate heat without a corresponding destruction of its
fixing property. A similar independence of the precipitation from
the complement-fixing property, in the presence of an antigen, has
been observed by Muir and Martin.88
85 Gay. Univ. of Cal Public, in Pathol, Vol. 2, No. 1, 1911.
86 Wassermann and Bruck. Mediz. KL, 1905, Vol. 1, No. 55.
87 Friedberger. Deut. med. Woch., 1906, No. 15.
88 Muir and Martin. Jour, of Hyg., Vol. 6, 1906.
FURTHER DEVELOPMENT OF KNOWLEDGE 193
Gay,89 90 also, though he was the first definitely to associate
precipitin formation with the alexin-fixing property and, indeed,
determined a rough parallelism between the amount of precipitate
and the degree of alexin fixation, has nevertheless recently declared
himself in favor of the assumption of the presence in protein anti-
sera of two antibodies, the alexin-fixing lysins and the precipi-
tins. This he does on the basis of certain experiments from which
he concludes that the antigen-antibody complex which fixes alexin
is distinct from the precipitin-precipitinogen complex, but is usu-
ally "brought down in its formation in such a way as to simulate
fixation by the precipitate." Nicolle91 goes even further than
this in declaring that the "coagulins" or precipitins are "anti-
corps bons," which prevent the action of the albuminolysin upon
the antigen, thereby inhibiting the liberation of poisonous cleavage
products.
It seems to the writer9293 that the assumption of a separation
between the precipitin and the albuminolysin is a needlessly compli-
cated interpretation of the phenomena. In order to elucidate this
point a comparison was made between the fixing properties of a
mixture of a protein (sheep serum) and its antibody and a mixture
of typhoid filtrate and antityphoid serum in which it is known that
both precipitins and antibacterial sensitizers are present. It was
shown that, as stated before, in the former mixture the alexin-fixing
property resided entirely in the precipitate, whereas in the latter
case both the precipitate and the supernatant fluid fixed alexin.
From this it seems to follow that immunization with the more com-
plex cellular elements has given rise to the precipitating antibody
present also in the antisheep serum, and, in addition to this, to sensi-
tizers which are not precipitable (remaining in the supernatant
liquid) and not present in the antisheep serum. The precipitates,
moreover, were found to fix "end-piece" and "mid-piece," frac-
tions of alexin, in the same way as these are fixed by sensi-
tized cells.
Without going into further complicated detail, it would seem to
us 94 to be justified that we look upon the so-called precipitins not as
separate antibodies but as identical with so-called albuminolysins.
They unite with the antigen, producing an alexin-fixing complex.
Since both reacting bodies are colloidal in nature, they precipitate
each other in the test tube, but, following the laws governing other
mutually precipitating colloids, they do so only when brought to-
89 Gay. Loc. cit.
90 Also Univ. of Cal. Publ. in PathoL, Vol. 2, No. 1, 1911.
91 Nicolle. Ref. in Bull, tie VInst. Past., Vol. 5, 1907.
92 Zinsser. Journ. Exp. Med., Vol. 15, 1912.
93 Zinsser. Proc. of Soc. of Exp. Biol. and Med., April, 1913.
94 Zinsser. Journ. of Exp. Med., Sept., 1913.
194 INFECTION AND RESISTANCE
gether in concentrations which lie within definite zones of relative
proportions. The visible precipitation would seem, therefore, to be
merely a secondary phenomenon, the essential one being the union of
an antigen with a sensitizer by which it is rendered amenable to the
action of the alexin. This would enable us to comprehend also the
experiments of Friedberger, discussed in the section on anaphylaxis,
in which it was shown that the action of alexin upon precipitates
gives rise to the formation of toxic bodies just as this occurs when
alexin acts upon sensitized cells. It leads, moreover, to a compre-
hension of the processes of the digestion of intravascularly intro-
duced foreign proteins, which are rendered amenable to the digestive
action of the alexin by the antibodies spoken of as precipitins, which
functionally and in structure are conceived as identical with other
sensitizers.
Dean,95 who has lately analyzed the relation between precipita-
tion and alexin fixation on the basis of extensive experimentation,
comes to the conclusion that the proportions of antigen and antibody
which are favorable for rapid and complete precipitation do not
favor the most complete alexin fixation. He states that the two reac-
tions do not run a parallel course but believes that this does not mean
that they are necessarily distinct phenomena. He says : "They rep-
resent two phases of the same reaction ... a flocculent precipitate
represents the final stage of a change which can be recognized in its
earliest and incomplete stage by means of a complement fixation."
Our view differs from this only in that we believe that the pre-
cipitation is merely a secondary, colloidal phenomenon, which may,
or may not, coincide with the phase of greatest alexin fixation, ac-
cording to other fortuitous conditions which may favor or retard
flocculation. Indeed, if our view be accepted, rapid compact pre-
cipitation may possibly be assumed to interfere with alexin fixation
in that it would inhibit perfect contact of the alexin with the antigen-
antibody complexes.
Another view of the mechanism of alexin fixation is that which
has been advanced by Neufeld and Haendel.96 These workers have
found that sensitized cholera spirilla will fix hemolytic complement
at 0° C., whereas the same bacteria at 37° C. will fix both the hemo-
lytic and the bactericidal complement. They conclude from this that
the fixation at 37° C. was brought about by virtue of the bactericidal
amboceptor, whereas at 0° C. fixation was brought about by an anti-
body which is distinct from amboceptor or sensitizer. They believe
from this and other observations, which we cannot consider in detail,
that alexin fixation may be brought about by a special fixing anti-
95 Dean. Zeitschr. f. Imm., Vol. 13, 1912.
96 Neufeld and Haendel. Arb. a. d. kais. Gcsund., Vol. 28, 1908.
FURTHER DEVELOPMENT OF KNOWLEDGE 195
body, the "Bordetscher Antikorper," which is not identical with any
of the other known antibodies.
In all experiments which deal with alexin fixation by specific
antigen-antibody complexes it is of the greatest importance that we
should guard against the errors easily introduced by fortuitous non-
specific antihemolytic agencies. Thus there are a number of factors
which will interfere with the functionation of alexin upon a sensi-
tized antigen, either by direct non-specific absorption of the alexin
itself or by producing physical conditions in the presence of which
alexin cannot act.
Thus many animal tissue cells, in emulsion, will absorb alexin,
and the same property may be possessed by tissue extracts. Von
Dungern 97 was the first to call attention to this, and his observations
have been variously confirmed. Muir 98 showed that the stromata of
hemolyzed red blood cells exert strong anticomplementary action,
and that this is due to a firm union with the complement. It is not un-
likely that the action of cells in this respect is referable to their
lipoidal contents. This suggestion was first made by Landsteiner
and von Eisler," who found that the petroleum-ether extracts of red
blood cells possessed strong anticomplementary action which, to a
limited extent, was specific toward the particular corpuscles from
which the extracts had been made. Similar observations have been
made by Noguchi,100 who speaks of the substance he extracts as "pro-
tectin." In general, the protective action of the lipoidal extracts
seems to depend largely upon cholesterin, and, since this substance is
present to some extent in many tissues, their antihemolytic action is
easily understood. In another section we have discussed the similar
neutralizing action of lipoidal substances upon poisons of various
kinds (saponin, tetanolysin, and snake poison), but, as we have noted
there, the neutralizing properties of the extracts do not, as a rule,
equal those of the whole tissues.101 It is not unlikely that in such
cases as Landsteiner suggests the potent agent is not the lipoid itself
but rather a lipoid-protein combination, a class of substances of which
we know very little, but the importance of which, in many phases of
serum reactions, seems assured.
We have already mentioned that yeast cells may absorb alexin.
And it has been found by Wilde 102 and others that almost all bac-
teria in emulsion may possess varying degrees of alexin-fixing prop-
erties even though unsensitized. There seems to be no regularity
either qualitatively or quantitatively in regard to this, but the fixa-
97 Von Dungern. Munch, med. Woch., Nos. 20 and 28, 1900.
98 Muir. "Studies in Immunity," London, Vol. 19.
99 Landsteiner and von Eisler. Wien. kl Woch., No. 24, 1904.
100 Noguchi. Journ. Exp. Med., Vol. 8, 1906, p. 726.
101 See also Ivar Bang, "Bioehemie der Lipoide."
102 Wilde. Berl. kl. Woch., 1901, Vol. 38, and Archiv f. Hyg., 39, 1902.
196 INFECTION AND RESISTANCE
tion is usually sufficiently marked to render the use of whole bacteria
unreliable for specific fixation experiments. For this reason, as we
will see, bacterial extracts must be used in such work unless careful
quantitative controls are made. Upon what this fixation depends it
is difficult to determine. It may be that it is purely non-specific and
due to absorption of the fine emulsion of the bacteria comparable to
that observed on the part of kaolin or quartz sand emulsions, or, pos-
sibly fixation by such bacterial emulsions may occur because of the
small amounts of normal sensitizer almost always present in the
serum employed as alexin.
Apart from the lipoids, a number of other substances have been
found to fix alexin and exert consequent antihemolytic action. Thus
Landsteiner and Stankovic,103 and Landsteiner and von Eisler 104
describe the anti-alexic action of various proteins coagulated or pre-
cipitated. They refer this action not to particular chemical struc-
ture but to the colloidal state, since they obtained similar antilytic
action with such inorganic emulsions as quartz sand and kaolin
(aluminium-orthosilicate). Since anticomplementary action has,
moreover, been noted in the case of a large number of extracts of
such materials as wool, leather, etc., it is clear that the methods of
alexin fixation, as applied to the forensic differentiation of blood,
must be carefully controlled with this point in view.105
Among the most practically important non-specific agencies
which fix alexin there are some which appear under certain condi-
tions in normal serum. Noguchi 106 has found that serum will often
develop anticomplementary properties as a consequence of heating
during the process of inactivation. On more detailed investigation
he determined that the anticomplementary action increased as the
serum was heated to about 90° C. Above this temperature it is de-
stroyed. He refers this property to the serum lipoids, since he was
able to remove it by extraction with ether, the ether extract possessing
the same anticomplementary power as the original serum.
Neisser and Doring 1 °7 have noticed anti-alexic or anticomple-
mentary properties of human sera which were destroyed on heating,
and which they associate with disease of the kidneys, since they
noted it in sera of uremic patients. Browning and McKenzie 1(
have observed a similar heat-sensitive anti-alexic action on the part
of normal serum, and the subject has been studied by Zinsser and
Johnston.109 It was found that all normal sera will develop anti-
103 Landsteiner and Stankovic. Centralbl f. Bakt., 1906, Vols. 41 and 42.
104 Landsteiner and von Eisler. Wien. kl. Woch., 1904, No. 24.
105 Uhlenhuth. Deut. med. Woch., 1906, Nos. 31 and 51, and Centralbl. f.
Bakt., 1906, I, Ref., Vol. 38.
106 Noguchi. Journ. of Exp. Med., Vol. 8, 1906, p. 726.
107 Neisser and Doring Berl. kl. Woch., 1901, No. 22.
108 Browning and McKenzie. Journ. of Path, and Bact., Vol. 13, 1909.
109 Zinsser and Johnston. Journ. of Exp. Med., Vol. 13, 1911.
FURTHER DEVELOPMENT OF KNOWLEDGE 197
alexic properties on preservation at room temperature within a few
days, and more slowly but no less regularly in the ice chest. This
anti-alexin is destroyed on heating to 56° C., and may be precipitated
out with the globulins of the serum. There appeared in these studies
no particular association between the anti-alexic property and ne-
phritis.
The action of alexin upon sensitized cells may be prevented, also,
by physical or chemical conditions without actual fixation or binding
of the alexin. We refer to the effects of the addition of salts, prob-
lems which have been considered above.
CHAPTER VIII
PEACTICAL APPLICATIONS OF THE COMPLEMENT-
FIXATION METHOD
THE WASSERMANN REACTION
THE principle of specific alexin fixation has been practically
utilized in the diagnosis of disease and in the forensic determination
of the nature of spots of blood or other protein material.
Soon after Bordet and Gengou's experiments Wassermann and
Bruck 1 showed that bacterial extracts could be successfully substi-
tuted for whole bacteria in these reactions. Citron,2 too, made sim-
ilar observations, and, indeed, we now know that the use of bacterial
extracts is more suitable for these experiments than are emulsions
of whole bacteria, since, as we have mentioned above, bacterial emul-
sions may often fix small amounts of complement of themselves
(without specific sensitization), thereby confusing the results of the
reaction.
On the basis of their experience with bacterial extracts Wasser-
mann and Bruck 3 then determined that complement fixation could
be carried out in tuberculosis when the various tuberculin prepara-
tions were used as antigen.4 These investigations fell into the period
during which active research upon the Spirochoeta pallida in syphilis
was going on, and it occurred to Wassermann that the technique of
complement or alexin fixation might be utilized in the diagnosis of
syphilis. Together with Neisser and Bruck 5 he subjected this idea
to experimental test. The publication of their first results appeared
in 1906. They used in their experiments the syphilitic monkeys
which were being observed in Neisser's clinic. Their method con-
sisted in mixing inactivated serum from syphilis-inoculated monkeys
with organ extracts, serum, etc., of syphilitic human beings, and
1 Wassermann and Bruck. M ed. Klinik, Vol. 55, 1905.
2 Citron. Centralbl. f. Bakt., Vol. 41, 1906.
3 Wassermann and Bruck. Deut. med. Woch., No. 12, 1906.
4 Complement fixation in tuberculosis is not yet on a practical or reliable
basis. Recent claims of Besredka (Ann. Past., 1913) for his new antigen
promise a successful technique, but no extensive confirmation has followed up
to the present time.
6 Wassermann, A. Neisser, and Bruck. Deut. med. Woch., No. 19, 1906.
198
PRACTICAL APPLICATIONS OF METHOD 199
adding a small amount of fresh guinea pig complement. After these
materials had been together for a certain time, sensitized red blood
cells were added. If the complement was bound during the first
exposure no hemolysis resulted and the reaction was regarded as
positive. From their results they drew the following conclusions :
1. Immune serum from monkeys, produced by treatment with
syphilitic material, will sensitize syphilitic material from human
beings or monkeys, so that an alexin-fixing complex is formed.
2. Complement fixation results only when the syphilitic immune
serum of monkeys is added to similar material from men or mon-
keys, but not when added to organ extracts of normal men or mon-
keys.
3. Normal monkey serum has no such action.
They concluded that their results justified them in assuming a
specific fixation due to specific antisyphilitic immune bodies in the
blood of the treated monkeys. They excluded experimentally the
possibility of fixation by a precipitin reaction resulting from the
treatment of the monkeys with human material. It might well have
happened that precipitins against human protein appearing in the
serum of the treated monkeys might subsequently react with the
human protein material used as antigen, a complement-fixing com-
plex resulting. This, however, was excluded by the fact that they
obtained positive reactions only when the human material was ob-
tained from luetic lesions.
The same authors, with Schucht,6 very soon after this, extended
their method to the diagnosis of syphilis in human beings. The
same thing had been done shortly before their publication appeared
by Detre7 on a smaller material. By these and many other investi-
gations it was very soon shown that syphilis may be reliably diag-
nosed by complement fixation when extracts of the syphilitic organs,
employed as antigen, are mixed with the inactivated serum of syphi-
litic individuals. It was incidentally shown by Wassermann and
Plaut 8 that the reaction could be obtained not only with blood serum
but also with spinal fluid in paralytic cases.
It was generally assumed, at this time, that the reaction in syph-
ilis depended, as in the case of other infections, upon the presence in
the syphilitic serum of specific antibodies. For it seemed reasonable
to suppose that the specific antigen obtained in the extracts was de-
rived from the extraction of large numbers of spirochetes demonstra-
ble in the extracted organs.
This, of course, is the most logical and simple theoretical concep-
tion of the reaction, and is justified on the basis of analogy. Un-
6 Wassermann, Neisser, Bruck, and Schucht. Zeitschr. f. Hyg., Vol. 55,
1906.
7 Detre. Wien. kl. ~Woch., Vol. 19, No. 21, 1906.
8 Wassermann and Plaut. Deut. med. Woch., No. 44, 1906.
200 INFECTION AND RESISTANCE
fortunately, however, it was soon found by a number of workers,
Marie and Levaditi,9 Weygant, Kraus and Volk, Landsteiner,
Miiller, and Potzl,10 and others that antigens perfectly capable of fix-
ing complement in the presence of syphilitic serum could be pro-
duced from normal organs.11
Theoretically it must be admitted that we are very much in the
dark at present. The fact, now entirely unquestionable, that the
sera of syphilitic patients will give fixation with antigens derived
from extracts of normal organs, as well as from those of syphilitic
organs, seems to throw doubt upon the simple specific antigen-anti-
body conception at first held.
In order to understand the questions involved in the theories of
the Wassermann reaction as at present conceived it will be necessary
to consider the types of antigen which are now employed.
Wassermann' s original method of antigen preparation consisted
in using the liver or spleen of a congenitally syphilitic fetus. The
organs were finely divided and emulsified in 4 to 6 parts of normal
salt solution. This mixture was shaken for 24 hours, eentrifugal-
ized, and the clear supernatant fluid used as antigen. Later the
specific organ substances were extracted by Forges and Meier 12 in
five times the volume of absolute alcohol for 24 hours. This alco-
holic extract was evaporated in vacuo and the residue taken up in
salt solution and shaken until an even suspension resulted.
After it had been discovered that normal organ extracts could
serve as antigen as well as the extracts of syphilitic organs, Land-
steiner, Forges and Meier, and others, introduced antigens produced
by alcoholic extraction of normal organs of animals and of man.
Landsteiner introduced the alcoholic extract of normal guinea pig
organs, especially extracts of the heart and liver, and Weil and
Braun 13 made use of extracts of normal human organs. There are
various methods of preparing extracts for this purpose. We may
mention, to illustrate these methods, the one suggested, first, we be-
lieve, by Noguchi, a procedure which is applicable to the extraction
of normal human organs (spleen), beef hearts, and guinea pig hearts.
The finely divided or triturated organ substance is shaken up with
five times its weight of absolute alcohol and allowed to stand in the
9 Marie and Levaditi. Cited from Mclntosh and Fildes' "Syphilis."
Longmans & Co., 1911, p. 94.
10 Landsteiner, Miiller, and Potzl. Wien. kl Woch., Vol. 20, 1907.
11 An extensive historical review of the development of the Wassermann
reaction is found in the book of Boas, "Die Wassermannsche Reaktion,"
Karger, Berlin, 1911. Since these earlier publications have appeared the
literature of the Wassermann reaction has become very extensive. It is
enumerated more fully than we can afford space for here in the book of
Noguchi ("Serum Diagnosis of Syphilis") and that of Boas, mentioned above.
12 Forges and Meier. Berl kl Woch., No. 15, 1908.
13 Weil and Braun. Berl. kl. Woch., No. 49, 1907.
PRACTICAL APPLICATIONS OF METHOD 201
incubator at 37.5° C., for from 5 to 7 days. At the end of this time
it is filtered through cheesecloth and then through coarse paper, and
the filtrate placed in a large crystallizing dish in which it is evapo-
rated at room temperature with the aid of an electric fan. A gummy
yellow residue is left, which is then taken up in as small a quantity
of ether as possible. This ether solution is then precipitated with
4 times its volume of acetone, in consequence of which there is a pro-
fuse precipitation of coarse white flakes. This acetone-insoluble
substance, which is at first white, later yellowish, in color, is the
stock antigen. A little of this is taken up in a very small quantity
of ether, and this ethereal solution is shaken up in salt solution until
the ether has evaporated or the material has gone into very fine col-
loidal suspension in the salt solution. This is the antigen ready to
be used.
It is immediately evident that these antigenic substances must
consist very largely of lipoidal extractives of the organ substances,
and it has been found that such antigen contains sodium oleate, leci-
thin and cholesterin. Indeed, Forges and Meier have claimed that a
1 per cent, solution of commercial lecithin may be used with success.
Browning and Cruikshank 14 have found further that the addition
of small amounts of cholesterin to syphilitic antigen very largely
increases its specifically diagnostic value, and this idea has since
been utilized more especially by Sachs,15 Walker and Swift,16 and
others. Sachs, especially, has obtained excellent antigens in the
following way: 1 gram of moist guinea pig heart substance was
extracted with 5 c. c. of alcohol and left at room temperature for
twelve hours or in the ice box for two days ; it was then filtered and
0.5 to 1 per cent, of cholesterin was added; frequently the alcohol
extract had to be diluted two or three times before use. Sachs and
Rondoni 17 have also recommended artificial mixtures of lipoids con-
taining sodium oleate, lecithin, and oleic acid.
The fact that cholesterin added to alcoholic organ extracts in-
creases the antigenic value of these for the Wassermann reaction is
all the more curious inasmuch as cholesterin alone has practically no
antigenic action. Walker and Swift have recommended an antigen
in which alcoholic extracts of human or guinea pig hearts were made
up to 0.4 per cent, of cholesterin, 0.4 per cent/having been found by
comparative test to be the most favorable concentration. Cholesterin-
liver extracts or even alcoholic extracts of syphilitic livers without
cholesterin were found to be inferior in specific antigenic value to
0.4 per cent, cholesterin-heart antigens. From the experience of
many investigators it now seems unquestionable that additions of
14 Browning- and Cruikshank. Journ. of Path, and Bact., Vol. 16, 1911-
15 Sachs. Berl kl Woch., No. 46, 1911.
16 Walker and .Swift. Journ. of Exp. Med., Vol. 18, 1913.
17 Sachs and Rondoni. Zeitschr. f. Imm., Vol. 1, 1909.
INFECTION AND RESISTANCE
cholesterin increase the delicacy of the reaction in that more cases
react positively with such an antigen than with the imcholesterinized
preparations. The experience of Hopkins and Zimmermann, how-
ever, would indicate that great caution must be exercised when the
reaction is done in this way, since occasional positive results are
obtained with cases clinically not syphilitic. These workers believe
that cholesterinized antigen is extremely useful, but advise its use
only parallel with the ordinary lipoidal antigens and together with
careful study of the clinical aspects of the case.
The fact that these antigens are non-specific in origin naturally
necessitates careful determination of their usefulness before they
are used. Before any antigen can be regarded as reliable, therefore,
a titration must be carried out in the following way : Two series of
tubes are prepared, in the first of which antigen and complement are
added to normal serum, and in the second the same substances are
added to known syphilitic serum. The antigen must, of course, be
such that in no test tube does it cause alexin fixation in the presence
of normal serum, but, in the quantities used, it must give fixation
regularly with syphilitic serum. An example of such a titration
may be tabulated as follows :
EXAMPLE OF ANTIGEN TITEATION
Antigen by Landsteiner's method : normal guinea pig heart
freed from fat and ground up in a mortar. To each gram is added
5 c. c. of absolute ethyl alcohol and the mixture allowed to extract
at 60° C. for 12 hours (or several days at 37.5° C.). It is then fil-
tered through paper. The following titration is then carried out :
A
Tube 1
Tube 2
Tube 3
Tube 4
Tube 5
Normal serum
0.2
0.2
0.2
0.2
Antigen
0 05
0 1
0 2
0 3
0 6
Alexin
0.1
0 1
0 1
0.1
0 1
B
Tube 1
Tube 2
Tube 3
Tube 4
Syphilitic serum
0 2
0 2
0 2
0 2
Antigen
0.05
0.1
0.2
0.8
Alexin
0 1
0 1
0 1
0 1
The volume in all of these tubes is brought to 3 c. c. with isotonic
salt solution. After one hour at 37.5° C., sensitized red cells are
PRACTICAL APPLICATIONS OF METHOD 203
added to each tube.18 If the antigen is suitable in that it does not
fix alexin by itself or in the presence of normal serum, hemolysis
will result in all of the tubes of series A. If it is suitable in that it
fixes in the presence of syphilitic serum, the tubes in series B will
show no hemolysis ; if there is slight hemolysis in B 1, it is inferred
that 0.05 c. c. of the antigen is insufficient, and the smallest amount
(0.1 c. c.), which completely fixes 0.1 G. c. of alexin in the presence
of the positive serum, is the quantity used. Again the antigen may
be able to cause hemolysis by itself if used in too large amounts. If
this is the case in tube B 4, then this antigen is suitable only in
amounts varying between 0.1 c. c. and 0.2 c. c.
The titration is done with varying quantities because too little
antigen might fail in fixing the alexin, even if the serum were posi-
tively syphilitic, whereas too much antigen might possess alexin-fix-
ing properties in itself, even in the presence of normal serum, or
possibly without any serum at all, an attribute which is not uncom-
monly possessed by lipoidal extracts.
It is thus seen that Wassermann reactions can be carried out
with antigens which do not contain extracts of syphilitic lesions or
of the micro-organisms which give rise to syphilis. This fact alone
would exclude the possibility of considering the fixation of comple-
ment as at present carried out in the Wassermann reaction as being
due to a specific antigen-antibody union.
This conclusion is strengthened by the recent discovery that a
specific antigen prepared from cultures of 8pirocJiceta pallida cannot
be successfully used in diagnostic Wassermann tests. The first in-
vestigations of this kind were made by Schereschewsky,19 who used
as antigen extracts of mixed cultures in which the spirochete was
present ; his results were inconclusive. Noguchi 20 later investigated
this phase of the problem, preparing his antigens by the extraction of
pure cultures and of syphilitic rabbit testicles in which the spirochetes
were very profuse. He found that positive tests with such an antigen
were obtained only in isolated cases of prolonged syphilis which had
been thoroughly treated, and that the ordinary Wassermann reaction,
as obtained in active cases, is not due to antibodies which combine
specifically with the pallida antigen. Craig and ISTichols 21 also have
found that cases of untreated syphilis which gave positive reactions
with syphilitic liver extracts gave absolutely negative results when
culture antigens were used.
18 Tube "5" is the antigen control which shows that the antigen in large
amounts is neither anticomplementary nor hemolytic by itself. It is well, in
addition, also to test out various amounts of the antigen and alexin, without
either normal or syphilitic serum, to determine the largest amount of antigen
which, by itself, is devoid of the actions mentioned above.
19 Schereschewsky. Deut. med. Woch., 1909, p. 1653.
20 Noguchi. Journ. A. M. A., Vol. 58, 1912.
21 Craig and Nichols. Journ. of Exp. Med., Vol. 16, 1912.
204 INFECTION AND RESISTANCE
From these results also we may infer that the Wassermann reac-
tion does not represent a fixation of alexin by the union of a specific
syphilitic antigen with antibodies found against the Spirochceta pal-
lida. Noguchi concludes that it is caused by "lipotropic" substances
in the sera of syphilitic human beings ; a conclusion which is justi-
fied by the fact that the antigens used, all of them, contain large
quantities of lipoids. It must be acknowledged, however, that we
have no definite information concerning the nature of the reaction
beyond this. Schmidt 22 believes that it is a colloidal reaction, and
depends upon the union of the serum globulins with the extract
colloids in the antigen. In normal serum such a union is prevented
by the albumins which act as a sort of protective colloid. In syph-
ilitic serum the globulins are increased quantitatively or are changed
qualitatively in the degree of their dispersion, or possibly in both
characteristics. He regards the serum globulins in the Wassermann
reaction as directly uniting with the extract colloid.
Levaditi and Yamanouchi 23 also conclude that the Wassermann
reaction depends upon the union of two colloidal substances — one a
non-proteid constituent of syphilitic serum (cholesterin derivatives
or fatty acids), the other the lipoidal constituents of the antigen.
Like others they found that the active substances in the antigenic
extracts are non-protein and alcohol soluble.
It is interesting to note, moreover, that Forges and Meier 24 ob-
served actual precipitation when syphilitic serum was added to
lecithin emulsions. In consequence, attempts have been made to
make the diagnosis of syphilis by direct precipitation of syphilitic
serum by such emulsions of lecithin and of sodium glycocholate
(Merck). The results of these investigations as well as those of
Klausner,25 who claims that syphilitic sera are more easily precipi-
tated by distilled water than are normal sera, have led to no diag-
nostically reliable results, but they have seemed to show that the
serum globulins are probably more plentiful and more easily pre-
cipitated out of syphilitic than out of normal sera.
The inference of many workers, therefore, has been that the
Wassermann reaction is primarily due to the precipitation of (prob-
ably) globulin by the lipoidal colloids of the antigen, the resulting
precipitate being capable of absorbing alexin. Jacobsthal 26 has ex-
amined mixtures of syphilitic serum and antigen by the ultramicro-
scopic method, and claims that precipitates are always present even
when they are not macroscopically visible. Bergel,26 who has re-
22 Schmidt. Zeitschr. f. Hyg., Vol. 69, 1911.
23 Levaditi and Yamanouchi. C. R. de la Soc. de BioL, 1907, Vol. 63, p.
740.
24 Forges and Meier. Berl kl. Woch., No. 15, 1908.
25 Klausner. Wien. kl Woch., No. 7, 1908.
26 Jacobsthal. Munch, med. Woch., 1910.
27 Bergel. Zeitschr. /. Imm., Vol. 17, 1913.
PRACTICAL APPLICATIONS OF METHOD 205
cently suggested the importance of specific lipase production as a
cause of hemolysis, suggests that the Wassermann reaction is due to
fixation exerted by the products of the action of a specific lipase
formed in the syphilitic body against "lues-lipoids." This theory is
open to objections similar to those mentioned above, namely, that
the antigen need not necessarily be a lues-lipoid, but may be derived
from normal organs. Other theories have been brought forward by
Bruck, Weil, Braun, Manwaring, and more recently by Rabino-
witch.28 The data supporting most of these theories are, as yet, too
speculative to justify our discussion of them at any length. The
only fact which seems established with any reasonable certainty is
the independence of the Wassermann test from a specific antigen-
antibody reaction in the usual sense.
Although the Wassermann reaction is thus apparently not based
on those principles in the investigation of which it was discovered,
its practical diagnostic value is not therefore diminished. . For its
proper performance any of the methods of antigen preparation con-
sidered above may be employed, provided that the usefulness of the
preparation utilized is carefully controlled in each case as indicated.
Since, of course, a hemolytic system is used in such tests as an in-
dicator, it is necessary also to titrate sensitizer and alexin.
From what has been said in another place concerning the quanti-
tative relations of alexin and amboceptor or sensitizer (see reference
to work of Morgenroth and Sachs, p. 163), it is evident that the use
of too strongly sensitized cells might result in hemolysis, if a slight
fraction of alexin were left unbound by a weak syphilis reaction.
Conversely the use of too large a quantity of alexin would result in
hemolysis, since, even if the amount of syphilitic fixation were con-
siderable, a sufficient excess of alexin might remain. The use of
uniform amounts of fresh guinea pig serum in each case does not
control this adequately, for different specimens of guinea pig serum
may vary considerably in alexin content. In consequence, titra-
tions of both sensitizer and alexin should be made. For practical
purposes it is quite enough to titrate the hemolytic sensitizer every
few weeks and use a stated amount in successive reactions. The
alexin or complement can then be titrated individually for each set
of reactions. Examples of such preliminary titrations follow:
Titration of Hemolytic Amboceptor or Sensitizer
Rabbit injected 3 times at 5-day intervals with washed sheep
corpuscles . . . . , 3, 4, and 5 c. c., and bled 10 days after the last
injection.29
28 Rabinowitch. Centralbl. f. Bakt., Orig., 1914.
29 In immunizing1 animals with blood cells for this or any other purpose
it is necessary to wash the cells very carefully in salt solution. Unless this is
206 INFECTION AND RESISTANCE
This serum is inactivated at 56° C. for 20 minutes.
Washed sheep corpuscles
5% emulsion
in salt solution
Sensitizer
Fresh
g- P-
serum
Hemolysis
1
1 c. c
0 01
0 1
-f- -4-4-
2
1 c. c
0.005
0.1
4-4-4-
3
Ice
0 003
0 1
4-4-4-
4
1 c. c
0.001
0.1
4-4-4-
5
Ice
0 0005
0 1
4-4-
6
1 c. c.
0.0002
0 1
=b
7
Ice
0 1
8
1 c. c
salt sol.
In this case 0.001 c. c. still causes complete hemolysis of 1 c. c.
of a 5 per cent, emulsion of sheep cells (volumetric measurement of
cells sedimented in the centrifuge), and this amount (-rsW. c. c.)
is called the "hemolytic unit" of sensitizer ; two units are then used
in the reactions.
Against these cells alexin can, in each case, be titrated as follows :
Alexin Titration:
Fresh Guinea Pig Serum Pipetted from Clot
Red cells
5% emulsion
Sensitizer
as above
determined
Guinea pig
serum
Hemolysis
1
2
3
4
1 C. C.
1 C. C.
1 c. c.
1 c. c.
2 units (.002)
2 units (.002)
2 units (.002)
2 units (.002)
0.1 c. c.
0.05 c. c.
0.025c. c.
0.01 c. c.
rf
The smallest amount of alexin which completely hemolyzes the
red cells (0.05 in this case) is the amount used. Since it is easier
to measure larger volumes with accuracy, the alexin is diluted 1 to
10 in salt solution before use. A typical Wassermann reaction can
then be carried out as follows :
done blood serum or plasma will be injected with them and the treated animal
will respond by the formation not only of hemolysin but of precipitins for
the serum proteins as well. When a subsequent hemolytic test is carried out,
a precipitin reaction between the precipitin in the antiserum and serum ad-
hering to the corpuscles will follow, and this, as we have seen, will fix alexin,
obscuring other reactions which may be under observation.
PRACTICAL APPLICATIONS OF METHOD
207
SCHEME FOR WASSERMANN TEST
ADAPTED TO ORIGINAL WASSERMANN SYSTEM AFTER SCHEME OF NOGUCHI
Test with known
Test with known
Test without serum
Test with
positive syphilitic
negative normal
to control efficiency
unknown serum
serum
serum
of hemolytic system
Serum 2 c. c.
« Serum .2 c. c.
Serum .2 c. c.
d
i
o _j_
_i_
*l
O Complement
„- O Complement
O Complement
O Complement
2 c3
.1 c. c.
.1 c. c.
.1 c. c.
.1 c. c.
11
4-
-}_
-|_
&|
Salt sol.
1 Salt sol.
Salt sol.
Salt sol.
•I
3 c. c.
3 c. c.
3 c. c.
3 c. c.
2.
4.
6.
8.
Serum .2 c. c.
Serum .2 c. c.
Serum .2 c. c.
O Complement
O Complement
O Complement
O Complement
.1 c. c.
d .1 C. C.
.1 c. c.
.1 c. c.
j_
d -U
-j-
-|-
d
* &
Antigen
™ Antigen
Antigen
Antigen
ef
(required amount
|
£*
in 1 c. c. salt sol.)
|
2£
fc-g
Salt sol.
3 +
Salt sol.
Salt sol.
Salt sol.
2 c. c.
2c. c.
2c. c.
2c. c.
1.
3.
5.
7.
O = Test tube.
Place in water bath at 40° C. for one hour, then add to all tubes red blood
cells and amboceptor. These are previously mixed so that 2 c. c. contains the
equivalents of 1 c. c. of a 5 per cent, emulsion of sheep corpuscles and 2 units of
amboceptor. Again expose to 40° C. If the serum tested is positive, tubes 1
and 3 should show no hemolysis, all the other tubes showing complete hemolysis
in one hour.
Since many human sera normally contain small amounts of antisheep
sensitizer, it is the habit of many workers to add the sheep corpuscles, without
the sensitizer or amboceptor, and incubate for a half -hour. If, at the end of this
time, no hemolysis has occurred either in the front or the back ^ row, then
amboceptor may be added. This technique avoids the possible error introduced
by an excess of amboceptor, a condition which easily occurs when any large
amount is normally present in the serum and in addition to this 2 units are
added as in the test described above.
The above represents the typical "Wassermann7' as at present
carried out in most laboratories. It may be carried out just as well
and with greater economy of material by using one-half the amounts
throughout. It is evident that the performance of the reaction calls
for experience of serum technique, and knowledge of such reactions,
so that fortuitous irregularities may be intelligently controlled. It is
our opinion that the performance of routine Wassermann tests by
workers without a thorough knowledge of the fundamental facts of
208 INFECTION AND RESISTANCE
serum phenomena is worse than useless in that insufficient attention
to special conditions and to details may easily result in a positive re-
action when syphilis is not present, and vice versa.
Recently Archibald McNeil and others have exposed the mix-
tures of complement, antigen, and patients' serum at refrigerator
temperature for a number of hours instead of in the water bath or
thermostat at 37.5° C., before adding the sensitized cells. It is a
curious fact, which has not yet been satisfactorily explained, that
such a procedure increases the delicacy of the reaction. It may be
that, when the tubes containing the antigen, patient's serum, and
alexin are left at incubator temperature, partial alexin fixation only
can take place during the brief period of 30 minutes to one hour,
which is usually employed. More prolonged exposure at this tempera-
ture would not be advisable on account of deterioration of the alexin.
On the other hand, at ordinary ice-box temperatures of about 8° to
10° C., the exposure can be continued for as long as 10 hours without
extensive complement deterioration, and meanwhile more complete
fixation can occur. This, however, is a surmise. The actual condi-
tions are not clear. As a matter of fact in our laboratory Dr. Otten-
berg, in 120 cases so far done in parallel series, one being exposed
for fixation for 30 minutes at 37.5° C., the other at 8° to 10° C. for
three hours, found discrepancies between the two methods in 15
cases. In all of these, positive reactions were obtained by the ice-box
method, whereas by the water bath method the results were negative.
Of these cases 7 were clearly unquestionable syphilitics, two were
treated syphilis, and four were probably syphilitic.
Many modifications of the Wassermann test have been suggested.
Probably the most important is that of Noguchi. The chief justifi-
sation for this modification is the fact that many normal human sera
contain hemolysins for sheep corpuscles. For this reason many
workers carry out the ordinary Wassermann technique without add-
ing antisheep sensitizer or amboceptor until they have first observed
whether or not the tested serum (in the "back row," without antigen)
will not hemolyze the corpuscles without such an addition, adding
the sensitizer only when this does not take place. This is advisable
since the presence of any considerable amount of normal antisheep
sensitizer in the human serum which is being examined (if added to
the amount used in the ordinary reaction, 2 units), may so increase
the total quantity that hemolysis wrill result even after most of the
alexin has been fixed. Noguchi excludes this uncertainty by avoid-
ing the use of the "sheep cell-antisheep sensitizer" system entirely,
substituting a hemolytic complex consisting of human cells and anti-
human sensitizer, produced by injecting washed human corpuscles
into rabbits.
His technique may be best illustrated in the following tabula-
tion:
PRACTICAL APPLICATIONS OF METHOD 209
Reagents
1. Sensitizer prepared by injecting washed human blood corpuscles into
rabbits.
2. 1 per cent, emulsion of washed human blood cells.
3. Alexin — fresh guinea pig serum diluted with one and one-half volumes
of salt solution, 40 per cent.
The reaction is performed in the following way :
Noguchi's Method of Complement Fixation for the Serum Diagnosis of Syphilis
Rear row |
Set for diagnosis
Test with the serum in
question
Positive control set
Test with a positive syphi-
litic serum
Negative control set
Test with a normal serum
a. Unknown ser-
um, 1 drop*
b. Complement,
2 units
O c. Corpuscle
susp., 1 c. c.
a. 'Positive syph.
serum, 1 drop*
b. Complement,
2 units
O c. Corpuscle susp.,
1 c. c.
a. "Normal serum,
1 drop*
b. Complement,
2 units
Oc. Corpuscle
susp., 1 c. c.
|
j|
k
1
.
ll
°1
1?
I"
1 Incubation at 37° C. for 2 hours longer,
then at room temperature.
| Front row
a. Unknown ser-
um, 1 drop*
b. Complement,
2 units
O c. Corpuscle
susp., 1 c. c.
+ Antigen
a. 'Positive syph.
serum, 1 drop*
b. Complement,
2 units
O c. Corpuscle susp.,
1 c. c.
+ Antigen
a. "Normal serum,
1 drop*
b. Complement,
2 units
Oc. Corpuscle
susp., 1 c. c.
-h Antigen
* When working with inactivated serum 4 drops (0.08 c. c.) should be em-
ployed. With cerebrospinal fluid, 0.2 c. c. (not inactivated) is used.
(Taken from Noguchi's " Serum Diagnosis of Syphilis," Lippincott, 1910,
p. 57.)
Bauer 30 has introduced a modification in which he utilizes the
presence of normal sheep sensitizer in many human sera. He per-
forms his tests without the addition of antisheep sensitizer at
first, adding this only to those tubes in which controls have shown
that no normal sensitizer is present. Stern,31 on the other hand,
utilizes the alexin normally present in human serum. The syphi-
litic serum to be tested is, therefore, not inactivated, and the
sheep cells are more heavily sensitized (9 to 12 units). It seems
to us that this method is objectionable chiefly because of the
anticomplementary action which develops in most normal human
sera if kept for a short time, and which can be removed only by
inactivation.
Other modifications of the Wassermann reaction are those of
30 Bauer. Semaine Medicate, 28, 1908.
31 Stern. Zeitschr. f. Imm., Vol. 1, 1909.
810 INFECTION AND RESISTANCE
Jacobaeus 32 and of Wechselman.33 It seems, however, that, as the
reaction is gaining in importance in clinical diagnosis, most labora-
tories are adhering to the original system used by Wassermann and
his associates, except for the substitution of the non-specific lipoidal
antigens for the originally employed organ extracts.
The value of the Wassermann test in the diagnosis of the various
stages of syphilis is a problem which can be approached only by
careful statistical analysis of the results obtained. This has been
done by various investigators, and some of the results have been
tabulated in the books of Noguchi, of Boas, and of Mclntosh and
Fildes. The figures we cite are those largely taken from Boas, as
summarized in F. C. Wood's "Chemical and Microscopical Diag-
nosis" (D. Appleton & Co., 1911), pp. 706 et seq.
Primary syphilis, 974 cases, 56.5 per cent, positive.
The reaction may appear before the primary sore, but this is
very rare. Usually it is positive in from 5 to 6 weeks after infection.
Secondary syphilis, 2,762 cases, 88 per cent, positive. In untreated
cases they are stated to be 100 per cent, positive.
Tertiary syphilis, 830 cases, 80 per cent, positive.
Tabes, 360 cases, 70 per cent, positive.
Dementia paralytica, 95 to 100 per cent, positive.
The tabulation on the following page, taken directly from Boas,
will give a comprehensive summary of this phase of the problem.
Since the reaction is not a specific antigen-antibody union but
depends on some substance liberated or produced by reason of the
syphilitic injection, it is not out of question that other infections
may give rise to a "positive Wassermann." And this, indeed, is
the case. It was claimed for a time that a positive reaction may
be obtained in tuberculosis, but this has been refuted by subsequent
experience, and the earlier positive results probably depended upon
faulty technique. There can be little doubt, however, that occasional
positive reactions are obtained in cases of leprosy, scarlet fever,
malaria, and trypanosoma infections.
The spinal fluid may be used instead of the blood serum in cases
of syphilis of the central nervous system, but even here, as Citron 34
has shown, the results with blood serum are more frequently positive
than those done with the spinal fluid itself. In isolated cases posi-
tive reactions have been obtained with ascitic fluids, pleural and
pericardial exudates. Bab 35 reports a case of positive reaction in
32 Jaeobaeus. Zeitschr. f. Imm., Vol. 8, 1911.
33 Wechselmann. Zeitschr. f. Imm., Vol. 3, 1909.
84 Citron. Deut. med. Woch., 1907, No. 29, p. 1165.
85 Bab. Munch, med. Woch., Vol. 46, 1907.
PRACTICAL APPLICATIONS OF METHOD
Table Compiled by Boas, loc. tit., p. 138.
Stage of disease
Number of
cases
Positive
reaction
Negative
reaction
Control cases (not syphilitic)
1,064
1
1 063
Induration
76
(scarlatina)
56
20
Secondary
Early untreated
269
269
o
Recurrent after treatment
199
187
12
Tertiary
No treatment of early tertiary mani-
festations
63
63
o
Treatment
20
16
4
Latent syphilis
Within 3 yrs. after infection
After 3 yrs
243
111
89
44
154
87
Tabes
Untreated
17
17
o
Treated .
26
11
15
Dementia paralytica
Serum
139
139
o
Spinal fluid
67
61
6
Congenital
With symptoms
54
54
o
Without symptoms
10
7
3
the milk of a syphilitic mother. Serum obtained at autopsy is not
suitable for the reaction, since this, for unknown reasons, may often
give a positive reaction in non-syphilitic cases.
COMPLEMENT OK ALEXIN FIXATION AS A METHOD OF
DETERMINING THE NATURE OF UNKNOWN
PROTEIN
FORENSIC ALEXIN FIXATION TESTS
Our preliminary discussions of the principles underlying alexin
or complement fixation have revealed that alexin is bound not only
by sensitized cells but also by the specific precipitates formed when
an unformed protein antigen is mixed with its specific antiserum.
This discovery, made by Gengou, was attributed by him, it will be
remembered, to the presence of "albuminolysins," or protein sensi-
tizers, antibodies which have been by many observers regarded as
separate from the precipitins, but which we believe, for stated rea-
sons (see p. 193), to be very probably identical with the precipitating
antibodies or precipitins. However this may be, when a dissolved
antigen is mixed with its antiserum alexin fixation is exerted by the
INFECTION AND RESISTANCE
complex, and this, even when the reacting quantities, antigen and
antibody, are so small that visible precipitation will not take place.
For this reason, it is plain, it should be possible by means of com-
plement fixation to detect amounts of a foreign protein too small to
be demonstrable by direct precipitation with an antiserum.
The method has, therefore, been suggested chiefly by Neisser and
Sachs 36 for the forensic determination of unknown proteins, as an
adjuvant to, and improvement upon, the forensic precipitin test. Our
discussion of the principles involved in the introductory paragraphs
of this chapter will render unnecessary an extensive discussion of
the reasoning upon which this reaction is based. It is well to remind
the reader, however, of the facts which we have discussed regarding
the quantitative proportions which govern the occurrence of precipi-
tation when an antigen, say human serum, is mixed with its antibody,
in this case antihuman rabbit serum. The actual precipitation may
be absent either when an excess of the antigen is used or when the
antigen is present in too small a quantity. Thus a given quantity
of the antiserum may precipitate strongly dilutions of the antigen
ranging from 1-50 to 1-10,000. No precipitation or, at least, a very
slight one only may occur when concentrations stronger than 1-50
are used and when the dilution is greater than 1-10,000. Neverthe-
less, in both cases, alexin fixation may be exerted by the complex
although no precipitation takes place. As Gay 37 has shown, com-
plement fixation may be exerted even when a formed precipitate has
been redispersed by the subsequent addition of more antigen. The
importance of the forensic reaction of Neisser and Sachs, however,
lies chiefly in its application to the detection of quantities of un-
known protein too small to be detected by precipitin reactions.
The tests are carried out by mixing a dilution of unknown pro-
tein with given quantities of antiserum, adding small quantities of
alexin (quantities determined best by previous alexin titration as
indicated in our section on the Wassermann reaction) ; these reagents
are left together for a given time at 37.5° C., and then sensitized cells
are added to determine whether or not the alexin has been bound.
The table on the following page, taken directly from the article
of Neisser and Sachs, loc. cit., will not only illustrate the method of
carrying out the reactions but will also give an indication of their
extreme delicacy.
It will be seen that 0.00001 c. c. of the normal human serum still
gave almost complete complement fixation of 0.05 c. c. of comple-
ment in the presence of 0.1 c. c. of the antihuman serum. The table
also shows that this reaction follows a general law of relative specific-
ity so often noted in other reactions, namely that, of all the ani-
mals tested, the serum of monkeys alone gave reactions with the
86 Neisser and Sachs. Berl kl Woch., Vol. 42, No. 44, 1905, p. 1388.
87 Gay. Univ. of Cal., "Publications in Pathology," 1912.
PRACTICAL APPLICATIONS OF METHOD
Table Taken from Neisser and Sachs, loc. cit., p. 1388
0.1 human antiserum + 0.05 complement and variable amounts of different
normal sera (brought to 1 c. c. volume with salt solution) ; the mixtures kept
1 hour at room temperature. Then added 1 c. c. 5 per cent, washed beef
blood + 0.0015 c. c. amboceptor and left 1-2 hours at 37° C.
The results are as follows :
Amounts
Hemolysis on addition of serum of:
normal
serum
Man
Monkey
Rat
Pig
Goat
Rabbit
Ox
Horse
0.01
0
0
*
-
•
-
-
0.001
0
0
0.0001
0
moderate
com-
com-
com-
com-
com-
com-
0.00001
slight
complete
" plete
" plete
plete
" plete
*" plete
^ plete
0.000001
complete
complete
0
complete
complete
j
J
„
„
human antiserum; and this in quantities as small as 0.001 cubic
centimeter.
The forensic complement fixation reaction of Neisser and Sachs
is both theoretically and practically valid. Its extensive use in many
investigations for theoretical purposes has well established its reli-
ability. However, it is more complicated and requires much more
experimental training and care than does the simpler precipitin test,
and it will rarely occur that an unknown protein is available in
quantities too small to permit of successful precipitation.
THE USE OF COMPLEMENT FIXATION TESTS IN THE DIAG-
NOSIS OF MALIGNANT NEOPLASMS
A great many attempts have been made to establish a method of
complement fixation by which a diagnosis of malignant tumors could
be made. It had been hoped that the substance of malignant tumors
might contain a form of protein or protein lipoid combination which
might represent substances specific for such tumors, and might there-
fore functionate as a specific antigen. On this basis it might be
possible that the serum of tumor patients would contain a specific
antibody which could react with a specific antigen in tumor extracts,,
with the resulting formation of an alexin-fixing complex.
No experimental facts have so far justified our assumption of
the presence of either specific antigen in tumor extracts, or that of
a specific antibody in the serum of such patients. However, we have
seen that the Wassermann reaction is a perfectly useful clinically
diagnostic method, in spite of the fact that the antigen need not be
specific, and the purely empirical basis on which the syphilis reac-
INFECTION AND RESISTANCE
tion is at present based has justified extensive attempts to establish
an analogous empirical method for tumor diagnosis.
The literature on this question is confusing. A number of ob-
servers using antigens variously prepared from tumor substances
have reported favorable results. Simon and Thomas 38 report many
positive reactions, as do Sanpietro and Tesa,39 and a number of
others. Clowes 40 has carried out a reaction on sarcoma rats and
obtained positive reactions in animals in which the tumors were small,
negative ones when the tumor had grown to a large size. Ranzi,
on the other hand, obtained negative results throughout. Ranzi 41
found that normal serum would often give complement fixation with
carcinoma extracts, also that many tumor extracts and sera of tumor
patients inhibited complement by themselves. The reactions were so
irregular that he assumed them to be without value. Recently the
subject has been very thoroughly investigated by v. Dungern.42
Von Dungern claims to have finally evolved a method by which
the diagnosis of malignant disease can be made with reasonable
accuracy. Like the Wassermann reaction his method is purely em-
pirical. He admits that probably it is not a specific antibody deter-
mination and depends rather upon the presence of pathological prod-
ucts of metabolism in the sera of tumor patients. The reliability of
his method depends upon the observation of a number of details
which he has determined empirically.
He obtains his antigen in a purely non-specific manner, using,
as just stated, for this reaction acetone extracts of human blood cells.
We take the description of the reaction entirely from his own article
in "Weichhardt's Jahresbericht." The antigen is prepared in the fol-
lowing way: Blood is taken from a vein, preferably from a para-
lytic patient, since v. Dungern claims that individual specimens of
blood vary, and he hafe had the best results with that of paralytic
cases. Clotting is prevented by sodium oxalate and the blood cells
are thoroughly washed in the centrifuge. To the sediment are added
19 volumes of pure acetone (Merck). This is allowed to stand three
days at room temperature and is occasionally shaken during this
time. It is then filtered, the acetone evaporated in the incubator at
37° C., and the residue taken up in 96 per cent, alcohol. This alco-
holic extract is diluted before use with four parts of salt solution. Of
this final preparation 0.8 c. c. is used in the individual test.
Particular precautions must also be taken in the handling of the
serum of the patient. In his earliest tests v. Dungern determined
38 Simon and Thomas. Journ. Exp. Med., Vol. 10, 1908.
39 Sanpietro and Tesa. Cited from v. Dungern in "Weichhardt's Jahres-
bericht," etc., Vol. 8, 1912, p. 163. '
40 Clowes. Journ. A. M. A., 1909, Vol. 52.
41 Ranzi. Wien. kl. Woch., 1906, p. 1552.
42 V. Dungern. Munch, med. Woch., Nos. 2, 20, 52, 1912; Berl. kl. Woch.,
1913, "Weichhardt's Jahresbericht," Vol. 8, 1912, p. 163.
PRACTICAL APPLICATIONS OF METHOD 215
that the inactivation of the tumor sera greatly diminishes their spe-
cific fixation properties, and for this reason he at first advised that
the serum be used unheated. He has found recently that the best
results are obtained when the serum is heated to 54° C., together with
a little sodium hydrate solution. He handles 'the blood in the fol-
lowing way: After being taken from the patient it is allowed to
stand 1 to 2 days in the refrigerator; just before use he adds two parts
of an -^5- NaOH solution with one part of serum and heats it for half
an hour at 54° C. As it is important that the sodium hydrate should
contain no sodium carbonate, he advises the use of the Kahlbaum
preparation. In setting up the test he uses graded quantities of the
mixture corresponding to 0.2, 0.1, 0.05, and 0.025 c. c. of the original
serum. To each of these quantities he adds the stated quantity, 0.8
c. c. antigen preparation described above, and the 0.05 guinea pig
complement. Controls must be set up with the antigen alone and
with the patient's serum alone to prevent error from independent
fixation by these substances. These reactions are allowed to stand
three hours at room temperature, and then one cubic centimeter of
a 5 per cent, solution of beef blood sensitized with two units of hemo-
lytic serum is added (as in the Wassermann reaction). It is im-
portant to use a strongly sensitizing serum, so that not too much of
the hemolytic rabbit serum must be added to the tubes. Experi-
ments done in this way with normal sera usually result in complete
hemolysis within one hour, although in certain other diseases, i. e.,
tuberculosis and syphilis, slight inhibition may result. However,
fixation with the patient's serum in quantities of 0.1 c. c. or less is, ac-
cording to von Dungern, fairly specific for malignant tumors, since
normal sera treated in the way described usually do not cause fixa-
tion in quantities of less than 0.2 c. c. and, in syphilis and tubercu-
losis, if fixation is at all present, it is usually not evident in quanti-
ties less than 0.1 c. c.
With a reaction so carried out von Dungern has examined 244
cases. The following tabulation states his results:
Malignant tumor of
No. of cases
Reaction positive*
Pharynx
3
3
Esophagus
6
6
Stomach
15
11
Rectum
14
10
Larynx
2
2
Tongue
5
5
Bladder
1
1
Breast
22
22
Uterus
10
10
Skin
8
7
Ethmoid bone
1
1
Upper maxilla . .
1
1
* Taken from von Dungern, "Weichhardt's Jahreabericht," Vol. 8, 1912, p. 174.
216 INFECTION AND RESISTANCE
We report von Dungern's results exactly as lie states them in
his last summary, since his well-known experimental ability necessi-
tates serious consideration of all of his work. We may say, however,
that a survey of the entire literature of complement fixation in the
diagnosis of malignant tumors does not yet justify our acceptation
of this method as of anything like the established value which the
similar method has attained in syphilis.
COMPLEMENT FIXATION IN GLANDEES
The diagnosis of glanders by the mallein test and by agglutina-
tion has been recently reenforced by the method of complement fixa-
tion. In carrying out these tests the method of preparation of the
antigen is of the greatest importance. The directions which we give
are those employed in the Diagnostic Laboratory of the New York
Department of Health, under the immediate supervision of Dr.
McNeil and Miss Olmstead, from whom we have our information.
The particular strain of glanders bacilli employed seems to be of
little importance. The organisms are grown on 1.6 per cent, acid
glycerin-potato-agar. This stock culture is transplanted every other
day. From it cultures are planted upon salt-free veal peptone agar.
It is of the greatest importance that this medium shall be neutral to
phenolphthalein. After twenty-four hours in the incubator the
growth is washed off with distilled water, which also should be neu-
tral, and the emulsion heated for from four to six hours at 80° C. in
a water bath. It is then filtered through a Buchner filter simply to
facilitate subsequent filtration through a Berkefeld "N" or "V"
filter. After filtration this antigen is again sterilized for one hour
at 80° C. and then on two successive days at 56° C. for one-half hour.
The fixation tests carried out with these antigens have yielded
excellent results as reported by Dr. McNeil 43 at the New York
Serological Society.
It is unnecessary to give further directions as to the technique of
this reaction, since it is simply that of complement fixations in gen-
eral, the chief difficulty being that of antigen preparation.
COMPLEMENT FIXATION IN GONOEEHEAL INFECTIONS
There are certain conditions following gonococcus infection of
the genito-urinary tract which are not easily distinguished from a
number of other tests unless the organisms can be cultivated or a
specific serum reaction can be applied. Most important of these are
gonorrheal rheumatism, salpingitis, and endocarditis. Complement
fixation with the sera of such patients, and an antigen produced from
43 McNeil, Archibald. N. Y. Serological Soc. Meeting, April 4, 1914.
PRACTICAL APPLICATIONS OF METHOD
gonococci, has been employed by many observers during recent years,
and promises to be of great value.
Here, too, the production of the antigen is the only feature of
the reaction which has offered difficulties. Since the researches of
Torrey have shown that not all races of gonococcus are antigenically
alike, it seems necessary to produce a polyvalent antigen. At the
Xew York Department of Health at present the antigen is prepared
by using the ten Torrey strains. Stock cultures are carried on neu-
tral veal agar and cultures are planted upon salt-free veal agar.
Twenty-four-hour growths, washed off in neutral distilled water, are
kept in a water bath at 56° C. for two hours, and then filtered
through, first, a Buchner and then a Berkefeld filter. They are then
sterilized for one hour.
COMPLEMENT FIXATION IN TUBERCULOSIS
Attempts to apply complement fixation to the diagnosis of tuber-
culosis have been made by as many as thirty or more investigators
with varying results. Wassermann and his collaborators attempted
it before they succeeded in developing the Wassermann reaction in
syphilis. Recently, intensive work has been done on the subject by
Besredka,44 Petroff,45 Craig,46 Bronfenbrenner,47 and Miller and
Zinsser.48 Results have warranted the application of the reaction to
clinical tuberculosis, although the actual degree of usefulness of the
reaction must still await the multiplication of cases tested. The diffi-
culty has of course consisted in the development of a suitable antigen.
The antigen of Petroff has consisted of a filtrate of potato broth.
Besredka has used a filtrate of cultures made upon egg broth. Craig
has used Besredka's antigen and suspensions of ground bacilli. Mil-
ler and Zinsser 48 have employed an antigen made by triturating liv-
ing and dead bacilli with crystals of table salt, then adding distilled
water to isotonicity. With all of the antigens favorable results have
been obtained. Our own results seem to check up with clinical
diagnoses in over 80 per cent, of the cases and so far have appeared
to give practically no positive results in negative cases and indicated
only active tuberculosis and not healed lesions.
The reactions are carried out by methods entirely analogous to
those employed in the Wassermann reaction,' but careful antigen
titrations must be done.
44 Besredka. Ztschr. f. Immunit., 1914.
45 Petroff.
46 Craig-. Am. Jour. Med. Sc., Dec.. 1915, p. 781.
17 Bronfenbrenner: Arch. Int. Med., 1914, No. 6, 786. Ztschr. f. Immunit.,
1914, XXIII, 2, 21.
48 Miller & Zinsser: Proc. Soc. E.rper. Biol and Med., 1916, 134. Jour.
Lab. and Clinical Med., Vol. I. p. 817.
CHAPTER IX
THE PHENOMENON OF AGGLUTINATION
WHEN bacteria are added to homologous immune serum there
occurs a peculiar agglomeration of the individual micro-organisms
into small clumps. The phenomenon is so general and so easily
observed that it is not surprising that it was noticed and reported by
a number of workers during the period of active investigation upon
serum reactions which preceded and followed the discovery of the
Pfeiffer phenomenon. Thus, in the years from 1891 to 1895,
Metchnikoff,1 Charrin and Roger,2 Isaeff and Ivanoff,3 Washburn,4
and several other workers made this observation with a variety of
bacteria and immune sera. But all of these observers failed to follow
up or analyze the process they incidentally noticed in the course of
other investigations. A thorough study of the phenomenon was not
made until 1896, when Gruber and Durham,5 in Vienna, in the
course of their studies upon bacteriolytic reactions with colon bacilli
and cholera spirilla, again noticed the agglutination of these bacteria
in homologous immune sera, recognized the specificity of the reaction,
and called attention to its apparent independence of other previously
studied serum phenomena.
The process known as agglutination consists in the following
train of occurrences. If we add to an even emulsion of bacteria a
small amount of homologous immune serum the micro-organisms
may be seen to collect rapidly in groups or masses, with a resultant
clearing of the fluid in which they have been suspended. The clumps
of bacteria gather in flakes which, not unlike flakes of snow,
sink to the bottom of the test tube. The speed and completeness
with which this phenomenon occurs depend, as we shall see,
upon the agglutinating strength and other qualities of the serum
which is employed, but the essential process of clumping is alike
in all cases.
There are a' large number of different methods by which this
• 1Metchnikoff. Ann. de I'lnst. Past., 1892.
2 Charrin and Roger. C. E. de la Soc. de Biol, 1889.
3 Isaeff and Ivanoff. Zeitschr. f. Hyg., Vol. 17, 1894.
4 Washburn. Journ. of Path, and Bact., 1896, p. 228.
6 Gruber and Durham. Munch, med. Woch., 1896.
THE PHENOMENON OF AGGLUTINATION
occurrence can be observed, each one particularly adapted to some
special purpose for which the reaction is carried out. Gruber and
Durham, who were investigating the properties of bacteriolysins
when they observed agglutination, naturally recognized the specific
nature of the reaction and proposed to make use of it for the purpose
of bacterial differentiation and species determination. For this pur-
pose, which has become one of the most important of the practical
applications of the agglutination reaction, the phenomenon is best
observed by the so-called "macroscopic method," in which a series
of serum dilutions are mixed, in small test tubes, with equal volumes
of emulsions of the bacteria. Thus, if we wish to determine the
nature of an unknown bacillus, suspected of belonging to the typhoid
bacillus group, by this meth-
od, we may determine its ag-
glutination in the serum of
an animal immunized with
a known strain of typhoid.
The tubes are incubated
after the mixtures have been
made, and the agglutination
which has taken place in the
various tubes is recognized
by a clearing up of the fluid
and the flaking of the bac-
teria after from one to three
hours. The test tube method
has the advantage of permit-
ting the use of larger quanti-
ties of reagents than can be
used in the other methods
described below, and therefore more exact quantitative measurements
can be made.
Although this method for the determination of bacteria has found
universal application, it is probably most frequently employed at the
present time for the rapid identification of colonies of doubtful
typhoid or dysentery, incident to the isolation of these organisms for
stools by such methods of plating as those of Conradi-Drigalski, of
Endo, or of Hiss. The suspicious colonies can thus be fished directly
to an agar slant, and the cultures, when developed, emulsified and
identified by agglutination. The advantages of such a method for
the determination of the biological interrelationship of the organ-
isms of a given group, like, for instance, that of the dysentery bacilli,
are obvious.
An ingenious use of this reaction was also made by Shiga when
he determined, among various bacteria isolated from the stools of
dysentery cases, the particular one which was specifically aggluti-
MICEOSCOPIC AGGLUTINATION.
220 INFECTION AND RESISTANCE
nated by the patient's serum, thus discovering the dysentery bacillus
which bears his name.
Within a few months after the publication of Gruber and Dur-
ham's work, Widal and, apparently independently of him, Griiii-
baum,6 by a process of reasoning the converse of that detailed above,
applied the reaction to the diagnosis of infectious disease.
It is obvious that a human being or an animal infected with a
given variety of bacteria may develop agglutinating properties
against these bacteria. It is of great value, therefore, to determine
the agglutinating power of the serum of a patient for the bacteria
which are known to cause the disease suspected in the particular case
in which a diagnosis is desired. This method has become a routine
measure in the early diagnosis of typhoid fever under the name of
"Widal" or "Gruber-Widal" reaction and, since the quantities of
serum which can easily be obtained from a patient are usually small,
it is convenient to carry out the reaction by the microscopic method.
This consists in mixing serum and bacterial emulsion in hang-drop
preparations and observing them with the microscope. An excellent
method, also, is the so-called Proescher 7 method in which serum and
bacterial emulsion are mixed in small watch-glasses or salt cellars.
Proescher used this method extensively in the study of staphylococcus
agglutinations. The mixtures in the salt cellars were set away at
37° C. for two hours, and then observed with a magnification of GO
to 70 diameters.
Close observation of the occurrence under the higher power of a
microscope shows that the bacteria, if motile, lose their motility, if
non-motile the Brownian motion is arrested. They are then rapidly
gathered in small clumps, isolated individuals between these clumps
being gradually drawn into them, until finally the fluid between the
masses is entirely clear. This complete clearing, of course, happens
only when there is not an excess of bacteria, for, like other serum re-
actions, this phenomenon is a quantitative one in which definite pro-
portions must be maintained.
Clinically the most frequent use of the agglutination reaction is
in the diagnosis of typhoid fever. The technique used for this test
is, in the large majority of cases, the microscopic hang-drop method.
In Germany the Proescher method is sometimes used, and the micro-
scopic method with dead organisms, as first introduced by Ficker, is
also not uncommon at the present day.
Since the serum of normal human beings very often contains
moderate agglutinating powers for the typhoid bacillus, the diag-
nostic value of the reaction in this disease depends upon the elimina-
tion of this error by sufficient dilution. If dilutions of the serum of
from 1-40 to 1-60 are used diagnostic errors on this point are
6 Griinbaum. Lancet, 1896, Vol. 2.
7 Proescher. Centralbl f. Bakt., Vol. 34, 1903.
THE PHENOMENON OF AGGLUTINATION
avoided, since the normal agglutinating power of human beings is
never such that typhoid bacilli will be clumped by it in these dilu-
tions within one hour. Prompt clumping in serum dilutions of 1-20
is fairly reliable, but does not absolutely exclude an unusually high
normal agglutinating power. In carrying out tests clinically dilu-
tions of 1-20, 1-40, and 1-80 are usually made and observed for one
hour. From such tests diagnosis can be made without danger of
error. In rare cases of icterus the agglutinating power for typhoid
bacilli may be increased. Just what is the cause of this is not cer-
tain ; Wood 8 reports cases in which agglutination of 1-40 was pres-
ent with slight jaundice (hepatic cirrhosis). On the other hand he
has frequently failed to notice agglutination in other cases of intense
jaundice. It is not impossible, as Wood suggests, that the occasional
presence of unusual agglutinating power in individuals with jaun-
dice has some relation to the frequent persistence of typhoid bacilli
in the gall bladder.
Occasionally it will be noticed that dilutions of the patient's
serum of 1-5 to 1-20 fail to agglutinate, while higher dilutions will
give positive tests. This is referable to the so-called "pro-agglu-
tinoid zone," the principles underlying which we shall discuss in an-
other place.
The Widal test in typhoid cases rarely appears before the end
of the first week, and, in the majority of cases, is present before the
end of the second week. It may proceed for months, although Wood
states that he has seen it disappear at the end of three to six weeks.
In paratyphoid fever the diagnosis can often be made by agglu-
tination, and in dysentery, as we have seen, the fact that the pa-
tient's blood agglutinated the bacteria was one of the important facts
utilized by Shiga in his discovery of the organism which bears his
name.
In pneumonia agglutination of the pneumococcus, isolated from
the patient's sputum by sera prepared by immunization with various
types of pneumococci, has become of considerable importance clin-
ically, since Neufeld and Haendel and, in this country, Cole, Dochez,
and Gillespie have determined that there are a number of different
types of this micro-organism. The use of pneumococcus serum in the
disease will be of value only if a serum is used which has been pro-
duced with an organism of the same type as the one infecting the
patient. Therefore, determinations of the type by highly potent
agglutinating sera give a finger-point to the variety of serum to be
used. Whatever may be the eventual outcome of the serum treat-
ment in pneumonia, no results whatever can be expected, according
to our present knowledge, unless such determinations are made. The
technique of agglutinations in pneumococcus work is facilitated by
8 Wood. "Chemical and Microscopical Diagnosis," Appleton & Co., p. 242.
222 INFECTION AND RESISTANCE
growing mass cultures of organisms, as advised by Hiss, in flasks
of glucose broth containing 1 per cent, calcium carbonate.
The same method of growing micro-organisms is useful in the
case of streptococcus agglutinations, since the insoluble calcium car-
bonate, if thoroughly shaken, breaks the chains of streptococci and
thereby facilitates judgment as to the reaction.
Agglutination reactions have been of considerable usefulness also
in the diagnosis of glanders in horses. The early work on this sub-
ject was done chiefly by MacFadyean,9 and the reaction has been
particularly studied by Wladimiroff.10 Since the serum of normal
horses will often agglutinate glanders bacilli in dilutions of as much
as 1-500, Wladimiroff advises making the positive diagnosis on dilu-
tions only higher than 1-1,000, since he states that normal horses may
occasionally reach an agglutination titre of 1-1,000. The same writer
states, moreover, that glanders bacilli are subject to great variations
in agglutinability, and that for this reason the choice of a suitable
strain is of great importance.
The motility of bacteria has absolutely no relation to the reac-
tion, and their agglutination is entirely passive.
Some of the earlier investigators of agglutination associated the
reaction with alteration in the flagellar mechanism of the micro-
organisms. It is now well known, however, that non-motile, as well
as motile, bacteria are subject to the phenomenon, and that no visible
change in the appearance or arrangement of flagella accompanies the
clumping. Although this is the case, observation of the motility of
such organisms as the bacillus of typhoid fever, while subjected to
the action of agglutinating serum, may be of great value in aiding
in the determination of the degree of completeness with which the
reaction is taking place.
Agglutination, furthermore, does not lead to the death of the
bacteria. Of course, whenever the reaction is carried out in un-
heated serum the concomitant effects of the bactericidal substances
bring about bacterial death. Agglutination does not, however,
depend upon the cooperation of alexin, and serum may be inactivated
- without interference with its power of agglutination. In such heated
serum clumping takes place without bactericidal effects, and, more
than this, the bacteria may grow, if exposed to proper temperature
conditions, when suspended in the serum. In fact, it is of consider-
able interest to carry out the reaction in this way, for the bacteria
growing in agglutinating serum form long convoluted threads and
skeins even when in ordinary culture they habitually occur as sep-
arate individuals. Thus colon bacilli, typhoid bacilli, pneumococci,
cholera spirilla, and other organisms, which ordinarily grow as free
single cells, or, at most, in chains of two or three, if kept in the
9 Macfadyean. Journ. of Comparative Path, and Ther., Vol. 9, 1896.
10 Wladimiroff. "Kolle u. Wassermann Handbuch," 2d Ed., Vol. 5.
THE PHENOMENON OF AGGLUTINATION
incubator for ten to twelve hours together with homologous serum,
will grow in long, delicate chains, like those of streptococci. This
form of reaction has been especially studied by Pfaundler,11 who
attributed particularly delicate specificity to it. However, the
"Thread Reaction" of Pfaundler, as it is sometimes called, is merely
another manifestation of the phenomenon of agglutination and sub-
ject to the same laws and limitations of specificity which apply to
other methods.
The purely passive role played by the bacteria in agglutination
is best shown by the fact that dead bacteria, killed in various ways,"*"
are specifically clumped just as are the living cultures.12 On this
fact depends the method spoken of as "Ficker's Reaction," in which
emulsions of typhoid bacilli, killed by formaldehyd or carbolic acid
(distributed commercially), are agglutinated in small test tubes by
the serum of typhoid patients. The original method of Ficker is
said to be a proprietary secret ; however, a number of other methods
which attain the same purpose are in use in various places. Volk 13
describes the method used in Vienna, and states that there carbolic
acid is used to kill the cultures. Similar to this is the method de-
scribed by J. H. Borden,14 who proceeds as follows :
The bacilli are grown on agar slants in large tubes for 24 hours.
They are then washed from the medium with a sterile mixture of
salt solution 450 parts, glycerin 50 parts, and 95 per cent, carbolic
acid 2.5 parts. After this solution has been kept a week it becomes
translucent and by this time the bacilli are dead. The preparation is
then ready for use and can be kept a long time in dark bottles in a
cool place. Borden very carefully controlled this bacterial emulsion
with positive and negative typhoid sera and found it reliable. The
great advantage of all these methods, of course, consists in the possi-
bility of furnishing the general practitioner with materials for clini-
cal agglutination tests in which the necessity of preserving and sus-
pending living cultures is eliminated.
The facts which we have just considered tend to show that agglu-
tination is not a vital phenomenon 15 dependent in any way upon
the living nature of the bacterial cell, but, like other phenomena of
antigen-antibody reactions, a purely chemical or physical process in
which the substance of the bacterial cell enters specifically into rela- ^
tion with the agglutinating factor of the serum. In uniformity with
other analogous reactions the antigenic substance is here spoken of ^
as "agglutinogen," the antibody as "agglutinin." N
11 Pfaundler. Wien. kl. Woch., 1898, and Centralbl. f. Bakt., I, Vol. 23,
1898.
12 Bordet. Ann. Past., Vol. 10, 1896.
13 Volk. "Kraus und Levaditi Handbuch," Vol. 2.
14 Borden. Medical News, N. Y., Mar., 1903.
15 Bordet. Ann. Past., Vol. 10, 1896.
INFECTION AND RESISTANCE
The agglutinogen, or agglutinin-inducing substance in the bac-
teria is apparently an inherent part of the bacterial protein, and
agglutinins may be produced in animals by injection not only of
living and dead whole bacteria, but by bacterial extracts, prepared
in various ways. And, furthermore, just as the agglutinins of serum
are absorbed out of a serum by the whole bacteria, they may be neu-
tralized by the dissolved bacterial extracts.
Just what the nature of the agglutinogen is has been a matter
of prolonged controversy, Pick 16 and others claiming that it is pos-
sible to obtain an agglutinogen by alcohol precipitation from old
bacterial cultures which, upon further purification, can be found to
give none of the usual protein reactions (Biuret, Millon). It is by
no means certain, however, that Pick's results are correct. In fact,
many objections have been advanced against them, and the accept-
ance of an antigen of non-protein nature is so opposed to all our
knowledge regarding antigens in other cases that Pick's results
should not be taken as final until very careful revision of the experi-
mental facts has been carried out. That the agglutinogen is, to a
certain extent, subject to dialysis has been claimed because of ex^
periments in which specific agglutinins have appeared in the sera of
animals into whose peritoneal cavities celloidin sacs, filled with bac-
teria, have been placed.17
There has been a great deal of discussion regarding the possible
localization of the agglutinogen of bacteria in the ectoplasmic layers
of the cells, and especially in the flagellar substance. We have seen
that, as a matter of fact, nonmotile bacteria are subject .to the phe-
nomena of agglutination just as are the motile forms, but numerous
attempts were made during the earlier stages of our knowledge of
this reaction to demonstrate that changes in ectoplasm and flagella
accompanied the actual agglutination. Gruber18 himself held the
opinion for a time that the clumping was due to an eotoplasmic
swelling which rendered the bacteria sticky, causing them to hold
together after chance approximation. He soon gave up this idea
himself, but a similar theory was for some time upheld by ISTicolle 19
and others.
Malvoz 20 in 1897 devised an ingenious method by which he be-
lieved that he could show that the agglutination of bacteria depended!
upon their ectoplasmic substances. He passed the typhoid emulsion
through Chamberland filters and, when the bacilli had been caught
16 Pick. "Hof meister's Beitrage," 1901, 1902.
17 This would be in keeping with Pick's work just referred to, and should
be subjected to the same criticism before final acceptance. For a more de-
tailed discussion of these conditions the reader is referred to the article by
Paltauf, "Kolle u. Wassermann Handbuch," Vol. 4, part 1.
18 Gruber. Munch, med. Woch., 1896.
19 Nicolle. Ann. de I'Inst. Past., 1898.
20 Malvoz. Ann. de I'Inst. Past., Vol. 11, 1897.
THE PHENOMENON OF AGGLUTINATION 225
upon the filters, he subjected them to prolonged washing. The ba-
cilli, now removed from the filter by passing fluid through in the
opposite direction, were no longer motile or agglutinable either
by formalin, safranin, or other chemical agents, nor by agglutinating
sera. Dineur,21 repeating the experiments of Malvoz, came to the
same conclusions. He decided that in agglutination the bacteria
became entangled with each other by means of the flagella. Harri-
son,22 in later studies working under Tavel, attempted to dissolve
out the ectoplasmic layers of bacteria with pyocyanase, and from his
experiments also came to the conclusion that the agglutinogen was
contained in the external layers. Similar results were obtained by De
Eossi.23
Further studies on the same problem are those of Smith and
Reagh.24 These investigators worked with two strains of bacilli,
both of which they regarded as belonging to the hog-cholera group,
though the one was motile and the other nonmotile. When rabbits
were immunized with the nonmotile bacillus an agglutinin was ob-
tained which acted upon this bacillus differently and less powerfully
than did the agglutinin produced with the motile one. Contact witK
the nonmotile bacillus did not deprive the serum produced with the
flagellated organism of the agglutinins for the latter. They con-
clude that two agglutinins were involved — one incited by the ecto-
plasm and flagellar substance, the other by the bacterial cell body
proper. Rehns as well as Paltauf have criticized these results by
questioning the species identity, of the two bacilli employed in the
experiments, referring the phenomenon to the occurrence of group
agglutination. •»
As a matter of fact our present knowledge of agglutination no
longer justifies the association of agglutination with flagella,/ !N"on-
motile as well as motile bacteria are readily agglutinated, and we
have much evidence which will be discussed presently which shows
that the agglutination reaction is governed by many of the laws
which obtain in colloidal flocculations.^ This, however, does not ex-
clude the possibility that it is the ectoplasmic zone chiefly which
takes part in the reaction.. Furthermore, loss of motility, which
always accompanies agglutination when a motile organism is under
observation, is an extremely valuable aid in guiding us in our judg-
ment of incomplete reactions..
That changes may be brought about in bacterial agglutinogen by
various methods of treatment has been shown by a number of work-
21 Dineur. Bull, tie I'Acad. de Med. de Beige, 1898, cited from Smith and
Beagh.
22 Harrison. Centralbl. f. Bakt., Vol. 30, I, Orig. 1901.
23 De Rossi. Centralbl. f. Bakt., I,Vols. 36 and 40.
24 Smith and Reagh. Journ. of Med. Ees., Vol. 10, 1903.
226 INFECTION AND RESISTANCE
ers, although the fundamental principles underlying such changes
are not at all clear.
Joos 25 was the first to study agglutination with particular refer-
ence to the effects upon the reaction of heating both the antigen and
the antibody. On the basis of extensive and complicated experiments
upon the agglutinin produced in horses by immunization with heated
and unheated typhoid bacilli, he drew the conclusion that agglu-
tinogen (agglutinin-inducing substance) in bacteria was not a single
element but consisted of at least two definite parts of which he speaks
as a and /3-agglutinogen. a-agglutinogen is a constituent of the
living bacteria, and is easily destroyed at 60° to 62° C. The /8-agglu-
tinogen is also present in normal bacilli, but is more heat-resistant
and will withstand 60° to 62° C. for several hours. The injection
of living unheated bacilli then induces the formation of both a
and /?-agglutinin, which have respectively a particular affinity for
a and /8-agglutinogens. The injection of heated bacilli, on the
other hand, induces the formation only of /?-agglutinjn . and not a
trace of a-agglutinin.. The a-agglutinin is considerably heat-
resistant, resisting 60° to 62° C., whereas the /?-agglutinin loses its
agglutinating property when heated to 60° C. The a-agglutinin
is entirely incapable of uniting with /?-agglutinogen.y However,
/?-agglutinogen can combine or react with both the a and p con-
stituents of the bacilli. For this reason Paltauf has spoken of agglu-
tinin produced with the heated bacteria as "umfanglicher." This
is a point of great interest, and if Joos is right is, of course, of con-
siderable practical importance.
However one may look upon these experiments, as well as the
similar ones of other workers, it seems established that heating bac-
teria leaves them still capable of inciting agglutinins powerfully
and rapidly, perhaps of an "umfanglicher" nature than those pro-
duced with the native cells.
Heating bacteria may also alter their agglutinability. Thus, ac-
cording to Eisenberg and Volk,26 heated above 65° C. the bacteria
no longer agglutinate in the presence of specific immune serum, but
still absorb agglutinin. Eisenberg and Yolk, therefore, distinguish
between a heat-sensitive constituent of the antigen, which is
particularly associated with the clumping, whereas the thermo-
stable substance represents the haptophore or combining portion..
It seems simpler, in this case also, to assume a change in the
colloidal stability of the bacteria by heating than to seek it in a
differentiation into combining and agglutinable parts of the same
antigen.
The points raised by Joos' work have been followed up particu-
25 Joos. Centralbl. f. Bakt., Vol. 33, 1903.
26 Eisenberg and Volk. Zeitschr. /. Hyg., Vol. 40, 1902.
THE PHENOMENON OF AGGLUTINATION
larly by Kraus and Joachim 27 and by Scheller.28 Scheller sum-
marizes the results of his work as follows : First, in agreement with
Joos he found that immune sera obtained by injection of bacteria
modified by heat vary considerably from each other. Secondly, im-
munization with living typhoid bacilli produces sera which agglu-
tinate living bacilli very highly and less highly bacilli heated to 60°
C. The titre of agglutinating serum is altered very little toward
living bacilli after heating to 60° to 62° C., but is sometimes dimin-
ished toward bacteria that have been heated. Bacilli jthat have been
heated to 100° C. but slightly agglutinate unheated serum. Sera
produced by the injection of typhoid bacilli heated to 60° to 62° C.
agglutinate with both living and heated bacilli. . Very important
furthermore in Scheller's work are the determinations that typhoid
bacilli which have been heated absorb agglutinins out of the sera
more actively than do the unheated bacteria, and that the highest
agglutinin titres can be obtained by agglutination with bacilli that
have been heated to 60° C. The analogy of Scheller's results with
similar work done in connection with the precipitin reaction is strik-
ing and will be referred to in another place.
Alterations in the agglutinability of bacteria may also occur spon-
taneously, without previous heating, as in the preceding experiments.
It has been frequently noticed that typhoid bacilli, recently culti-
vated out of the human body, will not agglutinate in sera which have
high agglutinating power for strains kept for some time on labora-
tory media. Much investigation has been focused upon the deter-
mination of the cause for this, and although many explanations have
been suggested no satisfactory solution has been found. • Most work-
ers who have attempted to attack this problem have based their rea-
soning upon the receptor conception of Ehrlich and have assumed
that such inagglutinable bacteria are characterized by a diminished
equipment in "receptors." Such strains have been especially well
studied in the case of typhoid bacilli and cholera spirilla. Inagglu-
tinable typhoid bacilli have been cultivated by many investigators
from the spleen, gall bladder, and urine of typhoid patients, and
many of these, when studied for prolonged periods, have been found
to regain normal agglutinability after several generations of culti-
vation upon artificial media.. Apparently some alteration of the
bacilli had taken place in the presence of the body fluids (immune
serum) which affected their sensitiveness to the agglutinins, i. e.,
their ability to unite with or absorb this antibody. The phenomenon
involves an important principle, emphasized some years ago by Pro-
fessor Welch, namely, that the bacteria may acquire a quasi-
immunity against the attacking forces of the body, a property which'
may be responsible for the increase of virulence noted when some
27 Kraus and Joachim. Centralbl f. Bakt., I, Vol. 36, 1904.
28 Scheller. Centralbl. f. Bakt.t Vol. 36, 1904, and Vol. 38, 1905.
228 INFECTION AND RESISTANCE
bacteria are repeatedly passed through the bodies of animals, and,
indeed, alterations of virulence signify biologically a process of
adaptation on the part of the bacteria just as increased immunity
indicates a similar process on the part of the invaded subject.
This lessened susceptibility to antibodies is noticeable not only in
strains cultivated from the body in disease, but can be produced arti-
ficially by cultivating the bacteria in inactivated homologous immune
serum. This has been accomplished by Walker 29 especially, and by
Miiller,30 with both typhoid bacilli and cholera spirilla cultivated
apon broth mixed with serum. Such strains not only increase
in virulence but lose in both agglutinability and susceptibility to
bactericidal effects. Sacqueppee 31 obtained similar results by keep-
ing the organisms in collodium sacs in the peritoneal cavity, and
Bail 32 found similar inagglutinability of typhoid bacilli taken from
the peritoneal exudates of guinea pigs dead of infection.
Zinsser and Dwyer 33 have noticed similar inagglutinability in
typhoid bacilli recovered from the peritoneal cavities of guinea pigs
injected with anaphylatoxin and bacteria. The anaphylatoxin in
these cases possessed distinct aggressive action, and the conditions
here were probably very similar to those observed by Bail.
There are various possible explanations, the most prevalent ones
all representing variations of the opinion that such inagglutinable
strains possess an inadequate receptor apparatus. Cole34 advances
this because he found these cultures possessed less power to absorb
agglutinin than others, and, injected into animals, produced sera
which were not highly agglutinating for the injected strain. Some
of Cole's experiments show clearly the variable agglutinability dis-
played by different strains of the same species. Thus the agglutina-
tion in the same serum
Of strain E = 1:8,000
Of strain H = 1:7,000
Of strain I = 1:4,500
Of strain W= 1:4,500
Of strain C = 1:4,000
The difference here between E and C actually amounted to a
relation of 1 to 2. A rabbit immunized with strain I furnished a
.serum which agglutinated strain E more powerfully than I itself.
Miiller's experiments have the same general significance. It has
;also been suggested that the inagglutinable bacteria, especially those
from the peritoneal exudate, which Bail found were neither agglu-
29 Walker. Journ. of Path, and Bact., Vol. 8, 1902.
30 Miiller. Munch, med. Woch., 1903.
31 Sacqueppee, Ann. Past., Vol. 4, 1901.
32 Bail. Archiv f. Hyg., Vol. 42.
33 Zinsser and Dwyer. Proc. Soc. for Exp. Biol. and Med.,, Feb., 1914.
34 Cole. Zeitschr. f. Hyg., Vol. 46, 1904.
THE PHENOMENON OF AGGLUTINATION 229
tinable nor absorbed agglutinin, may have taken up altered agglu-
tinin or agglutinoid. We will have occasion to recur to this problem
in connection with our discussions of the capsulated bacteria and of
virulence. The explanations given above do not seem on the whole
satisfactory, and the problem is an exceedingly complex one. It has
been found indeed that the acquired resistance of bacteria against
agglutinins is not at all unique, and that acquired resistance against
serum lysins may be observed.35 The extensive investigations of
Bail, Walker,36 and others, on the nature of changes in virulence in
many invasive bacteria, and the knowledge more recently gained on
the resistance to phagocytosis of virulent strains of streptococci and
pneumococci are facts closely related in underlying principle to the
inagglutinability of typhoid strains cultivated in immune sera.
That no two strains of bacteria of the same species are exactly
similar in their agglutinability in the same serum, moreover, is an
observation wrhich is made by every one who is in a position to carry
out routine Widal tests in hospital practice. The spontaneous ag-
glutination which occasionally occurs in the broth cultures of typhoid
bacilli used for this test in many laboratories 37 may often be re-
ferred, at least in the cases which have come to the writer's notice, to
an excessive acidity of the broth, a phenomenon which will be dis-
cussed in a subsequent paragraph. As far as the phenomenon of
variable agglutinability inherent in different strains is concerned,
however, it is of great practical importance in carrying out routine
Widal tests to bear this in mind and to control the strain of typhoid
bacilli employed in the reactions from this point of view. A strain
also which has been in use for a long time should be titrated with
agglutinating animal sera from time to time to determine whether or
not alterations in agglutinability have occurred.
That the reaction of bacterial agglutination was specific was
noted, we have seen, by Gruber and Durham from the very begin-
ning. The closer study of the reaction in its application to bacterial
identification has led to interesting data which have revealed certain
definite limitations of this specificity. It has been found, for in-
stance, that, while immunization with any given species of bacteria
gives rise to a very marked increase of agglutinins for this species,
there are formed at the same time, though to a lesser degree, agglu-
tinins for bacteria of other species. This has been referred to as
"group reaction/' and the agglutinins appearing in such sera are
spoken of by German observers as "Haupt Aggluiinine" and "Neben"
or "Mit Agglutinine" In English texts they are usually referred to
as "chief" or "major" agglutinins and "para" or "minor" agglu-
tinins. Although, as a general rule, such group-agglutinin formation
35 Eisenberg. Centralbl. f. Bakt., Vol. 34, p. 739, 1903.
36 Walker. Centralbl. f. Bakt., Vol. 33, 1903.
37 See section on Aggressins.
230 INFECTION AND RESISTANCE
is evident most markedly in the cases of biologically related micro-
organisms like the typhoid, paratyphoid, and colon bacilli, this is not
necessarily the case, and in some instances the immunization of an
animal with a given bacterium may produce minor agglutinins for
other bacteria that have no morphologically or culturally demonstra-
ble biological relation to that which reacts with the major agglutinin.
We may obtain the most graphic survey of these conditions by exam-
ining one of a number of experimental protocols in which such major
and minor agglutinin formation is illustrated. Thus in the work of
Hiss 38 on the dysentery bacilli the following relations were ob-
served :
A serum produced in rabbits by immunization with the Shiga
bacillus agglutinated the Shiga bacillus in dilutions of 1 to 20,000,
the "Baltimore," "Harris," "Gray," and "Wollstein" bacillus 1 to
1,200, the Y bacillus and others 1 to 200.
An Anti-Y bacillus serum, which agglutinated this bacillus 1 to
6,400, agglutinated the Baltimore bacillus 1 to 1,600, and the Shiga
bacillus 1 to 100.
Anti-"Baltimore" bacillus serum agglutinated this bacillus 1 to
3,200, and the Y bacillus 1 to 400, and the Shiga bacillus 1 to 100.
In this series there is fair correspondence between cultural bio-
logical relations and agglutination, and from many such investi-
gations it would seem that "group" agglutination might be taken
to represent a method of determining biological classifications simi-
lar to the zoological relations revealed by the precipitin reaction.
While, in a general way, this is undoubtedly true, nevertheless great
caution must be exercised in relying upon such evidence for classifi-
cation, since notable exceptions have been observed. Park,39 for
instance, cites a case in which a horse, immunized with a paradysen-
tery bacillus, agglutinated a colon strain in dilutions up to 1 to
10,000. Similarly Durham 40 found that two members of the colon
group — one saccharose fermenting — reacted almost identically with
the same agglutinating serum, while the agglutinations of two cul-
turally identical bacilli of the hog cholera group were entirely at
variance.
The cause for the phenomenon of group agglutination must un-
questionably be sought in the nature of the bacterial agglutinogens,
and it is but reasonable to assume that living cells so little differen-
tiated biologically and morphologically should have much in common
chemically as well. The bacterial cell, moreover, may contain sev-
eral antigenic complexes and, beside its specifically peculiar constitu-
ents, therefore, we may suppose that every bacterium contains addi-
tional antigenic substances which it has in common with other
38 Hiss. Jourri. of Med. Res., 13, N. S., Vol. 8, 1904.
39 Park. "Pathogenic Micro-organisms," 1910, p. 166.
40 Durham. Journ. of Med. Res., Vol. 5.
THE PHENOMENON OF AGGLUTINATION 231
to.ooo
ZONE OF
ABSOLUTS
SPECIFICITY
species. It is the specific antigen in response to which the "chief"
agglutinin is formed, while the others, present in smaller quantity, /
lead to the formation of the minor or paraagglutinins with an in-^
tensity proportionate to the amounts present in the bacterial cell.
Thus, as Durham expresses it, if we assume one micro-organism to
contain antigenic substances a, b, c, and d, and another d, e, f, and g,
the antibodies produced by injections of the former would react with
the common element d in the latter.
The diagnostic value of the specificity, however, is plainly not
affected by the phenomenon of group
agglutination, since the action of minor
agglutinins can be always easily elim-
inated by sufficient dilution. Thus if
we possess a typhoid-immune serum
which agglutinates the typhoid bacillus
in dilutions of 1 to 10,000, the para-
typhoid bacillus 1 to 1,000 and the
colon bacillus 1 to 100 (as in the fig-
ure), we may still utilize this serum for
the identification of suspected typhoid
cultures, as, let us say, in the isolation
of unknown bacteria from stools or
urine, by using potent sera in dilutions
as high or higher than 1 to 1,000, be-
yond which point the action of minor
agglutinins is eliminated. The dia-
gram illustrates our meaning in the
hypothetical case of a typhoid-immune
serum which agglutinates typhoid in
dilutions of 1 to 10,000, paratyphoid
bacilli 1 to 800, and colon bacilli 1 to
100. The relation of agglutination to biologic relationship is not~a
simple problem in that individual strains even of the same species
may vary considerably in agglutination by the same serum. Smith
and Reagh41 have studied particularly these conditions as they pre-
vail in the colon, hog cholera and allied. groups. They found that
biologic relationship usually may be concluded from close agglutina-
tion affinities, and that minor biologic differences such as colony
appearance, etc., do not exclude such affinities. On the other hand,
closely related bacteria vegetating on mucous surfaces (different
strains of diphtheria, dysentery, and colon bacilli) may vary con-
siderably in their agglutinative characteristics, while invasive species
show a greater homogeneity among their varieties or races. This
brings in another important feature — that is, the modification in
41 Smith and Reagh. "Studies from the Rockefeller Institute," Vol. 1,
1904, p. 270.
BOO
SOO
T
TYPHOJD PARATYPHOID COLON
DIAGRAMMATIC EEPBESENTATION
OF GROUP AGGLUTINATION.
232
INFECTION AND RESISTANCE
agglutinative characteristics induced in bacteria when they become
parasitic upon different hosts, and Smith and Reagh conclude that
such changes of parasitic habitat may modify the agglutinative prop-
erties (probably by adaptation to the peculiar reactions of each host),
some of them being weakened and others strengthened.
The animal species used for immunization indeed influences the
quantity and nature of the produced agglutinin considerably. For
instance, in Pfeiffer's 42 experiments, a dog, a chicken, and a rabbit
were immunized with the same strain of cholera spirilla. The
sera obtained from these animals agglutinated this and other strains
of cholera spirilla in an entirely irregular manner — showing that
the constitution of the agglutinins in each case was an absolutely
different one in regard to the relative concentrations of "major" and
"minor" constituents.
Castellani 43 found that the immunization of an animal with
two or more different species of bacteria results in the formation of
agglutinins against all of these, (^upposing, for instance, that species
A and B are used for treatment, agglutinins against both A and B
are formed in quantity, depending upon the intensity of the treat-
ment in each case. Now, if to the serum so produced an emulsion of
A is added, agglutinin A only will be removed, while agglutinin B
will remain in the serum almost undiminished. An example of this
is seen in the following protocol taken from Castellani's paper :
Titre of the serum
Titre after
absorption with
B. typhi
Titre after
absorption with
B. coli "31"
Titre after
absorption with
B. coli and B. typhi
B. typhi 4,000
B. coli'1 31" 1,000
B. typhi 0
B.co&"31"
1,000 > 300
B. typhi 4,000
B. coli" 31" Q
B. typhi 0
B. coli" 31" 0
In the preceding paragraphs, however, we have seen that im-
munization with a single organism, say B. typliosus, will induce the
formation of agglutinins, not only for this species, but also of para
or minor agglutinins for biologically similar strains as well. In
such cases, as Castellani showed, absorption of the serum with the
organism used for immunization takes out, not only the major ag-
glutinins, but rather all of the agglutinins, major and minor. Con-
versely, however, absorption of such a serum with the species ag-
glutinated by the minor agglutinins takes out these antibodies only^
leaving the major substances intact. These relations are well illus-
42Pfeiffer. Quoted from Paltauf in "Kolle u. Wassermann Handbuch,"
Vol. 4.
43 Castellani. Zeitschr. f. Ilyg., Vol. 40, 1902.
THE PHENOMENON OF AGGLUTINATION
trated by the two following protocols, also taken from Castellani's
paper :
Serum of rabbit No. 1 immunized with B. typhi only
Agglutination titre of serum
Titre after absorption with
B. typhi
B typhi 5,000
B. coli . 600
B. typhi 0*
B. coli 0
Serum of rabbit No. 7 immunized with B. typhi only
Agglutination titre
of serum
After absorption with
B. typhi •
After absorption with
B. coli
B.
B.
typhi 10,000
(heavy clumps)
coli 800
B. typhi 0
B. coli 0
B. typhi 10,000 ,
(small clumps)
B. coli 0 -
Note: All of these protocols are taken from Castellani's communication, loc. cit.
These facts, variously confirmed, tend to corroborate the concep-
tion of the production of major and minor agglutinins outlined
above.
It is of practical and theoretical importance to mention that
complete absorption of specific agglutinin by a single exposure to
homologous bacteria, however thickly emulsified, is not possible. It
is always necessary to absorb repeatedly, and even then a minute
trace of agglutinin may eventually remain. Eisenberg and
Yolk,44, 45 who have studied these conditions particularly, attribute
this to the nature of the union of agglutinogen with agglutinin,
which they conceive as following the laws of mass action — this ac-
counting for the persistence of a small "rest" of free agglutinin,
even after repeated absorption by partial dissociation. The prin-
ciple involved here is identical with that discussed in connection with
antigen-antibody union in general. /
It is not only in the sera of immunized animals, however, that
agglutinins are found. Just as the other antibodies, antitoxins, and
bactericidal sensitizers may be found in the blood of normal animals,
so agglutinins for various bacteria may be normally present. These
normal agglutinins do not in any respect, further than that of quan-
tity, differ from the immune agglutinins and follow the same laws
of specificity which have been described for the latter. It has been
shown a number of times that such normal agglutinins are not pres-
44 Eisenberg and Volk. Zeitschr. f. Hyg., Vol. 40, 1902.
45 Eisenberg. Centralbl. f. Bakt., Vol. 34, 1903.
INFECTION AND RESISTANCE
ent in the new-born animal, but are acquired later in life, possibly
because of the absorption of bacterial products from the intestinal
canal. It has been variously shown,46 too, that living bacteria them-
selves may enter the lymphatics and the portal circulation from the
intestine during apparently perfect health of the individual.
This subject is of interest, not only in connection with the ag-
glutinins, but has bearing upon the existence of normal antibodies
in general. Ruffer,47 who has studied particularly the penetration
of leukocytes and bacteria through the intestinal mucosa, demon-
strated micro-organisms in the sub-mucous lymph nodes of normal
rabbits, and Ribbert 48 and Bizzozero 49 have shown the presence of
bacteria in apparently normal mesenteric lymph nodes. Adami
and Nichols even claim that during health a certain number of liv-
ing bacteria enter the portal circulation from the intestine, and from
here may get into the systemic circulation, and are ordinarily de-
stroyed by either leukocytes, liver lymphatic organs, or the kidneys.
It is thus not surprising that normal agglutinins should occur,
and that they should be qualitatively identical with the so-called
"immune" agglutinins, since they probably arise by a sort of spon-
taneous immunization through the intestinal canal. From the in-
vestigations of Ford especially we may conclude that the immune
agglutinin may be regarded as merely a quantitative increase of the
normal antibody, if this has been present before immunization.
Ford 50 found that when an animal is treated with an agglutinating
serum an anti-agglutinin may be obtained which neutralizes the
action, not only of immune, but also of homologous, normal ag-
glutinin.
An interpretation of the process of agglutination, according to
the theory of Ehrlich, conceives it as a chemical union of agglutinin
and bacteria (agglutinogen). The agglutinin is regarded as con-
sisting of two atom complexes, one the "haptophore," having af-
finity for the bacterial protein, and concerned with the union, the
other the "ergophore" or "zymophore," by means of which the actual
agglutination is brought about after the union has taken place. Un- j
like the antibody concerned in the processes of hemolysis or bac- !
teriolysis, the agglutinins are not dependent in their action upon the ,
cooperation of alexin, and the agglutination power of a serum is
therefore not destroyed by inactivation or heating to 56° C., as is
the case with the former. Although the accurate point of thermal
destruction varies with different agglutinins (the agglutinins for the
Bacillus peslis and a few other bacilli are said to be destroyed at
46 Adami. Jour. Am. Med. Assoc., Dec., 1899.
47 Ruffer. Brit. Med. Journal, 2, 1890.
48 Ribbert. Deutsche med. Woch., 1885.
49 Bizzozero. Centralbl. f. d. Med. Wiss., Vol. 23, 1885, p. 49. Quoted
from Adami.
50 Ford. Zeitschr. f. Hyg., Vol. 40, 1902.
THE PHENOMENON OF AGGLUTINATION 235
56° C.), as a rule agglutinins will not disappear from serum until
the temperature is raised to between 70° and 80° C. Once de-
stroyed, however, no reactivation takes place upon the addition of
fresh normal serum. Ehrlich, for this reason, has conceived the
structure of agglutinins as ' Ilaptines of the Second Order," in
which he supposes that the zymophore and the haptophore groups are
inseparably connected, and in which we could assume an alteration
of the less stable zymophore group without interference with the
functional integrity of the haptophore group. Such an altered ag-
glutinin could be spoken of as "agglutinoid," and this could become
united with a bacterial cell without
inducing agglutination, but, by its
union, prevent subsequent combina-
tion of the cell with unaltered agglu-
tinin. This process of Aagglutinoid
Yerstopfung" has been held respon-
sible for the failure of agglutination
when bacteria that have been in con-
tact with heated serum are subse-
quently exposed to the action of ac-
tively agglutinating serum. It is
•^ i ; i . . . , , . , DIAGRAMMATIC REPRESENTATION OF
assumed that the agglutmoids which EHRLICH 's CONCEPTION OF THE
were present in the heated serum STRUCTURE OF AGGLUTININ.
have occupied the bacterial receptors
and have thereby prevented the union of these with the agglutinins
later added.
The work of Eisenberg and Yolk 51 has gone very thoroughly
into these conditions and forms the strongest bulwark of this point
of view. These workers showed that bacteria thus exposed have
not only become less sensitive to agglutinins, but have, at the same
time, lost much of their power to absorb agglutinins when compared
with normal bacteria. The same loss of agglutinating power which
is observable in heated agglutinating serum is evident to a lesser
extent in serum which has been preserved at room temperature.
Eisenberg and Yolk have shown that such serum, in addition to a
loss of its total agglutinin content, loses the power to agglutinate
in high concentrations. Thus a serum which has been preserved in
this way will no longer agglutinate bacteria in concentrations of 1
to 20, 1 to 40, or even 1 to 100,- but will agglutinate as before in
higher dilutions. This is taken to mean that the agglutmoids formed
during the period of standing possess a higher affinity for the bac-
terial antigen than do the true unaltered agglutinins. Since these
so-called "proagglutinoids" are relatively small in amount, their
action is masked when considerable dilution has sufficiently di-
minished their quantity, in proportion to the more plentiful un-
51 Eisenberg and Volk. Zeitschr. f. Hyg., Vol. 40, 1902.
236 INFECTION AND RESISTANCE
altered agglutinins. In support of this assumption it has been
further shown that sera which 'have been rendered inhibitory by
either of the methods named can be deprived of their inhibiting
characteristics by absorption with homologous bacteria. Together
these observations constitute a strong argument in favor of the ag-
glutinoid theory.
In practical experience the existence of such an inhibition zone
is of great importance, since freshly taken sera will occasionally
show this failure of agglutination in concentration, while strong
agglutination follows when the dilution is increased. In clinical
tests, as in the Widal reaction for the diagnosis of typhoid fever,
we not infrequently encounter sera which will give no agglutination
in dilutions of 1 to 20 and even 1 to 40, and the reaction would
therefore be regarded as negative unless the possibility of the pro-
agglutinoid zone were recognized and further dilutions carried out.
While there is no question of the accuracy of the experimental
data cited in the preceding paragraphs, the interpretation of the phe-
nomena on the basis of Ehrlich's haptine conception has not been
universally accepted.
The fact that the agglutinins lose their agglutinating power after
heating, while retaining their power to prevent the subsequent agglu-
tination of the bacteria, may be more simply explained in analogy
with the observation of Forges on the influence of heated serum upon
the agglutination of mastic suspensions. He found that unheated
serum will flocculate such suspensions, while heated serum of the
same concentration will prevent the flocculation, acting probably as
a protective colloid (see chapter on Colloids). In the same way the
heated agglutinating serum may prevent subsequent flocculation by a
similar protective action. We suggest this as a possible explanation
of the proagglutinoid phenomenon, although of course it is a mere
conjecture as opposed to the painstaking experiments of Eisenberg
and Volk. It has the advantages of simplicity, but does not, it is
true, explain the apparent specific absorption of the agglutinin-
inhibiting properties out of heated sera with homologous bacteria, as
claimed by these authors as well as Kraus and Joachim. The simi-
larity of the proagglutinoid phenomenon to the inhibition zones oc-
curring in the flocculation of colloids will be referred to in a subse-
quent paragraph.
In describing the investigations which led to the discovery of
the mechanism of "the lytic phenomenon, in the chapter on Cytolysis,
we mentioned that Bordet and others had noticed the frequent ag-
glutination of red blood cells, in the sera of animals treated with
such cells after the hemolytic property had been destroyed by heat-
ing to 56° C. Such hemagglutination is a phenomenon entirely
analogous to the agglutination of bacteria by serum, and hemag-
glutinins regularly result when an animal is systematically treated
THE PHENOMENON OF AGGLUTINATION 237
with the red blood cells of another species. Like the bacterial ag-
glutinins, the hemagglutinins are relatively thermostable and are
best observed, therefore, after the sera are inactivated. Otherwise
rapid hemolysis will often obscure the agglutination. Like other
agglutinins the hemagglutinins thus produced are specific, acting
only upon that variety of cells which are used in their production.
Moreover, certain sera may normally contain hemagglutinins for the
blood cells of animals of another species. An illustration of this is
the hemolytic and hemagglutinating property of normal goat serum
for rabbit cells — but there are many other examples which might be
cited. JSuch normal hemolytic and hemagglutinating properties for
the cells of other animals usually render thejsera toxic for these ani->
malsj and some observers have attributed the toxicity T;o this agglu-1
tinating action, believing that the clumped erythrocytes form emboli/
around which clotting is initiated. The writer's own investigations,
however, seem to show that this is not the case, since the toxicity of
such sera is completely removed after they have been heated, in spite
of the fact that the hemagglutinative property remains unchanged.
In discussing hemolysins, also, we called attention to the observa-
tion that the sera of animals may develop the property of hemolyzing
blood cells of other individuals of the same species when immunized
with such cells, and that on occasion such isolysins may be normally
present.
Analogous to iso-lysins, iso-agglutinins also jiaye been observed.
They were described first in human blood in lyoOTTndependently, by
Landsteiner,52 and by Shattock,. As the result of the work of a
number of men, in particular that of Landsteiner, of Ascoli,53 and
of Descatello and Sturlii,54 it became evident that all human blood
fell into four sharply separated and, for the individual, permanent
groups, according to the way in which they interagglutinate. The
members of the first group have blood cells which are not agglutinated
by the serum of any human blood. The serum of the members of
this group agglutinates the blood cells of persons belonging to any
of the other three groups. This group constitutes between 40 to 50
per cent, of all human beings. Members of the second group have
blood serum which agglutinates the cells of persons belonging to the
third and fourth groups ; while the cells of the second group are ag-
glutinable by the serum of the first and third groups. The third is
the reciprocal of the second, and the serum of the third group ag-
glutinates the cells of the second and fourth groups ; while the cells
of the third group are agglutinated only by the serum of the first and
second groups. The fourth group (which is very rare) is the recipro-
52 Landsteiner and Richter. Zeitschr. f. Med., 3, 1902.
53 Ascoli. Munch, med. Woch., 1901.
54 Descatello and Sturlii. Munch, med. Woch., 1902, 49, p. 1090.
238
INFECTION AND RESISTANCE
TABLE FOR ISO-AGGLUTININS *
Sera
I
II
III
IV
1
2
3
4
5
6
7
8
9
10
r
1
0
0
0
0
0
0
0
0
0
0
2
0
0
0
0
0
0
0
0
0
0
I'
3
0
0
0
0
0
0
0
0
0
0
4
0
0
0
0
0
0
0
0
0
0
[
5
+
+
+
+
0
0
0
+
+
0
fm
6
+
+
+
+
Q
0
0
+
+
0
i
7
+
+
+
+
0
0
0
+
+
0
\
8
+
+
+
• +
+
+
0
0
0
in\
i
9
+
+
+
+
+
+
+
0
0
0
ivj
10
+
*
i
t
*
+
+
+
+
0
* For this table as well as for much direct information concerning the iso-agglutinins and iso-
lysin we are indebted to Dr. Ottenberg of this laboratory.
cal of the first, the serum having no agglutinin, the cells being ag-
glutinated by the serum of any other group. (See table.) It is
evident from examining this grouping that the phenomena can be
explained (as Landsteiner has suggested) if it is assumed that there
are two agglutinins (a and ft) and two corresponding agglutinogens
present in the red cells (A and B). The blood of the first group pos-
sesses both agglutinins, but no agglutinogens, the blood of the second
group possesses agglutinin a? agglutinogen B, the blood of the third
group possesses agglutinin /?, agglutinogen A, the blood of the fourth
group possesses no agglutinin but both agglutinogens.
These agglutinins are present in weak dilution only, being gen-
erally active in dilutions only of 1-15 to 1-30. They are separately
absorbed from the serum by the suitable red cells (Hektoen).55 Ot-
55 Hektoen. Jour, of Inf. Dis.} 1907, p. 297.
THE PHENOMENON OF AGGLUTINATION 239
tenberg noticed that they were inherited, and this was also shown
in 1908 and in 1910 by von Dungern and Hirschfeld,56 who further
found that this inheritance followed the Mendelian law strictly.
The agglutinogens are the unit characters. The agglutinogens ap-
parently are present at an earlier embryonic stage than the ag-
glutinins. On account of their hereditary nature and permanence
for the individual the iso-agglutinins may possibly be of medicolegal
value. They may also be of some practical importance in selecting
donors for blood transfusion, as phagocytosis of red blood cells in the
circulation after transfusion, first described by Hopkins, was proved
by Ottenberg to occur only when the patient's serum was agglutina-
tive toward the donor's red cells, and several such transfusions have
had fatal results. Similar iso-agglutinins have been observed in the
blood of lower animals, in horses (Klein,57 1902) ; rabbits (Boycott
and Douglas,58 1910) ; cats (Ingebrigtsen) ; dogs, rats, and steers
(Ottenberg).59 The iso-agglutinins have been developed in dogs
(von Dungern and Hirschfeld).60 In most of the lower animals
they have occurred with peculiar irregularity, indicating probably
the presence of, not two, but of a larger number of agglutinins and
agglutinogens. In steers, however, they fall into simple groups,,
indicating the presence of only one agglutinin and agglutiuogen. In
many animals the agglutinins are entirely latent, and the biochemical
differences represented by the agglutinogens are present in the red
cells, and the correct agglutinin is developed by the animal only
when it is immunized with blood whose cells contain agglutinogen
not present in the animal's own blood cells.
The fundamental principle underlying all the Ehrlich hypotheses,
is the conception that these reactions take place as do purely chemical
reactions^ following the law of multiple proportions. Such reasoning"
has often necessitated the assumption of differences of affinity which,
critically examined, are really ex post facto explanations, forcing /
the phenomena to conform with the theory. As a matter of fact, ^/
the bodies which participate in the antibody-antigen reactions are
probably of the nature of the substances which are spoken of as col-^
Ipids, and it is therefore more than likely that the quantitative and
other relations governing the union of these reagents should be anal-
ogous to those governing colloidal reactions in general. The reaction
of agglutination, like that of precipitation, has lent itself particularly
to the study of the principles of the union from this point of view,
and the first and fundamental progress made in this direction is j
found in the, work of Bordet.
56 Von Dungern and Hirschfeld. Zeitschr. f. Imm., 4, 1909-1910, p. 53;
also ibid., 1910, p. 284.
"Klein. Wien. kl. Woch., 1902, p. 413.
58 Boycott and Douglas. Jour, of Path, and Pact., Jan., 1910.
59 Epstein and Ottenberg. Tr. N. Y. Path. Soc., 1908.
60 Von Dungern and Hirschfeld. Zeitschr. f. Imm., 1909, 1910, p. 531.
240 INFECTION AND RESISTANCE
Bordet 61 compared the formation of precipitates in bacterial
emulsions to the precipitation of such inorganic emulsions as clay in
distilled water, and noted that the precipitation of homogeneous
emulsions of such substances is "often controlled by such insignifi-
cant causes as the presence of salts." Applying this analogy to the
agglutination of bacteria, he performed the following experiment:
Cholera spirilla, emulsified in salt solution, were treated with homol-
ogous immune serum and, after agglutination had taken place, the
bacteria were thrown to the bottom by centrifugation and divided into
two parts. One part was again suspended in salt solution, and the
other was washed, and then suspended in distilled water. The bac-
teria in the tube of salt solution rapidly agglutinated, while those
in the distilled water, after thorough shaking, remained indefinitely
suspended in an even emulsion. If, however, to these unagglutinated
bacteria a small amount of pure sodium chlorid was added agglutina-
tion occurred.
The conclusions that can justly be drawn from this experiment
are, first, that the bacteria could not agglutinate, even though they
\/ had been bound to agglutinin, when salt was removed from the en-
vironment, and, second, that the addition of salt to such emulsions
brought about immediate agglutination. The same principle can be
demonstrated in other ways. If, for instance, a bacterial emulsion
is rendered free of salt by dialysis, and this is added to an aggluti-
nating serum similarly dialyzed, no agglutination occurs. The sus-
pension may remain evenly clouded indefinitely unless salt is added.
As soon as a little salt is added, however, perfect agglutination
occurs. To this technique the very obvious criticism may be applied
that perhaps the absence of salt has precipitated the agglutinins,
which, as we know, are precipitated with globulin, which is insoluble
in the absence of salt. However, this source of error is excluded by
the first experiment cited, and, moreover, it can be shown by the last
experiment that, even though the bacteria are not agglutinated in
the salt-free serum, they have nevertheless absorbed agglutinin. For,
if such a salt-free mixture is centrifugalized, the bacteria washed
and suspended in distilled water, and salt is then added, agglutination
occurs. The supernatant fluid of the original suspension, further-
more, can be shown to have been deprived of agglutinins by suitable
• experiment.
These facts, first observed by Bordet, and further elaborated by
the studies of Joos,62 Friedberger,63 and others, have been inter-
preted in different ways. Joos claims that there is a ^chemical union
between the bacteria and the salt, and bases this upon me" observation
that the salt added to a salt-free mixture cannot be demonstrated in
61 Bordet. Ann. de I'Inst. Pasteur, 1896, 1899.
62 Joos. Centralbl f. Bakt., 1, Vol. 33, 1903.
63 Friedberg-er. Berl kl Woch., 1902; Centralbl. f. Bakt., 1, Vol. 30.
THE PHENOMENON OF AGGLUTINATION
the supernatant fluid after agglutination has taken place. His ob-
servations in this respect have not found confirmation at the hands
of Friedberger and other workers, and it is generally agreed to-day
that the role of the salts is, as Eordet first' assumed it to be, a purely
physical ona Bordet's opinion is often spoken of as the "two phase" »
theory, in that he conceives the process of agglutination to consist of [
two distinct occurrences, first, an absorption of the agglutinin by •
the bacteria, and, second, an agglutination of the new complex by the «
salt. It is not the agglutinin which causes agglutination, but by
union with the agglutinogen forms a complex which is altered in
"colloidal stability/7 and therefore is flocculable by the electrolyte.
The opinion of Bordet becomes clearer as we consider the con-
ditions governing the flocculation of colloids in general. Without
wishing to enter in this place into detail regarding the nature of
colloidal suspensions, it nevertheless seems necessary in order to do
justice to this phase of the question to recall briefly the conditions
governing such flocculation. The so-called colloidal solutions are
not true solutions as the term is applied to dissociable substances, but
are looked upon as consisting of small particles in suspension. The
particles are similarly charged, as can be demonstrated by their
wandering when subjected to an electric current, and it is supposed
that it is this fact of similarity of charge which, in the "sol" state,
permits them to remain in suspension. For the similarity of the
charges of the individual particles prevents their mutual approxima-
tion.
The state of suspension of these substances, then, represents a
delicately balanced equilibrium between the two forces of electrical
repulsion and of surface tension, an equilibrium which may be
disturbed by the action of a number of factors. Thus, studies on
inorganic colloids have shown, long before these considerations were /
applied to the explanation of serum reactions, that the stability of \/
these suspensions could be disturbed both by electrolytes and by the
addition of other colloids. Thus they may be precipitated by vari-
ous salts, acids, and bases and, as Schultze 64 has shown, they react
with that ion of the electrolyte which carries an opposite charge to
that of the colloidal particles. For, although the colloidal units are
similarly charged, this may be either negative or positive, according
to the nature of the particular 'substance. In the case of the so-called
amphoteric colloids reaction may take place, according to Pauli,65
with both ions of the electrolyte.
The probable mechanism of the process is postulated by Pauli
in describing the precipitation of a colloidal metal by salts, acids, or
bases in the following way:
64 Schultze. Journ. f. prakt. Chem., 25, 1882, and 27, 1883.
5 Pauli. "Hofmeister's Beitra^e," 1905, and "Physical Chemistry in
icine," Wiley & Son, N. Y., 1907.
65
Medicine
INFECTION AND RESISTANCE
"The introduction of such electrolytes into a colloidal suspension
is of course accompanied by a certain amount of dissociation. In
consequence the weakly charged particles of the colloid collect ahout
the ions of opposite charge until a sufficient accumulation of the
particles leads to an electrical neutralization of the ion, and the ag-
gregation, if of sufficient size, will sink to the bottom, forming a
precipitate."
In regard to the mutual influences exerted upon each other when
two colloids are mixed, it has been shown by Biltz, Hardy, and
many other observers that oppositely charged 'colloids precipitate
each other, though this is not an absolute rule, as experiments by
Professor Stewart Young, of Stanford, have shown. Thus colloidal
gold and platinum will be precipitated by such colloids as ferric
oxid or aluminium oxid. When such a precipitating colloid is
added to another oppositely charged suspension in quantities too
small to bring about flocculation, moreover, the addition of a quan-
tity of salt, likewise too small to precipitate alone, will in many cases
bring about the flocculation.
These and other phenomena of colloidal reaction have found
close analogy in antibody-antigen studies, and have given support to
the interpretation of the latter in the sense of Bordet.
To return to the consideration of bacterial agglutination, we have
spoken of the dependence of the reaction upon the presence of salts,
and have seen that the researches of Friedberger and others have
/ refuted the assumption that the action of the salt in bringing about
agglutination depends upon chemical union of the salt with the
bacteria. It is probable, therefore, that here, as in other colloidal
precipitations, the function of the salt is to be regarded purely as
an electrophysical phenomenon.
The analogy becomes still closer when we consider the researches
of Bechold,66 Neisser and Friedemann,67 Sears and Jameson,68 and
others, which have shown that bacteria in suspension are to be com-
pared very closely with true colloidal suspensions in that the bac-
terial cells carry a definite and uniform electrical charge.
Bacteria in salt solution emulsion, for instance, wander to the
anode, thus giving evidence of their carrying a negative charge.
This charge may be altered by adding to the emulsions definite con-
centrations of acids or bases, a reversal of the charge taking place
under the influence of NaOH or other hydroxids. Just how this is
brought about is by no means clear, but it is not impossible that
there is a selective absorption of OH ions by the bacteria, which
therefore take on the charge of the ion.
66 Bechold. Zeitschr. f. physik. Chem., 48, 1904.
67 Neisser and Friedemann. Munch, med. Woch., Vol. 51, pp. 465, 827
1904.
68 Sears and Jameson. "Thesis for M. A. Stanford University," 1912.
THE PHENOMENON OF AGGLUTINATION
However this may be, and we must admit that explanations of
these phenomena are as yet largely speculative, a fact which interests
us particularly in connection with the phenomena under discussion
at present is the influence exerted upon the charge of bacteria by
exposure to the influence of serum. Bechold,69 as well as Neisser
and Friedemann,70 assert that bacteria which have absorbed ag-
glutinin no longer wander to the anode, but act as though they had
been deprived of electrical charge, and precipitate between the elec-
trodes.
Bechojo! has suggested, for this reason, that it may be possible
that bacteria in the normal condition are protected from the action
of the electrolyte by a membrane or coating of protoplasm which
acts as a protective colloid. The absorption of agglutinin may alter
this in such a way that they become amenable to the flocculating
effects of the salt ions. In keeping with such an opinion is the well-
known observation of the inagglutinability of capsulated organisms,
which, as Forges 71 has shown, become agglutinable as soon as the
capsules have been destroyed by heating in a weak acid.
That the absorption of agglutinin indeed alters the electric sta-
bility of the emulsified bacteria further appears from the fact that
"agglutinin" bacteria 72 are agglutinated by concentrations of salts
which are too slight to affect the normal micro-organisms. In this
respect there is close similarity between the flocculation of agglutinin-
bacteria and such emulsions as kaolin and mastic, whereas bacteria
without agglutinin require much higher concentrations of the salts
to produce like effects. The absorption of agglutinin may have re-
moved a factor which protected the bacteria against the influence of
the salt. On~fhe" other hand, it is equally just to assume — and this
is more likely and corresponds with Eordet's views — that the ab-\
sorption of agglutinin has "sensitized" the bacteria to the action of]
the electrolyte. The experimental facts upon which the above state-
ments are formulated are largely found in the important papers of
K"eisser and Friedemann — whose work brought out, likewise, inter-
esting analogies of the colloidal precipitations with the phenomenon
which we have described above as the proagglutinoid zone.
In regard to the greater amenability of agglutinin bacteria to
flocculation by electrolytes, the following protocol, adapted from the
work of these authors, will explain itself. They were tabulated from
experiments in which different quantities of normal T solution
of various salts were added, on the one hand to emulsions of unal-
69 Bechold. "Die Kolloide in der Biologie u. Medizin," Steinkopf, Dres-
den, 1912.
70 Neisser and Friedemann. Munch, med. Wocli., Vol. 51, 1904, pp. 465
and 827.
71 Forges. Ztschr. f. exp. Path. u. Therapie, 1905.
72 "Agglutinin" bacteria — bacteria which have absorbed specific agglutinin.
244
INFECTION AND RESISTANCE
tered bacteria, and, on the other, to bacteria which had absorbed
agglutinin. It is seen that, with some salts, agglutination of the
unaltered bacteria did not occur at all, whereas agglutination was
brought about in the treated bacteria with comparatively small
amounts ; in other cases the difference is a quantitative one only :
Protocol constructed from the tables of Neisser and Friedemann, loc. cit.
y solution of salt
Quantity of salt sol.
which brought about
agglutination of
1 c. c. of normal bacteria
in emulsion.
0 = no agglutination
by the salt
solution
Quantity of salt sol.
which agglutinated
1 c. c. of
agglutinin bacteria
in
emulsion
NaCl. .
o
025
NaNOg
o
.025
Na2S04
0
025
Rbl
o
.025
MgS04
0
.0025
ZnS04
01
001
CaCl2
0
.005
BaCl2 . . .
o
005
Cd(N03)2
.01
.001
CuS04
.0025
.0001
CuCl2
0025
.0005
Pb(N03)2
.0025
.0001
HgCL..
.0025
.0005
The analogy between the experiment tabulated in the preceding
protocol and the following from the work of the same writers is self-
evident. Just as the absorption of agglutinin by bacteria rendered
these more amenable to precipitation by salts, so the addition of
minute quantities of gelatin to mastic emulsions had a similar sensi-
tizing effect upon these.
1 c. c. mastic
1 o. c. mastic -f .0001 o.
0.
NaCl
10% solution
(1-10 original emulsion)
diluted to 3 c. c.
of a 2% gel. solution,
the whole diluted to 3 c.
c.
1.0
+ + +
+ + +
0.5
0
+ + +
0.25
0
+ + +
0.125
0
+ + +
0.05
0
0
0.025
0
0
Finally, one of the most important analogies yielded by the work
of the above investigators is illustrated in the following protocol :
THE PHENOMENON OF AGGLUTINATION
Colloidal iron
hydroxid
Precipitation
emulsion
of mastic
1 c. c.
1.
0
0.5
0
0.25
0
0.1
+-
h
0.05
+
0.025
++
+
0.01
++
+
0.005
++
+
0.0025
+-
h
0.001
0
Here we have an inhibition zone in the tubes containing the
highest concentrations, accurately analogous to the previously dis-
cussed proagglutinoid zone. It is a phenomenon similar also to the
inhibition zones noticed in precipitin reactions and observed, though
by a different technique, in bacteriolytic phenomena discussed in an-
other place in connection with the Neisser-Wechsberg notion of com-
plement-deviation or "Komplement Ablenkung." It seems to be a
universal fact governing the union of colloidal substances, that defi-
nite quantitative proportions must be maintained in order to lead
to reaction, this being, possibly, explicable on the basis that actual
union can take place only after disturbance of the electrical balance
which keeps the particles apart. These reactions will be found more
accurately discussed in another place. Whatever the mechanisms
may be, however, these and similar experiments have seemed to
render unnecessary and unlikely the assumption of proagglutinoids,
proprecipitoids, etc., to explain the inhibition zones so frequently
observed in all reactions of this kind.
A peculiar observation, which does not coincide with the above
interpretation of these phenomena, and the significance of which is
indeed doubtful, is one which Friedberger 73 made in researches in
which he confirmed the work of Bordet on the absence of agglutina-
tion in a salt-free environment. He found that not only the addition
of various salts would bring about agglutination under such condi-
tions, but that organic substances such as dextrose and asparagin
could be substituted for salts and had similar agglutinating effect —
although higher concentrations of these than of the salts were re-
quired. Were these substances at all dissociable it might be pos-
sible that they acted by a mechanism identical with that of the salts —
but since such substances as dextrose either do not dissociate at all or
do so to an infinitesimal degree only there does not seem any pos-
sibility of reconciling these results with Bordet's theory.
73 Friedberger. Centralbl. f. Bakt., 30, 1901.
246 INFECTION AND RESISTANCE
It is difficult to explain Friedberger's results. Possible impurity
of his preparations and the presence of traces of electrolyte seem
to be excluded by the fact that he was quite conscious of this possi-
bility of error and used only substances which yielded no ash on
combustion.
It may be that the results of Friedberger in which glucose and
asparagin were used may have brought about agglutination by an
entirely different mechanism from that which we are discussing -and
form no analogy to this.
In one of the preceding paragraphs we have mentioned the phe-
nomenon spoken of as "acid agglutination." By this is meant the
spontaneous clumping, not only of bacteria, but of small particles of
any kind, in suspension, in the presence of certain concentrations of
acid. Michaelis,74 Beniasch,75 and others who have studied this
phenomenon in detail have come to the conclusion that it is the con-
centration of the hydrogen ions which is responsible for the ag-
glutination. This explanation is also applicable to the agglutination
often observed about the anode when bacteria are subjected in sus-
pension to the action of a direct current. In such experiments the
organisms after concentrating at this electrode often flocculate, and
it is here, of course, that hydrogen ions are present in the greatest
concentration. How this takes place is problematical, but the reason-
ing of Pauli, if applied to this, would favor the assumption that the
weakly charged bacteria group themselves about the ions and, when
a sufficiently large aggregation has formed, fall to the bottom as
precipitate. This phenomenon of acid agglutination is of course
entirely different in nature from the specific serum agglutination
which we are discussing. Nevertheless, Schidorsky and Reim,76
Jaffe,77 and others have attempted to apply acid agglutination to the
isolation and differentiation of bacteria, on the conception that dif-
ferent species are agglutinated by varying concentrations of hy-
drogen ions. The former investigators, even, claim to have been
successful in isolating typhoid bacilli from the stools by this method
in that the typhoid bacillus was agglutinated by concentrations of
acid which had no effect upon the Bacillus coli. Sears 78 has gone
over this work carefully, and, while he has obtained results which
bear out the contention that the agglutination is probably due to the
'concentration of the H ions, his experiments have revealed an irregu-
larity in the behavior of bacteria of the same species in acid solutions
:and an overlapping of those of one species with those of another.
Therefore the use of acid agglutination for differential purposes
74 Michaelis. Folia Serol, 7, p. 1010, and Deutsche med. Woch., 37, 969.
75 Beniasch. Zeitschr. f. Imm., Vol. 12, 1912.
76 Schidorsky and Reim. Deutsche med. Woch., Vol. 38, p. 1125.
*7 Jaffe. Arch. f. Hyg., Vol. 76.
78 Sears. Proc. Soc. of Exp. Biol and Med., 1913.
THE PHENOMENON OF AGGLUTINATION 247
seems to us entirely hopeless. And indeed it would be surprising if
any such distinctive and regular reaction differences between simple
bacterial cells, after all chemically and physically so essentially alike,
could be found.
CHAPTER X
THE PHENOMENON OF PRECIPITATION
(Precipitins)
THE establishment of the agglutinin reaction as a constant and
specific serum-phenomenon by the work of Gruber and Durham led
immediately to assiduous investigation of the many problems sug-
gested by it, and among them, as we have seen, the question of the
nature of the agglutinogen. It was found that agglutinins could be
produced, not only by the injection of whole bacteria, but equally
as well by treatment with dissolved bacterial extracts or with filtrates
from old broth cultures. This naturally led to the thought that there
might be a definite reaction if such extracts (instead of the bacteria
themselves) were added to agglutinating sera in vitro. Rudolf
Kraus -1 was the first to perform this very logical experiment. He
was working with broth filtrates of Bacillus pestis and of the cholera
spirillum, and found that when he mixed the perfectly clear filtrates
of such cultures with their respective antisera the mixtures would at
first become turbid and finally show a light flocculent precipitate. He
named the reaction the "precipitin reaction" and, in analogy to
agglutinins, spoke of the bodies in the serum which caused the pre-
cipitation as "precipitins/" The reaction was found, like that of
agglutination, to be specific; the cholera serum gave no precipitate
with the plague extract and vice versa, and Kraus, after extending
his observations to other bacteria, pointed out the practical diagnostic
possibilities of his discovery.
Though Kraus7 first observations were made entirely with bac-
terial culture filtrates and antibacterial sera, it was soon discovered
that his results were merely isolated instances of a broad biological
law, and that specific precipitins were produced whenever animals
were treated with injections of any kind of foreign protein. Thus
Tschistovitch,2 in 1899, found that the blood serum of rabbits im-
munized with eel-serum gave specific precipitates when mixed with
eel-serum, and Bordet 3 obtained analogous results by treating rab-
bits with defibrinated chicken blood and with milk. Thus rapidly
1 R. Kraus. Wien. Idin. Woch., No. 32, 1897.
2 Tschistovitch. Ann. de I'Inst. Past., 13, 1899.
3 Bordet, Ann. de I'Inst. Past., Vol. 13, 1899, pp. 225-273.
THE PHENOMENON OF PRECIPITATION 249
the discovery of Kraus was developed into the generalization that
the sera of animals that have been treated with foreign proteins of
any kind — bacterial, animal, or vegetable — will develop the property
of causing precipitates when mixed with clear solutions of the re-
spective antigens.
The substances which, after injection into the animal body, lead
to the formation of precipitating antibodies are spoken of in the
language of immunology as "precipitinogen." In the case of bac-
teria it has been shown that, while the injection of the whole bac-
terial cell — dead or alive — will lead to precipitin formation, bacterial
extracts produced in a variety of ways will lead to the same result.
Such precipitinogen extracts can be obtained by allowing the bacteria
to grow in flasks of slightly alkaline bouillon, keeping them in the
incubator for from three weeks to three months, and then filtering
them through Berkefeldt candles. Again, useful extracts can be
more rapidly produced by growing large quantities of bacilli on
agar, emulsifying in salt solution, and shaking in any one of the
ordinary types of shaking machine for 48 hours or longer. On filter-
ing an extract is obtained which will form precipitates with homol-
ogous immune serum, or will incite precipitins when injected into
animals. In fact, any one of the customary vigorous methods of
extracting bacterial or other cells will yield precipitinogen. A rela-
tively purified precipitinogen in the form of a dry, water-soluble
powder has been obtained by Pick by the precipitation of culture
filtrates with alcohol.
Regarding the chemical nature of the precipitin-inducing sub-
stances, or precipitinogens, the same problems have arisen which
have been discussed in connection with antigens in general. We may
say that all soluble native proteins possess precipitin-inducing prop-
erties. Yet this does not sufficiently define the term, since many
observations have been published which show that physically and
chemically altered proteins may still induce specific precipitins; a
few investigators, furthermore, have claimed that they have produced
non-protein precipitinogen by various methods of breaking up the
molecule of the original antigen. In the section on agglutination
we have seen that moderate heating (56-65° C.) rather increases
than decreases the agglutinogen characteristics of bacteria, and it
is equally true that such heated bacteria or bacterial extracts may
induce precipitins. However, regarding the action of higher de-
grees of heat (boiling) upon precipitinogens in general we will have
more to say in another place.
Of more immediate, indeed of fundamental, importance is the
problem of a non-protein antigen. The most important claims in
this regard have been made by Pick,4 Obermeyer and Pick,5 and by
4 Pick. "Hof meister's Beitrage," Vol. 1, 1901.
5 Obermeyer and Pick. Wien. klin. Woch., 1904, p. 265.
250 INFECTION AND RESISTANCE
Jacoby. 6 7 Jacoby, working with a vegetable antigen, ricin, fount
that by trypsin digestion he could obtain a substance which still
retained antigenic properties, but no longer gave any of the pro-
tein reactions. Obermeyer and Pick, by the same method, claim,
that they have produced a non-protein precipitinogen from egg al-
bumen. On the other hand, others have had negative results, and
Kraus8 himself, after reviewing the evidence on both sides, comes
to the conclusion that available data do not justify us in separating
^ the antigenic properties from the protein molecule. In unpublished
^-experiments which the writer carried on in the laboratory of Profes-J
sor Friedemann in Berlin also attempts to produce a non-protein\
precipitinogen from horse serum by bacterial putrefaction were en-£
tirely negative. The putrefaction of the serum, though carried out <L
in dialyzing bags for the removal of diffusible products, was ex- .i
tremely slow, and when finally the Biuret reaction disappeared the J
serum was no longer precipitable by potent antisera. Hxpwever, the )
flaw in these experiments is that the true tesj_of the j)resence_pf
precipitinogen is not the precipitable character of the solution SPV
question, since actual precipitation is dependent, as we shall see,
upon many modifying secondary factors, but rather the ability o&
the substance to induce precipitins in treated animals.
The fact that ISTicolle,9 and later Pick,10 were unable to obtain
alcohol-soluble substances from bacteria and bacterial extracts which
were still precipitable might also be taken to point toward the non-
protein character of the precipitinogens, suggesting that these sub-
stances may be of a lipoidal nature. However, as Landsteiner 11
points out, mere solubility in organic solvents can no longer be taken
as a proof of lipoidal character, since it is more than probable that
non-lipoidal substances may go into alcoholic and other organic solu-
tion when lipoids, such as lecithin, are present. Thus Miiller 12
found that the antigen of typhoid bacilli was soluble in chloroform
in the presence of old preparations of lecithin. Pick and Schwartz,13
who had previously studied similar antigen solubilities in the pres-
ence both of lecithin and of other organ lipoids, suggest that possibly
such solutions represent lipoid-protein combinations — colloidal "so-
lutions"— which permit the presence of protein mechanically or
chemically united to the lipoid in the organic solvents-^— alcohol,
chloroform, etc. Here, too, then there is no evidence for the ex-
istence of non-protein precipitinogen.
6 Jacoby. "Hofmeister's Beitrage," Vol. 1, 1901.
7 Oppenheimer. "Hofmeister's Beitrage," Vol. 4, 1904, p. 259.
8 Kraus in "Kolle u. Wassermann Handbuch," Vol. 4, p. 605.
9 Nicolle. Ann. de I'Inst. Past., 12, 1898.
10 Pick. "Hofmeister's Beitrage," Vol. 1, 1901.
11 Landsteiner. "Weichhardt's Jahresbericht," Vol. 6, 1910, p. 214.
12 Miiller. Zeitschr. f. Imm., Vol. 5, 1910.
13 Pick and Schwartz. Biochem. Zeitschr., Vol. 15, 1909.
THE PHENOMENON OF PRECIPITATION 251
Of importance in connection with the problem of the nature of
precipitinogen, also, is the claim of Myers,14 that specific precipitins
may be produced in rabbits by treatment with Witte peptone, a sub-
stance complex in constitution, but consisting' largely of albumoses.
This observation has failed of confirmation in the hands of Ober-
meyer and Pick, Michaelis,15 ISTorris,16 and others, and cannot, there-
fore, be accepted as an established fact.
Whichever method of precipitinogen production is used bacterial
precipitins appear in the serum of the immunized animal only after
careful and continued immunization, usually later than the demon-
strable appearance of the bactericidal or agglutinating properties of
the serum. The most convenient material for such immunization
consists of salt solution emulsions of agar cultures, killed at 60° to
TO0 C. These may be injected subcutaneously, intraperitoneally, or
intravenously, the last method leading to the most satisfactory and
rapid results and, therefore, best employed unless great inherent
toxicity of the particular bacteria contraindicates. When rabbits
are used it is generally necessary to inject 3, 4, or 5 times at 5 or 6-
day intervals, and to bleed the animals on the 8th or 9th day after
the last injection.
The bacterial precipitins so produced are, as we have said above,
specific — but, again, specificity, as in the case of agglutinins, is
limited by the so-called "group reactions." In the chapter dealing
with agglutination we have seen that the serum of a typhoid-immune
animal which agglutinates typhoid bacilli strongly will also aggluti-
nate, though far less powerfully, paratyphoid bacilli and, in some
cases, even colon bacilli, this appearance of "minor" agglutinins
being probably due to a close group relationship of these bacteria to
the typhoid bacillus. In the case of bacterial precipitins the same
thing is true, and has been made the subject of special studies by
Zupnik,17 Kraus,18 Nor r is,19 and others. As in the case of ag-
glutination, however, this fact does not in any way interfere with
the practical value of the specificity of the reaction because elimina-
tion of the secondary group reactions, which in agglutination is
obtained by dilution of the antiserum, can liere be obtained, as Kraus
points out, by diminishing the quantity of the undiluted precipitat-
ing serum added to the bacterial filtrates. Thus, while one volume
of serum added to one, two, or three volumes of culture filtrate may
still give error due to non-specific group reactions, a proportion of
14 Myers. Centralbl f. Bakt., Vol. 28, 1900.
15 Michaelis. Deutsche med. Woch., 1902.
16 Norris. Jour, of Inf. Dis., Vol. 1, 1904.
17 Zupnik. Zeitschr. f. Hyg., 49, 1905.
18 Kraus. Wien. klin. Woch., 1901, No. 29.
19 Norris. Jour, of Inf. Dis., Vol. 1, 1904.
INFECTION AND RESISTANCE
one part of serum to 8 or 10 parts of the filtrate will usually elimi-
nate all secondary reactions and prove strictly specific.
An illustration of such an elimination of "partial" or "minor"
precipitins by diminution of the amount of the homologous anti-
serum is given in the following table taken from the work of
ANTICOLI RABBIT SERUM
TABLE III
The precipitating action of the anticoli rabbit serum upon its corresponding
filtrates and upon the nitrates of B. N° 1 (hog cholera) and B. typhosus.
Coli filtrate Anticoli serum
0 . 5 c. c. 0 . 05 Cloudiness in all tubes in 1 hour at 37.5° C. which
0.5 c. c. 0.10 increases rapidly. Six hours well-marked precipita-
0.5 c. c. 0. 15 tion — most copious in tube containing 0.25 serum.
0.5 c. c. 0.25 Fluid in all tubes becomes clear.
B. N°l filtrate Anticoli serum
O.Sc.c. 0.10 At 6 hours a slight precipitate in the form of fine
0 . 5 c. c. 0 . 25 granules appears on the sides of the tubes. After
24 hours the precipitate in the tube containing
0.25 c. c. serum compares in amount to that
formed in the homologous filtrate with 0.05 c. c.
of serum.
B. typh. (Coll)
filtrate Anticoli serum
O.Sc.c. 0 . 10 Similar reaction obtained to that with B. N° 1 filtrate.
0.5 c. c. 0.25
B. typh. (Pfeif-
fer) filtrate
6.5 c. c. 0.10 Similar delay in reaction as obtained with B. typh.
O.Sc.c. 0.25 Coll.
And, indeed, though the great practical value of the precipitin
reaction has not been in the special field of bacteriology, it has been
successfully utilized in the diagnosis of glanders by Wladimiroff,21
and constitutes a valuable adjuvant to our methods of determining
the biological relationship between micro-organisms.
The production of precipitins against unformed proteins, egg
albumen, animal sera, etc., is much more easily accomplished than
the production of bacterial precipitins, and three intravenous injec-
tions of from 2 to 5 c. c. of the protein at 5 or 6-day intervals usually
give rise to a formation of potent precipitins. When a small quan-
tity of the serum of such an animal, taken 9 or 10 days after the
third injection, is mixed in a test tube with an equal quantity of a
20 Norris. Joyr. of Inf. Dis., Vol. 1, 1904, p. 472.
21 Wladimiroff. "Kolle u. Wassermann Handbueh " article on "Glanders."
Vol. 5, 2d Ed.
THE PHENOMENON OF PRECIPITATION
253
dilution of the protein which was injected, turbidity and rapid floccu-
lation will result. In tests of this kind, unlike the bacterial precipitin
tests in which the delicacy of the reaction is ordinarily determined
by diminution of the amounts of antiserum, the same object may be
more conveniently attained by dilution of the antigen. Thus, in test-
ing the precipitating potency of, let us say, the serum of a rabbit
immunized with sheep serum, we would proceed by setting up a
series of small tubes, each of which contains a constant amount of
antiserum (precipitin), but a progressively diminishing amount of
antigen in the same volume — i. e., in dilution with iso tonic salt solu-
tion. The following example will make this clear :
Anti sheep serum
from rabbit
Sheep serum 0.5 c. c.
of following dilutions:
Precipitation
0.5 c. c.
+
1:10
=t
0.5 c. c.
+
1:100
+ + +
0.5 c. c.
-j-
1:500
+ + +
0.5 c. c.
0.5 c. c.
0.5 c. c.
+
+
+
1:1,000
1:5,000
1:10,000
—
In this test it will be noticed that the strongest concentration of
the antigen (1:10) gave a relatively slight precipitation only. This
phenomenon is analogous to the inhibition zones noticed in the case
of agglutination and other antibody reactions and will be more espe-
cially discussed in a succeeding paragraph.
The delicacy of the above example, moreover, is by no means
unusual, and sera have been obtained by careful immunization with
which the specific antigen could be detected in dilutions as high as 1
to 100,000 (Uhlenhuth). A serum which will react with antigen
dilutions of 1 to 10,000 and 1 to 20,000 is not at all uncommon nor
difficult to obtain. Apart from the additional advantage of the
specificity of the reaction, therefore, this biological method of de-
tecting proteins is more delicate than that of any of the known
chemical methods; neither the Biuret nor Millon's reaction will far
exceed a delicacy of 1 to 1,000. By a modification known as the
method of Complement or Alexin-fixation, furthermore, the delicacy
of the biological reactions can be still further enhanced. This is
discussed in detail in another place (see page 212).
The practical value of the precipitin reaction, however, lies, not
in the mere detection of protein, but, by virtue of its specificity,22 in
the determination of the variety of protein under examination. And
22 Wassermann and Schiitze. Deutsche med. Woch., 1900, Vereinsbeilage,
p. 178; Berl. kl. Woch., 1901; Deutsche med. Woch., 1902; Bordet, Ann. Past.,
Vol. 13, 1899; Nolf, ibid., Vol. 14, 1900; Fish, Medical Courier, St. Louis,
1900, cited from Wassermann.
254 INFECTION AND RESISTANCE
here again the specificity, like that of bacterial precipitation, ag-
glutination, and other serum tests, is relative rather than absolute.
Thus a serum which has been obtained by the immunization of an
animal with human serum may react, not only with human serum,
but also with relatively higher concentrations of the sera of some of
the higher apes. However, such non-specific partial reactions can be
eliminated entirely by employing higher dilutions of antigen. Thus
Uhlenhuth,23, 24? 25 on the basis of a large experience, has established
a -standard of antigen dilution at 1 to 1,000, beyond which no "para"
or "minor" precipitation will occur. Since potency far exceeding
this is easily procured, absolute specificity can be ensured by the
very simple precaution of a sufficient dilution.
The most important practical use for the reaction has been found
in forensic medicine, where it is possible in this way to determine
the species of animal from which have emanated the blood, sperm,
etc., found in spots on wearing apparel, weapons, or other articles.
The extensive investigations of Nuttall 26 upon this subject have inci-
dentally been of much value in furnishing a further method for the
determination of zoological species relationships. Nuttall carried
out 16,000 precipitin tests, with precipitating sera, upon 900 speci-
mens of blood which he obtained from various sources. He not only
confirmed many of the accepted zoological classifications, but shed
much light upon a number of disputed points. In working out the
tests upon monkeys he found that the reactions carried out with anti-
human serum become weaker as the species examined is farther re-
moved from man zoologically. Thus as we read down the column
from man to the hapalidsB the precipitate becomes less and less in
amount.
Nuttall's Tests with Antihuman Serum. (Nuttall, loc. cit., p. 165.}
ANTIHUMAN PRECIPITATING SERUM
Tested against Precipitate
34 Specimens human blood 100%27
8 Simiidse, 3 species 100%
36 Cercopithecidae 92%
13 Cebidse 78%
4 Hapalidse 50%
2 Lemuridse 0
23 Uhlenhuth. Deutsche med. Woch., 1900, 1901; Rob. Koch Festschrift,
1903.
24 Uhlenhuth and Weidanz. "Kraus u. Levaditi Handbuch," etc., Vol. 2,
1909.
25 Uhlenhuth and Weidanz. Loc. cit., where other publications are sum-
marized.
26 Nuttall. "Blood Immunity and Blood Relationship," Cambridge Uni-
versity Press, 1904.
27 The percentages refer to the volume of precipitate formed on standing
for a given time, the amount formed by the antiserum with its specific antigen
being taken as 100 per cent. Antigen dilutions correspond throughout.
THE PHENOMENON OF PRECIPITATION 255
In another series he finds:
ANTIHUMAN PRECIPITATING SERUM
Tested against Precipitate
Man 100%
Chimpanzee (loose precip.) 130%
Gorilla 64%
Ourang 42%
Cynocephalus mormon 42%
Cynocephalus sphinx 29%
Ateles 29%
Among the primates the highest figures with antihuman serum are
given by the chimpanzee. Other bloods than those of the primates
gave slight reactions or none whatever with the antihuman serum.
In addition to these results the relationships within the dog
family, the horse family, and many other kinships similar to these
were confirmed. In every case the precipitin reaction was con-
sistent with the results of other methods of classification, and Nut-
tali's work is an extremely valuable aid to zoologists in disputed
questions of animal relationships.
These facts are the more surprising in that they demonstrate
species differences between the proteins of various animals which
are not determinable by known chemical methods. How funda-
mental these differences are and how delicate the reaction, is further
shown by experiments of Uhlenhuth, in which he obtained a specific
antihare serum by treating rabbits' with hares' blood, an astonishing
result in view of the close zoological relations between these animals.
Isoprecipitins, that is, precipitins resulting from the treatment
of animals with blood of another individual of the same species, have
also been described by Schiitze and others. They are not, however,
regular in their appearance, nor are they very potent when obtained.
Since the reaction is equally applicable to vegetable proteins,
similar investigations on the interrelationship of different varieties ,
of wheat have been carried out by Magnus.28
The methods of performing precipitin tests for forensic or other
purposes is extremely simple. Nevertheless, there are a number of
theoretical considerations which we must take up in order to make
clear the limitations of accuracy and conditions of control which are
involved in these reactions. From our discussion of the nature of
precipitinogen it follows that blood stains, etc., on linen or articles
of any kind will be suitable for precipitin tests even after they have
been exposed for considerable periods to unfavorable conditions, that
is, an environment in which they are subjected to exposure to light,
moderate heat, or drying. Thus blood spots, etc., if kept dry and in
28 Magnus. Cited from Uhlenhuth, loc. cit.
£56 INFECTION AND RESISTANCE
the dark, may give positive reactions even after years, as experi-
ments by Uhlenhuth have shown. Meyer 29 claims even to have
obtained a precipitation with extracts of the material of mummies.
One of his specimens was a mummy dating back to the first Egyptian
Empire (5,000 years), the other about 2,000 years old. Pieces of
the leg and neck muscles of these specimens were chopped up finely,
extracted for 24 hours with salt solution, then filtered until clear.
With antihuman serum they gave turbidity after one hour at
37.5° C.
Under conditions of putrefaction, of course, the precipitinogen
is more rapidly destroyed, though blood putrefies with surprising
slowness, even if, as in our own experiments, the conditions of mois-
ture, temperature, and reinoculation with putrefactive bacteria are
constantly observed. Under such conditions a weak reaction may be
obtained after as long as a month or six weeks.
In carrying out the tests with any material it is first necessary
to get it into clear solution, a result which is best accomplished by
soaking it in a small quantity of isotonic salt solution. Preliminary
to this it is always necessary to scrape off a bit of the specimen and
examine it microscopically to discover, if possible, whether blood cells,
sperm, or other cellular constituents can be detected. The infusion in
salt solution should be continued for several hours — if necessary for
12 to 24 hours. After the first few hours in the incubator the material
should be placed at room or refrigerator temperature so that the
yield in unchanged protein may not be diminished by the action of
bacterial growth. After extraction the solution may be filtered in
order to clear it, but often mere centrifugation suffices for this pur-
pose. The concentration of antigen in such an extract is always an
uncertainty, but may be determined with sufficient accuracy for
practical purposes by shaking and observing the formation of a
lasting foam. Protein solutions will show foam on shaking in dilu-
tions as high as 1 to 1,000, and if the original amount of salt solution
used, in washing out the material is properly gauged to the amount
of blood available in the stain, and the solution shaken and observed
for the formation of foam, it is usually a simple matter to obtain a
final concentration approximating one to one thousand.30
The antiserum which is used should be of such a potency that
preliminary titration with the specific antigen, diluted 1 to 1,000,
should give an almost immediate cloudiness at room temperature.
By testing this serum against a number of other varieties of
29 Meyer. Munch, med. Woch., Vol. 51, No. 15, 1904.
30 If there is enough material, .the amount of dissolved protein can be
also approximately gauged by adding to a little of it a drop of acid, boiling
and observing the. heaviness of the cloud which forms. A control test of a
known dilution of the suspected variety of blood can be made at the same
time and the heaviness of this cloud compared with that in the test solution,
THE PHENOMENON OF PRECIPITATION 257
protein — dog serum, beef serum, etc. — it must be determined that
the precipitin in this case is strictly specific.
The reaction can be observed with greater delicacy if it is first
set up by the method recommended by Fornet and Miiller,31 which
we may speak of as the "ring test." The antiserum is put into the
tubes and the solution to be tested is allowed to flow slowly over this
—as in Heller's nitric acid albumin test. - At the line of contact be-
tween the two a fine white ring will rapidly appear, thickening and
growing heavier as the preparation is allowed to stand. After taking
the final readings from such a test, let us say after an hour or so, it
is well to shake up the tubes, set them away in the ice-chest, and again
read the amount of precipitates formed in the various tubes the next
morning. Since every test of this kind necessitates a number of
controls, the following example will serve as a basis for discussion :
Forensic Blood Examination
Material: Blood spot on trouser pocket, washed up in salt solution. Clear
after paper filtration.
Antiserum: Rabbit treated with three intravenous injections, 2, 5, and 5 c. c.
of human serum at six-day intervals; bled on tenth day after last injection.
This serum has been titrated against human serum and gives precipitation
in dilutions up to one to ten thousand. With one to one thousand there is
clouding which begins in three minutes and is very distinct in eight minutes,
at room temperature.32
Test Jtw****
Tube 1. Known human serum 1 to 1,000. . 1.0 c. c. + Antiserum 0.2 c. c. f u
Tube 2. Unknown solution to be tested. ... 1.0 c. c. + Antfserum 0.2 c. c.
Tube 3. Unknown solution to be tested 1.0 c. c. + Normal rabbit
serum 0.2 c. c.
Tube 4. Salt solution 1.0 c. c. + Antiserum .... 0.2 c. c. —
Tube 5. Unknown solution 1.0 c. c. -f- Salt solution. . .0.2 c. c. —
In this test, if the original material was human blood, tubes 1
and 2 should show ring formation within 5 minutes — while the
other tubes remain clear. In addition to these controls it is well to
be sure that the test extract is neither strongly acid nor alkaline, and
that, as Uhlenhuth suggests, the material from which it is extracted
does not contain other substances which can give precipitates by
themselves when added to serum. This is especially necessary in the
case of cloth fabrics, and a recent instance in our own experience
has suggested to us the possibility that such materials may also con-
tain colloidal dye stuffs or other extractable substances which can
cause inhibition of the precipitation. In an apparently positive
case the reactions with a blood extract from trouser cloth were suffi-
ciently heavy, but regularly delayed, as in the flocculation of such
31 Fornet and Miiller. Zeitschr. f. Hyg., Vol. 66, 1910.
32 A mixture of too specific antisera should never be used, since such
sera may often precipitate each other for reasons that are discussed below.
258 INFECTION AND RESISTANCE
colloidal suspensions as arsenic trisulphide in the presence of a
protective colloid.
In the ordinary criminal or civil case which would come under
consideration for precipitin tests the spots or stains are made by
blood as it flows from the wound and unchanged by chemical or
physical agencies except as these are encountered afterward, by
exposure. In the case of meat inspection, in which the precipitin
test is useful in detecting admixtures of horse flesh, dog flesh, or
other less desirable varieties of meat, in sausages, chopped meat,
etc., it often happens that such procedures as heating or smoking
may vitiate the results of precipitin reactions. It is of practical im-
portance, therefore, that we should know exactly what the effects of
heating (boiling) may be upon precipitinogen. Moreover, this ques-
tion possesses considerable theoretical interest since the coagulation
of proteins by heat seems to involve chiefly a physical rather than a
chemical change.
Cohnheim 33 says in discussing this question : "It is still unclear
what the changes are that take place in coagulation. It may be that
there is merely an intramolecular 'Umlagerung' — or there may be
cleavage ; or the process may be comparable to the flocculation of col-
loidal clay emulsions by salts. .. . . With coagulation all proteins
have lost the differences which they possess in the native state in
respect to solubility or precipitability by salts. Physically all coagu-
lated proteins are alike ; they are no longer native proteins, and with-
out further decomposition are insoluble. Chemical differences, how-
ever, variations of composition, and the cleavage products which
they yield still distinguish them."
The question has been experimentally approached by Obermeyer
and Pick 34 in connection with their general investigations upon the
influence of chemical and physical alterations upon precipitinogen.
They found that precipitin produced with unchanged (native) beef
serum does not react with heated beef serum, even if immunization
was prolonged and a very potent serum was produced. On the other
hand, when animals were immunized with beef serum which had
been boiled for a short time ("Kurz aufgekocht"35) the precipitin
so produced reacted, not only with native beef serum, but also pre-
cipitated the boiled serum and a whole row of split products which
give no reaction to normal precipitin. The "coctoprecipitin" so
produced, furthermore, was found by them to be specific, acting
only upon beef protein or its derivatives.
33 Otto Cohnheim. "Chemie der Eiweiss Korper Vieweg Braunschweig,"
1900, p. 8.
34 Obermeyer and Pick. Wien. kl Wocli., 12, 1906.
35 Sera or other proteins used in such tests are boiled in dilutions of 1
to 10 or more, in order to avoid the formation of heavy flakes which cannot
be injected. Boiled in sufficient dilution, an opalescent suspension is formed
which easily passes through the syringe.
THE PHENOMENON OF PRECIPITATION 259
It is immediately evident that these investigations are closely
analogous to those of Joos and others on the agghitinins. The anti-
serum produced with the heated antigen here again reacts both with
the native and with the heated antigen, whereas the antiserum pro-
duced with the native unheated antigen reacts only with the un-
heated. The "heat-precipitins" therefore may be also called "um-
fanglicher" — the term applied by Paltauf to the agglutinins pro-
duced with heated bacteria.
Schmidt,36 who has studied the problem extensively, finds that
heating serum protein to 70° C. for as long as 30 to 60 minutes
alters its precipitability by "native precipitin" (precipitin produced
by immunization with native unheated serum) only in so far as it
diminishes the delicacy of the reaction by 10 to 30 per cent., and'
that heating to 90° C. for as long as an hour does not render it en-
tirely non-precipitable, so that protein so treated may yet be detect-
able by ordinary specific precipitins produced by injections of un-
heated serum, though the delicacy of the reaction is lessened. Boil-
ing, according to Schmidt, renders the antigen no longer precipitable
by such "native precipitin," but, on the other hand, it does not seem
to destroy its antigenic property of inciting precipitins on injection
into animals. Fornet and Miiller, on the other hand, claim that even
boiled protein can be detected by "native precipitins," though the
reaction is only about one-tenth as delicate as it is with unheated
protein.
Schmidt studied these relations especially as they affect -the per-
formance of specific precipitin reactions in the identification of
boiled meat. He found that when he immunized rabbits with serum
protein that had been heated at TO0 C. for 30 minutes the antiserum
so obtained gave strong and practically useful reactions with its
specific antigen even if this had been boiled. Since "native pre-
cipitin" gives weak reactions only with such a boiled protein,
Schmidt recommends the use of the "70° precipitin" (produced by
injections of heated serum) for tests in which a heated antigen is to
be identified.
He states, however, that very prolonged heating may so com-
pletely coagulate the antigen that none of it can be gotten into "solu-
tion" (suspension), and in such cases results can be obtained neither
with the "native" nor with the "70° precipitin." He has at-
tempted, therefore, to find a method whereby even such entirely
insoluble proteins may be identified, and claims to have succeeded
by preparing what he calls his "heat-alkali-precipitin." 37 He di-
36 Schmidt. Biochem. Zeitschr., 14, 1908; also Zeitschr. f. Imm., Vol.
13, 1912.
37 "Native precipitin" = preeipitin produced by injections of normal un-
heated serum.
"70° precipitin" — precipitin produced by injections of serum heated
to 70° C. for 30 minutes.
260
INFECTION AND RESISTANCE
lutes serum with equal parts of isotonic salt solution and heats it to
70° C. for 30 minutes in a water bath. To 60 c. c. of such a solution
he now adds 10 c. c. of f NaOH, and continues heating for 15 to
20 minutes. At the end of this time he neutralizes with HC1, cools,
and injects 20 c. c. intraperitoneally into rabbits. (The neutraliza-
tion is not absolutely necessary.) Five or more injections. yield a
serum sufficiently potent for use.
A precipitin so produced will, according to Schmidt, react spe-
cifically with heated proteins, and also with protein which has been
solidly coagulated and brought into solution by means of NaOH and
heat. It will not, however, react with normal unheated antigen.
He tested this by coagulating horse serum by boiling for 3 hours.
The coagulum was washed with salt solution, dried, and powdered.
Tests were then made to prove that this powder was entirely in-
soluble in NaCl solution. A little of it was then treated with 10
c. c. of salt solution containing enough NaOH to correspond to an
^ solution. The exposure was continued for 20 minutes in a water
bath at 60° to 70° C. Before the entire mass was dissolved the solu-
tion was filtered and neutralized with -^ HC1.
The rather complicated relations described by Schmidt are easily
surveyed in the following protocol taken from his work :
TABLE I
(W. A. Schmidt, Zeitschr. /. 7mm., Vol. 13, 1912, p.
173)
Solution
of
Native
precipitin
Heat (70°)
precipitin
Heat-alkali-
precipitin
Native serum
Strong reaction
Good reaction
0 (very slight
70° serum (heated 30 min.)
100° serum (heated 30
min )
Good reaction
0
Strong reaction
Good reaction
turbidity)
Strong reaction
Strong reaction
70° serum treated with
NaOH (used to produce
heat-alkali-precipitin) .
Boiled insoluble serum,
brought into solution
with NaOH
0
0
0
0
Strong reaction
Good reaction
.Native serum treated with
« NaOH in the cold
0
o
Good reaction
Schmidt speaks of the "heat-alkali-precipitin" also as "alkali-
albuminate-precipitin." It can be produced only if the NaOH treat-
ment of the serum is cautiously performed. If the sodium hydroxid
is allowed to act too vigorously in strong concentrations or for too
long a time the antigen is completely destroyed, is no longer pre-
THE PHENOMENON OF PRECIPITATION 261
cipitable, and no longer produces precipitin when injected into
animals.
The striking feature of these experiments is that they show a
gradual alteration of the protein first by heat, then by alkali and
heat, in such a way that the antigenic properties are changed but
not destroyed. Each precipitin, moreover, seems to react most
strongly with the particular antigen-alteration which produced it,
and, according to Schmidt, retains its species specificity. This is
not the case with the iodized proteins and nitroproteins and diazo-
proteins produced by Obermeyer and Pick.38 Here iodized beef
protein injected into animals produced a precipitin which reacted
with the iodized protein, not only of the beef, but also similarly
altered proteins of other animals — and the same was true of the
nitro and diazo modifications.
Although the experiments of Schmidt have great theoretical
value, their practical utilization must depend upon the degree of
specificity possessed by the heat-precipitins or the heat-alkali-pre-
cipitins. In Obermeyer and Pick's original investigations we have
seen that they found the precipitin produced with heated serum as
strictly specific as that induced by native serum. This has also been
the experience of Schmidt. Fornet and Miiller,39 on the other hand,
report that the precipitins produced by them with heated muscle-
protein were not as strictly specific as those produced with the un-
heated — in that the former gave precipitates, not only with homol-
ogous protein solutions, but with foreign proteins in moderate con-
centration as well. In experiments carried out by the writer with
Ostenberg 40 it was attempted to determine whether or not precipi-
tins could be produced by injecting animals with protein that had
been boiled, and if so what the action of these substances would be
upon boiled proteins. Contrary to the results of Fornet and Miiller,
it was actually found that sera boiled for 3 to 5 minutes injected into
rabbits induced precipitins which acted upon boiled proteins, but at
the same time it was determined that the antibodies so produced
were no longer strictly specific. The protocol given at the top of the
next page will illustiate these experiments.
Summarizing these results together with those of Fornet and
Miiller and of Schmidt it would seem that the injection of boiled
proteins induces precipitins which .no longer act on native antigen,
which act powerfully on boiled antigen, but are no longer strictly
specific. This seems to us of great theoretical interest as showing
an alteration by heating in the species adherence of the antigen.
Practically, therefore, precipitins produced with boiled protein are
of little value, and forensic determinations of boiled proteins should
38 Obermeyer and Pick. Wien. klin. Woch., No. 12, 1906.
39 Fornet and Miiller. Zeitschr. f. Hyg., Vol. 66, 1910.
40 Zinsser and Ostenberg. Proc. N. Y. Pathol. Soc., 1914.
INFECTION AND RESISTANCE
Experiments on Cocto-precipitin. Table II (March 23, 1913).
Cross titrations — dilutions of sera in salt solution boiled 5 minutes,
precipitated with antisera produced by injections with similarly boiled material.
The readings here indicated were taken by "ring" test at the end of 30
minutes.
Beef
Beef
Beef
Dog
Dog
Dog
Sheep
Sheep
Sheep
serum
serum
serum
serum
serum
serum
serum
serum
serum
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
vs.
Dilution
anti-
anti-
anti-
anti-
anti-
anti-
anti-
anti-
anti-
beef
dog
sheep
dog
beef
sheep
sheep
dog
beef
precipi-
precipi-
precipi-
precipi-
precipi-
precipi-
precipi-
precipi-
precipi-
tin
tin
tin
tin
tin
tin
tin
tin
tin
:20
+
+
+
+ +
+
+ +
+ +
+
:50
+ + +
+
+ + +
+ +
+ +
+ + +
+
+ + +
:100
.cnn
+ + +
_| 1
+
+ +
+
+
+ +
+
+ +
.OUU
:1,000
T T"
±
—
•§•
•§•
±
±
±
Controls of
boiled serum
alone*
1:20
—
1:50
_ —
1
.
Serum
control
* These controls were necessitated by the fact that the boiled serum suspensions were them-
selves turbid and occasionally showed slight settling on standing.
be done, as advised by Schmidt, by the "70° precipitins," or with
native precipitin as practiced by Fornet and Miiller.
The specificity which is the basis of the practical value of the reac-
tions that we have discussed is spoken of as "species" specificity
since it has been found that the blood serum of rabbits or other ani-
mals into which the serum of another animal has been injected
reacts, not only with the homologous blood serum, but also with
extracts of the various organs of the particular species of animal
which furnished the serum. Thus if we immunize rabbit, let us say,
with sheep serum the resulting precipitin will react, not only with
sheep serum, but also with extracts of sheep spleen, sheep liver, etc.
It seems that every species of animal possesses throughout its tis-
sues a particular variety of protein, fundamental to its general
metabolism and peculiar to its species. On the other hand, we have
seen in the preceding discussions how chemically slight the changes
in a protein may be which can alter materially its antigenic nature,
and it is a logical deduction that different organs of the same animal
might contain antigenic constituents qualitatively different from the
general serum protein. There are undoubtedly in many organs
protein complexes which are peculiar to them and not present in
other organs, and it would be reasonable to expect therefore that
immunization with separate organ substances would lead to the pro-
duction of sera of specific precipitating power for the protein of that
particular kind of organ. This is not ordinarily obtainable, how-
THE PHENOMENON OF PRECIPITATION 263
ever, because it has been impossible to isolate from organs their pe-
culiar, characteristic proteins, and immunization of animals with
organ extracts or solutions has necessarily implied the injection of
much blood protein and other albuminous material of a character
general to many organs of the animal, i. e., to the species. These
quantitatively overshadow the organ-specific substances which may be
present, and give rise, therefore, to a "species" precipitin. That
"organ specificity/7 however, is a fact has been shown by the experi-
ments of Uhlenhuth with the protein of the crystalline lens of the eye.
Immunization with this substance induces a precipitin which does
not react with the serum of the animal from which the lens was taken,
but does react, not only with the crystalline lens proteins of this spe-
cies of animal, but also with crystalline lens proteins in general,
though taken from another animal species. Analogous to this are the
experiments of von Dungern and others upon the protein derived
from the testicle.
In both of these cases, as well as in other less sharply defined
examples, the specificity is attached, not to the species of animal,
but rather to the nature of the organ from which the particular
protein is derived. These facts — first ascertained by means of the
precipitin reaction — have been recently confirmed by means of the
reaction of anaphylaxis by Uhlenhuth and Haendel, and by Kraus,
Doerr, and Sohma. (See chapter on Anaphylaxis.) They have
been discussed, moreover, in connection with the problem of spe-
cificity in general.
Biologically they probably signify that, although there are fun-
damental species differences between the general body proteins of
various animals, there are still, in certain highly specialized organs,
varieties of protein which, possibly because of functional exigencies,
have developed similar chemical characteristics. These have been
determinable by our present methods, however, only for organs like
the lens, the testicle, and the placenta from which the organ-specific
protein can be gotten in a relatively pure state. The pathological
importance of these phenomena lies in the fact that, although guinea
pig serum injected into a guinea pig will not give rise to antibodies,
lens protein apparently will do so — -an observation which opens the
possibility of autocytotoxins. The significance of this is indicated in
such investigations as those of Homer,41 who, using the complement-
fixation technique to determine antibody, found that the serum of
adult human beings possessed antibodies for their own lens protein,
but that such antibodies were absent in the sera of children.
The study of agglutination and that of precipitation reveal,
throughout, a close similarity between the two reactions, and indeed
in physical principles they are probably the same, although the one
41 Rb'mer. Klin. Monatsbl. f. Augenlieilkunde, Sept., 1906. Ref. from
"Weichhardt's Jahresber.," Vol. 2, 1906, p. 348.
264 INFECTION AND RESISTANCE
(agglutination) consists in the flocculation of large particles in sus-
pension— the bacteria — while in the other the precipitation is one
of smaller units — the precipitable colloidal particles of the protein
solutions. This phase of the subject will be more thoroughly dis-
cussed directly.
Meanwhile, it is noticeable also that, even without drawing the
physical parallel between the two reactions, there is much in the
behavior of the antibodies — the agglutinins and the precipitins as
conceived by Ehrlich, which led him and his school to attribute to
them a similar receptor structure. Like the agglutinins, the pre-
cipitins are not inactivated by 56° C., but when once rendered in-
effectual by higher temperatures (70° C. or over) they can no longer
be reactivated by the addition of fresh normal serum. For this
reason chiefly Ehrlich has conceived that both agglutinins and pre-
cipitins are "haptines" of the second order.
ZYMOPHOfSE
CELL OR ('BOW CELL
OTHER ANTIGEN
SCHEMATIC KEPRESENTATION OF EHRLICH 's VIEWS ON THE STRUCTURE OP PRE-
CIPITINS.
Ehrlich assumes that when dissolved protein substances — ordi-
narily suitable for body nutrition — are injected into animals, they
become anchored to the cells by such receptors of the second order.
When overproduction occurs in response to repeated stimulation of
the cells by consecutive injections (see Side-Chain Theory), these
haptines of the second order circulate as agglutinins or pre-
cipitins. Since they act without the apparent cooperation of alexin,
he supposes that they carry within themselves the "zymophore," or
ferment groups, by means of which the agglutination or coagulation
is accomplished. It is this zymophore group which, it is assumed,
accomplishes the digestion of the foreign protein before its assimila-
tion, when these receptors are still parts of the living cell.
Thus the conception of precipitins is identical with that formu-
lated by the same school concerning the agglutinins, and the deduc-
tions from these premises have been essentially similar. Thus, anal-
ogous to the conditions prevailing in agglutination, Pick,42 and
Kraus and v. Pirquet 43 have shown that when precipitating serum
is inactivated by heat, and then is added to bacterial filtrates, it will
42 Pick. "Hofmeister's Beitrage," Vol. 1, 1902.
43Kraus and von Pirquet. CentralU. /. Bakt., Vol. 32, 1902.
THE PHENOMENON OF PRECIPITATION 265
prevent their subsequent precipitation by active precipitin. An
illustration of this is found in the following protocol taken from the
paper by Kraus and v. Pirquet (loc. cit., p. 69).
(a) 5 c. c. cholera filtrate + 0.5 c. c. inactiv. (60°) cholera serum = no precipitate
after 10 hours at 37° C.
After 10 hours add 0.5 c. c. active cholera serum = no precipitate.
(b) Omitted.
(c) Omitted.
(d) 5 c. c. cholera filtrate + &-5 c. c. active cholera serum = after 10 hrs.
typical precipitate.
From this it was concluded that heat may destroy the zymophore
or coagulating group of precipitins, leading to the formation of
"precipitinoids" which, like agglutinoids, may have a higher affinity
for the antigen than is possessed by the uninjured antibody.
Subsequently there were opposed to these views the physical in-
terpretations which have been outlined sufficiently in the section on
Agglutination (see p. 240). In the case of precipitation the anal-
ogy between colloidal reactions and the serum phenomena is fully as
striking as in the former, an analogy in the delineation of which the
first credit belongs to Landsteiner,44 45 and important further contri-
butions have been made by Neisser and Friedemann, Porges, Gen-
gou, and a number of others. As in agglutination and colloidal floc-
culation, the presence of salts (electrolytes) fundamentally influ-
ences the occurrence of precipitin reactions; and in both colloidal
and precipitin reactions the relative concentration of the reacting
bodies is paramount in determining whether or not precipitation
takes place. In this connection the most frequently observed inhibi-
tion occurring in serum precipitations is that which is caused by an
excess of antigen. An example of this is as follows :
Sheep serum 0.5 c. c.
Antisheep rabbit serum
Precipitate
1:10 + 0.5 c.c. —
1:50 + 0.5 c.c.
1:100 + 0.5 c.c. ++•
1:500 + 0.5 c.c. + + +
1:1,000 + 0.5 c.c. ++
1:5,000 + 0.5 c.c. -f
This is entirely analogous to the inhibition which may occur
when, let us say, a weak gelatin solution is added to a colloidal sus-
pension of arsenic trisulphid ; or blood serum is added to mastic or
arsenic suspensions. In both cases inhibition zones appear which
44 Landsteiner and Jagic. Munch, med. Woch., No. 18, 1903; No. 27,
1904; Wien. kl. Woch., No. 3, 1904.
45 Landsteiner and Stankovic. Centralbl f. Bakt., Vols. 41 and 42, 1906.
266 INFECTION AND RESISTANCE
show that the relative quantities of the two reacting bodies are quite
as significant as their chemical or physical constitution in determin-
ing the occurrence of flocculation. This, according to Bechold, Bil-
litzer,46 and others depends upon the fact that the reason for floccu-
lation is one of electrical charge. One hydrosol — say arsenic tri-
sulphid — can be flocculated by the oppositely charged colloidal alu-
minium hydroxid, but this will occur only when the quantitative
relations are properly adjusted. If one or the other is in excess, no
flocculation may occur, and, if subjected to a direct current, both
colloids, though ordinarily wandering in opposite directions, will
now wander in that of the one which is now present in the largest
amount. We will not elaborate here upon the causes for this, since
they have been indicated in the section on Agglutinins, and are set
forth more accurately by Prof. Young in the special chapter on Col-
loids.
This effect of quantitative proportions would explain not only
the absence of precipitation in the presence of too much antigen, but
also the converse phenomenon, already mentioned, that precipitation
may be inhibited when the precipitin is in excess.
The fact that heated precipitating serum when added to its an-
tigen not only does not cause flocculation, but may even prevent sub-
sequent precipitation by active precipitin, also finds its analogy in
colloidal reactions in the so-called protective colloids. Thus arsenic
trisulphid may be protected from precipitation by gelatin, if a small
amount of gum arabic is added, and the analogy has been brought
even closer by Porges,47 who showed that heated serum will protect
mastic suspension from precipitation by normal serum. This obser-
vation of Porges is so closely similar to the results obtained by Kraus
and v. Pirquet and others on the inhibition of precipitation by heated
precipitating serum that it would seem, on first consideration, effec-
tually to refute the conception of "precipitoids."
However, it does not explain the specificity of such inhibition on
the part of heated precipitating serum, as reported by Kraus and v.
Pirquet, an observation which is one of the strongest arguments in
favor of the derivation of the inhibiting factor from the specific
precipitin (a precipitoid).48
In spite of the strong evidence in favor of the colloidal inter-
pretations, such contrary evidence, brought forward by careful and
46 Billitzer. Cited from Bechold, "Die Kolloide, etc.," p. 79.
47 Porges. Chapter on "Colloids and Lipoids" in "Kraus u. Levaditi
Handbuch," Vol. 1.
48 Although normal sera may gradually precipitate on standing, this
takes place much more rapidly in precipitin-sera, The spontaneous precipi-
tation of normal sera as well as of those under consideration is analogous
to what Bechold and others call the "ageing" (altern) of colloidal suspen-
sions, which, though originally stable, will eventually settle out, even in the
presence of protective colloids.
THE PHENOMENON OF PRECIPITATION 267
experienced workers, must be borne in mind and positive acceptance
of the colloidal explanations, however attractive, must be withheld
until much further investigation has been done.
Another important and interesting phase of the study of precipi-
tins is that associated with the occasional presence in the same serum
of remnants of antigen and of precipitins which, though present
side by side, do not unite to form precipitates. This condition
is frequently seen in such sera as those produced by Fornet and
Mliller 49 for rapid precipitin production for forensic work, a
method in which the foreign serum is injected into rabbits in large
amounts (2 to 10 c. c.), on consecutive days, and the animals are
bled 6 to 8 days after the last injection. That such sera contain
both antigen and antibody is shown by the fact that, though clear
when taken, they will show precipitation not only when mixed with
dilutions of the antigen, but also when added to homologous precipi-
tating sera.50
This phenomenon has been noticed by Linossier and Lemoine,51
Eisenberg,52 Ascoli,53 and others, and has been extensively studied
by von Dungern.54 Gay and Rusk 55 have recently observed it in
connection with the rapid method of precipitin production of Fornet
and Miiller, and have noted that such sera, although containing both
antigen and precipitin, do not possess complement-fixing properties.
According to Uhlenhuth and Weidanz,56 the antigen may persist in
the sera of protein-immunized animals, in demonstrable amounts,
as long as fifteen days after the last injection, and it is constantly
present during this period, but in progressively diminishing amounts.
We are thus confronted by the apparently paradoxical phenom-
enon of the presence in these sera, side by side, of an antigen and its
homologous precipitin, incapable of reacting with each other, al-
though each of them readily reacts with precipitin or antigen, re-
spectively, when these are added from another source.
Many attempts have been made to account for this. A number
of observers, notably Eisenberg, have concluded from extensive an-
49 Fornet and Miiller. Zeitschr. f. biol. Technik u. Methodik, Vol. 1,
1908.
so j>or instance, a rabbit was injected on three consecutive days with
sheep serum. It was bled on the fifth day after the last injection. The
serum was clear when taken, but a precipitate was formed when it was
added to sheep serum and also when it was added to serum from another
rabbit similarly treated and containing sheep serum precipitin.
51 Linossier and Lemoine. C. R. de la Soc. de Biol., 54, 1902.
52 Eisenberg. Centralbl. f. Bakt., 34, 1903.
53 Ascoli. Munch, med. Woch., Vol. 49, No. 34, 1902.
54 Von Dungern. Centralbl f. Bakt., 34, 1903.
55 Gay and Rusk. "Univ. of Cal. Public, in Pathology," Vol. 2, 1912.
56 Uhlenhuth and Weidanz. "Praktische Anleitung zur Ausfiihrung,
•etc.," Jena, 1909.
268 INFECTION AND RESISTANCE
alyses of quantitative relationships, both of agglutinin and precipitin
reactions, that these take place according to the laws of mass action.
In consequence, in addition to the combined precipitin-antigen com-
plex present in all mixtures of the two, there should also be present
free dissociated fractions of each, in amounts dependent upon rela-
tive concentrations. This might explain conditions such as those
described above.
Von Dungern, whose paper forms one of the most extensive studies
of the phenomenon with which we are concerned, does not believe
that precipitin reactions can follow the laws of mass action, and
explains the simultaneous presence of precipitin and antigen in the
same serum by assuming a multiplicity of precipitins. He believes
that every proteid antigen contains a number of related partial an-
tigens which give rise in the immunized animal each to a partial
precipitin. In sera in which both antigen and precipitin are found
side by side and free, he believes that the antigen is of a nature that
has no affinity for the particular partial precipitin present with it.
He says : "Auch hier handelt es sich nicht um zwei reaktionsf ahige
Korper, deren Verbindung aus irgend welchen Griinden unterbleibt,
sondern um Substanzen, welche keine Affinitat zu einander besitzen.
Die betreffenden Kaninchen haben zu dieser Zeit noch nicht alle
moglichen Teilprazipitine gebildet, sondern nur einzelne derselben.
Diese zunachst produzierten, nur auf bestimmte Gruppen der prazi-
pitablen Eiweisskorper passenden Partialprazipitine sind es, welche
nach der Absattigung aller zur Verfiigung stehenden zugehorigen
Gruppen der prazipitablen Substanz in Serum nachweisbar werden.
Daneben bleibt aber ein anderer Teil der prazipitablen Substanz,
der keine Affinitat zu dem gebildeten Prazipitin bestizt, bestehen,
solange bis ein anderes Partialprazipitin von den Kaninchenzellen
geliefert wird, welches sich mit Gruppen der in Losung geliebenen
Eiweisskorper vereinigen kann."
Zinsser and Young57 have also studied these phenomena and
have attempted to explain them on the basis of protective colloidal
action. In considering the theories that have been advanced to ex-
plain these occurrences, the conception of mass action as accounting
for the simultaneous presence of the two reacting bodies in the same
serum seemed entirely incompatible with our own observations and
with those of Gay and Rusk, that these sera do not of themselves
fix alexin. Were the conception of the manner of union of these
two reagents, according to the laws of mass action, representative
of the true state of affairs, it would be necessary to assume the pres-
ence, in such sera, not only of the two reacting bodies free and disso-
ciated, but also of a definite quantity of the united complex of the
two, a state of equilibrium being established. If this were the case
the sera should, in agreement with all experience on the phenomenon
57 Zinsser and Young. Jour, of Exp. Med., 1913, Vol. 17.
THE PHENOMENON OF PRECIPITATION 269
of complement fixation, exert definite complement-binding power.
Moreover, it has not been experimentally shown that colloidal sub-
stances react in accordance with the laws of mass action as observed
for simpler chemical substances.
As regards the opinion of von Dungern, this seemed incom-
patible with another occurrence, observed by many writers, namely,
that such sera, although clear at first, eventually, after prolonged
standing, do actually precipitate spontaneously; that is, the union
of the precipitin and the precipitinogen does actually take place, but
goes on with extreme slowness.
Now a notable and strange feature of this phenomenon is the
fact that two such sera, both containing antigen and precipitin, but
neither of them precipitating by itself, will precipitate each other
when mixed. For this reason Uhlenhuth has advised against the
use of mixtures of precipitin sera for forensic tests. For it is not
unusual that precipitin sera, even when produced by the slow method,
may contain traces of antigen, and this may lead to precipitate
formation if such a serum is mixed with another homologous pre-
cipitin and thereby simulate a positive forensic test.
In seeking analogy for this serum phenomenon with the various
colloidal suspensions, the problem consisted in protecting two
mutually precipitating colloids by a third, and this in such propor-
tions that the mixing of two such protected suspensions, each con-
taining all three of the elements, would be followed by precipitation.
This was obtained by the use of gum arabic, gelatin, and arsenic tri-
sulphid. Thin emulsions of gelatin will precipitate arsenic tri-
sulphid suspensions. Small amounts of gum arabic will act as a
protective agent, preventing the precipitations.
The amount of the protecting substance necessary to prevent
precipitation in any one mixture varies apparently with every
change in the relative proportions of the two. Thus a considerable
number of mixtures of the three can be made which will remain
stable for days, the actual and relative quantities of the three
ingredients differing in each of the mixtures. When two such mix*
tures are poured together, in many cases precipitation will result,
varying in speed and completeness, according to the particular quan-
titative relationship arrived at in the mixture.
An example of such an experiment follows :
Two solutions of colloidal arsenic sulphid were prepared, one containing
1 gm. per liter, the other containing 5 gm. per liter. With Kahlbaum's "Gold-
ruck" gelatin a solution containing 1 gm. per liter was prepared. A solution of
gum arabic was prepared which contained 10 gm. per liter, this being made
stronger than the gelatin solution to avoid too great dilution in the final mixtures.
The gelatin solution was prepared twenty-four hours before being used, as
freshly prepared gelatin has but slight precipitating power for arsenic sulphid,
this power appearing to increase greatly with the ageing of the solution.
270 INFECTION AND RESISTANCE
For the purpose of demonstrating this analogy two protected
solutions were prepared as follows :
Solution 1. — This consisted of 2 drops of gum arabic, 2 c. c. of gelatin, and
5 c. c. of the weaker arsenic solution.
Solution 2. — This consisted of 10 drops of gum arabic, 1 c. c. of gelatiD,
and about 4 c. c. of the stronger arsenic solution.
In each case the arsenic sulphid was added until there were signs
of increasing opalescence or turbidity, this being done in order that
the two solutions should each be as little overprotected as possible.
Portions of the two solutions were then mixed in equal propor-
tions. In the course of a few minutes the mixture was noticeably
more turbid than either of the original solutions. This turbidity
continued to increase quite rapidly, and on the following morning
after about sixteen hours of standing, the mixture was found to be
completely flocculated out, while the original protected mixtures re-
mained unprecipitated and showed about the same degree of opales-
cence as on the preceding night. The same condition of affairs was
found to have persisted after five days. On the fifth day the less
concentrated of the clear protected suspension began to settle out,
and was completely precipitated within twenty-four hours. The
other remained clear for four days more, but on the ninth day it
began to precipitate slightly, the precipitation remaining incom-
plete.
In these cases it appears, therefore, that a complete analogy to
the observed conditions of the serum reactions has been found, and
that all data observed in connection with sera in which antigen and
precipitin are found side by side without reacting can be most simply
explained on the conception of protective colloid action. Moreover,
the chemical nature of the substances involved seems to add weight
to this point of view.
These relations have been gone into here at some length, since
they seem to us to possess considerable theoretical and practical sig-
nificance. For it may be that the presence of a protective colloid
may, by inhibiting the union of antigen and precipitin within the
body, protect the animal from intoxication during the early stages
of immunization when antigen and antibody are present simulta-
neously for longer or shorter periods. Were union between the two
possible at such times in the circulation, an assumption necessitated
both by the hypotheses of mass action and of multiplicity of precip-
itins, there would probably be an absorption of complement by these
complexes, with, as shown by Friedberger, a consequent formation of
powerful toxic products. (See chapter on Anaphylaxis. ) It is
not impossible by any means, therefore, that the injection of anti-
gen in an animal in which such a balance has been established may
THE PHENOMENON OF PRECIPITATION 271
lead to a sudden elimination of the colloidal protective action, union
of the antigen and antibody, and, by the mechanism just outlined,
a naphy lactic shock.
The fact, moreover, that mere heating will change the precipi-
tating action, which certain sera have on inorganic colloids, to a
protective one seems to show that this latter function may justly
be associated with delicate physical or chemical alterations of animal
sera.
Furthermore, this point of view is strengthened by the fact that
the mutual precipitation of sera, such as those described, takes place
slowly, as does the mutual precipitation of two protected colloidal
mixtures, in contradistinction to the more rapid precipitation which
takes place when any of these sera is added to an antigen dilution,
where the element of protection may be assumed to be practically
eliminated by more extensively changed quantitative relations.
This point of view has lately been disputed by Weil, who has gone
back to the older view of von Dungern, largely on the basis of pre-
cipitation experiments he carried out with crystallized egg albumen.
Weil claims that if a pure protein, like crystallized egg albumen, is
used for immunization, antigen and antibody are never found simul-
taneously in the blood stream. If this were true it would indeed con-
stitute a very important contradiction of our point of view. In con-
sequence Bayne-Jones carried out similar experiments in our labora-
tory, and found that even with the purest obtainable recrystallized
egg albumen both antigen and antibody are at times demonstrable
in the blood. He showed this both by precipitation and by comple-
ment fixation. We do not consider the question closed, however, be-
cause it is indeed true that when working with a purified protein it
is more difficult to demonstrate the two substances in any quantity,
than it is when the crude serum antigen- is used. This may be, of
course, due to the fact that pure egg albumen may be more rapidly
assimilated and remains in the circulation for a less extensive period
than does the crude antigen. However, further experiments in this
direction will unquestionably clear the matter up because it is a
simple question of fact amenable to experiment.
As far as the separation of a pure albumen from an albumen
globulin mixture is concerned, we feel rather doubtful whether such
a sharp separation can be made and are inclined to lean towards the
view expressed by Wells some years ago, that the various proteins
(that is, albumens and globulins) of the same animal material are not
sharply separable but shade one into the next as in a spectrum. This
view is the more likely when we consider that our methods of separa-
ting these various proteins are purely physical methods of heating
and of precipitation by salts.
CHAPTER XI
PHAGOCYTOSIS
EARLY investigations into the fate of bacteria within the infected
animal body were largely carried out by pathological anatomists, and
the observation of the presence of micro-organisms within the cells
of the animal and human tissues was definitely made as early as
1870. Hay em,1 Klebs,2 Waldeyer,3 and others, saw leukocytes con-
taining bacteria but failed to interpret this in the sense of possible
protection. The process was regarded rather as a means of trans-
portation of the bacteria through the infected body, or it was as-
sumed that possibly the micro-organisms had entered these cells be-
cause erf the favorable nutritive environment thus furnished.
The first to suggest that such cell ingestion might represent a
method of defence was Panum,4 who referred to it as a vague possi-
bility. A similar but more convinced expression of this opinion
was made in 1881, according to Metchnikoff,5 by Roser in explaining
the resistance of certain lower animals and plants against bacteria.
But Roser brought no experimental support for his contention, and
little attention was paid to his assertion.
The significance of cell ingestion as a mode of protection against
bacterial invasion, therefore, was hardly more than a vague sugges-
tion when Metchnikoff, who, though a zoologist, had become intensely
interested in the problem of inflammation, began to experiment upon
the cell reaction which followed the introduction of foreign material,
living or dead, into the larvae of certain starfishes (Bipinnaria).
Pathologists, at this time, held complicated views of inflamma-
tion which involved complex coordinated reactions of vascular and
nervous systems, and Metchnikoff's primary purpose was to observe
reactions to irritation in simple forms devoid of specialized vascular
or nervous apparatus. He noted in these transparent, simple forms
of life that the foreign particles were rapidly surrounded by masses
of ameboid cells and reached a conclusion which, in his own words,
is expressed as follows:
1 Hayem. C. E. de la Soc. Biol, 1870.
2 Klebs. Pathol. Anat. der Schusswinden, 1872.
3 Waldeyer. Arch. f. GynekoL, Vol. 3, 1872.
4 Panum. Virch. Arch., Vol. 60, 1874.
5 Metchnikoff. "L'lmmunite dans les Maladies Infectieuses."
PHAGOCYTOSIS
"L'exsudat inflammatoire doit etre considere comme une reac-
tion centre toutes sortes de lesions et Fexsudation est un phenomena
plus primitif et plus ancien que le role du systeme nerveux et des
vaisseaux dans rinflamrnation." 6
He compared the process of cell ingestion or phagocytosis of for-
eign particles, as here observed, to that taking place in the most
simple intracellular digestion which occurs in unicellular forms, a
hereditary cell function now specialized in certain mesodermal cells,
and passed on in the evolution of higher forms to other specialized
cells. And indeed in animals of the most complex structure the
leukocytes which carry on this phagocytic process may be considered
as, in a way, representing a primitive form of cell, since they are
only nucleated elements of the body which wander from place to
place, and are anatomically independent of nervous control. In
1883, at the Naturalists' Congress in Odessa, Metchnikoff 7 first
expressed his views and communicated the first of the splendid re-
searches upon which our modern conception of phagocytosis is based.
His earlier studies were carried out with a small crustacean, the
daphnia, in which he studied the reaction which followed the intro-
duction of yeast cells. He observed the struggle which ensued be-
tween the ameboid leukocytes of the crustacean and the infecting
agents and determined that complete enclosure of the yeast within
the leukocytes assured protection to the daphnia, while a failure of
this process, either from fortuitous causes or because of too large a
quantity of the infecting agents, resulted in disease and rapid death.
This early work of Metchnikoff forms the beginning of a long
train of investigations to which we owe most of the basic facts we
possess concerning the role of the phagocytic cells in the protection
of the body against infection. Just as the various serum phenomena,
of which we have spoken, have a general biological significance apart
from their importance in relation to bacterial invasion, so the process
of phagocytosis must be looked upon as an attribute of the animal
and vegetable cell which has important physiological bearing entirely
apart from infection.
In fact, the ingestion of bacteria and other foreign particles by
the leukocytes and other phagocytic cells of higher plants and ani-
mals is entirely analogous to the intracellular digestive processes
which take place, as the ordinary manner of nutrition, among the
unicellular forms. Among the rhyzopods, in general, food is taken
in by means of the ingestion of other smaller forms of life, bacteria,
6 Inflammatory exudation should be considered as a reaction against all
sorts of injuries, and exudation is a phenomenon more primitive and ancient
than are the parts played by nervous system and blood vessels in the process
of inflammation.
7 Metchnikoff. Arb. a, d. zool. Inst., Wien, Vol. 5, 1883.
274 INFECTION AND RESISTANCE
algse, etc. (or particles of dead organic matter), into the cell bodj-
of the protozoon.
These materials are gradually engulfed by the body of the ameba,
which flows about them with its pseudopods, and within the cyto-
plasm undergo gradual digestion. The process has been carefully
studied by Mouton.8 In symbiotic cultures of amebse with colon
bacilli on agar plates, the bacteria are taken up in large numbers
and about them are formed small vacuoles. That the digestion takes
place in a slightly acid medium with the vacuoles can be proved by
adding a drop of neutral red to the hang-drop preparation of ameba3
and observing the brownish-red color taken by the materials in the
vacuoles. Mouton was able to obtain a digestive ferment from the
amebse, by glycerin extraction, which exerted strong proteolytic action
upon various albuminous substances, liquefied gelatin, and digested
dead colon bacilli in vitro, acting best in slightly alkaline, but also
in slightly acid, reactions. It is plain, therefore, that the most prim-
itive form of digestion is an intracellular one carried on by ferments
comparable in every way to the secreted digestive enzymes which
accomplish the same purpose outside of the cells in higher animals.
In essence, however, there is no fundamental difference physiolog-
ically between intra- and extracellular digestions, and the intracellu-
lar manner of assimilating solid nutritive particles may be retained
in forms much higher in the scale of evolution than the rhizopods.
It has been studied by Metchnikoff and others in certain of the flat
worms (Dendrocelum lacteum) in which typical phagocytosis is car-
ried on by the cells of the intestinal mucosa. Many of these plan-
aria obtain their nourishment by sucking the blood of higher ani-
mals. Placed under a microscope after feeding, it may be seen that
the foreign blood cells are rapidly taken up by the intestinal epithe-
lial cells, which engulf them by means of pseudopodia not unlike
those of the ameba. After ingestion, here, too, the blood cells are
surrounded by vacuoles within which their gradual disintegration or
digestion is accomplished. Similar intracellular digestion seems to
be general among the coelenterates, and has been thoroughly studied
by Metchnikoff in the actinia. Here the food particles are carriecf
by the tentacles into the esophagus, and are taken up by the endo-
dermal cells of the so-called "mesenteric filaments/7 where they are
digested by a trypsin-like enzyme. In these animals digestion is
entirely intracellular, though the ingesting cells are the parts of a
specialized tissue. In other forms, still higher in the scale, although
there is persistence of intracellular digestion, the extracellular
process begins to be developed. Thus in certain mollusca the solid
food is taken into the intestinal canal, where it first undergoes a
preliminary digestion by secreted intestinal juices. After it has
8 Mouton. C. E. de I'Acad. des Sciences, Vol. 133, 1901.
PHAGOCYTOSIS 275
been reduced to small amorphous particles in this way, these are
seized by the ameboid cells, and intracellular digestion completes the
process which has been begun extracellularly.
As we study the process among higher animals, it appears that,
among vertebrates, the intracellular methods of digestion have been,
at least for normal metabolism, entirely displaced by the extracellu-
lar as it occurs in the intestine, where solid particles are rendered
completely amorphous, dissolved, and reduced to a diffusible condi-
tion by the digestive juices before they are offered to the cells for
utilization. However, the capacity for intracellular digestion is not
entirely lost, and is retained of necessity in certain body cells. For
were there not such an emergency arrangement the body would lack
an available mechanism with which to meet such accidents as ex-
travasations of blood, or the entrance of bacteria and other foreign
solid particles into the tissues. It seems reasonable to classify both
the phagocytic action of body cells and the formation of antibodies
in the blood plasma, primarily as emergency devices for the diges-
tion of foreign materials both formed and unformed which, under
abnormal conditions, penetrate into the physiological interior of the
body (blood stream or tissue spaces), and must be disposed of.
In the lowest animals the single cell is called upon to perform
all necessary functions. In the course of evolution, however, as the
body becomes more and more a community of many cells, a division
of labor takes place which is expressed morphologically in the differ-
entiation of tissues and organs, and physiologically in the adaptation
of individual tissue cells to the performance of specialized functions.
Nevertheless, it is necessary, both for certain normal processes, as
well as for provision against such complex emergencies as those
mentioned, that certain cells of the complex community should retain
the primitive abilities of the more independent cells of the lower
forms. Thus, among many animals, the phagocytic action of cells
performs definite services in the course of normal development.
This is seen most markedly in some insects (diptera) in which the
destruction of larval organs, useless to the adult animal, may be en-
tirely accomplished by the action of phagocytic cells, and a similar
process may accompany the transformation of the tadpole to the adult
in many amphibia.9 In higher animals the removal of extravasations
of blood is accompanied by a train of occurrences which is readily
subjected to study.10 In such cases the leukocytes rapidly enter the
area of extravasation and an engulfment of the blood cells occurs,
followed by a process of digestion entirely analogous to the digestion
of similar blood elements by the various forms of intestinal hem-
amebse. In the latter case it is a process of normal digestion, in the
9 See Henneguy. "Les Insectes," Paris, 1904, p. 677.
10 Langhans. Virchow's Archiv, Vol. 49, 1870.
276 INFECTION AND RESISTANCE
former an emergency procedure carried out by virtue of the retained
ancestral characteristics of the special phagocytic cells.
The leukocytes, whose chief functions seem to be associated with
such processes of intracellular digestion, may, therefore, be looked
upon as cells retaining primitive characteristics for definite physio-
logical purposes. We shall see, however, that, to meet exceptional
conditions, the process of phagocytosis may be carried out also by
many other cells which are associated ordinarily with functions en-
tirely apart from this phenomenon.
During normal life in higher animals, too, constant destruction
of red blood cells by phagocytosis takes place in the spleen and liver,
and is described by Dickson 1 1 as occurring in the bone marrow as
well ; and similar phagocytosis of red cells is seen in the hemolymph
nodes. It is claimed by Metchnikoff, furthermore, that many of the
degenerative and retrogressive processes which take place in the
human body are carried on by the mechanism of phagocytosis. The
rapid return of the puerperal uterus to the normal state is explained
in this way, and work by Helme 12 seems to show that there is an
actual phagocytosis of the hyperplastic uterine musculature during
this period. The atrophic changes of senility, too, are attributed by
Metchnikoff 13 14 to the same processes. The involution of the
ovaries is accompanied by active phagocytosis of portions, of this
organ, and Metchnikoff claims further to have shown that the de-
generation of the nervous system during old age is accomplished by
the phagocytosis of nerve cells by phagocytic elements derived either
from the leukocytes or the neuroglia, or from both.15 The whitening
of the hair, both in human beings and in old animals (dogs), is simi-
larly due, he claims, to phagocytosis of the pigment by cells which
wander in from the root sheaths. It is, up to the present time, im-
possible to determine the stimulus to which this phagocytosis is due.
Since the subject is a very important one, many studies have been
made to determine which cells of the body of higher animals can
take in and digest foreign particles and to classify them according
to this power. Metchnikoff has distinguished between the "motile"
and "fixed" phagocytes, the former the leukocytes of the circulating
blood, the latter certain connective tissue cells, endothelial cells,
splenic pulp cells, and certain cellular elements of the lymph nodes,
11 Dickson. "The Bone Marrow," Longmans, Green, London, 1908.
12 Helme. Transact. Roy. Soc. of Edinburgh, Vol. 35, 1889. Cited from
Metchnikoff.
13 Matschinsky. Ann. de I'Inst. Past., Vol. 14, 1900.
u Metchnikoff. Ann. de I'Inst. Past., Vol. 15, 1901.
15 That the leukocytes are concerned in the destruction and resorption
of dead tissues has been shown by Leber especially (Leber, "Die Entstehung
der Entziindung," Leipzig, Engelmann, 1891). An accumulation of leukocytes
about a bacterial focus or from any other stimulus is followed by tissue
lysis due to leukocytic enzymes.
PHAGOCYTOSIS
277
POLYNUCLEAR LEUKOCYTES TAKING UP STA-
PHYLOCOCCI.
the neuroglia tissue, and, in fact, all phagocytic cells which are
ordinarily confined to some definite localization in the body. Among
phagocytic cells Metchnikoff further distinguishes between "micro-
phages," by which he designates the polymorphonuclear leukocytes
of the circulating blood and
"macrophages." The ma-
crophages include the fixed
cells mentioned above, to-
gether with the large mono-
nuclear elements of the
blood, in short, all phago-
cytic cells except the micro-
phages.
Although no absolute
functional differentiation is
possible between the two, it
is true, in a general way,
that the microphages are
concerned primarily with
the phagocytosis of bacteria
and especially of those
which invade acutely, while
the macrophages are con-
cerned especially with the resorption of cellular detritus, foreign
bodies, and such bacteria as are more chronic in their activities, or
are peculiarly insoluble.
On the other hand, micro-
phages may take up foreign
particles and bacteria of all
kinds under suitable condi-
tions, and no sharp line can
be drawn between the two
varieties in this respect.
Metchnikoff further be-
lieves that the two classes
of phagocytic cells differ in
the nature of the protective
substances they secrete and
furnish in the blood
plasma. This, however, is
a problem concerning which
there is much difference of
opinion and which calls for
KUPFER CELLS CONTAINING MALARIAL PIG- ~prmrfltp rliomiccinTi in an
MENT. UNGRAMMATICALLY DRAWN FROM
A SECTION OF MALARIAL LIVER KINDLY other place.
FURNISHED BY DR. E. LAMBERT. The property of phago-
278
INFECTION AND RESISTANCE
EAT LEPROSY BACILLI GROUPED IN THE EEMAINS OF
DEAD SPLEEN CELLS GROWING IN PLASMA.
cytosis is therefore
an attribute of a
considerable num-
ber of different va-
rieties of cells. In
the circulating
blood the polynu-
clear leukocytes are
the most actively
motile and phago-
cytic elements. The
eosinophile cells
may also take up
foreign particles
and bacteria, as
may also the large
lymphocytes. The
Drawn after illustration in Zinsser and Carey, Journal small lymphocytes
of the A. M. A., Vol. 58, 1912. and mast cells are
either entirely inac-
tive in this respect, or, at least, possess phagocytic powers under ex-
ceptional circumstances only. This does not mean, however, that
these last-named cells may
not accumulate at the point
of invasion nor that they
may not play an important
part in the defence of the
body. It is well-known, of
course, that, in tuberculosis
and a number of other con-
ditions, the lymphocytes
may form the majority of
the cellular elements which
accumulate at the site of
the lesion. Among the
fixed cells of the body it is
probable that phagocytosis
may be carried on by cells
<of many different origins,
though the identification of
PHAGOCYTOSIS or SENSITIZED PIGEON COR-
PUSCLES BY ALVEOLAR CELLS OF LUNG.
Drawing made after photomicrograph pub-
lished by Briscoe, Journal of Path, and
Bact., Vol. 12, 1908.
•cells in tissues is often a
purely morphological prob-
lem, and therefore fraught
with many possibilities of
error. Probably the most active fixed tissue cells are the endothelial
cells of the blood vessels and those which line the serous cavities, the
PHAGOCYTOSIS
279
sinuses of the lymphnodes, and of the spleen. However, there are
many other cells in addition to these which may be phagocytic. The
writer, with Carey,16 has observed the active phagocytosis of leprosy
bacilli by cells, probably of connective tissue origin, growing from
plants of rat spleen in plasma. Phagocytosis by the cells lining the
alveoli of the lungs has been observed by Briscoe.17 This author made
the interesting observation that in cases of mild infection such cells
can free the lungs of micro-organisms entirely without aid from the
leukocytes of the circulating blood. It is these cells, too, which, in
the ordinary conditions of life, take up the inhaled particles of dust
and are, therefore, often spoken of as dust cells. The origin of the
dust cells has often been the subject of controversy. In the embryo
the alveoli of the lung, like the bronchi, are lined with columnar cells
which are transformed into flattened epithelium as the alveoli ex-
pand at the first inspirations after birth. These flattened cells,
which constitute the alveolar or dust cells, are probably of epithelial
origin, and as such are probably the only epithelial cells which act
as phagocytes under ordinary conditions. Although no positive
general statement is justified, we can yet say with reasonable accuracy
that among the phagocytic fixed tissue cells the most important are
the connective tissue and endothelial cells.
The type of phagocy-
tosis and the variety of cell
whic,h participates in it
seem to depend to a great
extent upon the nature of
the substance which incites
the process, or rather at
which the process is aimed.
Thus the large cells which,
in tissues, take up the lep-
rosy bacillus, those which
are characteristic of tuber-
culous foci, or those caused
by blastomycetes, or by for-
eign bodies, all have special
appearances which are suf-
ficiently characteristic to
have diagnostic value.
However, it is difficult to determine with certainty the origin
of the cells which participate. The chemical nature of the substances
taken up, moreover, often complicates the phagocytic process in such
a way that different cellular elements are enlisted in succession in
order that the ingested substances may be disposed of. Thus tubercle
16 Zinsser and Carey. Jour. A. M. A., March, 1912, Vol. 58.
17 Briscoe. Jour, of Path, and Bacter., Vol. 12, 1907.
GIANT CELL IN TUBERCULOSIS.
280 INFECTION AND RESISTANCE
or leprosy bacilli which are injected into an animal may be at first
taken up by polynuclear leucocytes or microphages, by which they may
even be carried into the lymph channels and distributed, perhaps to
the detriment of the host. But these cells, probably because they lack
a lipolytic ferment by means of which the waxes of the acid-fast or-
ganisms can be digested, cannot destroy the bacteria, which are then
attacked by other cellular elements at the site of their final deposit.
That fixed tissue cells as a matter of fact play a very important
role in the disposal of invading bacteria is becoming more and more
clear. Kyes ls showed a few years ago that the immunity of the
pigeon to pneumococcus infection is largely due to active removal of
the bacteria by the Kupfer cells in the liver. Bull's 19 observations
on the intravascular agglutination of typhoid bacilli which are then
phagocyted by cells in the liver and spleen point in the same direc-
tion, and recently Hopkins and Parker 19a in our laboratory have
shown that streptococci injected into rabbits and cats are rapidly re-
moved from the circulation, the removal being to a great extent due
to phagocytosis carried on by the endothelial cells in the lungs and
by similar cells in the liver and spleen.
In many such cases the further resolution of the foreign sub-
stance is accomplished by an important type of phagocytosis which is
characterized by the formation of the so-called giant cells. These
cells are of varying appearance in different conditions and locations.
Thus the giant cells which form about foreign bodies, such as the
small cotton fibers occasionally left in wounds, or injected particles of
paraffin or iron splinters, etc., are quite characteristic and distinct
from the giant cells of tuberculous foci, or of rhinoscleroma, glanders,
or leprosy. They are all large cells, containing often numerous
nuclei which form either by the fusion of several cells, as claimed by
Borrell,19b Hektoen,19c and others, or by the cleavage of the nuclei
alone, without coincident divisions of the cytoplasm.
Although it is, of course, impossible to decide definitely upon
purely morphological grounds, the researches of Hektoen especially
would lead one strongly to favor the former view. It is equally
difficult to decide the origin of giant cells, and endothelial, connec-
tive tissue, and even leucocytic origin has been claimed for them.
Yet in no case has it thus far been possible to actually observe their
formation by a method which could positively decide this point.
In order to gain a clear conception of the participation of phago-
cytes in the response of the body to injury or invasion, it will be
useful to follow out the process of inflammation as it occurs in the
18 Kyes. Journal of Infectious Diseases, Vol. 18, 1916, p. 272.
19 Bull. Journal of Exp. Med., Vol. XX, p. 237.
19a Hopkins and Parker. Paper in Manuscript.
19b Borrell. Ann. de I'Inst. Past., 7, 1893.
19c Hektoen. Jour. Exp. Med.} 3, 1898, p. 21.
PHAGOCYTOSIS
281
higher animals. Inflammation may be incited by a large number of
agencies — chemical irritants, mechanical injury, or even by the in-
troduction of inactive and isotonic substances such as broth or salt
solution.20 21 Yet in these cases the response, though essentially
'similar in principle to that following invasion by bacteria, lacks
certain features especially interesting in the present connection, and
it will be most profitable for our purpose to consider in detail the
result of infection with pathogenic micro-organisms.
If an emulsion of pyogenic staphylococci is injected into an ani-
mal subcutaneously the site of injection will soon become reddened
and swollen and microscopic examination will show, within a few
hours, a swelling and engorgement of the blood vessels.
The injected cocci will be found to lie partly scattered in the
tissue spaces, in part within polynuclear leukocytes and connective
tissue cells which have begun to ingest them. The tissue spaces
will be swollen and
stretched by the
exudation of blood
serum from the ves-
sels. This condi-
tion will begin in
from 4 to 6 hours
after injection and
increase during the
next 24 hours in ex-
tent and severity,
according to the
quantity and viru-
lence of the cocci
injected. The con-
ditions which pre-
cede the wandering
of the p o 1 y m o r-
phonuclear leuko-
cytes out of the ves-
sels have been care-
fully studied in such thin tissues as the mesentery of a frog after
injury by trauma or acid. Within the vessels of the affected area
there is at first an acceleration of the blood stream, then a dilatation
of the capillaries and a slowing of the current. Leukocytes may now
be observed moving more slowly than the main stream, and keeping
close to the periphery along the walls of the vessels. Here and there
they seem interrupted in their movements and adhere to the vascular
wall. A little later these cells appear to pass through the wall of the
20 See Adami. "Inflammation," Macmillan, London, 1909.
21 See Adami, loc. cit.
DIAGRAMMATIC EEPRESENTATION OF LEUKOCYTES WAN-
DERING THROUGH CAPILLARY WALLS.
Adapted from Eibbert, "Lehrbuch der Allgemeinen
Pathologie," p. 337.
282 INFECTION AND RESISTANCE
vessel by sending out pseudopodia which slowly penetrate it. Adami
states that if, at this stage, the tissues be excised, fixed in osmic
acid, and stained, leukocytes may be seen crowding the inner sur-
face of the vessel in all stages of transition from its anterior to
the lymph spaces on the outside.
In the staphylococcus infection, after from 12 to 4$ hours, there
will be seen the results of an active and destructive struggle between
the invading bacteria and the defending cells. In the center of the
area of invasion tissue has been destroyed and disintegrated. Amid
the necrotic detritus, closely packed, lie leukocytes and cocci and
active phagocytosis has taken place. In some cases the intracellular
bacteria appear swollen and disintegrating, in others the leukocyte
itself, overcome by the larger number of bacteria it has taken in,
becomes vacuolated, indefinite in outline, and apparently is being
itself destroyed. The presence of blood serum, which is aiding in
the destruction of bacteria both by its bactericidal powers and its
reenforcement of the phagocytic process, renders this mass fluid or
semi-fluid, and the whole mixture constitutes what is known as pus.
Around the periphery cocci and leukocytes become more scattered
and sparse, and bacteria, together with leukocytes, loaded with cocci,
may be seen lying within large mononuclear cells (macrophages).
Whether the process goes on to further extension or is eventually
walled off into a distinct abscess by the formation of granulation
tissue and new connective tissue depends upon the balance of forces
between attacking agent and defensive factors.
If we inject a similar emulsion of cocci into the pleural or peri-
toneal cavity of an animal a process similar in principle may be
observed.
Normally the peritoneum contains a small amount of this serous
fluid and a moderate number of white blood cells, chiefly lympho-
cytes. When any substance, broth or salt solution, an aleuronat or
a bacterial emulsion, is injected into the peritoneal cavity, there
follows a brief period during which there is a diminution of the
free cellular elements in the peritoneal fluid. At this time there is
a clumping of cells in the folds of the omentum and mesentery, a
transient stage of flight away from the point of injury. This, how-
ever, is soon over. Within one to two hours an active immigration
of leukocytes into the serous cavity occurs and if, during the next
12 to 24 hours small quantities of fluid are, from time to time, with-
drawn with a capillary pipette, a rapid and constant increase of
leukocytic elements, chiefly of the microphage or polynuclear type,
is observed. If the injected substance has been a sterile, harmless
fluid, a gradual return to normal within 48 hours then ensues. If,
however, we have injected bacteria, a struggle similar to the one
described above takes place within the peritoneum, and active
phagocytosis of the micro-organisms takes place.
PHAGOCYTOSIS 283
Let us suppose that the injected bacteria have been small in
quantity and moderate in virulence. In such a case a rapid phago-
cytosis gradually rids the fluid of micro-organisms and within 24
hours after injection few, if any, free bacteria are visible.
A little exudate taken at this time shows large numbers of micro-
phages varyingly crowded with well-preserved and disintegrating
bacteria. Some of the phagocytes, having literally taken up more
than they can digest, are vacuolated and disintegrating, but, in gen-
eral, the victory lies with the cells. A little later large mononuclear
elements appear, and here and there will be seen to take up dead
leukocytes together with ingested cocci. In this way gradually a
cleaning out of the peritoneum takes place, the animal recovers, and
the peritoneum returns to normal.
Let us suppose, on the other hand, that the bacteria injected are
in larger doses and of greater virulence. In such a case, after a
period of active phagocytosis, there may be a gradual increase of
bacteria over leukocytes. The phagocytic cells are found to be under-
going degeneration in larger numbers, the free bacteria increase,
and the impending death of the animal can often be foretold by the
appearance of the exudate. Finally, the peritoneal fluid may con-
sist chiefly of free and rapidly multiplying bacteria with a practical
absence of phagocytic cells.
In all of the processes so far as described the burden of the
defence has fallen upon the microphages or polynuclear leukocytes,
while the macrophages — endothelial and connective tissue cells —
have taken a purely secondary part in the reaction, forming, to some
extent, a second line of defence, or, more probably, taking part only
in the final removal of degenerated and disintegrating combatants
and tissue detritus. In order to obtain a complete conception of
phagocytosis in its entire significance it will be necessary to consider
a further example, namely, the process which takes place within tis-
sues in the course of the efforts of macrophages to remove bacteria
and other substances which, either because of their insolubility or for
other unknown reasons, are refractory to the attacks of the mi-
crophages. Since we are interested in this subject chiefly from the
point of view of the defence against bacteria, we may illustrate this
process best by the description of the reaction which takes place when
tubercle bacilli become localized anywhere within the animal body.
When tubercle bacilli are injected into the peritoneum they are
actively taken up by the polynuclear leukocytes just as are other
bacteria and many entirely inactive solid particles. A similar inges-
tion by microphages may take place in the folds of the intestinal
mucosa if tubercle bacilli are fed to guinea pigs. However, this
preliminary phagocytosis is probably of but secondary significance
in the combat of the body against tuberculosis, since it has still to
be shown that polynuclear leukocytes are capable of digesting and
284 INFECTION AND RESISTANCE
destroying acid-fast bacilli. Indeed, much evidence tends to show
that the ingestion of tubercle bacilli by microphages may be a detri-
ment to the host, since the bacilli by this means are carried through
the lymphatics and variously distributed throughout the body. Poly-
nuclear leukocyte extracts, though containing, as we shall see, pro-
teolytic enzymes, do not, according to Tschernorutzky, contain any
lipase, and it may well be that for this reason they are unable to
attack the waxy substances which form an integral part of these or-
ganisms. This is in keeping with the observations made by Terry
in our laboratory, that rat leprosy bacilli may be kept within leu-
kocytes for weeks without losing their acid-fast properties, whereas
the same bacilli, as the writer and Gary found, were rapidly disin-
tegrated in spleen cells growing in plasma. Moreover, it is well
known that the estimation of tuberculo-opsonin contents of the sera
of tuberculous patients has been peculiarly unsatisfactory in throw-
ing light on the progress of the disease. It would seem, therefore,
that in this disease, as well as in others caused by acid-fast organ-
isms, the microphages play only an unimportant part in the defence
of the body.
On the other hand, when tubercle bacilli are deposited either in
a lymphnode (through the vehicle of leukocytes) or in a capillary
anywhere by the blood stream, a train of cellular changes is initiated
in which the predominant part is played by the macrophages. The
tubercle bacilli so deposited are rapidly surrounded by large mono-
nuclear cells, probably endothelial in origin. Some of the micro-
organisms may even be phagocyted and taken into these cells. These
cells, spoken of as "epithelioid cells," surround the clump of bacteria
in more or less concentric rings, and around these there is an accumu-
lation of leukocytes, largely of the lymphocyte variety, with an ad-
mixture of a very few microphages. Then by the fusion of endothe-
lial cells, or possibly by division of the nuclei of some of these cells
within the individual cell bodies, giant cells are formed which take
up the bacilli. The further progress of the tubercle now greatly
depends upon the balance of power. Often such a tubercle may
heal, possibly because of complete intracellular digestion of the ba-
cilli. On the other hand growth and multiplication may lead to a
slow and dry necrosis of the center of such a mass of cells, leading
to the condition spoken of as caseation. Epithelioid cells lose their
outlines and staining properties, and go to pieces. The center of the
lesion is a grumous mass, the periphery shows a few giant cells and
connective tissue proliferation.
It is always surprising to those who study these lesions for the
first time how rarely they succeed in finding tubercle bacilli in
microscopic sections prepared from such tubercles by the ordinary
Ziehl-E"eelsen method of staining. Repeated and careful examina-
tion of such material may fail to reveal any acid-fast organisms,
PHAGOCYTOSIS 285
though inoculation into guinea pigs is nevertheless successful, pro-
ducing typical tuberculosis. Much 22 has studied this peculiar state
of affairs particularly and has shown that, although such lesions may
show no tubercle bacilli by the Ziehl-Neelsen carbol-fuchsin method,
staining by a modified Gram technique will reveal numerous Gram-
positive rods and granules which have lost their acid-fast properties.
This, too, if true, and the evidence is very much in its favor, would
point to an ability of the macrophages to digest the waxy substance
•of the tubercle and other acid-fast bacilli, a property not possessed
by the microphages. It may, of course, mean on the other hand that
the tubercle bacilli in the lesion have not developed the waxy condi-
tion.
ClIEMOTAXIS AND LEUKOCYTOSIS
The part played by the phagocytic cell in the defence of the body
against the entrance of bacteria and other foreign substances consists,
then, of two functionally different phases. The first is an active
motion of the cells toward the point attacked, and their accumulation
about the noxious agent, the second consists in the act of ingestion
itself.
The motion of the leukocytes toward the invading substances
indicates a sensibility on the part of the cell to changes in its environ-
ment incited by the foreign agent, and since the stimuli most likely
to reach the leukocytes and bring about this alteration in the direc-
tion of their movements are chemical in nature, the phenomenon is
spoken of as "chemotaxis." This term was borrowed from Pfeffer,23
who studied similar phenomena in connection with many freely
motile plant cells, spermatozoa, and bacteria. Since the change of
direction brought about in a moving cell by such influences may be
such as either to attract or to repel, the term "positive chemotaxis"
is used to designate the former and that of "negative chemotaxis"
the latter.
The property of chemotaxis is of vital interest in the present
connection, since, whatever may be our opinion regarding the relative
values of phagocytosis and serum protection in immunity, the great
importance of the phagocytic process cannot be questioned, and any
agency which repels the approach of the phagocytes must be a detri-
ment, while any factor which attracts them is, of necessity, a power-
ful means of defence. In the investigations upon the nature of infec-
tious diseases attention 'has been concentrated upon the phenomenon
of phagocytosis, and the relations governing the act of ingestion
have been very thoroughly studred. The details of the chemotactic
phenomenon, however, though of equal importance, are much more
22 Much. "Beitrage zur Klinik der Tuberk.," Vol. 8, 1907, Hft. 1 and 4.
23 Pfeffer. "Untersuch. a. d. Botan. Inst. Tubingen," Vols. 1 and 2, 1884
sincl 1S88.
286 INFECTION AND RESISTANCE
obscure. A large part of our sparse knowledge in this connection,
moreover, has been gained by studies not related to infection.
The stimuli which determine the motion of cells are, of course,
not necessarily chemical, and extensive studies have been made upon
the effect of light waves in this connection. Although these inves-
tigations are of great biological importance, they have little direct
bearing upon the problems of tropism as related to bacteria and leu-
kocytes and cannot therefore be considered here.
Some of the earlier researches upon chemotaxis were those made
by Stahl 24 upon the slime-molds or myxomycetes. These organisms
possess the power of ameboid motion, and were observed by Stahl to
move toward or away from any given region, according to the nature
of the substances with which they came in contact. Pfeffer sub-
jected this phenomenon to closer analysis. Working with the sperma-
tozoa of ferns, swarm spores, bacteria and infusoria, he elaborated
an ingenious technique by means of which he was enabled to de-
termine directly the negative or positive chemotactic properties of
various substances in solution upon these motile forms. His tech-
nique was exceedingly simple. Capillary glass tubes, about 8 to 10
mm. long and 0.1 mm. in diameter, were sealed at one end in the
flame, and then dropped into a watch-glass. The solution which was
to be tested was poured over the tubes and the watch-glass then
placed under the bell of an air-pump. When the air was evacuated
and pressure reduced the tubes became partly filled up with the
liquid. They were then removed, washed in water, and placed under
a cover slip under which a preparation of the motile cells was swim-
ming. Positive chemotaxis was indicated by entrance of the cells
into the tubes, negative, by their refusal to enter. Failure of the
solution to exert any chemotactic influence resulted in their moving
into and out of the tubes indiscriminately.25
By this technique a large number of interesting observations were
made which threw much light upon the causes underlying the move-
ments of plant cells. For instance, in investigating the spermatozoa
of the ferns it was found that they were attracted strongly by malic
acid and its salts, while no other substance investigated approached
these compounds in the intensity of positively chemotactic stimula-
tion. From this Pfeffer concludes that the bursting of the fern
archegonia is accompanied by the liberation of malic acid, this at-
tracting the male to the female cell.
Similar experiments have been carried out since then by numer-
ous naturalists, among them Buller,26 Lidforss,27 and Jennings,28
24 Stahl. Botanische Zeitung, 1884.
25 Buller. Annals of Botany, Vol. 16, No. 56, 1900.
26 Buller. Loc, tit.
27 Lidforss. "Jahrbiicher f. wissensch. Botanik," 41, 1904.
28 Jennings. "Behavior of Lower Organisms/' Columbia Univ. Press,
Macmillan, 1906.
PHAGOCYTOSIS 287
and it has been found that in addition to malic acid compounds many
other substances, organic and inorganic, occurring in plant cells and
cell-sap exert positive chemotactic power. Lidforss has shown, for
instance, that calcium chlorid in 0.1 per cent, solution may strongly
attract plant spermatozoids (equisetum — horsetail). When the solu-
tion is concentrated to 1 per cent., attraction is still exerted, but the
spermatozoids immediately lose their motility upon entrance into the
fluid.
The same worker has shown that a substance which is positively
chemotactic for one variety of plant cell may be negatively chemo-
tactic for another, showing a certain selective variation which should
be of great biological importance. Thus capillaries with a 1 per cent,
solution of potassium malate actively attracted the spermatozoids of
marchantia (a liverwort), while not a single spermatozoid of equi-
setum would enter these tubes. Low 29 has applied these methods of
study to the investigation of the chemotaxis of mammalian sperma-
tozoa and found that these cells were actively attracted by weakly
alkaline solutions.
Studies upon the factors determining the movement of bacteria
and ameba3 toward some substances and away from others have been
numerous, and are valuable for the understanding of leukocytic
chemotaxis, because they have led to the formulation of a number of
important general theories. The fact that the motions of bacteria
in suspensions are, to a certain extent, determined by the negative
electrical charge which they all carry in neutral media, has been
touched upon in the section on agglutination. Attempts on the part
of Young and the writer to determine whether the attraction of
leukocytes toward bacteria might be due to the carrying of an elec-
tropositive charge by the white cells have met with no result, owing
so far to the failure to elaborate a reliable technique. However, this
thought is not an impossible one and should be borne in mind.
That certain bacteria will wander actively toward a source of
oxygen was shown by Engelmann' s 30 classical experiment in which a
diatom, half in the shade and half in the light, was surrounded by an
emulsion of bacteria, and these were seen to collect about the lighted
half only, where oxygen was being liberated by virtue of the chloro-
phyll. The extreme delicacy of chemotactic reactions is illustrated
in these experiments in that Engelmann calculated that the bacteria
reacted to one one-hundred billionth of a milligram of oxygen. The
selective reaction of bacteria to various chemical substances, further-
more, has been shown by allowing different solutions to diffuse into
bacterial emulsions from capillary tubes, and by observing attraction
or repulsion from the point of contact.
The chemotaxis of leukocytes has opposed more difficulties to
29 Low. Sitzungs Berichte kais. Akad. d. Wiss., Wien, Vol. 3, Abf. 3.
80 Engelmann. Arch. f. d. ges. PhysioL, Vol. 57, p. 375.
288 INFECTION AND RESISTANCE
direct study, since the conditions within the living body are subject
to a large number of modifying factors, and experiments upon the
isolated cells, in vitro, even under conditions of the most careful
technique, are fraught with much unavoidable injury to the cells.
However, enough has been learned to indicate that these cells are
subject to the phenomena of chemotaxis or tropism just as are inde-
pendent unicellular forms, and that they may be attracted or re-
pelled by a variety of organic and inorganic substances. Leber 31
was one of the first to study this in his work upon inflammation.
He found that leukocytes were actively attracted by powdered cop-
per and mercury compounds, but not by powdered gold or iron. He
also observed that dead bacteria exerted a similar positive chemo-
tactic influence, and Buchner 32 later succeeded in extracting sub-
stances from various bacteria which possessed similar properties. It
appears, from these and other investigations, that the power of stim-
ulating positive chemotaxis is a general property of bacterial pro-
teins, equally evident in bacterial extracts, dead bacteria, or the liv-
ing organisms. It is likely, therefore, that the attraction of leu-
kocytes toward the point of bacterial invasion is, in part at least,
due to the properties of the bacterial proteins themselves. That this,
however, is not the whole story is evident from the work of Massart
and Bordet,33 who showed that the products of cell destruction and
disintegration possess similar positively chemotactic properties. This
is true not only of the products of disintegrated tissue cells, but of
those of the destroyed leukocytes themselves. Thus it appears that
when any injury of tissue takes place, a stimulus which attracts
leukocytes results, even when the injury is not accompanied by bac-
terial invasion. This would explain the participation of leukocytes
in reactions to injury, and in inflammations not of bacterial origin,
and their local accumulation following the injection of insoluble
inorganic substances.
When bacteria are actually present, however, the added stimulus
due to the diffusion of bacterial proteins probably increases the
process to a degree often sufficient to meet the added requirements
for protection. Following this, both the destruction of tissues, of
bacteria, and of leukocytes may together exert a cumulative chemo-
tactic power which continues the process proportionately with the
extent of the lesion.
It is of the utmost importance, therefore, to ascertain whether
or not any substances derived from bacteria may, under any circum-
stances, exert a repellent or negatively chemotactic power. If we
infect an animal intraperitoneally with virulent bacteria, in doses
31 Leber. FortscJir. der Med., 1888; also "Die Entstehung der Ent-
ziindung," Engelmann, Leipzig, 1891.
32 Buchner. Berl. klin. Woch., Vol. 27, No. 30, 1890.
33 Massart and Bordet. Ann. de I'Inst. Past., Vol. 5, 1891.
PHAGOCYTOSIS 289
sufficient to lead to death, and examine the peritoneal exudate just
before the lethal outcome, we may observe that leukocytes are gradu-
ally disappearing, and that finally but a few will be present and the
fluid will be swimming with free micro-organisms. In the same way
it is well known that the diminution of leukocytes in the circulating
blood — or even the failure of these cells to increase in the circulation
in the course of such diseases as pneumonia, or general infections
with staphylococci or streptococci — is seriously prognostic of fatal
outcome. The conditions here observed point strongly to the ex-
istence of substances of negative chemotactic influence which protect
the bacteria, not from phagocytosis itself, but from that necessary
forerunner of phagocytosis, the approach of the leukocyte. It is
necessary to draw this distinction since these phenomena are not
merely, as often believed, "antiopsonic," but in truth largely "anti-
chemotactic." It is true that Kanthack,34 and more especially
Werigo,35 have denied the existence of negatively chemotactic bac-
terial products, the latter basing his assertion upon the observation
that active phagocytosis occurs in the lungs, liver, and spleen of
animals dying of infection with virulent germs. However, the* argu-
ments of these authors are not conclusive and the mass of experi-
mental and clinical evidence which points to a direct failure of
leukocyte accumulation in the presence of virulent bacteria in the
animal body would alone suffice to render such conclusions unlikely.
Moreover, strong evidence in favor of the existence of negatively
chemotactic influences, is brought by the extensive experiments of
Bail upon the so-called aggressins, discussed in another place, and
such observations as those of Vaillard and Vincent 36 and Vaillard
and Rouget,37 which showed that the injection of a little tetanus
toxin together with tetanus spores would prevent the ingestion of the
spores by leukocytes, and thereby furnish an opportunity for germi-
nation and consequent fatal toxemia.
Similar observations have been made by Besson 38 in the case of
the bacillus of malignant edema by the use of the original technique
of Pfeiffer. Capillary tubes containing the toxin remained free of
leukocytes after subcutaneous introduction into guinea pigs, while
similar tubes containing the culture medium alone, or the bacilli and
their spores, attracted leukocytes in considerable numbers.
It is possible, of course, to interpret such phenomena as due to a
failure of positive chemotaxis rather than to an active negative
chemotaxis.
Although the phenomena of chemotaxis are most easily studied
34 Kanthack. Quoted from Adami, loc. cit.
35 Werigo. Ann. de I'Inst. Past., Vol. 8, 1894.
36 Vaillard and Vincent. Ann. de I'Inst. Past., Vol. 5, 1891.
37 Vaillard and Rouget. Ann. de I'Inst. Past., Vol. 6, 1892.
38 Besson. Ann. de I'Inst. Past., Vol. 9, 1895.
290 INFECTION AND RESISTANCE
in extravascular inflammatory changes, there is none the less a
regular and apparently purposeful attraction or repulsion of leuko-
cytes evident in the circulating blood during infectious diseases.
That infection of the body with many micro-organisms results in the
increase of leukocytes, and that in others there is either no increase
or even a decrease, is too well known and too generally applied in
diagnosis and prognosis to warrant our giving up much space to a
review of the facts. Nevertheless, the causes which lead to a leuko-
cytosis in the one case, a leukopenia in the other, are still very
obscure and deserve discussion.
I In the first place it is by no means certain whether a leuko-
cytosis signifies an active discharge of new leukocytes from the bone
marrow or whether it means simply an altered distribution in that
the phagocytes accumulated in the lymphatic and other organs are
attracted by chemotaxis into the peripheral circulation. Studies of
the bone marrow during infection as well as the occasional appear-
ance of myelocytes and other cells ordinarily found only in the bone
marrow during health would point toward a participation of active
bone-marrow hyperplasia in the increase of peripheral leukocytes.
There is no good reason to doubt, moreover, that a chemotactic stimu-
lus exercised in the circulation should withdraw leukocytes from
any place of accumulation to the circulation. Probably both proc-
esses take part. When bacteria are injected into the circulation of
an animal there is, at first, a moderate diminution of the leukocytes
just as there is after injection of bacteria or other substances into
the peritoneum. This is soon followed in most cases by a rapid and
progressive increase, in which, whenever the leukocytosis is one of
considerable degree, the polynuclear leukocytes preponderate. The
extensive clinical study of the white cells in infectious disease of the
human being give us more material for reasoning in this respect
than we have available from animal experiment. Infection with
invasive bacteria such as the pneumococcus (and Neufeld and others
have shown that most lobar pneumonias are accompanied by pneu-
mococcus bacteriemia), streptococci, staphylococci, and others is
always accompanied by an increase of the leukocytes, while, in
typhoid fever, influenzal infection, tuberculosis, and a number of
other infections, the leukocytes do not increase and may even de-
crease. How are we going to account for this ? That all these bac-
teria contain a substance which is positive in its chemotactic effects
is easily demonstrated by injecting them into the peritoneum and
observing an accumulation of leukocytes and a consequent phago-
cytosis, even in the cases of those organisms which do not call forth
a leukocytosis in the blood of the diseased human being. Thus it
has been our experience as well as that of others invariably to ob-
serve the rapid and complete polynuclear phagocytosis of both
leprosy bacilli and tubercle bacilli after injection of these micro-
PHAGOCYTOSIS 291
organisms into the peritoneal cavities of guinea pigs. Yet a chronic
tuberculous peritonitis or pleurisy is characterized usually by an
exudate which contains but few polynuclears and relatively many
lymphocytes. A final explanation of these conditions is not pos-
sible at present. No adequate explanation for the selective accu-
mulation of lymphocytes and the absence of polynuclear cells about
tuberculous foci has yet been advanced. The absence of polynuclear
leukocytosis may possibly be due to the great insolubility of these
bacilli, in consequence of which little or no leukocytosis-stimulating
substances are liberated.
Pearce 39 has suggested a similar reason for the absence of poly-
nuclear accumulations about chronic localized lesions of any kind
in which tissue encapsulation may prevent the contact of the inciting
agents with the body fluids and there is a consequently slow or slight
production of such chemotactic stimulating materials.
In typhoid f ever, where the slight primary leukocytosis is rap-
idly succeeded by a leukopenia with a relative lymphocytosis, the
conditions are somewhat different. Here, as in some other infec-
tions, as Friedberger and others have shown, we are dealing with a
generalized infection by an organism which is easily subject to the
action of alexin with consequent production of anaphylatoxin. (See
chapter on Anaphylaxis.) This poison, it seems, exerts a nega-
tive chemotaxis, and probably during the height of the disease,
therefore, leads to the low leukocyte count observed. That this is
at least likely seems to follow from the studies which have been
made upon the nature of the typhoid poisons, and also from the
observation of Gay and Claypole, that typhoid-immune rabbits react
to the infection of typhoid bacilli with a rapid and powerful increase
in the polynuclear leukocytes, whereas similar injections into the
normal animal lead to leukopenia.
If the supposition regarding tuberculosis, made above, is correct,
it would follow that a sudden and considerable increase in the poly-
nuclear leukocytes in a case of tuberculosis would indicate a dis-
charge of organisms into the circulation and a tendency toward gen-
eralization of the infection in this manner. (See Weigert's view of
the manner in which tuberculosis may spread by the destruction of
the wall of a vein by a localized lesion.) However, although specu-
lation in the absence of experimental proof is justified, it must not
be forgotten that the problems of selective chemotaxis are too ob-
scure to permit of our laying much weight on any of these views.
Gabritchewsky,40 who investigated this subject extensively, has
classified various substances according to their positively, negatively,
or neutral chemotactic activities. It is not necessary to recapitulate
these, but it is interesting to note that he found that some substances
39 Pearce. Jour. A. M. A., Vol. 61, 1913.
40 Gabritchewsky. Ann. Past., Vol. 4, 1890.
INFECTION AND RESISTANCE
which were positively chemotactic in certain concentrations became
neutral or even negative when the concentration was altered.
We have seen that the action of the leukocytes in moving toward
some substances and away from others is entirely analogous to simi-
lar phenomena occurring among lower, unicellular forms of life, and
the explanations applied to the apparently conscious acts of the
ameba, such as the motion toward and the engulfment of food, have
been applied to the activities of the leukocytes as well. Many of
the theories developed concerning the free living forms, however,
have been easily excluded in the case of the leukocytes, because of the
environment in which their activities are developed. Thus the many
interesting reactions of paramecia and other organisms to light
(heliotropism) have little bearing upon this subject, and the views
based on the theories of orientation may be excluded on the ground
of the symmetry of the normal leukocyte. The observations of
Garrey,41 that indicate that it is the dissociated ions of various acids
and bases which are responsible for the directive stimuli exerted
upon certain flagellates, may yet result in throwing some light upon
leukocytic movements, especially if we can come to accept the con-
ceptions of ion-proteins upheld by Loeb 42 and his pupils. However,
the facts concerning these phenomena, as well as the possibility,
previously mentioned, of the opposite electrical charges carried by
the leukocytes and the substances attracting them cannot be regarded
at present as more than interesting thoughts. Of more than merely
speculative interest, however, are the views of chemotaxis which are
based upon the study of conditions of surface tension. In order to
consider these properly it will be useful to review briefly the funda-
mental principles governing these conditions.
The molecules of any fluid are held together by mutual attraction
due to the force generally spoken of as cohesion. This force is ex-
erted by like molecules upon each other in solids more strongly than
in liquids, and in gases less strongly. Since we are dealing in this
•connection with occurrences taking place in liquids, we will restrict
our consideration to these. The force of cohesion is influenced in
a number of ways. Thus, for instance, heat reduces it, and this
is the cause that solids are converted into liquids and liquids into
gases, provided of course that the heat brings about no chemical
change. In large masses of fluids the force of gravitation over-
•comes that of cohesion and larger masses of liquid assume the shape
•of the containing vessel. In smaller masses the force of cohesion
lends to bring about the spherical shape. This comes about in
the following way: Within the interior of a drop of liquid all the
molecules attract each other, and since the force of attraction is
equal in all directions it neutralizes itself, and the molecules are
41 Garrey. Am. Jour, of Phys., 3, 1900.
42 Loeb. Am. Jour, of Phys., 3, 1900.
PHAGOCYTOSIS 293
uninfluenced by it, mobile and free. The molecules on the surface
are in a different condition, however. They are subjected to the
force of cohesion from within, but not from without, and are there-
fore drawn strongly toward the center. The result is the same as
though the drop were subjected to pressure from without and the
surface layers were in a state of compression. There is in conse-
quence a constant tendency of all the surface molecules to be drawn
toward the center and a resulting tendency to a diminution of the
surface area. It is as though the surface of such a drop were a thin,
elastic membrane which tended to contract and diminish in size and
surface. The force with which this takes place is spoken of as sur-
face tension,43 and the energy underlying it is called, by Ostwald,
surface energy. Since a drop of one fluid suspended in another with
which it cannot mix is relieved of the disturbing factor of gravitation,
its surface tension has the effect of contracting the small mass into a
form which, for the given volume, will expose the smallest possible
surface, and this is, of course, the sphere. It is for this reason that, if
we shake up such systems as water and chloroform, or oil and water,
the chloroform or the oil will be distributed through the water as
small droplets. The degree of surface tension of any fluid is meas-
urable by a number of reasonably accurate methods which may be
found in any text-book of physics and which we need not consider
here. It is of course dependent in each case upon the nature of the
surrounding medium. We have taken into consideration above only
the force which is exerted within the drop by the cohesion, that is,
the attraction toward the center. This would be uninfluenced from
without only in a vacuum. In nature the surface molecules, though
forcibly drawn toward the center, are also affected from without by
the attraction exerted by the molecules of the substances surrounding
the drop. There is a constant balance, therefore, at any part of the
surface of a drop of fluid between the cohesion tension from within
and attractions from without. The resultant of the two forces de-
termines the surface tension, which will be greater or less in inverse
ratio to the attraction from without for any given drop, and a varia-
tion of the external attraction at different points on the periphery of
the drop will naturally influence the shape of the drop. For a relief
of attraction at one point would tend to permit that part of the sur-
face to retract, and an increase in this attraction would tend to
allow it to bulge, with the formation of a sort of pseudopod.
In studying the importance of surface tension 44 in determining
the motions of unicellular organisms a number of important attempts
have been made to imitate cell motion by means of the suspension of
various substances of strong cohesive properties in liquid media. The
43 Michaelis. "Dynamik der Oberflachen," Steinkopf, Dresden, 1909.
44 For a thorough discussion of this phenomenon see also Gideon Wells,
"Chemical Pathology/' Saunders, 1911.
294 .INFECTION AND RESISTANCE
idea was suggested by Quincke,45 and later by Biitschli,46 but has
been most expensively studied by Rhumbler.47 The result has been
the production of a number of "artificial amebse" which in almost
all respects behave like the living organisms. Thus if a small mass
of mercury is placed into a dish filled with water acidified with
nitric acid, and a small crystal of dichromate of potassium is dropped
near the mercury, the dichromate* will -dissolve and a yellow cloud
will gradually diffuse from it toward t*he mercury. As soon as the
yellow cloud touches this it will begin to show change of form and
to elongate in the direction of the dichromate, often moving to-
ward it. The motion of the quicksilver will resemble with con-
siderable accuracy that of an ameba moving toward a particle of
food or sending out pseudopodia. A more striking and complete
imitation is that obtained by Rhumbler when he placed a drop of
clove oil into a mixture of alcohol and glycerin. The changes of sur-
face tension produced upon the surface of the clove oil by the alcohol
give rise to movements in the oil entirely analogous takjthose of mo-
tile cells in favorable media. The similarity has been ^extended
even to the processes of engulfment of the food as observed among
amebse. Thus a drop of chloroform in water will flow about a
particle of shellac and dissolve it. If a piece of glass coated with
shellac is placed in contact with the drop it will engulf it,
but will cast out the glass after the shellac coating has been dissolved
away.
The similarity between phenomena purely referable to surface
tension and those taking place in the living cells is therefore very
striking and has been clearly analyzed in regard to its bearing
upon leukocytic chemotaxis by Gideon Wells in his "Chemical Path-
ology.77 The chemotactic substances, diffusing to the leukocyte, will
lower its surface tension on the side at which they come in contact.
Pseudopodia will be thrown out on this side in consequence, and the
leukocyte will move in this direction. The motion will be continued
in this direction as long as the concentration of the chemotactic sub-
stance, and therefore the diminution of surface tension is greater on
this side than on other parts of the periphery, until a point is reached
at which the chemotactic substance is equally diffused on all sides, and
motion will cease. The actual engulfment may then occur or the
nature and concentration of the chemotactic substance may be so
great that injury is done to the leukocyte. Whether or not the purely
physical explanation of chemotaxis tells the whole story it is of course
not possible to decide. At any rate, it furnishes a rational basis for
45 Quincke. Quoted from H. G. Wells, "Chemical Patholo£rv," Saunders.
1907.
46 Biitschli. "Untersuch. uber mikroskopische Schaume und das Proto-
plasma." Leipzig, 1892. See also H. G. Wells, loc. cit., pp. 220 et seq.
47 Rhumbler. Arch. f. Entwickelungs Mechanik, 1898.
PHAGOCYTOSIS 295
the study of the phenomenon more promising than any of the others
so far offered.
It is true, on the other hand, that such a theory in no way ac-
counts for the apparently selective positive chemotaxis which is
exerted by different substances. Thus the preponderance of poly-
nuclear leukocytes in foci and serous cavities containing organisms
like staphylococci, meningococci, streptococci, and others is in con-
trast to the lymphocytic accumulation in the pleural, subarachnoid',
and peritoneal spaces infected with tubercle bacilli. Some writers
have spoken, therefore, of active and passive leukocytosis according
to whether or not the cells attracted seemed to possess ameboid
motility. That surface tension phenomena alone do not account for
this is clear. But it must be remembered that even tubercle bacilli,
though eventually attracting few polynuclears and many lympho-
cytes, will cause an active polynuclear accumulation in the perito-
neum and pleura when first injected, and are actively phagocyted.
Later when the lesion is established and the bacilli are lodged in the
tissues the polynuclears give way to the lymphocytes, which even
then never accumulate in such proportion as do the microphages in
acute suppurative lesions. It may well be that the chemotaxis origi-
nally attracting the polymorphonuclear leukocytes is the same in
every case, but that a continued irritant, especially one well sur-
rounded by tissue elements as are the organisms within the tubercles,
may cease to exact any chemotactic influence, the accumulation of in-
active lymphocytes possibly being due to a progressive death of these
cells carried into the neighborhood of the lesion by the normal circu-
lation of the lymph.
CHAPTER XII
THE EELATION OF THE LEUKOCYTES AKD OF
PHAGOCYTOSIS TO IMMUNITY
IN MetchnikofFs earliest work upon the daphnia or water flea
he observed clearly that there was a direct relation between the de-
gree of phagocytosis and the outcome of the infection. When
phagocytosis of the invading yeasts was energetic and complete the
daphnia recovered. When the yeast cells penetrated the intestinal
wall of the daphnia in large numbers, and were enabled to multiply
before the phagocytic cells could accumulate in sufficient numbers to
engulf them, then the body of the daphnia was soon swamped with
the parasites and death ensued.
This simple observation fostered the thought that the basic prin-
ciple underlying all processes of immunity was represented in this
struggle between the invading bacteria and the phagocytic cells.
To the activity of the latter, entirely, he attributed the power of
resistance.
In support of this contention Metchnikoff and his pupils have
marshaled many facts, most of which are set forth in his classical
work "L'Immunite dans les Maladies Infectieuses." It will be
manifestly impossible here to do more than outline the plan of
study which these investigations have followed and the conclusions
to which they gave just foundation.
The original study upon the infectious disease of daphnia led to
analogous experiments upon higher animals and, by the prolonged
and patient investigations of Metchnikoff and his pupils, it was
shown that, throughout the field of infectious disease, there was a
striking parallelism between the resistance of the infected subject
and the degree of phagocytosis which occurred.
Earlier studies concern themselves chiefly with the natural im-
munity possessed by many animals against certain infection. The
infectious disease which at this time had been most thoroughly
studied was anthrax, and Koch had shown that frogs and other cold-
blooded animals were markedly resistant against this micro-organism.
Taking advantage of this observation, Metchnikoff studied the phago-
cytosis of anthrax bacilli in frogs and found that it took place rap-
idly and effectively, all of the injected bacilli being soon engulfed
by the accumulating cells. Similarly, active phagocytosis of anthrax
296
RELATION OF LEUKOCYTES TO IMMUNITY 297
bacilli was demonstrated in such naturally resistant animals as dogs
and chickens, while almost no cell ingestion occurred in delicately
susceptible animals like guinea pigs and rabbits. Rats, on the other
hand, more resistant to anthrax than guinea pigs, less so than dogs,
showed a degree of phagocytosis intermediate between that observed
in the cases of the other animals mentioned above. And yet, in these
more susceptible animals, the normal bactericidal action of the blood
upon anthrax bacilli, though never extreme, was often more marked
than that of the naturally immune animals mentioned above.
It is well known, for instance, that the serum of dogs possesses
almost no bactericidal properties for anthrax bacilli,1 although the
animals are highly resistant to this infection, while the serum of
rabbits is probably more strongly bactericidal for these bacilli than
the serum of most other animals, and yet rabbits are extremely
susceptible. That the lack of bactericidal powers of the serum is
not always a sign of susceptibility on the part of the animal was
shown as early as 1889 by Lubarsch. (We must remember, how-
ever, that lack of bactericidal power does not necessarily mean lack
of sensitizer. For bacteria may be sensitized without being killed
extracellularly as can be shown by the alexin-fixation reaction.)
The study of anthrax infections was a peculiarly fortunate
choice of subject, since in this bacillus resistance to serum lysis is
especially well marked and phagocytosis seems indeed to be the chief
mode of bacterial destruction. Studies analogous to those originally
made with anthrax, however, were subsequently carried on with
streptococci, pneumococci, and staphylococci chiefly by Bordet,2
Marchand,3 and others, and results coinciding with those of Metchni-
koff were obtained. In every case naturally resistant animals
showed marked phagocytosis, and susceptible ones failed to show it
to the same degree. It is a strong support of the same opinions, too,
that Marchand's studies, later extensively confirmed, showed that
the more virulent and invading strains of streptococci, the less active
is the phagocytosis — a converse, but equally conclusive, observation.
Further support for this point of view is manifold and cannot be
considered with anything like completeness. We may refer briefly,
however, to the experiments of Vaillard, Vincent, and Rouget 4 with
tetanus, and those of Leclainche and Vallee 5 with symptomatic
anthrax, because they are especially valuable in illustrating the
importance of phagocytosis in another class of infection. The pois-
ons of these micro-organisms are extremely toxic for rabbits, and if
1 Petterson. CentraM. f. Bakt., 1, 39, 1905.
2 Bordet. Ann. de I'Inst. Past., Vol. 11, 1897.
3 Marchand. Archiv. de med. Exp., Vol. 10, 1898.
4 Vaillard, Vincent, and Rouget, Ann. de I'Inst. Past., Vols. 5, 6, 1891-
1892.
5 Leclainche and Vallee. Ann. de I'Inst. Past., Vol. 14, 1900.
298 INFECTION AND RESISTANCE
a small amount of culture material, together with agar, broth, or
any foreign substance which may inhibit or divert phagocytosis from
the spores, is injected into these animals rapid proliferation and
death with toxemia result. If, on the other hand, the spores are
carefully washed of foreign material and toxin rapid phagocytosis
results and the animals recover.
The parallelism which was followed out so extensively between
natural immunity and phagocytosis was even more closely marked
in the case of artificially acquired immunity. The first observations
of this kind made by Metchnikoff, again on the subject of anthrax
infection, were carried out by the active immunization of rabbits.
The subcutaneous injection of virulent anthrax bacilli into normal
rabbits is usually followed by a rapid growth of the bacteria, with
much serous exudation but hardly any leukocytic accumulation. In
immunized animals, on the other hand, the bacilli are taken up by
hosts of phagocytes, just as this occurs in naturally resistant dogs or
other animals. Similarly Bordet 6 has shown that cholera spirilla
injected into the blood stream of cholera-immune animals are taken
up by leukocytes even before they can be subjected to lysis by the
circulating lytic antibodies.
It would add little to clearness were we to multiply the examples
in which it has been demonstrated that the acquisition of increased
resistance is accompanied by enhancement of the phagocytic process.
This statement may be regarded as an axiom, and indeed our later
discussions of the opsonins and bacteriotropins will show clearly why
such a state of affairs is to be expected. Taken by itself, however,
it does not necessarily prove that the destruction of the invading
germs is entirely due to the leukocytes. It might still be possible
that the bacteria are injured or even killed by the antibacterial
serum constituents before they can be taken up and carried away
by the cellular elements; the phagocytes then would act only as
scavengers for the removal of the dead bodies. Indeed, this opinion
was long held by a number of the adherents of the purely humoral
school. However, such a point of view is no Tonger tenable — espe-
cially in the light of the later opsonin studies just referred to.
Moreover, long before these more recent . studies it was clear that
bacteria may often grow within the leukocytes — finally destroying
these — and that they may even remain fully virulent after ingestion.
For, as Metchnikoff showed, if guinea pigs were injected with a
little of the exudate formed after the injection of anthrax bacilli
into immunized rabbits (an exudate in which there were no longer
any extracellular bacteria because of energetic phagocytosis) death
often resulted. It was clear, therefore, not only that the ingested
bacteria were still alive, but that they were, at least in part, still
fully virulent.
6 Bordet. Ann. de I'Inst. Past., Vol. 9, 1895.
RELATION OF LEUKOCYTES TO IMMUNITY 299
A further method of investigation employed by Metchnikoff in
Ms endeavors to prove his point consisted in the attempt to demon-
strate that virulent bacteria could be protected from destruction in
the bodies of resistant animals if the leukocytes could be held at bay.
This resulted in a number of ingenious experiments, the most con-
vincing of which is the one carried out with anthrax bacilli and
frogs by Trapetznikoff.7 Anthrax spores were inclosed in little sacks
of filter paper and these were introduced subcutaneously into frogs.
In consequence the spores, bathed in the tissue fluids, but protected
from phagocytosis, developed into the vegetative forms, multiplied,
and remained virulent for several days. Taken up by the frog's
phagocytes under ordinary conditions, they would rapidly have been
taken up, digested, and destroyed. Here again it was shown that the
body of fluids alone were unable to dispose of the bacteria and that
the natural resistance of the frog was due entirely to phagocytic
processes.
Other experiments have been aimed at a general reduction of
phagocytic activity by the use of narcotics. Thus, Cantacuzene 8
showed that animals treated with opium are very much more sus-
ceptible to infection than are normal controls. And since opium
markedly inhibits the activity of the white cells it may possibly be
that these experiments furnish a further support for Metchnikoff s
opinion. At any rate, it is worth noting that, even though these
experiments are not convincing in their assertion that the increased
susceptibility was due entirely to the interference with the leuko-
cytes, they indicate very definitely the inadvisability of using mor-
phin and similar narcotics in infectious diseases.
It is quite clear at any rate, then, that the process of phagocy-
tosis increases in energy as immunity is acquired and, so far, Metch-
nikoff' s assertions are entirely upheld by later knowledge. In his
contention that all properties upon which the resistance of the ani-
mal against infection depends center directly or indirectly in the
phagocyte, however, many subsequent amendments have been neces-
sary, which will become self-evident in the following discussions of
individual phases of the destruction of invading bacteria.
We cannot at the present time attempt to correlate these extreme
views of Metchnikoff with the equally extreme opinions of those in-
vestigators who formerly attributed immunity entirely tojjthe prop-
erties of the body fluids, assigning to the cellular activities a merely
secondary role. Many of the apparently opposed contentions have
become reconciled, and we now realize that neither process alone tells
the whole story, both being parts of the complicated correlated proc-
esses which together constitute the mechanism of resistance. It was
7 Trapetznikoff. Ann. de I'Inst. Past., 1891.
8 Cantacuzene. Ann. de I'Inst. Past., Vol. 12, 1898.
300 INFECTION AND RESISTANCE
indeed to the eager controversy between the two schools that we owe
much of the clearness of conception which recent years have given.
After the bacteria are taken up by phagocytes they undergo a
gradual disintegration or dissolution comparable to that by which
a particle of food is digested within the cell body of a rhizopod.
With the exception of such particularly insoluble micro-organisms as
the tubercle bacillus, the leprosy bacillus, blastomyces, and a few
others, there is in all cases an eventual complete resolution of the
bacterial body. As in amebse the digestion takes place often after
the formation of digestive vacuoles, and by staining at this time with
neutral red it may be demonstrated that the process goes on in a
weakly acid environment.
Metchnikoff naturally assumed, therefore, that the intracellular
digestion of bacteria by microphages (polynuclear leukocytes), or of
cellular elements, etc., by macrophages, was a process carried on
most probably by enzymes, and that these enzymes were identical
with the bactericidal bodies described as "alexin" and "sensitizer"
or "amboceptor" in the blood stream. To follow without confusion
the development of his ideas, however, it is necessary to bear in mind
that much of his earlier work was done at a time when the discovery
of the cooperation of two substances in bacteriolysis and hemolysis
(the amboceptor and the complement) had not yet been made by
Bordet, and when the bactericidal effect of normal serum was at-
tributed entirely to a single substance — the alexin of Buchner.
Buchner 9 himself had suggested that alexin was an enzyme-like
body probably derived from the leukocytes.
In his experiments Buchner had noticed that exudates, produced
by intrapleural injections of aleuronat in rabbits and dogs, possessed
a bactericidal value for Bacillus coli which exceeded the bactericidal
power of the blood serum itself. The influence of active phagocytosis
could be excluded by the fact that the leukocytes of the exudate had
been killed by repeated freezing and thawing. Similar results were
obtained by Hahn 10 with B. typhosus.
Denys and Kaisin,11 working along similar lines, found that the
pleural exudates of rabbits, obtained by the injection of dead staphy-
lococci and freed of cells by centrifugalization, were more highly
bactericidal for staphylococci than the blood serum of the same ani-
mals, but found also that the inactivated exudate could not be reacti-
vated by the addition of leukocytes. Denys offered as an explanation
for these phenomena that the living leukocytes in the original exudate
secreted alexin or complement which enhanced the bactericidal
activity of the exudate, that the leukocytes, subsequently added -to
9 Buchner. Munch, med. Woch., No. 24, 1894.
10 Hahn. Archiv f. Hyg., Vol. 25, 1895.
"Denys and Kaisin, Denys and Havet. La Cellule, Vol. 9, 1893; VoL
10, 1894.
RELATION OF LEUKOCYTES TO IMMUNITY 301
inactivated exudate, however, had lost vitality during the process of
isolation and washing, and no longer possessed secretory power.
Hankin,12 Kanthack and Hardy 13 had gone even farther than
this, and had attributed the production of alexin to the eosinophile
leukocytes particularly.
Metchnikoff,14 basing his opinion on his own studies, those of
his pupils, and many other investigations similar to those mentioned
above, came to the conclusion that, under ordinary conditions, the
normal blood contains no free bactericidal substances. He assumes
that these substances are entirely intracellular, being constituents of
the various phagocytic elements, by means of which the cells digest
the foreign elements they take up. He believes that there are essen-
tially two varieties of such digestive enzymes or "cytases" — just as
there are two varieties of phagocytes. The microphages, chiefly con-
cerned in the digestion of bacteria, secrete the bactericidal alexin, or,
as Metchnikoff calls it, "microcytase." The macrophages, the large
mononuclear lymph and endothelial cells, primarily concerned in the
phagocytosis of cellular elements (red cells, etc.), contain another
variety of digestive enzyme, the "macrocytase," or cytolytic (hemo-
lytic) alexin. The supposition that the hemolytic "cytase" is largely
derived from the macrophages was based particularly upon the in-
vestigations of MetchnikofFs pupil, Tarassewitch,15 who found that
the extracts obtained from lymph nodes, and other organs rich in
macrophages, possessed hemolytic properties. Both this work and
the preceding studies regarding the extraction of alexin from poly-
nuclear leukocytes will be more fully discussed below.
Maintaining that these cytases are purely intracellular under
ordinary conditions, Metchnikoif believes that, in normal animals,
the destruction of invading bacteria or of injected cellular substances
(blood cells, etc.) is accomplished entirely by the phagocytic process,
with subsequent intracellular digestion. In immunized animals,
however, there is present in the circulating blood another substance,
not identical with the cytases, but also derived from the leukocytes or
from the blood-forming organs — the afixateur" (Ehrlich's "ambo-
ceptor" — Borders "sensitizer"). This specific "fixateur" sensitizes
the bacteria or other antigens to the action of the cytases. For his
assumption regarding the origin of this sensitizer he finds support
largely in the researches of Pfeiffer and Marx, and others mentioned
in our section on the origin of antibodies, as well as in the simi-
lar investigations of Deutsch,16 carried on under MetchnikofFs per-
sonal supervision.
12Hankin. Centralbl f. Bakt., Vol. 12, 1892.
13 Kanthack and Hardy. Proc. Eoy. Soc., Vol. 52, 1892.
14 Metchnikoff. Ann. de VInst. Pasteur, Vol. 7, 1893; Vol. 8, 1894; Vol.
9, 1895.
15 Tarassewitch. Ann. de VInst. Past., Vol. 16, 1902.
16 Deutsch. Ann. de VInst. Pasteur, Vol. 13, 1899.
302 INFECTION AND RESISTANCE
Final digestion of the sensitized antigens (bacteria or blood
cells), however, can take place only under the influence of the
cytases intracellularly, unless by previous leukocytic injury these
enzymes have been liberated into the blood stream.
While it is admitted, then, that bacteria may be killed and di-
gested both intra- and extracellularly in the animal body, the cytases,
which accomplish this, are regarded as the same in both cases, being
derived from the phagocytic cells. In immunized animals "fixateur"
may be produced under the stress of active immunization and fur-
nished to the blood stream by the blood-forming organs. By this
substance bacteria and cells may be sensitized. However, the enzyme
by which digestion is actually accomplished, "cytase" or alexin, is
not present free in the blood even in immune animals unless it has
become free and extracellular by injury to the leukocytes.
How, then, on this basis does Metchnikoff account for the Pfeiffer
phenomenon, in which the extracellular destruction of bacteria takes
place rapidly in the peritoneal exudate ? His explanation is the fol-
lowing: When bacteria or other substances are injected into the
peritoneum there is a preliminary injury of leukocytes (phagolysis),
and by this alexin or cytase is liberated. When cholera spirilla, for
instance, are injected into the peritoneal cavity of an immunized
guinea pig there follows a short period during which few if any
leukocytes are present in the exudate, but many may be found gath-
ered in motionless clumps in the folds of the peritoneum and mesen-
tery, incapable of phagocytosis and apparently injured. If such
leukocytic injury can be avoided Metchnikoff claims that the extra-
cellular lysis of cholera spirilla will fail to take place. Thus if
sterile broth or salt solution is injected into the peritoneum of a
guinea pig a preliminary phagolysis will be followed by an accumu-
lation of leukocytes. If cholera spirilla are now introduced no
extracellular digestion is seen, but, instead of this, rapid phagocytosis
takes place. This he says is due to the fact that the freshly accumu-
lated, healthy phagocytes, collected in response to the preliminary
broth injection, are not easily injured and do not undergo phago-
lysis; no cytase is liberated and, in consequence, no serum bacterio-
lysis can take place. In the same way he claims that the hemolysis
of red blood cells (goose blood) in the peritoneum of specifically
immunized guinea pigs may be prevented if, by a previous injection
of broth, healthy leukocytes have been caused to accumulate. In
such a case the goose blood corpuscles are rapidly ingested by the
phagocytes and no hemolysis occurs.
It is self-evident that the validity of this interpretation of the
occurrences is strictly dependent upon the demonstration that the
circulating blood normally contains no alexin or complement. This
is rigidly maintained by MetchnikofF, and is indeed one of the most
important uncertainties of serology. He asserts that alexin appears
RELATION OF LEUKOCYTES TO IMMUNITY 303
in the blood serum only because changes in the leukocytes take place
during coagulation. It is not, by any means, settled that Metchni-
koff is right in this — in fact, more recent investigations seem to show
that he is wrong, and that we may assume definitely that alexin is
present in considerable amounts in the circulating blood plasma of
normal animals.
Metchnikoff s denial of this is based chiefly on the experiments
of Gengou. Gengou 17 took the blood from various animals into
paraffined tubes and centrifugalized it at low temperature before it
could clot. This freed the plasma from the cells before clotting,
though coagulation of course took place as soon as this plasma was
removed to tubes and kept at room temperature. The serum ex-
pressed from this clotted plasma he compared for alexin contents
(bactericidal properties) with that obtained from clotted whole
blood.
He found that, in all cases examined (dogs, rabbits, and rats),
the plasma contained practically no bactericidal substances, or at any
rate an incomparably smaller amount than was present in the serum
obtained from the clotted blood.
These experiments of Gengou would be conclusive in establishing
MetchnikofFs theory if they were confirmed by other observers.
This, however, has not been the case. Petterson 18 found no differ-
ence between the bactericidal properties of serum and oxalate plasma,
and Lambotte 19 arrived at similar results when he compared serum
with the coagulable plasma obtained by tying off a section of a vein
and centrifugalizing the blood without opening the vessel. Hew-
lett,20 Falloise,21 Schneider,22 and more recently Dick 23 and Addis,24
whose work has been done with particular attention to technical
accuracy, fail to confirm Gengou, finding no appreciable difference
between plasma and serum.
In favor of Gengou's results are the investigations of Herman 25
and the more recent ones of Gurd.26 Further supporting Gengou's
conclusion is the observation recorded by a number of workers (Wal-
ker,27 Longcope,28 and others) that the complement or alexin con-
tents of serum will increase somewhat as the serum is allowed to
17 Gengou. Ann. de I'Inst. Past., Vol. 15, 1901.
18 Petterson. Arch. f. Hyg., Vol. 43, 1902.
19 Lambotte. Centralbl f. Bdkt., Vol. 34, 1903.
20 Hewlett. Zeitschr. f. Heilkunde, 24, 1903.
21 Falloise. Bull, de I'Acad. Eoy. de Med,, 1905.
22 Schneider. Arch. f. Hyg., 65, 1908.
23 Dick. Jour. Inf. Dis., Vol. 12, 1913.
24 Addis. Jour. Inf. Dis., Vol. 10, 1912.
25 Herman. Bull, de I'Acad. Roy. de Med., 1904.
26 Gurd. Jour. Inf. Dis., Vol. 11, 1912.
27 Walker. Jour, of Hyg., 3, 1903.
28 Longcope. Med. Bull. Univ. of Pa., 1902, Vol. 15, p. 331,
INFECTION AND RESISTANCE
stand on the clot. This observation, too, has been rendered incon-
clusive by contrary reports from other investigators. Longcope,29
further, has found that alexin was more plentiful in the blood of
individuals suffering from leukemia — in which of course a larger
percentage of leukocytes is present in the circulation. This, too,
has been contradicted by other workers, but even if upheld would not
influence the possibility of there being alexin in the normal circula-
tion. On the whole Gengou's contentions with their consequent
bearing upon Metchnikoff's theory cannot be accepted as final. In
fact, the greater part of available experimental evidence seems to
point to the actual presence of alexin in the normal circulating blood.
This seems also indicated by the unquestionable fact that active
phagocytosis may take place in the circulation of an animal and, as
we shall see below, free alexin is probably necessary (as opsonin) in
this process. Further evidence in this direction also is furnished by
the immediate anaphylactic shock which follows the injection of
antigen into the blood stream of a sensitized animal, a process in
which we have much reason to believe that alexin takes an active
part. However, the problem is a difficult one, and, while we
favor the opinion that free alexin is present in the intravascular
blood, we must admit that a crucial experiment has not yet been
formulated.
Now, as regards the apparent extraction of alexin from leuko-
cytes and lymphatic organs, we have already called attention to the
fact that most of the researches associating these cells with the bac-
tericidal substances were carried out before the dual mechanism of
sensitizer and alexin in bacteriolysis had been fully worked out. In
consequence conclusions were formulated from the mere facts of the
presence of bactericidal or hemolytic properties in cell-extracts with-
out the further determination of heat stability or the possibility of
reactivation. Most of the earlier work also was done without suffi-
cient attention to the separation of the cells and the serum of leuko-
cytic exudates. The first one to do this carefully was Hahn,30 who,
like his predecessors, concluded that the bactericidal leukocytic sub-
stances, undoubtedly encountered by him, were identical with alexin.
Doubt was first cast upon this by Schattenfroh,31 who worked with
leukocytes suspended and extracted in physiological salt solution.
He found that bactericidal substances were, indeed, obtained in this
way, but that, unlike alexin, these substances were thermostable,
withstanding exposure to a temperature of 56° C. and destroyed only
by exposure to temperatures as high as 75° C. to 80° C. for thirty
minutes.
29 Longcope. Jour, of Hyg., Vol. 3.
30 Hahn. Arch. f. Hyg., Vol. 25.
31 Schattenfroh. Arch. f. Hyg., Vols. 31 and 35; 1897.
RELATION OF LEUKOCYTES TO IMMUNITY 305
Moxter,32 a little later, working with cholera spirilla, also came
to the conclusion that the leukocytic bactericidal substances were not
identical with those found in the blood serum. Petterson,33 too, made
thorough investigations into the nature of the bactericidal substances
extracted from the leukocytes, and, working chiefly with B. proteus
and B. anthracis, found such substances in the leukocytes of dogs,
rabbits, and guinea pigs active against the bacteria mentioned above,
but failed to find them active, at least in guinea pig and cat leuko-
cytes, against B. typhosus or the cholera spirillum. He expresses the
opinion that bactericidal leukocytic substances are normally given
up to the blood in minute quantity only or not at all, and that these
substances hold no definite relationship to the bactericidal sub-
stances found in blood serum. In a later investigation he showed
that the "endolysins," as he now calls the leukocytic bactericidal
substances, may, like many enzymes and serum bacteriolysins, be
precipitated out of solution with alcohol and ether ; but he separates
them absolutely from serum lysins and complement. The latter,
while they may be, in part at least, secreted by the leukocytes, are,
according to Petterson, induced easily to come out of the cells during
life by slight injury or other stimulation, while the endolysins them-
selves are abstracted from the cells only after extensive extraction or
maceration.
Schneider 34 has come to similar conclusions and speaks of the
endocellular bactericidal substances as "Leukine." In a recent in-
vestigation of the same subject the writer 35 has in all essentials con-
firmed Schattenfroh's original conclusions regarding the heat sta-
bility of the extracted leukocytic bactericidal substances, and has
shown that after inactivation by heat these substances are not reacti-
vable by the addition of fresh leukocytic extracts, and that the yield
obtained from the leukocytes of immunized animals is not greater
than that obtained from normal leukocytes.
It appears, therefore, that, contrary to MetchnikofFs first sup-
position, the enzymes which bring about the digestion of phagocyted
bacteria within the cell are not identical with those which bring
about a similar extracellular digestion. In addition to the demon-
stration of a definitely different structure possessed by the bacteri-
cidal leukocytic extracts, as evidenced by their heat stability, we have
the negative evidence that neither true alexin nor sensitizers have
ever been successfully extracted from such cells.
It is still possible that this may eventually be done — but, al-
though indirect evidence like that of Denys, Longcope's observations
in leukemia, and the occasional increase of the alexic, powers of serum
32Moxter. Deutsche med. WocJi., 1899, p. 687.
33 Petterson. Centralbl f. Bakt., i, 39, 1905; 46, 1908.
34 Schneider. Archiv f. Hyg., Vol. 70, 1909.
35 Zinsser. Jour. Med. Res., Vol. 22, 1910.
306 INFECTION AND RESISTANCE
after standing on the clot points to a probability of this, no direct
evidence has so far been satisfactorily produced. In the hope that
the leukocytes would give up alexin — possibly as a secretion as sug-
gested by Denys — the writer, with Cary, some years ago kept
washed leukocytes at 37.5° C. in Locke's solution, but was unable
to find any evidence of alexin production within 48 hours.
The apparent extraction of hemolysin from macrophages by
Tarassewith, moreover, has met with a similar refutation. Korschun
and Morgenroth 36 have shown that these hemolytic extracts are ex-
tremely heat resistant, are alcohol and ether soluble, and do not act
as antigens. They are quite different from the serum hemolysins,
therefore, and probably closely related to the hemolytic lipoidal
substances described by Noguchi and others.
Further strong arguments against the assumption of the pres-
ence of hemolytic alexin in the body of the macrophages have been
advanced by Gruber 37 and by Neufeld.38 Gruber showed that no
extracellular hemolysis takes place when leukocytes are brought
together with sensitized red blood cells, and Neufeld showed that
even after the phagocytosis of such sensitized cells the hemolysis is
very much slower, and of a different character from extracellular
serum hemolysis. In the intracellular digestion there are no mere
solution of the hemoglobin and formation of a shadow form
(stroma), but there occur a gradual degeneration with the forma-
tion of a granular detritus of hemoglobin.
It is probable, then, that the digestion of bacteria and cells
within the phagocytes is carried out by substances not identical with
those taking part in serum lysis. It is not unlikely that the intra-
cellular process is a quite complicated one, not depending on a single
enzyme.
In addition to the bactericidal substances extracted from leuko-
cytes a number of true enzymes have indeed been obtained by various
investigators. We have mentioned in another place that one of the
earliest observations in this respect was that of Leber,39 who noticed
that sterile pus could liquefy gelatin. It may be commonly observed
in paraffin or celloidin sections of staphylococcus abscesses that a
ring of apparently digested or degenerating tissue is formed about
an accumulation of leukocytes — in foci in which the bacteria may
be too sparse to be held accountable for the changes. These leuko-
proteases have subsequently been carefully studied by Opie,40 Joch-
mann and Miiller,41 and a number of others.
36 Korschun and Morgenroth. Berl klin. Woch., 1902.
37 Gruber. Quoted from Sachs, in "Kraus u. Levadiii Handbuch," Vol. 2,
p. 991.
38 Neufeld. . Arb. a. d. kais. Gesundh. Ami., Vol. 28, 1908.
39 Leber. "Die Entstehung der Entziindung," Leipzig, 1891.
40 Opie. Jour. Exp. Med., Vol. 7, 1905; Vol. 8, 1906; Vol. 9, 1907.
41 Midler and Jochmann. Munch, med. Woch., Nos. 29 and 3~L, 1906.
RELATION OF LEUKOCYTES TO IMMUNITY 307
Opie found that two distinct proteolytic enzymes could be ex-
tracted from the cells of exudates obtained by turpentine injections.
One — peculiar to the polynuclear leukocyte, and similar to one pre-
viously described by Miiller 42 — acts in a weakly alkaline medium.
The other, present particularly in exudates containing a predominat-
ing number of mononuclear cells, acts in a weakly acid reaction.
Tschernorutski also found proteolytic ferments in both micro- and
macrophages, but found no lipase in the polynuclear extracts. This
seems particularly interesting in view of the great resistance to
intracellular digestion noticed in acid-fast bacteria, a point of some
importance in connection with the destruction in the body of such
micro-organisms as the bacilli of tuberculosis and leprosy.43 Joch-
mann 44 states that the action of the leukoprotease, which acts in an
alkaline medium upon casein, results in the formation of tryptophan
and ammonia, and believes it to be functionally very similar to
trypsin. It is interesting to note that the most active protease is
obtained from pus as it forms about acute infections or other stimuli
which lead to the accumulation of polynuclear leukocytes, whereas
it is apparently completely absent from tuberculous pus.
The question immediately arises, are these leukoproteases identi-
cal with the bactericidal substances extracted from leukocytes as de-
scribed above ? For it might well be that bacterial death resulted
merely from the digestive action of the enzyme upon the bacterial
protein. Jochmann,45 who has approached this problem experi-
mentally, has answered it in the negative. By repeated alcohol
precipitation of glycerin extracts of leukocytes he obtained an en-
zyme preparation which possessed absolutely no bactericidal prop-
erties, though it was still actively proteolytic.
Not only did this relatively pure ferment possess no bactericidal
action, but bacteria actively proliferated when suspended in it. Joch-
mann believes that living bacteria are not amenable to the enzyme
possibly because of their possession of an "antiferment," at least
this would follow in some cases from the experiments of Kantoro-
wicz.46
The leukoproteases, therefore, appear to possess no direct sig-
nificance in bacterial immunity. Their function seems rather to
lie in the resorption of dead tissues, fibrin, blood clots, etc. Friedrich
Miiller 47 has pointed out their possible importance in the rapid de-
struction and liquefaction of the massive fibrinous exudates remain-
ing after the crisis in lobar pneumonia.
42 Miiller. Congr. f. inn. Med., Wiesbaden, 1902.
43 Zinsser and Gary. Jour. A. M. A., 1911.
44 Jochmann. Leucozyten Fermente, etc., "Kolle u. Wassermann Hand-
buch," 2d Ed., Vol. 49, 2.
45 Jochmann. Zeitschr. f. Hyg., 61, 1908.
46 Kantorowicz. Munch, med, Woch., No. 28, 1909.
47 Friedrich Miiller. "Verhand. d. Kongr. f. inn. Med.," 1902.
308 INFECTION AND RESISTANCE
From the discovery of antibacterial properties in the extracts
of leukocytes it is but a logical step to the attempt to utilize these
substances therapeutically. Petterson 48 was probably the first to
study this phase of the problem systematically in connection with
anthrax infection in dogs and rabbits. In preliminary studies he
claimed to have determined that when leukocytes are left in con-
tact with serum for four hours or longer there develops in the mix-
ture a bactericidal power far superior to that which is possessed
by these elements when separately kept in salt solution and mixed
only just before the bactericidal tests. He attributes this to the
fact that in dogs, at least, the leukocytes furnish bactericidal sub-
stances to the serum — an assumption which is entirely in accord
with the earlier opinion of Denys and Kaisin,49 which we have men-
tioned in another place. In direct continuance of these experiments
he injected leukocytes into dogs at the same time at which he in-
fected them with anthrax and observed a moderately protective in-
fluence, which, however, he admits was not very great. He followed
this work in 1906 with similar observations on the protective influ-
ence of leukocytes in intraperitoneal infections of guinea pigs with
typhoid bacilli. In these experiments 50 he made the curious ob-
servation that, although such protective influence was unquestionable,
the guinea pig leukocytes contained no bactericidal substances active
against typhoid bacilli. In consequence he concluded that the de-
struction of these bacteria in the guinea pig was due entirely to the
serum-antibodies absorbed by the microorganisms before phago-
cytosis, even when the actual destructive process took place intra-
cellularly. The protective effect following on the injection of the
leukocytes he attributed to an indirect influence of the leukocytic
substances in stimulating the more rapid accumulation of alexin or
complement in the peritoneum, with consequently more powerful
phagocytosis. Following this, in 1908, Opie 51 carried out experi-
ments in which he observed that leukocytes injected intrapleurally
into dogs, together with tubercle bacilli, exerted a distinct protec-
tion.
In the same year extensive observations on the protective prop-
erties of leukocyte extracts were published by Hiss.
Hiss52 worked at first with extracts of dog, rabbit, and guinea
pig leukocytes; later he confined himself entirely to rabbit leuko-
cytes. He extracted the leukocytes at first by repeated freezing and
thawing in physiological salt solution, but the technique of hi» sub-
sequent work was uniformly as follows: Intrapleural injections of
48 Petterson. Centralbl f. Bakt., Vol. 36, 1904.
49 Denys and Kaisin. "La Cellule," Vol. 9, 1893.
50 Petterson. Centralbl f. Bakt., Vols. 40 and 42, 1906.
51 Opie. Jour. Exp. Med., 1908.
52 Hiss. Jour, of Med. Res., new series, Vol. 14, 1908.
RELATION OF LEUKOCYTES TO IMMUNITY 309
aleuronat emulsions were made in rabbits and, after about 24 hours,
the resulting exudates were taken away with sterile pipettes and
centrifugalized before clotting could take place; the serum was de-
canted and the leucocytes then emulsified in distilled water, in quan-
tity about equal to the amount of serum poured off. In this the leu-
cocytes were allowed to stand for a few hours at incubator tempera-
ture, and then in the ice-box until used. For his experimental work
in both animals and man, in most instances, not only the clear super-
natant fluid was injected, but the cell residue as well.
With leucocytic extracts so prepared Hiss treated staphylococ-
cus, typhoid bacillus, pneumococcus, streptococcus, and meningococ-
cus infections in- rabbits and obtained results which justified him in
concluding that the leucocyte extract exerted strong protective action
in all of these cases. Many of his animals survived infections fatal
to controls even when the treatment was delayed as long as 24 hours
after infection. Subsequently Hiss and Zinsser 53 treated series of
patients, ill with pneumonia, meningitis, and staphylococcus infec-
tions, with leucocyte extracts prepared by the method of Hiss, and
felt that they were justified in concluding that in many cases, at
least, the course of the disease was favorably influenced by the leuco-
cyte extract. Favorable results have since then been obtained also by
Lambert in erysipelas, and by Hiss and Dwyer in a variety of con-
ditions.
While there seems to be little question about the actually favor-
able influence of the leucocyte extract, both in experiments with
animals and in the treatment of human cases, there has been consid- •
erable difficulty in determining the reasons for this influence. In .
subsequent studies Hiss and Zinsser (loc. cit.) were able to show that
/-"the extracts did not favor phagocytosis and that the moderate bacteri-
cidal properties possessed by the leucocytic substances could not ac- »
count for their effectiveness. There did seem to be a more rapid
v accumulation of phagocytes in the peritoneal cavities of guinea pigs
infected with cholera spirilla when leucocyte extract was injected .
with the bacteria, and it is not impossible, in fact, it seems probable
to the writer, from subsequent experience, that the protective prop- «
erties of the leucocyte extracts are attributable, in part at least, to «
their positively chemotactic effect.
We are inclined to believe at present that the beneficial effects of leucocyte
extracts are based on the same principles as those which determine the reactions
following on the injection of bacterial and any other protein.
In this connection a very interesting problem has arisen — namely, that spoken
of as the phenomenon of Specific Hyperleucocytosis. Bordet, as early as 1896,
made the following statement, "Active immunity has also other characteristics, in
that there is an increase of the number of leucocytes, that is, an 'exaltation' of
the chemotactic sensibilities of the leucocytes." He suggests herein that an im-
^ munized animal may respond with a more powerful leucocytic reaction to the in-
" jection of the infectious agent than would a normal animal similarly treated. This
53 Hiss and Zinsser. Jour, of Med. Res., new series, Vol. 14, 1908.
310 INFECTION AND RESISTANCE
idea has recently found experimental elaboration in the work of Gay and Claypole,
who found that the reinjection of immune animals with the homologous bacteria
produced a specific hyperleucocytosis, that is, typhoid immune animals receiving
typhoid bacilli would respond with counts ranging up to 150,000 leucocytes per
cu. mm., whereas the normal animals rarely showed more than 40,000 to 50,000.
This observation would tend to indicate a great advantage of the specific over
the non-specific methods of treatment.
Unfortunately the results of Gay and Claypole M have not found confirmation.
McWilliams55 in similar experiments found no differences in the degree of leuco-
cytosis between normal and immune animals in response to the injection of bacteria
and reported that the same degree of response followed in typhoid immune animals
when injected with Bacillus coli as when typhoid bacilli were administered. In
part, this is also stated to be the experience of Jobling and Petersen.
The question is such an important one that the writer, with Dr. Tsen,56 re-
investigated it in connection with work on the therapeutic effect of leucocytic ex-
tract. Our conclusions showed little agreement with those of Gay and Claypole.
We found that when homologous Gram-negative bacilli are injected into im-
munized animals there seems to be a definitely higher leucocytosis in the immu-
nized animals than in normal controls similarly treated. The contrasts in our
experiments, however, were nothing like as striking as those reported by Gay and
Claypole. Indeed the contrast in general is so slight and so irregular that in the
case of the Gram-negative bacilli we were at first inclined to agree with McWil-
liams. There was, however, a sufficiently definite difference in an average of many
counts to convince us that this was more than coincidence.
In the case of the Gram-positive cocci there was a more marked difference, in
that the immunized animals reacted more promptly and very much more ener-
getically than did the normal animals.
It seems reasonably clear, then, that an animal reacts more energetically as
far as its mobilization of leucocytes is concerned when reinjected than does a
normal animal treated with the same variety and quantity of bacteria.
The reaction is dependent upon a number of factors, chief among which are:
(1) the condition of the animal (loss of weight, etc.) ; (2) the amount of bacteria
injected, and (3) the interval between injections. These factors all very naturally
signify the necessity of avoiding too profound an intoxication of the animal.
When immune animals are treated with heterologous bacteria — that is, when
prodigiosus bacilli or colon bacilli are injected into typhoid immune animals and
vice versa — there seems to be no specific difference in response. That is, the in-
jection of colon bacilli into typhoid animals has shown no marked difference in
leucocytic response from that observed when typhoid bacilli were injected into a
typhoid animal. In this our figures correspond with those of McWilliams. They
are also in keeping with the clinical experience of Kraus, Ichikawa, and others
which have been mentioned above.
The injection of leucocytic extract does not arouse as vigorous a leucocytie
Tesponse as does the injection of bacillary protein.
In reading these facts superficially they at first seem to be contradictory in
significance, inasmuch as specificity seems to exist in the fact that typhoid
immune rabbits or streptococcus immune rabbits respond somewhat more vigor-
ously than do normal controls injected with the same substance.
However, we think that these relations are explained by the fact that animals
that have reacted to such organisms as the typhoid bacillus, etc., develop a certain
amount of non-specific tolerance against the proteose-like substances which are
probably responsible for a not unimportant part of the symptoms caused by the
bacteria. Such tolerance has been shown in the experiments made by the writer
with Dwyer.
To summarize, therefore, we do not think that at present a specific leucocytosia
In the sense of Gay and Claypole has been demonstrated, but believe that an ani-
mal, immune to one microorganism, will have a slight non-specific increase of
resistance to other organisms.
This does not mean that the immunity is non-specific. The destruction of
living bacteria is still a purely specific process, and this, of course, would
determine the occurrence of outcome of an infection.
5* Gay and Claypole. Arch, of Int. Med., V, XIV, 1914, p. 662.
55 McWilliams. Journal of Immunology, Vol. I, No. 2, 1916, p. 159.
56 Zinsser and Tsen. Journal of Immunology } Vol. II, No. 3, 1917, p. 247.
CHAPTER XIII
FACTOID DETEKMINING PHAGOCYTOSIS
OPSONINS, TBOPINS
FROM the very beginnings of his researches upon phagocytosis
Metchnikoff recognized that the process was profoundly influenced
by the properties of the fluid constituents of the blood plasma in
which the phenomenon occurred. Both he and his pupil Bordet,1
at this time working in MetchnikofFs laboratory, noticed that the
phagocytic activity of leukocytes was greater in immune than in
normal sera and associated this with the specific properties of the
immune substances or antibodies in these sera ; Metchnikoff himself
interpreted the phagocytosis-enhancing power of the serum as a
stimulation of the leukocytes and referred to the serum constituents
by which this effect was produced as "stimulins." A closer analysis
of the factors involved in this interrelationship, however, was not
attempted at this time by him or his pupils, although indirect ref-
erence was made to it in a number of articles emanating from this
school in the course of investigations on kindred problems of phago-
cytosis. Thus Gabritschewsky,2 in 1894, published a paper on
"Leukocytose dans la Diphtheric/7 in which he concluded that the
poison of diphtheria bacilli, among other harmful effects, diminished
the phagocytic power of the leukocytes, and that one of the beneficial
influences of the curative serum was to render these and other cells
""less sensitive to the bacterial poisons." This may be interpreted
as indicating an assumption that the action of an immune serum in
increasing phagocytic activity rested rather upon its influence upon
the bacterial products than upon any stimulation of the phagocytes
themselves. However, in diphtheria the action of the leukocytes
was, even at this time, recognized as a merely secondary one, and
Gabritschewsky's results did not materially influence the "stimulin"
conception.
The first extensive investigation which occupied itself directly
with these problems was that of the Belgian bacteriologists Denys
and Leclef.3 The publication of these workers deals primarily with
1 Bordet. Ann. de I'Inst. Past., 1895.
2 Gabritschewsky. Ann. de I'Inst. Past., 1894.
3 Denys and Leclef. La Cellule, 11, 1895.
312 INFECTION AND RESISTANCE
the nature of streptococcus immunity in rabbits. It established, first
of all, the paramount importance of phagocytosis in the resistance
of animals against these bacteria, and made clear that the destruction
of bacteria was carried out equally as well by the leukocytes of nor-
mal as by those of immune animals, but was powerfully enhanced
when either normal or "immune" leukocytes were combined with
immune serum. Their work, therefore, indicated again that the in-
creased phagocytosis of virulent bacteria, taking place in immune
animals, depended clearly upon alterations in the functions of the
serum rather than in those of the cells, and they suggested that
the influence of this serum was not necessarily one of leukocyte
stimulation, but might rather consist in action upon the bacteria,
rendering them less resistant to phagocytosis. They say in sub-
stance: "A notre avis, on pourrait tout aussi bien admettre que la
substance vaccinante ou antitoxique agit, non pas sur le leukocyte,
mais sur un poison renferme dans le corps du microbe ou dissous
dans le milieu, et qui preserve le micro-organisme centre les at-
teintes du leukocyte." 4
In this statement we have, in brief, the distinct formulation of
our present view of the "dpsonins." 5
Observations with pneumococci and streptococci carried out
after this by Marchand 6 and by Mennes,7 whose investigations we
cannot discuss in detail, beside confirming most of the observations
of Denys and Leclef, brought out especially the relation of the
virulence of micro-organisms to phagocytosis, showing that very
virulent strains were taken up to a slight degree only in the presence
of normal serum, but were subject to active phagocytosis when im-
mune serum was employed. This, too, seemed to point primarily to
the fact that the serum influenced rather the bacteria than the
phagocytes, although no convincing proof is brought for this in their
publications. Though much that had bearing indirectly on this
problem was written during the following years, no definite progress
was made beyond the results of Denys and his pupils until 1902,
* In our opinion one can just as well believe that the vaccinating or anti-
toxic substance acts not upon the leukocyte but upon a poison inclosed within
the body of the bacteria or dissolved in the medium, which preserves the
micro-organism against the attacks of the leukocyte.
5 Denys formulated this view with still greater clearness and positiveness
at the Congress of Hygiene held at Brussels in 1903. We take our citation
from the discussion on opsonins by Gruber (3d meeting Freie Vereinigung f.
Mikrobiol., Vienna, 1909, Centralbl. /. Bakt., I Ref., Vol. 44, Suppl. p. 3).
Following is Denys' statement : 1. The phagocytosis in immune sera is de-
pendent upon substances which are precipitated with the euglobulins.
2. These substances cause phagocytosis by inciting a physical alteration of
the micro-organisms. 3. These substances are specific.
6 Marchand. Arch, de Med. Exp., 1898.
7 Mennes. Zeitschr. f. Hyg., Vol. 25.
FACTORS DETERMINING PHAGOCYTOSIS
when Leischman 8 introduced a technique by means of which it be-
came possible to observe the process of phagocytosis with fresh serum
and leukocytes in vitro.
By utilizing this technique and improving upon it Wright and
Douglas in the following year (1903) evolved a method by means of
which phagocytic activity could be quantitatively measured with
reasonable accuracy. They worked at first with staphylococcus
phagocytosis by human leukocytes in the presence of human citrate
plasma, a research undertaken primarily because Wright,9 in collabo-
ration with Windsor, had previously determined that human blood
serum possessed practically no bactericidal power for this organism,
and that phagocytosis was probably the chief mechanism of protection
which the human body possessed against these bacteria. The re-
searches of Wright and Douglas 10 were carried out chiefly by mixing
equal volumes of bacteria, serum, and leukocytes (in citrate sus-
pension),11 allowing these elements to remain together at 37.5° C.
for varying periods, then staining on slides and determining the
degree of phagocytosis by counting the numbers of bacteria taken up
by each polynuclear leukocyte. Though many technical difficulties
had to be overcome, and although the method at its best still permits
of much personal error, careful work and untiring repetition made
possible a considerable degree of accuracy, and definite facts regard-
ing the mechanism of phagocytosis, heretofore merely suspected,,
could be recorded. The most important result of these investiga-
tions was the unquestionable establishment of the function of serum
in the process of phagocytosis, namely, that it in no way "stimu-
lated" the leukocytes in the sense of Metchnikoff, but rather acted
entirely upon the bacteria, preparing them for ingestion. For this
reason Wright coined the word "opsonins" (o^oi/«o= I prepare food)
for the serum constituents which brought about this effect, believing
them to be new antibodies, entirely distinct from the other serum
antibodies heretofore recognized.
Wright and his followers now concluded that the role of the
leukocyte in taking up bacteria was entirely dependent upon the
opsonin contents of the serum. In a menstruum containing no
serum, or in a serum in which the opsonins had been destroyed by
heat, they found practically no phagocytic action on the part of
washed serum-free leukocytes, and they, therefore, doubted the oc-
currence of spontaneous phagocytosis on the part of leukocytes
themselves.
8 Leischman. Brit. Med. Jour., Vol. 2, 1901, and Vol. 1, 1902.
9 Wright and Windsor. Jour, of Hyg., Vol. 2, 1902.
10 Wright and Douglas. Proc. Roy. Soc., 72, 1903, 73 and 74, 1904.
See also Wright, "Studien liber Immunisierung," Fischer, Jena, 1909.
11 At first bacteria were merely mixed in equal volumes with citrated
whole blood.
INFECTION AND RESISTANCE
In this point it is not unlikely that Wright is mistaken, since
•other observers, notably Lohlein,12 have observed the phagocytosis of
various bacteria by washed leukocytes in indifferent, opsonin-free
media. Although we may take it as assured that such spontaneous
phagocytosis may take place (Metchnikoff and a number of others
having obtained results similar to those of Lohlein), this is prob-
ably never very intense.
In fact, Wright, in some of his later work, does not insist rigidly
upon the non-occurrence of spontaneous phagocytosis, but attempts
to associate such phenomena with the salt contents of the medium
in which it occurs. Together with Reid,13 he determined that spon-
taneous phagocytosis of tubercle bacilli unquestionably takes place,
is most intense at a concentration of about 0.6 per cent. JSTaCl, and
diminishes as the concentration is increased. This, as we shall see,
has bearing on the possible physical explanations advanced to ac-
count for opsonic action, and has its parallels in experiments on the
influence of electrolytes on agglutination and precipitation.
The fact remains that Wright demonstrated by his work that
Metchnikoff's original view, which interpreted the difference be-
tween susceptibility and immunity as a difference between the in-
herent phagocytic powers of the leukocytes, is incorrect, and that
the essential regulating influence affecting phagocytosis rests upon
the action of the serum upon the bacteria.
The following experiment from the work of Hektoen and Rue-
diger -14 illustrates this point with exceptional clearness. It shows
that human leukocytes in the presence of normal defibrinated blood
will take up bacteria energetically. When the leukocytes, however,
are washed free of blood and added to untreated bacteria phago-
cytosis is practically nil. If, however, such washed leukocytes are
mixed with bacteria that have been previously in contact with serum
active phagocytosis will take place. In other words, the bacteria
have been altered by the serum in such a way that they are now
amenable to phagocytosis by washed leukocytes. The serum then
acts upon the bacteria and not upon the leukocytes.
TABLE II
Phagocytosis by Human Leukocytes of Sensitized Bacteria
Average
Phagocytosis
Human leukocytes (defibrinated blood) + Staphylococcus aureus 22 .
Human leukocytes (washed in NaCl solution) + Staphylococcus aureus. 1 . 2
Human leukocytes (washed in NaCl solution) -f- Staphylococcus aureus
(treated with human serum) 10 .
Human leukocytes (defibrinated blood) + Streptococcus 300 22 .
12 Lohlein. Centralbl. f. Bakt., 38, 1906, Beihef t, p. 32 ; also Munch, med.
Woch., 1907, p. 1473.
is Wright and Reid. Proc. of Eoyal Soc. B., Vol. 77, 1906.
14 Hektoen and Ruediger. Jour. Inf. Dis., Vol. 2, 1905, p. 132.
FACTORS DETERMINING PHAGOCYTOSIS 315
Average
Phagocytosis
Human leukocytes (washed in NaCl solution) -f- Streptococcus 300 .... 1 .
Human leukocytes (washed hi NaCl solution) + Streptococcus (treated
with human serum) 14 .
Human leukocytes (washed in NaCl solution) + Streptococcus (treated
with guinea pig serum) 12 .
Human leukocytes (washed in NaCl solution) -f- Streptococcus (treated
with rabbit serum) 14 .
Wright and Douglas' 15 work was done at first with normal
serum or normal citrate plasma, and in this case they found that the
opsonins were essentially unstable, being easily weakened by ex-
posure to light, or heat, and even when preserved in sealed tubes in
the dark they diminished noticeably on standing for 5 or 6 days.
Other writers who have worked with the opsonic substances in nor-
mal serum have confirmed this instability of the normal opsonin,
although even Wright himself admits that heating to 60° C. does
not entirely destroy the opsonic power, though it reduces it to a
minimum. A protocol from Wright and Douglas' first paper will
best illustrate the degree of reduction of opsonic power resulting
from the exposure of normal serum to 60-65° C. for 10 to 15 minutes.
A. Unheated serum Wright — Staphylococcus suspension 1 vol. — Blood cells
Wright 3 vols.
(1) Phagocytic average 20 cells 17 .4
(2) Phagocytic average 20 cells 19.8
B. Heated serum as above.
(1) Phagocytic average 52 cells 0.6
(2) Phagocytic average 46 cells 3.4
The experiments just cited refer only to the opsonic powers of
normal serum. When an animal is immunized with any particular
micro-organism or other cellular antigen, such as red blood cells,
etc., a marked specific increase of opsonins occurs, but unlike the
opsonins of normal serum these newly formed elements in the im-
mune serum seem to possess a much greater resistance to heat.
Neufeld and Rimpau,16 who have studied these constituents of
immune serum with especial thoroughness, have shown that heating
to 62° to 63° C. for as long as three-quarters of an hour does not
destroy them, and that such sera may be preserved for as long as
several years without their complete disappearance.17
We may accept as definitely determined, therefore, that there is
a qualitative difference between the serum components which initiate
phagocytosis in normal serum (normal opsonins) and those which
carry out the same function to a much more powerful degree in
is "Wright and Douglas. Cited in Wright, "Studien iiber Immun., etc.,"
p. 9.
16Neufeld and Rimpau. Deutsche med. Woch., No. 40, 1904; Zeitschr.
f. Hyg., Vol. 51, 1905.
17 Leishman. Trans. London Path. Soc., Vol. 56, 1905.
316 INFECTION AND RESISTANCE
immune serum. This is the more surprising since, in the case of all
other antibodies (lysins, agglutinins, etc.), it has been shown that in
structure and mode of action the antibodies of immune serum are
in every way qualitatively similar to the corresponding ones of nor-
mal serum,18, 19 representing merely a specific quantitative increase
of substances originally present in small amount.
This difference between the normal and immune opsonic sub-
stances has added much difficulty to the investigation of the nature of
these bodies, and we may approach the problem with greater clearness
by considering them separately, at first, attempting to define the
relations between them after we have set down the facts ascertained
in connection with each.
In their earliest investigations upon the normal opsonins Wright
and Douglas 20 regarded them as new antibodies, separate and dis-
tinct from those already known. There is no convincing proof of
this, and a number of other interpretations of the observed phe-
nomena are possible. Indeed, the burden of proof is rather upon
those who would establish the existence of a new antibody, for before
this can be done it must be shown that the new function is not merely
another property of the serum constituents already known. For, as
Gruber has justly said, "One of the most important attributes of the
natural scientist is economy of hypotheses." And in the case of the
normal opsonins there are many good reasons for regarding them as
possibly identical with known serum constituents. The two possi-
bilities suggested have been (1) Are the opsonic substances identical
with the alexin or complement? or (2) Do they represent the com-
bined action of the normal sensitizer of the serum activated by the
alexin ?
The similarity of normal opsonin with alexin or complement has
been brought out especially by Muir and Martin,21 by Baecher,22
and by Levaditi and Inmann.23 The fact that both are thermolabile
has been mentioned above.
In addition to this, as Muir and Martin24 have shown, all antigen-
antibody complexes which absorb alexin out of serum at the same
time remove the normal opsonin. Thus sensitized red corpuscles,
sensitized bacteria, and specific precipitates added to normal serum
take out its opsonic substances. From this fact they also concluded
that the normal opsonins like alexin were non-specific. For just as
18 Dean. Proc. Royal Soc., 76, 1905.
19 Neufeld and Hiine. Arb. a. d. kais. Gesundh. Amt.} Vol. 25, 1907.
20 wright and Douglas. LOG. cit.
21 Muir and Martin. Br. Med. Jour., Vol. 2, 1906; Proc. Royal Soc. B.,
Vol. 79, 1907.
22 Baecher. Zeitschr. f. Hyg., Vol. 56, 1907.
23 Levaditi and Inmann. C. R. de la Soc. Biol., 1907, pp. 683, 725,
817, 869.
24 Muir and Martin. Loc. cit.
FACTORS DETERMINING PHAGOCYTOSIS 317
the alexin of a serum may serve to, activate a considerable variety of
sensitized antigens, so the opsonic action of a normal serum may
functionate upon a large variety of bacteria. Muir and Martin were
probably wrong in this and, as we shall see below, normal opsonins,
like normal sensitizers, may be regarded as specific.
Similar to the observations of Muir and Martin are those of
^vTeufeld and Hiine,25 which showed that yeast cells will absorb both
alexin and opsonin out of serum.
A further similarity between the two serum constituents is the
fact that both are absent from the normal fluid of the anterior cham-
ber of the eye, but they together appear in it after injury (puncture
for the first removal of fluid). A like parallelism between the ab-
sence and presence of both has been shown for edema fluids.
Furthermore, phosphorus poisoning which reduces alexin likewise
reduces opsonin.
Although this parallelism is very striking, it does not on this
account mean that necessarily the two are identical. It may signify
merely that the alexin is a necessary participant in normal opsonic
action, essential in that it activates a thermostable opsonic constitu-
ent just as it activates hemolytic or bactericidal sensitizer.
This opinion has been expressed by Levaditi, Neufeld,26 Dean,27
and others, and indeed it is a conception which seems most logical.
For the procedures which remove both alexin and opsonin, as stated
above, do not, as a matter of fact, remove all the opsonic action.
(Although Neufeld maintains this.28) Studies of Hektoen and
others have definitely proved that, though reduced to almost nil,
nevertheless heated serum shows definite though slight opsonic action
as compared with indifferent menstrua such as salt solution. A
similar slight remnant of opsonic action after absorption of normal
serum with sensitized cells, bacteria, and precipitates is evident in the
protocols of Muir and Martin. The significance of this point be-
comes immediately clear when we consider the properties of the bac-
teriotropins or immune opsonins, which are heat stable and capable
of initiating opsonic action in the entire absence of alexin or comple-
ment. It is possible, therefore, that there may be present in normal
serum a slight amount of specific thermostable opsonin, which,
though capable of acting feebly by itself, is nevertheless powerfully
activated by alexin — just as bactericidal or hemolytic antibody is
similarly activated.
One of the most thorough studies upon this question is that of
25 Neufeld and Hiine. Arb. a. d. kais. Gesundh. Amt., Vol. 25, 1907.
26 Neufeld. "Kolle u. Wassermann's Handbuch," Erganzungsband 2,
p. 313.
27 Dean. Brit. Med. Jour., 2, 1907, p. 1409.
28 In fact he states that heated normal serum may be used as a control
in opsonic experiments instead of salt solution,,
318 INFECTION AND RESISTANCE
Cowie and Chapin.29 Dean 30 had previously shown that, although
heated immune serum was capable of exerting opsonic action by
itself, this action could nevertheless be enhanced by the addition of
a little diluted fresh normal serum. The particular significance of
Dean's work will be discussed later. Cowie and Chapin, however,
carried on similar experiments with normal serum in which they at-
tempted to reactivate heated normal serum by the addition of small
amounts of diluted fresh serum, by itself but slightly opsonic. One
of their experiments may serve to illustrate this point, as follows :
Experiment 10. June 13, 1907
Phagocytic
count 31
1. Unheated serum 15 . 44
2. Salt solution 0. 18
3. Heated serum, 57° C 1 .08
4. Diluted serum (1 :15) 1 .56
5. Heated serum 57° C. + diluted serum (1 :15) 12 . 40
6. Unheated serum + unheated serum 16 . 08
This experiment and others like it seem to demonstrate clearly
that the opsonic action of normal serum, though dependent largely
upon alexin, is nevertheless also dependent upon a heat-stable body,
comparable to the sensitizer or amboceptor, in that it is reactivable to
almost the full power of the original condition (before heating) by
slight amounts of alexin — in themselves almost inactive.32
These findings were later confirmed by Eggers,34 and it is plain
from this work that the apparent opsonic inactivation of normal
serum by heat depends upon the destruction of the heat-sensitive
constituent only — the heat-stable substance — surely involved in the
process, remaining intact, and reactivable.
Closely associated with this phase of the problem is that of the
specificity of the normal opsonins. For if, as at first supposed, the
normal opsonins are, like complement or alexin, non-specific, the
above amboceptor-complement structure of this mechanism would
be rendered unlikely. Earlier work upon this question was con-
tradictory. Bulloch and Western,35 working with staphylococci and
29 Cowie and Chapin. Jour. Med. Ees., Vol. 17, 1907, pp. 57, 95 and
213.
30 Dean. Loc. cit.
31 Phagocytic count = average number of bacteria in each leukocyte.
32 In earlier experiments Hektoen and Ruediger 3S did not succeed in
reactivating heated sera and concluded that normal opsonins had the hypo-
thetical structure of toxins in that they possessed a haptophore and an
opsonophore group. From this point of view Hektoen has subsequently
receded largely because of work done under his own direction.
33 Hektoen and Ruediger. Jour. Inf. Dis., 1905.
34 Eggers. Jour, of Inf. Dis., Vol. 5, 1908.
35 Bulloch and Western. Proc. Roy. Soc. B., 77, 1906.
FACTORS DETERMINING PHAGOCYTOSIS 319
tubercle bacilli, found that each of these organisms absorbed out
separately specific opsonins from normal serum, leaving those for
other bacteria but slightly reduced. Slight reduction of the opsonic
action for other micro-organisms might easily be explained by a
partial removal of complement which is bound to take place in such
experiments. Simon, Lamar and Bispham,36 and some others failed
to find any such specificity. Russell,37 Axamit and Tsuda,38 and a
number of others obtained similar negative results — in that a num-
ber of different bacteria seemed to absorb opsonins out of normal
serum indiscriminately and without specificity. On the other hand,
more recent careful work by Rosenow,39 by Macdonald,40 and by
Hektoen 41 has upheld the original contention of Bulloch and West-
ern. The work of Rosenow, in which pneumococci were shown to
absorb out their specific opsonins from normal human serum, taking
out in part only those for streptococci, staphylococci, and tubercle
bacilli, is especially convincing, and the experiment of Hektoen with
normal hemopsonins (opsonins which cause the phagocytosis of red
blood cells) bear him out.
It seems fair to conclude, therefore, that normal opsonins de-
pend upon the cooperation of a heat-stable and a heat-sensitive body.
The heat-stable body, analogous to normal sensitizer or amboceptor, is
specific and reactivable by the heat-sensitive body which appears to
be identical with alexin. This statement merely asserts the facts of
the dual mechanism of the process without assuming necessarily the
identity of the heat-stable body with sensitizer or that of the heat-
sensitive one with alexin, though this seems extremely probable.
This question we will discuss again more particularly in connec-
tion with the bacteriotropins or immune opsonins.
Further proof for such a complex constitution of the normal
opsonins has been adduced by means of absorption experiments at
0° C. — by Cowie and Chapin. In our discussion of the lytic anti-
bodies we have seen that sensitizer or amboceptor may be absorbed
from serum by its specific antigen at 0° C. — but that the attachment
of alexin takes place only when the temperature is raised above this.
Practically no alexin is bound at the low temperature. Cowie and
Chapin, applying this method of investigation, showed :
1. That normal human serum may have its opsonic power for
staphylococci removed by absorption with staphylococci at 0° C.
2. Serum so treated retains the power of reactivating the op-
sonin of heated normal serum.
36 Simon, Lamar, and Bispham. Jour. Exp. Med., Vol. 8, 1906.
37 Russell. Johns Hopk. Bull., Vol. 18, 1907.
38 Axamit and Tsuda. Wien. klin. Woch., Vol. 20, No. 35, 1907.
39 Rosenow. Jour. Inf. Dis., Vol. 4, 1907.
40 Macdonald. Quoted from Hektoen, loc. cit.; Aberdeen Univ. Studies,
Vol. 21, 1906, p. 323.
41 Hektoen. Journ. Inf. Dis., Vol. 5, 1908.
320 INFECTION AND RESISTANCE
3. Staphylococci so treated are more easily subject to phago-
cytosis in the presence of dilute normal serum, or normal serum
which has been inactivated by contact with staphylococci in the cold,
than are the same bacteria untreated.
Kurt Meyer 42 has carried out similar experiments with paraty-
phoid bacilli and normal serum, and, though his work is less exten-
sive, he reaches the same conclusion as Cowie and Chapin.
We may accept, therefore, as fairly well established that the
opsonic power of normal serum depends upon a complex mechan-
ism consisting of (a) a thermostable substance comparable to
amboceptor or sensitizer, probably specific, but present in very
small amount, and (b) a thermolabile substance probably identical
with alexin or complement which powerfully, but non-speci-
fically, enhances the slight opsonic power of the thermostable
substance.
In considering this conception, together with the subsequent dis-
cussion of bacteriotropins or immune opsonins, it will be well to
remember that in normal inactivated sera the thermostable opsonic
constituent differs in its action from the bodies we speak of as ambo-
ceptors or sensitizers in that it may functionate for phagocytosis by
itself — entirely without alexin — while neither bactericidal nor he-
molytic effects can be brought about by sensitizer alone. Does this
definitely exclude the identity of this thermostable opsonic substance
and sensitizer ? It is indeed an argument against identification, but
in opsonic action, we must remember, there is merely a sensitization
to the action of the phagocyte. This phagocyte may in itself be
capable of furnishing a small amount of substance comparable in
action to alexin — in fact, we have seen that the origin of alexin from
leukocytes is still suspected by a number of workers. At any rate
the phagocyte is a living cell which may well be capable of supplying
in itself to some degree the necessary activation, and therefore the
difference cited above is not necessarily a proof that the normal
thermostable opsonic constituent is different from normal sensitizer
or amboceptor.
The difference between the opsonic action of normal serum and
that of immune serum, then, is the fact that heating to from 56° to
60° C. almost completely destroys the former, whereas it has but
slight if any diminishing effect upon the latter. The immune op-
sonins, or, as Neufeld and Kimpau have called them, bacteriotropins,
therefore are thermostable. This was determined as early as 1902
by Sawtschenko,43 and was subsequently studied with great accuracy
42 Kurt Meyer. Berl klin. Woch., 1908, p. 951.
43 Sawtschenko. Ann. de I'Inst. Past., Vol. 16, 1902, quoted from Levaditi.
FACTORS DETERMINING PHAGOCYTOSIS
by Neufeld and Rimpau,44 Neufeld and Topfer,45 Dean,46 Hektoen,47
and others. It was shown that when an animal is immunized with
any given bacterium or other cellular antigen (blood corpuscles,
etc.) opsonic substances specific for the particular antigen appear in
considerable quantities, and these are but slightly, if at all, dimin-
ished when the serum is heated ; Neufeld and Hiine 48 found that
heating for as long as three-quarters of an hour to 63° C. did not
noticeably reduce the activity of the bacteriotropins of immune
serum, and that, again, unlike the normal opsonins, prolonged pres-
ervation, under sterile conditions, changes them but slowly.
These facts alone indicate a close similarity between the bac-
teriotropins and the other well-known thermostable constituents of
immune sera, and the question here again immediately arises whether
we are to regard them as identical with any of the other specific
antibodies or as distinct substances independent of these.
It was suggested early in these investigations by Muir and Mar-
tin that bacteriotropins might be identified with agglutinins, inas-
much as they possessed resistance to heat, were active without ap-
parent dependence upon alexin, and could not, at least in the earlier
studies, be reactivated by the addition of fresh normal serum when
once inactivated. The supposition was that for this reason the bac-
teriotropin might have a structure like the hypothetical "haptines
of the second order77 which Ehrlich attributes to the agglutinins.
This supposition has found no experimental support in that ag-
glutination and bacteriotropic effects did not run parallel. We our-
selves are not ready to admit that such lack of parallelism is proof
against their identity. However, since it is very probable that both
agglutination and precipitation are merely phenomena of colloidal
flocculation effects which follow certain quantitatively adjusted com-
binations of antigen and specific antibody, and that it is not at all
necessary to assume separate agglutinating or precipitating serum
constituents, this problem becomes merely another version of the
question of the identity of bacteriotropins and sensitizer or ambo-
ceptor.
Apart from thermostability, further similarity lies in the fact
that bacteriotropins are strictly specific and may be specifically ab-
sorbed out of immune sera by their respective bacteria.
Like amboceptor or sensitizer they are specifically increased to a
powerful degree by the treatment of animals with any given micro-
organism and may be incited not only by the injection of bacteria
44 Neufeld and Rimpau. Deutsche med. Woch., No. 40, 1904; Zeitschr,
f. Hyg., 51, 1905.
45 Neufeld and Topfer. Centralbl. f. Bakt., 1, 38, 1905.
46 Dean. Proc. Roy. Soc. B.} 76, 1905.
47 Hektoen. Jour. Inf. Dis., 3, 1906, and loc. cit.
48 Neufeld arid Hiine. Arb. a. d. kais. Gesundh. Amt., Vol. 25, 1907.
INFECTION AND RESISTANCE
but by that of blood cells as well. In spite of these points of likeness,
however, Neufeld 49 and his associates maintain rigidly that the two
substances are not the same and that the bacteriotropins are distinct
and independent antibodies.
Among the reasons advanced in support of this opinion are the
facts that certain immune sera, both antibacterial and hemolytic,
may contain bacteriotropins without containing lysins and vice versa.
That this is undoubtedly true has been shown not only by ISTeufeld
and his associates but by Hektoen 50 and others, and it is likewise a
fact that in sera in which both functions are demonstrable they fre-
quently do not run quantitatively parallel. These are unquestionably
strong arguments, but their force is somewhat weakened, as Levaditi
has pointed out, by the fact that there are many varieties of bacterial
immune sera which undoubtedly sensitize the specific bacteria (as
can be shown by alexin fixation), but which do not lead to bacterio-
lysis. Wassermann 51 also attaches little value to the lack of parallel-
ism between the lytic and opsonic functions, expressing the belief
that the solubility of the particular antigen may determine whether
sensitization leads to phagocytosis or to lysis. With bacteria like
the cholera spirillum rapid lysis takes place, but when, as in pneumo-
cocci or streptococci, there is great resistance to lysis, sensitization
may lead to delayed lysis anticipated by leukocytic accumulation,
phagocytosis, and intracellular digestion.
It by no means follows from mere lack of parallelism, therefore,
that the two serum functions are dependent upon separate antibodies,
although the argument is sufficiently strong to impose conservatism
in identifying them.
Another important argument advanced against the identification
of bacteriotropins with the bactericidal sensitizers or amboceptors is
the fact that the former lead to phagocytosis without the participa-
tion of alexin, whereas the latter become active for lysis only when
alexin is present.
This point also has constituted Neufeld's strongest support for
maintaining that the bacteriotropins or immune opsonins are entirely
distinct from the normal opsonins. It is true, indeed, that immune
serum, unlike normal serum, may opsonize powerfully even after
heating to temperatures which destroy alexin.
If we regard the heat-stable lytic antibody as an amboceptor in
the strict sense of Ehrlich, as a specific "Zwischenkorper" with a
complementophile group, this argument would have considerable
weight. Even in this case, however, strong sensitization of the bac-
teria may make them amenable to the living cells — the phagocytes —
4&Neufeld and Topfer. Centralbl. f. Bakt., 1, 38, 1905.
50 Hektoen. Jour, of Inf. Dis., 6, 1909.
51 Wassermann. Deutsche med. Woch., Vol. 33, No. 47, 1907.
FACTORS DETERMINING PHAGOCYTOSIS 323
which in itself may furnish a slight amount of alexin or alexin-like
substances.
We may regard the action of the immune serum upon the antigen
as rather a sensitization in the sense of Bordet, and it does not seem
logical to assume that the heat-stable bodies, similar in other respects,
are different merely because they can sensitize bacteria both to the
action of an alexin and to that of a living cell, which in itself surely
contains a number of different enzymes, comparable functionally to
alexin, though possibly not identical with it.
Indeed, the experiments of Dean have given much positive evi-
dence in favor of regarding the immune opsonins or bacteriotropins
as true amboceptors or sensitizem Dean 52 found that, although
heated immune serum may unquestionably opsonize by itself, its
action may be still further enhanced by the addition of a little diluted
normal serum (compare these results with those of Cowie and Chapin
on normal opsonins). Hektoen's 53 experiments with hemopsonic
immune sera are analogous. We cite one of these as illustrating the
point in question :
TABLE I
Phagocytosis of Goat Corpuscles under the Influence of Goat-blood-immune Rabbit
Serum, and Normal Guinea Pig Complement (Table from Hektoen, loc. cit.)
Immune serum
Normal guinea pig serum
Phagocytosis
.001 4.
.001 + -01 20.
.01 0
Here, therefore, the diluted immune serum, but slightly cyto-
tropic in itself, was powerfully activated by a diluted unheated nor-
mal serum, which in itself was entirely inactive.
Indeed, an experiment by Neufeld himself, with Bickel,54 points
in the same direction. They found that, when a heated specific hemo-
lytic serum was added to the homologous cells in such small quanti-
ties that it no longer exerted cytotropic (opsonic) action, the addi-
tion of a small amount of alexin, too small to lead to hemolysis of
the cells (and not by itself cytotropic or hemopsonic), caused active
phagocytosis. Analogous experiments upon bacterial antisera were
carried out by Levaditi and Inmann. It thus appears that, even in
the case of the immune opsonins or bacteriotropins, we may think of
a participation of two substances — a sensitizer-like one and one com-
parable to alexin or complement. We may, at least, infer that the
full opsonic action both of normal and immune sera is dependent
52 Dean. Proc. Royal Soc. B., 79, 1907.
53 Hektoen. Jour. Inf. Dis., Vol. 6, 1909, p. 67.
5- Neufeld and Bickel. Arb. a. d. kais. Gesundh. Ami., Vol. 27, 1907.
INFECTION AND RESISTANCE
upon the cooperation of two such bodies. It is likely, therefore, that
the mechanism of normal and of immune opsonic action may, after
all, differ only in quantitative relations between the two.
For assuming this to be an antibody-alexin mechanism like hemol-
ysis, we may recall the work of Morgenroth and Sachs on the rela-
tions between amboceptor and complement in hemolysis. There we
saw that a large amount of amboceptor would cause hemolysis in the
presence of a small amount of complement and vice versa. There-
fore, here, too, in normal serum the small quantity of amboceptor or
specific thermostable opsonin (bacteriotropin) may act very power-
fully in the presence of the alexin. When the latter is destroyed,
however, the minute quantity of specific thermostable opsonin is
hardly enough to do more than initiate slight phagocytosis of com-
paratively non-resistant bacteria, whereas the large amount of spe-
cific sensitizer left in immune sera after inactivation may still lead
to strong bacteriotropic action. In outlining this explanation we
have consistently drawn upon the analogy between thermostable op-
sonin and amboceptor or sensitizer. Whether or not these two sub-
stances are identical is by no means positively determined and must
be considered an open question for the present. However, from the
above, it seems to us that much testifies in favor of such an identi-
fication.55
The preceding discussions have ignored the possibility that apart
from opsonic or bacteriotropic action on the bacteria there may be
a difference in phagocytic energy which depends upon inherent prop-
erties of the leukocyte itself.
Indeed, the technique by which the researches of Wright and
his followers were carried out does not in any way take into account
the source of the leukocytes as a possible variable factor. There is,
however, a considerable amount of evidence which points to differ-
ences in phagocytic powers residing in the leukocytes themselves
independent of the serum. Park and Biggs 56 have demonstrated
such differences for the leukocytes of normal persons in the phagocy-
tosis of staphylococci, and more extensive researches have been made
with similar results, in the case both of staphylococci and tubercle
bacilli by Glynn and Cox.57
The last-named authors, moreover, recognized the necessity, in
making such investigations, of experimenting with leukocyte emul-
sions containing approximately the same number of cells, for, as
JFleming 5^Piad shown, if unequal leukocytic emulsions are used,
55 Pfeiffer (quoted from P. Th. Miiller) regards opsonic action as due
to a combined action of amboceptor and complement and speaks of it as an
"Andauung" of the bacteria for the leukocyte — which we may translate best
as a partial predigestion.
56 Park and Biggs. Jour. Med. Ees., Vol. 17, 1907.
57 Glynn and Cox. Jour. Path, and Bact., 14, 1910.
58 Fleming. Practitioner, London, Vol. 80, 1908.
FACTORS DETERMINING PHAGOCYTOSIS 325
less phagocytosis per cell occurs in the emulsion containing the
greater number of leukocytes. This phase of the subject has been
taken up most thoroughly by Hektoen 59 and his associates, and Rose-
now 60 has made careful comparative studies on pneumococcus
phagocytosis, in which he standardized the leukocytic suspensions by
actual cell counts. His work as well as that of Tunnicliff,61 of the
same school, has shown definitely that the inherent phagocytic power
of leukocytes may vary not only in health and disease, but differences
may exist between the cells of apparently normal people. Tunni-
cliff showed, for instance, that at birth the leukocytes are less active
than in adult life.
For accurate experimental work, therefore, as well as in theoret-
ical reasoning upon problems of phagocytosis, it is necessary to bear
in mind the possible inherent variations in the leukocytes themselves.
Of the three factors concerned in the process of phagocytosis,
then, we have considered two, the serum and the leukocytes. The
former we have seen exerts a powerful determinative influence on
the process, the latter a less marked influence, though still definite
and measurable. We have still to discuss the bacteria themselves
as variable factors in determining the degree to which phagocytosis
may take place.
This problem was first investigated by Denys and Marchand in
connection with their work upon streptococcus immunity, and was
further studied in detail by Marchand. Marchand 62 showed that
leukocytes would readily take up non-virulent streptococci in the
presence of normal serum, but that under similar conditions virulent
streptococci were not phagocyted at all or to a very slight degree
only. He determined further that this resistance to phagocytosis
remained unchanged after the virulent organisms had been killed
by heat, and washed clean of culture fluid. It seemed, therefore,
that the resistance depended upon a condition of the bacterial body
and not upon substances secreted and given off to the environment.
These experiments, as well as similar work by Mennes,63 Gruber and
Futaki,64 and others makes it clear that differences in virulence
between different species of bacteria, as well as between different
strains of the same micro-organism, depend, at least in part, upon
the resistance which the bacterial bodies oppose to ingestion by the
leukocytes. We must distinguish clearly here between these appar-
ently purely "antiopsonic" bacterial properties and those supposedly
"antichemotactic" substances which are conceived as a cause for
59 Hektoen. Jour, of A. M. A., Vol. 57, No. 20, 1911.
60Rosenow. Jour, of Inf. Dis., 7, 1910.
61 Tunnicliff. Jour. Inf. Dis., 8, 1911.
62 Marchand. Arch, de Med. Exp., No. 2, 1898.
63 Mennes. Zeitschr. f. Hyg., Vol. 25, 1897.
64 Gruber and Futaki. Munch, med. Woch., 1906.
INFECTION AND RESISTANCE
virulence by Deutsch and Feistmantel 65 and by Bail 66 in his so-
called "aggressins." The latter are supposed to be secreted bacterial
substances by means of which the leukocytes are held at bay. The
properties we are, at present, considering are probably in no way
antichemotactic, but oppose purely the actual ingestion by the
leukocyte, nor do they seem to depend upon the secretion of sub-
stances which injure the leukocytes. For, in the first place, a
profuse accumulation of leukocytes may follow the injection of
virulent micro-organisms, and Denys (quoting from Gruber) has seen
active phagocytosis of virulent pneumococci, but none of virulent
streptococci when antipneumococcus serum was injected with the
mixture.
Eosenow 67 has carried out a thorough investigation dealing with
these relations in pneumococcus infection. Seventy-five strains of
this organism were all found non-phagocytable when first isolated
and the resistant condition was associated with virulence for rabbits
and guinea pigs. It was found, moreover, that the resistance to
phagocytosis was dependent upon the inability to absorb opsonin.
For, while phagocytable non-virulent pneumococci absorbed specific
opsonin from serum, the virulent ones failed to do this in proportion
to the degree of their virulence. Furthermore, extraction of the
bodies of the virulent organisms in KaCl solution yielded a substance
which inhibited the action of pneumococcus opsonin — a true anti-
opsonin — which he speaks of as "virulin." This discovery, if con-
firmed, would supply us with a very simple explanation for some
phases of the problem of virulence. It is, indeed, likely that the
antiopsonic property is closely bound up with chemical and struc-
tural changes which take place in the bacterial cell as it adapts itself
to the parasitic conditions. This is plain from the fact that pneumo-
cocci and some other bacteria will rapidly lose their virulence when
cultivated on artificial media devoid of animal serum, will retain it
longer if grown on some serum media, and will rapidly regain it if
passed through animals. The formation of a capsule is unquestion-
ably a morphological evidence of such a change. Habitually capsu-
lated bacteria, like the Friedlander bacillus, and Streptococcus muco-
sus, are of fairly constant virulence, while in other micro-organisms
like the pneumococci, anthrax bacillus, plague bacillus, and certain
other streptococci, the formation of a capsule goes hand in hand with
an increase of virulence. By the aid of this morphological earmark
of virulence, moreover, Gruber and Futaki have obtained further
65 Deutsch and Feistmantel. Quoted from Sauerbeck. Lubarsch und
Ostertag, Vol. 2, 1906.
66 Bail. Arch. f. Hyg., Vol. 52, 1905.
67 Rosenow. Jour. Inf. Dis., Vol. 4, 1907.
FACTORS DETERMINING PHAGOCYTOSIS 327
proof that the resistance to phagocytosis in these cases is due to the
nature of the bacterial cell body rather than to any secreted anti-
opsonic substances. For, after the injection of anthrax bacilli into
guinea pigs, they saw that leukocytes would take up uncapsulated
bacilli, apparently picking them out of the midst of surrounding
capsulated organisms which they were unable to ingest.
CHAPTER XIV
THE OPSONIC INDEX AND VACCINE THERAPY
WEIGHT'S 1 investigations upon phagocytosis were, indirectly,
the outcome of his earlier work upon antityphoid vaccination. His
purpose .in these studies had been a purely practical one, and he
had attempted to obtain a guide for the dosage and the interval be-
tween injections by measuring the bactericidal and agglutinating
powers of the blood serum. In the case of typhoid immunization
this was indeed a practicable method of control, since the bacteri-
cidal power of the blood serum rose directly as the immunization of
the patient was attained. In the cases of many other bacteria, how-
ever, this method of study was not practicable, and Wright, as others
before him, did not find a regularly increased specific bactericidal
power in the blood sera of immunized animals or of patients con-
valescing from infections with such bacteria as the staphylococcus,
streptococcus, Micrococcus melitensis, the Bacillus pestis, and a num-
ber of others. In fact, together with Windsor,2 he showed that nor-
mal human blood has practically no bactericidal power for pyogenic
staphylococci and that antistaphylococcus inoculations or recovery
from an infection do not result in the production of such proper-
ties in the serum. These determinations are practically identical
with Nuttall's 3 earlier studies on the same bacteria and, indeed, cor-
respond with the data obtained by Metchnikoff and his followers in
their work on anthrax infection. For, in discussing these investi-
gations, we saw that very often the serum of a comparatively resist-
ant animal is less potently bactericidal than that of a more suscep-
tible one. We need only recall the difference between rabbits and
dogs in this respect. The serum of the former is more strongly bac-
tericidal than that of the latter, and yet rabbits are the far more
susceptible animals. These relations have been studied with great
care, also, by Petterson.4 It was logical in such cases to look for
the cause of resistance in the activity of the phagocytes, and this,
we have seen, Metchnikoff did successfully in a large series of cases,
both as regards natural and acquired immunity.
1 Wright. Lancet, 1902; Practitioner, Vol. 72, 1904.
2 Wright and Windsor. Jour, of Hyg., Vol. 2, 1902 ; and Wright,
Lancet, 1900 and 1901.
3 Nuttall. Zeitschr. f. Hyg., Vol. 4, 1888.
4 Petterson. Centralbl. f. Bakt., Vol. 39.
328
OPSONIC INDEX AND VACCINE THERAPY 329
Yet the controversy between the strictly humoral and the cellular
schools was by no means regarded as closed, especially since, in such
cases as typhoid infection, the parallelism between increased resist-
ance and extracellular bactericidal power was so plainly evident,
while in this disease particularly (for technical reasons which will
become clear as we proceed) no such parallelism with phagocytosis
could at first be shown. It was because of such apparent confusion
that Leishmann 5 undertook to study again the relation of phago-
cytosis to active immunity, chiefly upon staphylococcus cases that
were being "vaccinated" therapeutically by Wright himself.
In order to obtain a numerical measure of the degree of phago-
cytosis, he developed a simple technique which, though crude, served
to give him the information he sought. It consisted in taking small
quantities of the blood of patients and mixing these in capillary
pipettes with equal volumes of bacteria suspended in salt solution.
The mixtures were then placed on slides, covered with a cover-
slip, and incubated at 37° C. for varying periods. At the end of
incubation the preparations were smeared upon slides and stained
by Leishmann's modification of the Romanowski method, the num-
ber of bacteria in a large series of leukocytes counted and an average
taken.
This method had many serious flaws, chief among them being
the liability to coagulation of the preparations and the fact that, in
each test, the fluid constituents as well as phagocytes, both of them
variable factors, came from -the same individual. While, therefore,
it was possible to estimate an increase or decrease of general phago-
cytic power, it was impossible to analyze this in reference to its de-
pendence either upon the condition of the cells, on the one hand, or
that of the plasma or serum, on the other. Moreover the relation of
the number of leukocytes to that of bacteria in individual tests neces-
sarily differed, and this, we have seen, adds a variable factor which
renders it impossible to compare any two experiments with accuracy.
In spite of these difficulties, however, Leishmann succeeded in
establishing, in a number of cases of staphylococcus infection, that
an increased resistance was accompanied by an increased energy of
phagocytosis.
Leishmann, however, went no further than this, and interpreted
his results on the basis of the "stimulin" theory of Metchnikoff.
The subsequent studies of Wright, which began at the point at
which Leishmann stopped, have been described in the preceding chap-
ter and had, as their main result, we have seen, the discovery of the
opsonins and the final confirmation of Denys' conception of the true
mechanism of cooperation between serum and leukocytes in phago-
cytosis. In order to carry out these studies the technique of Leish-
5 Leishmann. Br. Med. Jour., 1, 1902 ; Transact. Lond. Path. Soc., Vol.
56, 1905.
330
INFECTION AND RESISTANCE
mann was quite inadequate, and Wright's first task was to modify it
in such a way that reasonably accurate comparative estimates of
phagocytosis could be made.
It is necessary to outline Wright's method briefly in this place
in order that we may consider possible sources of error and obtain
a clear understanding of the conclu-
sions he based on his observations.
Wright recognized that the deter-
mination of the degree of phago-
cytosis, induced by the opsonin of
WR.GHT CAPSULE ro* TAKING any given serum in a single test, is
BLOOD TO OBTAIN SERUM FOR by itseli oi no value, since the actual
OPSONIC TESTS. number of bacteria taken up by each
leukocyte, apart from the opsoiiic
contents of the serum, depends also upon such purely technical fac-
tors as the concentration of the bacterial emulsion, the relative num-
ber of leukocytes, and the length of time of incubation. Two
individual tests, therefore, carried out with the serum of the same
patient at the same or at different times, with different bacterial
emulsions or leukocytes in each,
would give variable results, even
though the opsonin contents them-
selves were entirely alike.
In order, therefore, to obtain
a, relative estimate of the opsonic
contents of any serum it is neces-
sary to compare the phagocytic ac-
tivity induced by this serum with
the similar power of another sup-
posedly normal serum, both tests
being carried out, under exactly
similar conditions, with the same
bacterial emulsion and the same
leukocytes. The average number
of bacteria found in each leuko-
cyte in each one of the prepara-
tions is then the "phagocytic in-
dex." The relation of the phago-
cytic index of the unknown serum METHOD
to that of the supposedly normal
serum constitutes what Wright has
called the "opsonic index."
Instead of using the whole blood of the patient Wright takes a
small amount of blood in glass capsules, allows it to clot, and uses
the expressed serum in his test. For comparison with this he em-
ploys a "pool" of a number of specimens of serum from supposedly
OF PRODUCING AN EVEN
EMULSION OF BACTERIA FOR OP-
SONIN DETERMINATION.
OPSONIC INDEX AND VACCINE THERAPY 331
normal individuals. By the use of such a serum mixture any slight
possible variations from the normal in any one of the sera are likely
to be equalized, and a closer approach to a normal standard is at-
tained.
The leukocytes used in both tests are the same and taken, as a
rule, from the blood of the worker or from some other supposedly
healthy person. They are obtained by taking 15 or 20 drops of
blood from the finger or ear into 5 to 10 c. c. of sodium citrate solu-
tion, in which the blood does not clot. Brief centrifugalization throws
down the blood cells, with a thin, buffy coat of leukocytes on top, and
these are gently taken off with a pipette. This constitutes the leu-
kocytic cream of Wright's experiments, and furnishes a uniform leu-
METHOD OF TAKING UP EQUAL VOLUMES OF LEUKOCYTES, BLOOD SERUM AND
BACTERIAL EMULSION IN WRIGHT'S TECHNIQUE FOR OPSONIC-INDEX DETER-
MINATION.
kocyte factor for the two tests which are to be compared. The bac-
teria are obtained by emulsifying carefully in salt solution. It is
very important to obtain an emulsion free from clumps and neither
too thick nor too thin, a result which can be secured only by experi-
ence.
Equal quantities of serum (unknown and normal "pool" respec-
tively) are mixed with equal quantities of the bacterial emulsion and
the leukocytes in capillary pipettes, and the mixtures are incubated
for fifteen to thirty minutes under exactly similar conditions. At
the end of this time smears are made upon slides, the preparations
stained, and the numbers of bacteria in a hundred or .more leuko-
cytes counted in each of the two experiments. The average is taken,
and from the phagocytic indices thus obtained the opsonic index is
calculated. For instance, if
Phagocytic index (normal pool) = 8
Phagocytic index (patient's serum) = 6
then the opsonic index (patient's serum) = 0.75. Or, if the phago-
cytic index of the normal pool had been 10. and that of the patient's
serum 15., then the opsonic index of the patient's serum, higher than
normal, would be 1.5.
For the insurance of accuracy in carrying out this method Wright
calls especial attention to the caliber of the capillary pipettes that
are used, the concentration of the sodium citrate solution, which
should be 1.5 per cent., and the freshness of the leukocytes. But
it is still necessary to remember that with the greatest care in tech-
INFECTION AND RESISTANCE
nique uncontrollable sources of error influence this method. Most
important among them are the differences necessarily existing be-
tween different normal sera used for comparison and differences
in the agglutinative powers of the sera used in the two specimens.
For it is plain that different degrees of agglutination may bring
about great variations in the number of bacteria with which the
individual leukocyte comes into contact.
Wright's method has also been particularly unsatisfactory in
taking the opsonic index against such bacteria as the typhoid bacillus
and the cholera spirillum, or-
ganisms which are very rap-
idly digested after being
taken up by the leukocytes.
In consequence, even after as
short an incubation time as
five or ten minutes, the in-
gested bacteria are partly dis-
integrated, are stained in-
distinctly, and cannot be
counted with accuracy. In
order to avoid this source of
error Klien 6 has devised a
modification which depends
upon gradual dilution of the
serum in a series of pha-
gocytic tests with the
LEUKOCYTES CONTAINING BACTERIA. DRAW- same leukocytic and bacterial
ING OF FIELD AS SEEN IN WRIGHT'S mrml«irm« Tn tliic wav IIP
METHOD OF OPSONIC-INDEX ESTIMATION, emulsions. In tllia way ne
determines the degree of di-
lution of the serum to be
tested at which phagocytosis no longer exceeds that taking place in
salt solution alone. The degree of dilution at which this result was
obtained has been called by Simon the "coefficient of extinction." A
comparison of sera with regard to this value, it is clear, furnished an
estimate of their quantitative opsonic properties quite as instructive
as the direct estimations by the Wright method, and in our opinion,
at least, more reliable. Though also subject to some of the objections
advanced against the Wright method, it has the definite advantages
mentioned above, and is not so closely dependent upon irregularities
in counting, agglutinin influences, and differences in relative pro-
portions of bacteria and leukocytes employed. Jobling 7 has used
this method with success for the standardization of antimeningitis
serum.
A further modification suggested by Simon, Lamar, and Bis*
6 Klien. Johns Hop. Hosp. Bull., Vol. 18, 1907.
7 Jobling. "Studies from the Rockefeller Inst.," Vol. 10, 1910, p. 614.
OPSONIC INDEX AND VACCINE THERAPY 333
pham 8 depends upon a combination of the dilution method and a
modification in the method of counting. They make comparative
tests of the same serum, diluted from 1 to ten to 1 to one hundred in
salt solution, and estimate the opsonic power, not by determining
the average number of bacteria to the leukocyte, but by taking a per-
centage of the total number of leukocytes which take part in the
phagocytosis, that is, contain any leukocytes at all. The bacterial
emulsion for this method should be so thin that, in normal serum,
only about 50 per cent, of the leukocytes will contain bacteria.
That Wright's method, or any of the others, gives absolutely
accurate results will hardly be claimed by any one who has worked
upon opsonic-index estimations. There are certain uncontrollable
variable factors, some of which have been pointed out above; and,
apart from these, the delicacy of the technique is such that reliable
results can ordinarily be 'obtained only by trained workers after con-
siderable practice and experience. Even in such hands the per-
centage of personal error is more likely to be above than below 10
per cent. For ordinary clinical purposes, therefore, in the control
of cases the estimation of the opsonic index is not often practicable.
On the other hand, there can be little doubt about the fact that
careful comparative estimation, by Wright's method and by some
of the modifications, carried out by workers with experimental train-
ing and consequent attention to extensive controls, have yielded
results of sufficient accuracy to permit the recognition of definite
facts concerning opsonins. It is beyond question, therefore, that the
conclusion regarding the relation of opsonic fluctuations to clinical
conditions and the general significance of opsonins emanating from
laboratories like those of Wright, E~eufeld, Hektoen, and some others
may be accepted as fact — especially since in most essentials such
workers have agreed. In consequence we are now in possession of
knowledge regarding the opsonic constituents of the blood in health
and disease, and in the course of active immunization with bacterial
vaccines, which is of the greatest practical importance. We may
summarize the results of such investigations by saying that in many
of the infections of man the resistance of the patient is roughly pro-
portionate to the opsonic index — and that properly spaced inocula-
tion with suitable quantities of dead bacteria (vaccines) will raise
the opsonic index and lead to recovery in many of the localized
subacute and chronic conditions.
As to the usefulness of the treatment in various infections and
the limitations within which we may hope for results opinions differ,
and these will be discussed more fully below. Before we proceed
to this, however, it will be useful to consider the studies upon which
8 Simon and Lamar. Johns Hop. Hosp. Bull, Vol. 17, 1906; Simon,
Lamar, and Bispham, Jour. Exp. Med., Vol. 8, 1906 ; Simon, Jour. A. M. A.,
Vol. 48, 1907, p. 139.
334
INFECTION AND RESISTANCE
the parallelism between opsonic index and clinical condition was
founded.
Wright's own earlier studies were made chiefly upon staphylococ-
cus infections and tuberculosis. Since then the method has been
applied to almost all known infections with varyingly successful
results.
One of the first steps in determining such a parallelism between
the resistance of a patient and the opsonic index consisted, of course,
in comparing the index of the sera of normal individuals with that
of patients suffering from infection. Wright and Douglas did this
in a large series of studies. In the case of staphylococcus infections
the following experiment will illustrate their results:
TABLE I
(Wright and Douglas, Proc. Royal Soc., Vol. 74, 1904.)
Showing the ratio in which the phagocytic or opsonic power of the
patient's blood stood in each case to the phagocytic or opsonic power of the
normal individual who furnished the control blood. (The phagocytic power
of the control blood is taken in each case as unity.)
Initials of Patient
Form of Staphylococcus Invasion
Opsonic Index
E. E.
Furunculosis
0 48
F. F.
Sycosis
0 49
J. E.
Acne
0 64
J. H.
Furunculosis
0 87
W. B.
Acne
0 55
E. H.
Acne
0 82
W. H.
Furunculosis . . .
0 79
R. G.
Furunculosis
0 7
G. L
Acne and sycosis
0 74
S. C.
Furunculosis .
0 87
W. L.
Furunculosis
0 88
W. P.
Furunculosis
0 39
S. F.
Very aggravated sycosis
0 1
E. F. D.
Acne
0.73
D. C.
Sycosis
0 8
J. M.
Acne
0.48
W. M.
Sycosis
0.37
E. P.
Acne
0 6
M. S
Pustular affection of lips
0 6
F. V.
Repeated sta/ph infection . . .
0 47
In this series, as in others investigated by Wright and his col-
laborators, staphylococcus infection was uniformly associated with
a low index. He concludes that there is probably a causative rela-
tion between the two facts, in that under conditions of depressed
phagocytic powers staphylococci may gain a foothold, while under
OPSONIC INDEX AND VACCINE THERAPY 33*5
ordinary normal conditions they would fall prey to phagocytic de-
struction soon after entering the body.
The study of the opsonic index during the treatment of such
cases with dead staphylococcus cultures (usually with the organisms
cultivated from the patient's own lesions — "autogenous vaccines")
revealed a striking coincidence between the rise of the opsonic index
and improvement in the clinical conditions. A number of further
interesting and practically important points were brought out by
the systematic study of these relations which may be illustrated by
5% 6 769/0//IZ/3/4/S/6/7/8/9202/&23&&Z&
2.0
/.8
/S
tz
10
$
8
7
.6 \
J
r
S
$
s
/
1
c|
0
1
1
?
^
«Q
/
$
£
*
^
J
/
|
v)
1
/
^
$
f\
i
/
£
j
\
^
/
0
(V
I
/
A
x
^
/
f
X*
*
y
>
y
1
^
^
V
/
|
o
^
5
CURVE I. — RESULT UPON OPSONIC INDEX OF VACCINE TREATMENT IN Two CASES OF
CHRONIC STAPHYLOCOCCUS FURUNCULOSIS.
(After Wright and Douglas, Proc. Eoyal Soc., Vol. 74, 1904, p. 156; also from
" Studies on Immunity," p. 41.)
reproducing a plan of the opsonic index curves constructed from
cases.
The curve shown above, and taken from a paper by Wright and
Douglas, illustrates the course of the opsonic fluctuations in the
case of a medical student who had suffered for four years from boils.
When first seen the opsonic index (1. being normal) was 0.6,
and there were 2 boils on the neck. For 3 days after this there was
a spontaneous rise in the index accompanied by an improvement of
the lesions.
On the third day 2 billion staphylococci were injected. This
was followed by an immediate drop of the phagocytic power — (the
negative phase) ; together with this a new boil began to form. Soon,
however, the opsonic power began again to rise, this time consid-
erably above normal, reaching its highest point on the 8th day, when
<Q
5
K
'OO
07A
AO
07,
(h
7|7/
It
OCS
o/ttw
02
Q
I
i
D
•a
336
OPSONIC INDEX AND VACCINE THERAPY 337
it again began to diminish. A second inoculation on the 12th day
was followed by a similar preliminary negative phase, then a steady
and rapid positive phase, which was accompanied by cure.
Another curve — Curve 2 of the same publication (Wright and
Douglas, Proc. Royal Soc., Vol. 74, 1904, p\ 156) — is similar. This
case suffered from severe sycosis (barber's itch), had been ill for
17 months, and had been unsuccessfully treated during this time with
antiseptics. Staphylococci were isolated from a hair follicle, and
from this the vaccine was made which was used in the treatment.
Here the originally low opsonic index (0.8) rose after the first in-
jection without a preliminary negative phase — but after the second
treatment a sharp fall preceded the subsequent rise. Finally a sus-
tained high index accompanied complete cure.
The rise and fall of the opsonins after the injection of bacteria
is entirely analogous to the similar fluctuations of other antibodies
after antigen injections. Measurements of this kind are numerous
in the literature. Thus Saloinonsen and Madsen, measuring the
antitoxin contents of the blood and milk of a mare which were being
immunized by injections of diphtheria toxin, obtained the following
curve, which is entirely similar in essential features to those con-
structed for the opsonic index by Wright and Douglas :
/90
180
Jj
I
a
\
§
75
1
1
i
^
^
>
-i
M
g
i
_^
160
150
140
/SO
-
,
'
\
3
VF
50
k
|
f
y
,\
3
fit
1
^
J
1
j
\\
y
\
^
A,
3
LI
UO
110
100
90
80
r
/"
v
2
-
III
\
-•'
--
i
\
1
/
j
..,
-
V
A
/
1
"--
i
A
' J
70
60
50
4n
«i
i
I— •-
/
X1A7
/ro
K/C
Pot
v£^
o/^
THE
/w.L
.
1
....
...
4/V7
/rox
1C
POM
£*•
Of
THC.
/*4>o
%
2 4 6 8 JO /2/4I6 /6
CURVE DESCRIBING QUANTITATIVE MEASUREMENTS OF ANTITOXIN IN A MARE IN
EESPONSE TO TOXIN INJECTIONS.
(Taken from article by Salomonsen and Madsen, Ann. de I'Inst. Pasteur, Vol. 11,
1897, p. 319.)
Results having the same general significance are apparent in the
measurements made upon a tetanus toxin goat by Ehrlich and
Brieger,9 and in the observations upon the fluctuations of bacteri-
9 Ehrlich and Brieger. Zeitschr. f. Hyg., Vol. 13.
338
INFECTION AND RESISTANCE
'0 II
£ 13/4/5
PROLONGATION OF THE NEGATIVE PHASE DUE TO
Too VIGOROUS TREATMENT WITH TYPHOID
VACCINE.
(After A. E. Wright, Brit. Med. Journ., May
9, 1903. Also from "Studies on Im-
munity," p. 179.)
cidal power of the sera of patients treated with typhoid vaccines
made by Wright10 himself. Similar, again, are the various ag-
glutinin curves constructed by Jorgensen and Madsen n and others.
Apart from the purely theoretical value of such measurements,
they demonstrate features which are therapeutically of the greatest
importance. They show that in all processes of active immunization
the injection of antigen is followed almost immediately by a rapid
decline of specific antibodies in the blood serum. This "negative"
phase, as it is called, is probably due to a neutralization of existing
antibodies and lasts for
varying periods, which
must, of course, depend
upon complex relations be-
tween the degree of resist-
ance (or amount of anti-
body constituents of the
serum), the quantity of
antigen injected, and the
general recuperative pow-
ers of the subject. There-
fore, without some control
like that furnished by the
measurement of opsonins
or other antibodies it is impossible to determine whether the negative
phase has ended or is still in progress unless the clinical condition is
of such a nature or location that degrees of improvement or exacerba-
tion are well marked and easily observed. Even then clinical
observation alone is at best not an absolutely reliable guide.
The practical importance of the question lies in the harm which
may accrue to the patient if a second injection is practiced before
the cessation of the negative phase. Wright himself accentuates
this danger by expressing the opinion that, in typhoid inoculations,
an excessive dose administered to a patient in the physiological con-
dition of the negative phase may be followed by a prolongation of
this phase into a period of several months.
In the case of successive inoculations, as in vaccine treatment, a
too rapid repetition — i. e., a repetition of injection during such a
period of depression — leads to what Wright speaks of as a "summa-
tion of the negative phase," which obviously may seriously aggra-
vate the condition of the case. '
It is to such a cumulation of the negative phase that Wright
attributes the failures attendant upon the use of tuberculin during
the early days after its introduction, since injections at this time
10 Wright. Practitioner, Vol. 72, 1904, p. 118.
11 Jorgensen and Madsen. Festschrift. Serum Institut. Kopenhagen,
1902.
OPSONIC INDEX AND VACCINE THERAPY 339
were carried out without any control of serum reactions in the
patient and with comparatively large doses.
The danger to be carefully avoided, therefore, is a too rapid
succession of inoculations and too large a dosage, since both of these
procedures may be followed by cumulation of the "ebb tide of im-
munity," and great harm may result. On the other hand, if the
treatment is so spaced and measured that the successive inoculations
are given just before the positive phase has ended — in other words,
just before the apex of the curve is reached — a moderate negative
phase may be then followed by a second positive phase still higher
than the first, and corresponding improvement will result. It is even
possible to occasionally obtain a summation or cumulation of the
positive phase — in which the negative phase will be entirely sup-
pressed. This is illustrated in the following curve, in the case of
the first and second inoculation indicated on the chart. This case,
too, was a staphylococcus infection occurring in a laboratory at-
tendant :
STAPHYLO-
OPSONIC 20,
'6&Z8&X3/ I
IS l6l7t8'9&ZI
STAPHYLOCOCCUS INDEX AS DETERMINED BY WRIGHT IN A CASE OF ACNE TREATED
WITH STAPHYLOCOCCUS VACCINES.
Note summation of positive phase after third injection. (After A. E. Wright,
"Studies on Immunity," p. 348.)
Such a summation of positive phase, though of course the ideal
to be aimed at, cannot be produced with regularity, however carefully
we may attempt to control the treatment. It is worth mentioning,
moreover, a fact which should become evident from the preceding
and is too often overlooked, that a summation of the negative phase
can certainly be attained by the frequent repetition of larger doses.
This is practiced not infrequently in the false hope of hastening the
acquisition of immunity, and does harm more often than good.
Ordinarily the opsonic index when raised to a level considerably
above normal will gradually recede to the normal or even to a sub-
340
INFECTION AND RESISTANCE
normal condition. In isolated cases, however, especially in tubercu-
losis, the index may remain high for periods as long as a month.
This Wright speaks of as a sustained "high tide" of immunity.
These laws of fluctuation are all of them entirely analogous to those
long well known in the- cases of other antibodies, for even in diseases
in which the immunity following an attack — (typhoid fever, cholera,
plague, and others) — is continued through life the antibodies disap-
pear from the blood after varying periods and we are forced to seek
the cause of the permanently high resistance, not in the circulating
blood, but in the ultimate physiological units — the cells and tissues.
According to Wright also, the treatment with vaccines may
be either reenforced or entirely replaced by a process of autoin-
oculation from the patient's own lesion by increasing the local cir-
culation, thereby throwing more of the specific antigen into the
blood stream.
This reasoning has been applied, not only to the treatment of
tuberculosis and other conditions, but has been utilized to explain
fluctuations in the opsonic indices of untreated patients under the
influence of unusual motion of the diseased parts — as in walking or
other exercise. Wright's meaning is well illustrated by the follow-
ing curve of opsonins in a case of gonorrheal polyarthritis in which
massage of the joints resulted in reactions similar to those ordinarily
elicited by vaccine injections:
J/V/C
rx
40
|Bi
ji
3.5
iO
25
?0
=
I
lj
t
4-
/-
/.f
/<9
^~
5!
JL
S
I
i;
~v
(95
^N
V
Ez
i ~
^6
27
2fl
29
30
3/
1
2
3
4
5
6 ?•
a 5
/t7
7
12
13
/4
/?
/6
/7
/<9
/P
2f
^/
22
23
f\2S\Z6Z
7Zfl
29
30/
OPSONIC CURVE IN A CASE OF GONORRHEAL ARTHRITIS IN WHICH AUTO-INOCULA-
TION BY MASSAGE WAS PRACTICED.
(After Wright, Douglas, Freeman, Wells and Fleming, "Studies on Immunity,7'
p. 373.)
A further modification of the vaccine treatment of Wright origi-
nated in the observation that the exudate present in many infected
foci is often very much less rich in opsonins than is the blood serum
of the same patient. This is not unlikely to be due to an absorption
of the antibodies by the bacteria — as well as by the tissue detritus in
the lesion. But Wright has interpreted it as a purely specific ab-
OPSONIC INDEX AND VACCINE THERAPY 341
sorption by the bacteria, and has utilized it for diagnostic purposes.
Thus, with Reid,12 he has examined in this way the comparative
amounts of tubercle bacillus-opsonins in the blood, and in the local
exudates (peritoneal fluid) in cases suspected of tuberculosis, and
has determined the tuberculous nature of the condition by showing
a discrepancy between the two. These results have not been uni-
versally confirmed.13 But therapeutically, because of this supposed
lack of opsonin in the fluid of lesions, Wright has advised the in-
crease of the local flow of lymph by poulticing, heat, drainage, Bier's
cups, X-rays, Finsen light, and other means of accomplishing this
purpose.
All that has gone before (most of it taken directly from the
staphylococcus studies of Wright and his immediate followers) has
tended to show a very close correspondence of clinical improvement
with the increased opsonin contents of the blood.
As applied to other infections, such as gonococcus arthritis, colon
bacillus cystitis, localized pneumococcus lesions, and many other
conditions of a localized character, observations of a similar general
significance have been made. Such reports have been made, apart
from the Wright school, by Emery,14 Potter, Ditman, and Bradley,15
Potter,16 Tunnicliff,17 Whitfield,18 Cole and Meakins,19 and many
others, and we may say with reasonable accuracy that, in localized
infections particularly, there is much evidence to show that clinical
improvement and rise of the opsonic index go hand in hand.
There have been many exceptions to this — which, in view of the
complicated factors involved in immunization, as well as the diffi-
culty of the technique, is not surprising.
In tuberculosis — in which many of Wright's earlier studies were
made — the parallelism has not been so consistent. Thus even the
early work of Bullock 20 showed that, in contrast to similar staphy-
lococcus investigations, the tuberculo-opsonic indices of patients may
occasionally be higher than normal, and similar observations were
made by Lawson and Stewart 21 in cases of acute pulmonary tubercu-
losis.
Various investigations, too numerous to be reviewed in detail in
12 Wright and Reid. Lancet, 1906 ; Proc. Royal Society, Vol. 77, 1906.
13 Opie. Assoc. of Am. Phys., Washington, 1907.
14 Emery. "Immunity, etc.," Lewis,, London, 1909.
15 Potter, Ditman, and Bradley. Journ. A. M. A., Vol. 47, 1906, p. 1722.
16 Potter. Jour, of A. M. A., Vol. 49, 1907, p. 1815.
17 Tunnicliff. Jour, of Int. Dis., Vols. 4 and 5, 1907 and 1908. ,
18 Whitfield. Practitioner, May, 1908.
19 Cole and Meakins. Johns Hop. Hosp. Bull, Vol. 18, 1907.
20 Bullock. Transact, of Lond. Path. Soc., Vol. 56, 1905, and Lancet,
1905, Vol. II, p. 1603.
21 Lawson and Stewart. Lancet, 1905, Vol. II, p. 1406.
342 INFECTION AND RESISTANCE
this place, indicate in a general way that localized tuberculosis of
the skin, joints, intestines, and glands, with the patient quiet and at
rest, is apt to show a low index, while a high index may, under
such conditions, often point to an active pulmonary lesion. Accord-
ing to Wright, this depends upon the following factors: In a
localized lesion, with the body at rest — and when systemic symptoms
such as fever are absent — the focus is, very probably, quiescent and in
but slight communication with the circulation, even though it may
be slowly progressive. In such cases little or no antigen is being
discharged and, in consequence, no antibody formation is stimulated.
Indeed, even the small amount of antibody which is present comes
into but indifferent contact with the lesion because of its compara-
tive insulation from the body fluids. Such a lesion may be benefited
by rest, in that spreading is inhibited, and autointoxication, with
the production of a negative phase, prevented ; but it cannot be com-
pletely cured unless the antibodies are increased. This can be ac-
complished by carefully controlled vaccinations with tuberculin. At
the same time more effective contact of these antibodies with the
lesion may be attained by local applications, X-ray, etc. Or, again,
the same purpose may be accomplished by carefully controlled and
graded motion or massage of the diseased part — which may be used
both to increase the opsonin contents by auto-inoculation and to en-
hance the local circulation. If this is done with care it may serve to
substitute entirely for the treatment with vaccines.
On the other hand, such treatment with auto-inoculation, it must
be remembered, is entirely uncontrollable as to dosage, and, there-
fore, not to be generally recommended.22 In active pulmonary tu-
berculosis, when there are systemic symptoms such as rise of tem-
perature, the body is very probably already receiving excessive
amounts of antigen and vaccine treatment of any kind may be
dangerous.
However we analyze the work done on tuberculo-opsonins — and
the investigations on this subject are far too numerous to be here
reviewed — -we are forced to the conclusion that in this disease the
opsonic fluctuations are far more irregular than in most other condi-
tions. Much,23 for instance, found no regular differences between
the tubercle bacillus opsonins of healthy and of diseased individuals,
.and Koehlisch 24 obtained similar results, adding the important ob-
servation that animals that show a high natural resistance to the
liuman type of the tubercle bacillus invariably show an opsonic
index much lower than that of man.
We may question with much justice, therefore, whether in the
22 Meakin and Wheeler. Br. Med. Jour., 2, 1905.
23 Much. Munch, med. Woch., p. 496, 1908.
24 Koehlisch. Zeitschr. f. Hyg., Vol. 68, 1911.
OPSONIC INDEX AND VACCINE THERAPY 343
case of this bacillus opsonic investigations can be looked upon as
indicators of immunity with as much confidence as in cases of other
bacterial invasions. It is true, indeed, that tubercle bacilli — as well
as leprosy, rat leprosy, and other acid-fast bacteria — are eagerly
taken up by polynuclear leukocytes when they are injected into the
peritoneal cavity of a guinea pig or rat or other experimental ani-
mal. On the other hand, we have much evidence which seems to
show that such phagocytosis is not in these cases a direct method
of bacterial destruction. In another place we have cited the experi-
ments of Tschernorutski,25 which showed that polynuclear leuko-
cytes, though containing other ferments, were devoid of lipase. And
Carey and the writer — experimenting with rat leprosy bacilli — found
that these acid-fast bacteria were not disintegrated within leukocytes
in the course of weeks, while they were often subject to rapid de-
struction in the presence of living spleen cells in plasma. Further-
more, in the discussion of the tuberculin tests we have reviewed
evidence which points to the fact that in the reactions to
tubercle bacilli we have probably to deal more particularly with
sessile receptors on fixed tissue cells than with specific circulating
antibodies. Bartel and Neumann 26 have concluded that the phago-
cyte which takes up tubercle bacilli represents only a preliminary
vehicle by which the micro-organisms are conveyed to the spleen and
lymphatic tissues, in which actual destruction then takes place.
While no final conclusions can be drawn from the available evidence,
all these data render it uncertain whether the opsonic index as de-
termined for polynuclear phagocytosis may be at all regarded as a
reliable indication of increased or diminished resistance, and on
this basis the control of therapy in tuberculosis by opsonin estima-
tions is of course placed upon an uncertain basis.
We have then very briefly traced the work done upon opsonin
determinations from the purely practical point of view. There is
of course no question about the scientific accuracy of the observa-
tions upon which rests our knowledge of the opsonic properties of
blood serum. There is also no doubt concerning our ability to in-
crease the immunity of an individual by systematic treatment with
vaccines made of pure cultures of bacteria. However, the work of
Wright has concerned itself with two distinct questions which must
be separately answered. Briefly stated these are: 1. What is the
value of opsonic estimations in controlling the therapeutic vaccina-
tions of patients? 2. To what degree and in which particular
conditions may the process of vaccination (active immunization) be
regarded as a hopeful method of therapy ?
25 Tschernorutski. Hoppe-Seyler's Zeitschr. f. Phys. Chem., Vol. 75,
1911.
26 Bartel and Neumann. Wien. kl Woch.} Nos. 43 and 44, 1907; Cen-
tralbl f. Bakt., Vol. 48, 1909.
344 INFECTION AND RESISTANCE
The first question has, in part, been answered in the preceding
paragraphs. Reasonably accurate comparative estimations of the
opsonic properties of serum can unquestionably be made by Wright's
method, or some of its accepted modifications, in the hands of trained
workers who look upon each estimation as an experimental problem
and have time for control and repetition. That even in such cases
the matter is difficult is amply testified to by such reports as that of
E. C. Hort,27 who states that two of the most skilled experts 28 in
London, working with samples of the same serum taken before and
after vaccination, reported — "the one that the index was raised, the
other that it was lowered by the treatment." This, and similar ex-
periments of other observers, do not, of course, invalidate the results
obtained in special researches like those of Wright, Neufeld, and
others, but they do indicate that the control of clinical cases by
opsonic estimations is not a matter that can profitably be made a
routine procedure by which the treatment of the cases can be regu-
lated. As a problem of clinical research in a given series of patients
opsonin studies are unquestionably valuable and the comparative data
so obtained have proved, and will continue to prove, of great value.
But we cannot hope as yet, it seems to us, to utilize this method, ex-
cept in cases in which much time and care can be centered upon a
few patients under the best conditions. Opinions essentially similar
to this have been expressed by experienced clinicians (Potter,29 for
instance), who have followed out series of cases on which systematic
opsonin determinations were made.
As to the opsonic index in tuberculosis, we believe that the ex-
perimental evidence at present available does not show that such
measurements are reliable measures of resistance, and, in this dis-
ease, even when the index is taken with a degree of care which
precludes gross error, it is doubtful whether its estimation is of as-
much value in controlling treatment as are the data obtained by
skilled clinical observation.
This leaves us, therefore, for the control of vaccine treatment in
the routine work of the clinic only the information gleaned from
such indications as alterations in any visible or palpable lesions,
general systemic symptoms, temperature, leukocytosis, etc. Since
these will present such manifold and variable pictures in different
conditions, generalization is useless.
The second question concerning the value of vaccine treatment
in infectious disease of human beings cannot be so briefly answered,
and is one of the greatest importance in medicine. It is well known
27 Hort. Br. Med. Jour., Feb., 1909, p. 400.
28 Quoted from Adami, Trans. Amer. Phys. & Surg., Vol. 8, 1910. See-
also Pearson, Biometrica, 1911.
29 Potter. Loc. cit.
OPSONIC INDEX AND VACCINE THERAPY 345
that tuberculin therapy has come into carefully controlled use in
recent years only, although it was introduced early in the history
of specific therapy by Koch. The misuse and failure of this treat-
ment during the years following its introduction are easily explained
by the defective knowledge of antibody reactions and the general
principles of immunity — a condition which was removed only by
the subsequent assiduous work of numerous investigators. At the
present time the value of this method of treatment is being acknowl-
edged, though its limitations and possible dangers are properly
recognized. The Wright method of vaccine treatment is also an
unquestionably powerful therapeutic weapon, and yet, owing to com-
mercialization, unskilful application, and, more especially, because
of extensive attempts to apply it in unsuitable cases, it may easily,
like tuberculin therapy, enter into a period of neglect and disrepute.
It is very necessary to accentuate at the present time that the active
immunization of human beings with any form of bacterial product
is a serious procedure which requires painstaking and skilled control,
and should not be undertaken without the same degree of preliminary
experience and study which is considered prerequisite in any other
branch of specialized medicine.
Any opinion expressed regarding the ultimate value of a method
of treatment which is still undergoing active clinical investigation
must of course be purely tentative. Moreover, there are so many
differences of judgment that we wish to emphasize the purely per-
sonal character of the views expressed.
In passing judgment upon the value of active immunization in
man we must distinguish sharply between active immunization which
is prophylactic and that which is carried out after the disease has
gained a definite foothold in the body. In the former case we are
dealing with a new method and with one upon which the very foun-
dations of our knowledge of immunity have been built. It is the
method of Jenner in small-pox. It is that of Pasteur in chicken
cholera, in anthrax, and in many other infections. It has been used
as a routine in animal experimentation in laboratories since the first
days of the systematic study of infections. There is no question
about its being a rational and logical procedure. The immunity
which can be easily conferred upon a healthy individual in this
way need not be extensively above the normal in order to protect
from invasion by the small numbers of pathogenic germs which
may gain entrance under conditions of accidental, spontaneous in-
fection.
The possibilities of the method were recognized by Ferran, a
pupil of Pasteur, who applied it to cholera, and, since his time, it has
been extensively attempted in many of the infectious diseases which
occur epidemically, and therefore justify attempts in this direction.
346 INFECTION AND RESISTANCE
In essence also Pasteur's method of active immunization in rabies
represents such prophylactic vaccination, since, in this case, although
treatment is begun after infection has taken place, nevertheless the
process of immunization is carried out during the incubation period
before active manifestations of the disease have set in. Prophylactic
vaccination, therefore, is a valuable procedure which has reaped
remarkable results of recent years, especially in protection against
typhoid fever. In a subsequent section this phase of vaccination
is more extensively discussed, and we may therefore leave it for the
present.
In this place we are more particularly concerned with the prob-
lem of the treatment of existing disease with vaccines prepared
from the bacteria by which the disease is caused. In how far this is
justifiable or even logical is a question which depends upon the con-
ditions of each individual case. We can approach the problem best
by roughly classifying the various forms in which infection occurs
in the human being.
When bacteria gain entrance into the tissues of the human body,
granted that the organisms are pathogenic, an immediate struggle
ensues between the offensive properties of the micro-organisms and
the defensive powers of the tissues. The factors which determine
the outcome of such a combat have been more fully considered in
Chapter I. Briefly, if the defensive powers of the body greatly
preponderate the result is localization and rapid destruction of the
micro-organisms — with cure. In such a case any form of treatment
is unnecessary. On the other hand, the balance of power may be
turned in the opposite direction, in which case the infectious process
becomes rapidly generalized, the bacteria enter the blood stream
and lymphatics, and the defensive powers are overwhelmed. In,
such a case also active immunization with vaccines is entirely use-
less.
There are cases, however, in which the struggle is a more equal
one, and in which the infectious process is held in check by the
defenses, so that it takes a slow, chronic, localized form, and spreads,
if at all, very slowly. What is it in such a case that prevents com-
plete healing of the process ? The answer to this may be found both
in local and in systemic causes. Locally the lesion, after the pre-
liminary skirmishes, may become encapsulated either by fibrin
formation, clot, or other tissue changes so that, as Wright suggests,
the fluid constituents of the blood-plasma cannot easily approach the
organisms in the lesion. The same effect may result from internal
pressure by fluid and possibly by the presence of considerable quan-
tities of tissue detritus, by which protective serum constituents are
fixed and thus diverted from the bacteria. Against these factors, of
course, no form of immunization can be of value. Wright recog-
nizes this, and suggests the use of surgical evacuation, Bier's method,
OPSONIC INDEX AND VACCINE THERAPY 347
X-rays, Finsen light, heat, and a number of other localized methods
of increasing the blood supply. This, too, may be the reason for
the benefits derived from wet dressings, in that they keep the tissues
macerated, soft, and moist. At any rate, it is a matter of local
surgical treatment. At the same time, however, there may be sys-
temic causes which prevent the complete healing of such lesions,
namely, an insufficient supply of circulating antibodies, opsonic or
bactericidal substances. These may be sufficient to hold the lesion
in check, but since small quantities of bacteria only are in contact
with the blood stream, relatively small amounts of antigen are ab-
sorbed and antibody formation is consequently deficient. Here we
have an ideal condition for vaccine therapy. By isolation of the
organisms from the patient's lesion, for which, in this case, there
is time, and the careful immunization of the patient with these
organisms, the immunity may be considerably increased and cure
effected.
Closely related to this type of lesion are those conditions in
which there are localized infections which heal rapidly but recur in
quick succession again and again. Such are the common cases of con-
secutive crops of boils ; and not dissimilar are the manifestations of
erysipelas where the lesion extends along the edges while it heals in
the center. There is in this type, probably, a very close balance
between protection and offense ; the defensive reaction is sufficient to
overcome the localized lesion, but insufficient to set up a permanent
systemic protection. A certain amount of local immunity acquired
by the tissues of the affected areas may suffice to throw slight weight
into the balance on the side of protection, enough at least to decide
the struggle ; and this element of locally acquired tissue resistance is
in all probability also the cause for the failure of these lesions to
recur immediately in the same area. Here, too, treatment with vac-
cines is not illogical and may yield good results if properly carried
out.
In generalized systemic infections we must sharply distinguish
between cases of acute sepsis in which the bacteria are actively grow-
ing and multiplying in the circulation and cases in which blood cul-
tures are positive only because the bacteria are being constantly
discharged into the circulation from a focus in the tissues. In the
former the defenses of the body are overwhelmed by an extensive
flooding with the bacteria, and vaccines, if not harmful, are, at any
rate, utterly useless since the antigen is already so extensively dis-
tributed throughout the tissues that if the body were capable of
responding with sufficient antibody formation this would unques-
tionably occur without the small additional amount furnished in the
bacterial emulsion. Vaccination in such cases is entirely analogous to
an attempt to stimulate a degenerated heart muscle with strychnin —
the whipping of a tired mare.
348 INFECTION AND RESISTANCE
Such cases of septicemia, however, are not in our opinion the
most common ones in the human being. It is probable that all
localized infections of more than a very trifling nature discharge
living bacteria into the circulation from the very beginning. How-
ever, in most cases the bacteria, though able to hold their own in
their entrenched position at the focus where accumulated offensive
factors and local injury reenforce them, are yet rapidly destroyed
when, in small detachments, they get into the open circulation where
the plasma antibodies and phagocytes are freely active. There are
cases which take a middle course between such purely localized
lesions and the acute septicemia, conditions in which a well-estab-
lished focus continues to furnish bacteria to the blood stream as fast
as they are destroyed. An example which illustrates our meaning
well is that of the so-called subacute endocarditis caused by the
Streptococcus viridans and its close biological kin, where blood cul-
tures are often consistently positive for a long period or may show
occasional intervals in which the blood is bacteria-free. The focus
on the heart valves apparently can continue uncured in spite of a
relatively high or at least normal systemic resistance to the micro-
organisms. If, as we ourselves have done, we isolate the organisms
by blood culture from such cases, and then measure the opsonic prop-
erties of the patient's own serum against them, using the patient's
own leukocytes, we may often find that active phagocytosis takes
place, in a degree equal or even superior to that taking place in the
serum of normal individuals. Neither does there seem to be a dimin-
ished phagocytic power of the patient's own leukocytes. For a long
time these conditions may continue, with a constant destruction of
bacteria in the blood and a corresponding renewal of the supply from
the lesion. The same condition can be observed in rabbits in which
chronic endocarditis with persistently positive blood culture has been
produced by injections of these bacteria. In such animals measure-
ments similar to those described above have been made by Miss Gil-
bert in our laboratory, and it has seemed as though persistently
positive blood cultures could be obtained only when a localized focus
was set up in the animals. Unless this is the case the blood cultures
rapidly become negative.
Conditions essentially similar may exist in any ' other form of
severe localized infection. Positive blood cultures do not necessarily
mean a multiplication of the bacteria in the blood stream and a rapid
overwhelming of the body. We have had occasion to see a number of
cases of bacteriemia in which the focus of infection was surgically
accessible ; and in some of these cases early removal of the focus an(
purely surgical treatment resulted in a clearing up of the infection
Similar experiences have been reported by Libman and a number of
others, and for this reason general septicemia, if not fulminating,
may still be less desperate than ordinarily supposed.
OPSONIC INDEX AND VACCINE THERAPY 349
Now, having outlined the conditions obtaining in such cases, let
us briefly consider whether, under the circumstances, vaccine therapy
may logically be regarded as a hopeful form of treatment. We may
assume, on the one hand, that the bacteria, being consistently present
and destroyed in the blood, should furnish antigen sufficient to
stimulate the body tissues to their utmost reactive ability. This
would seem a strong argument against vaccine therapy. On the
other hand, we must take into consideration another phase of the
subject, one which has some experimental justification. In discus-
sing the origin of antibodies in another section it will be remembered
that we called attention to the fact that many different tissue cells
probably participate in the production of these protective reaction-
bodies. We cited an experiment of Wassermann and his pupils in
which they proved that antibodies were produced most energetically
in the tissues about the point of injection of the antigen, namely, in
the place at which it came into most concentrated contact with the
cells. They injected bacteria into the subcutaneous tissues of the ear
of a rabbit, measured the progressively increasing appearance of
antibodies in the blood stream, and then amputated the ear. A
sudden drop of antibody contents followed, showing that the supply
of antibodies had largely emanated from the tissues surrounding
the injection point. Park 30 has pointed out another reason why
vaccine treatment may be expected to exert beneficial action in such
cases. He calls attention to the fact that when very large amounts
of antitoxin are added to toxin before injection no antibody produc-
tion results, and assumes that in chronic or subacute general infec-
tions the circulating bacteria are in contact with specific antibodies,
partially "sensitized," and therefore not efficient as antigen. In
consequence the injection of homologous unsensitized bacteria may
hasten antibody formation. This assumption of Park is theoretically
valid, but it is not in accord with the more recent experiments of
Metchnikoff and Besredka, who claim to have obtained the best
results in prophylactic typhoid vaccination by the injection of sensi-
tized bacteria.
Thus the use of vaccines in the subacute or chronic cases of in-
fection with bacteria in the blood stream may be theoretically justi-
fied, and no one can say at the present time whether or not it has
therapeutic promise. At any rate, it cannot be absolutely condemned
on theoretical grounds.
Like so many other phases of this question, it must be answered
ultimately by clinical experience, for in experimentation upon ani-
mals, while it is easy to produce a purely localized lesion followed
"by rapid healing, or a generalized lesion leading to rapid death, it
is not easy to produce prolonged infections with anything like regu-
larity, and there are so many modifying accidental factors which
30 Park. Trans, of Amer. Phys., Vol. 8, 1910.
350 INFECTION AND RESISTANCE
influence the course of such infections in animals that the results
of vaccine treatment in them are difficult to judge.
In acute diseases which run a definite course, typhoid fever,
pneumonia, dysentery, cholera, plague, and a number of other con-
ditions, vaccine treatment during the course of the disease has not
much theoretical justification. In typhoid fever, especially, specific
antibodies appear in the blood in amounts enormously increased
above the normal at periods when the patient is still actively ill in
spite of the fact that the blood stream has been freed of the micro-
organisms. Whatever may be our opinion as to the continuance of
the disease after bacteria have been driven out of the blood stream,
the use of vaccines can only tend to further increase of antibodies
which are already present in amounts far exceeding normal. In
pneumonia the microorganisms seem curiously resistant against the
attack of the serum antibodies, and in spite of the presence of large
amounts of antigen both in the lungs and, for a time, in the circula-
tion the development of immunity is delayed until just before or near
the crisis. Since this, however, is usually only a matter of 7 or 8
days, it is hardly likely that the injection of vaccines during this
period could markedly alter the ultimate outcome. In a later sec-
tion we shall see that vaccine treatment in typhoid fever is neverthe-
less being extensively tried and gives reactions of a nature which
cannot entirely be explained on the basis of the above considerations.
As we have said before, the opinions expressed above are given,
with the purpose of stating as clearly as we can the logic of vaccine 32
therapy as we see it at present. The next ten years of clinical ex-
perience may largely modify these views. One thing is certain, how-
ever, and that is that the problem can only be settled if treatment by
this method is undertaken with the guidance of an accurate bacteri-
ological diagnosis, and with bacteriological control of the individual
case, so that, when occasion arises, estimations of antibodies can be
made.
To protest against the random use of commercial stock vaccines
without laboratory diagnosis and without control is almost a plati-
tude.
In the case of tuberculosis the problem had been actively investi-
gated before Wright, and there seems little question that tuberculin
therapy properly and cautiously applied has an established value in
the treatment of initial and localized tuberculous disease. Whether
81 See also Theobald Smith, Jour. A. M. A., Vol. 60, 1913, and R. M.
Pearce, Jour. A. M. A.} Vol. 61, 1913.
32 For discussion of various clinical applications of vaccine treatment
see symposium on vaccine treatment, Trans, of Ass'n of Amer. Phys. and
Surg., Vol. 8, 1910.
OPSONIC INDEX AND VACCINE THERAPY 351
or not its use in actively progressive tuberculosis may or may not be
hopeful, in which particular cases, and by what methods, it is to be
applied, these are problems that we have neither the space to deal
with nor the experience to summarize properly. They constitute a
special field of clinical research, a survey of which may be obtained
in such works as that of Bandelier and Roepke,33 or the more espe-
cial experimental studies of Denys.34
THE PRODUCTION AND STANDARDIZATION OF VACCINES
Vaccines in the sense of Wright consist merely of killed cultures
of the bacteria with which the patient is infected. In all cases it is
extremely desirable to make such vaccines "autogenous," by whicK
we mean that the organism used is one which has been isolated from
the case. The difference between various strains of the same species
of bacteria seems to make this imperative whenever it is at all pos-
sible. The recent investigations of Neufeld and Haendel in de-
termining that there are a number of types of pneumococcus which
are antigenically distinct illustrates this point. The same principle
is made clear by the recent work of Rosenau on the streptococcus-
pneumococcus group. Especially important is Rosenau's observa-
tion that a pneumococcus which he had been able to transform cul-
turally by special methods was found to be altered also in its reaction
to agglutinins.
In the development of prophylactic methods of vaccination
against epidemic disease like typhoid, cholera, plague, etc., many
different methods of antigen preparation have been developed. In
typhoid prophylaxis the bacteria have been used dead, living, and
sensitized, and even extracts have been employed. In cholera the
early use of living cultures by Ferran has given way, in the hands
of Kolle and others, to that of dead bacterial emulsions. In plague
and a number of other conditions the impression seems to be general
that the bacteria should be used in the living, but attenuated, state.
Special methods which have been developed in these cases are dis-
cussed in another section.
In treatment of developed diseases with vaccines the method most
commonly used is that which has been introduced by Wright, namely,
the use of dead cultures. In his earlier experiments Wright culti-
vated the bacteria on agar slants for about 24 hours, then washed off
the growth with 10 c. c. of sterile salt solution. It will be well to
describe in detail the preparation of such a vaccine.
The bacteria must be isolated from the patient by the usual
method of plate cultivation and colony fishing on suitable media.
33 Bandelier and Roepke. "Lehrbuch der spez. Diagnostik und Therapie
der Tuberkulose," 6th Ed., Kabitzsch, Wiirzburg, 1911
34 Denys. "Le Bouillon Filtre," Louvain, 1903.
INFECTION AND RESISTANCE
We do not think that any satisfactory substitute for careful isolation
by plating has been devised. After a pure culture of the organism
has been obtained this is grown on relatively large surfaces of agar,
glucose-agar, or ascitic agar, as the case may require. These culti-
vations may be made in Kolle flasks or, as Wright and
others have suggested, on large agar surfaces obtained
when the culture medium is allowed to harden in a
square 3-oz. medicine bottle laid on its side. Any de-
vice of this kind in which a large surface of agar is
exposed may be used.
After suitable growth of the micro-organisms has
taken place, 24-48 hours, the growth is gently washed
off with 10 c. c. or more of sterilized salt solution.
Care must be taken to do this in such a way that no
agar is drawn away with the emulsion. The thick
emulsion so obtained is removed from the culture bottle
with sterile nipple pipettes or Pasteur pipettes and
transferred to a sterile thick-walled test tube into which
glass beads have been placed. By drawing out the
neck of this test tube in the flame a glass capsule is
formed, in which we now have our so-called stock
emulsion. (See figure.)
The next thing to be done is to standardize this
stock emulsion, or, in other words, determine approxi-
TO HOLD STOCK mately the number of bacteria to the cubic centimeter.
VACCINE There are a number of methods by which this can be
EMULSION ,
FROM WHICH clone.
D i L u T i ONS The method most extensively used by Wright and
^Q followers was that in which the bacteria are counted
against red blood cells. The bacteria in the capsule
are snaken thoroughly with glass beads so that clumps
until may be broken up and even distribution obtained. A
cooled little of the emulsion is then put into a clean watch
it may crack glass, a step which can be accomplished most easily by
when quickly breaking off the tip of the drawn-out part of the cap-
sule, tilting it very gently and heating the closed end
over a small flame, so that some of the emulsion will be
driven out by the expanding air. With a nipple pipette marked
about an inch from the tip, as in the taking of an opsonic index, a
little of the emulsion is drawn up. This is placed into another clean
watch glass and is mixed with about 2 volumes of salt solution and
one volume of blood from the finger, these quantities being measured
with the same nipple pipette. We then have a mixture in which,
in a total of 4 volumes, there are equal parts of blood and of bac-
terial emulsion. After this emulsion has been thoroughly mixed by
drawing in and out through the nipple pipette smears are made on
CAPSULE MADE
OF TEST TUBE
ARE MADE.
(It is well to
open to
lary tip
it has
OPSONIC INDEX AND VACCINE THERAPY 353
slides and stained with Jenner or any other suitable blood and bac-
terial stain. Under the field of the microscope the ratio between
the bacteria and blood cells is then determined, and from our
knowledge of the number of the red blood cells in this blood to each
c. mm. we can easily calculate the number of bacteria to the c. mm.
or c. c.35
A more accurate method of enumerating the bacteria in a sus-
pension to be used for vaccine is by direct count of an accurately
made dilution in a hemocytometer chamber, as was first suggested
by Malory and Wright in 1908.36 The bacterial suspension is diluted
in blood-counting pipettes,
1-20 to 1-100 dilutions of
thick bacterial suspensions
being as a rule satisfactory.
As a diluent one may use
either salt solution or some
dilute anilin dye, such as one
made by mixing one part
alcoholic methylene blue
with 40 parts of 1 per cent,
carbolic acid. The dilute
suspension is then placed in
an ordinary Thoma-Zeiss
chamber, which was de-
signed for counting blood
platelets and has a depth of
0.02 mm. This enables one
to use an oil immersion lens
or high power dry system
with a short working dis-
tance. From such a count one may readily estimate the num-
ber of bacteria in the original suspension ; for example, if 20
squares in the Helber-Zeiss chamber are counted the result gives the
number of bacteria in 0.001 c. mm.37
Another method of standardization of vaccines wThich is suffi-
ciently accurate for clinical purposes is that of Hopkins, which con-
sists in measuring the volume of the sediment 38 after centrifugaliz-
ing the preparation under standard conditions in a graduated tube.
The tubes may be made with a capacity of 10 to 15 c. c. with a capil-
35 For such counts it is convenient to contract the field of the microscope
by using- a diaphragm or simply marking a circle on the eyepiece with a
grease pencil.
36 Malory and Wright. "Pathological Technique," 4th Ed., New York,
1908.
37 Glynn, Powell, Rees, and Cox. Jour, of Path, and Bact., Vol. 18, 1914,
p. 379.
38 Hopkins. Jour. A. M. A., 1913, Vol. 60, p. 1615.
MICROSCOPIC FIELD AS SEEN IN STANDARDIZA-
TION OF VACCINES BY WRIGHT'S METHOD.
354 INFECTION AND RESISTANCE
lary tip about one inch in length, having a capacity of about 0.05
c. c. graduated in 0.01 c. c. The bacterial suspension, after being fil-
tered through sterile cotton to remove fragments of the agar or other
foreign bodies, is centrifugalized in such a tube for half an hour at
about 2,800 revolutions a minute. The supernatant fluid and bac-
teria are removed down to the 0.5 c. c. mark and the sediment resus-
pended in 5 c. c. sterile salt solution by means of a capillary pipette
which gives a 1 per cent, suspension. 0.05 c. c. of streptococci sedi-
mented in this way represent quite constantly 16 mm. of dried bac-
terial substance. The number of organisms per cubic centimeter
contained in 1 per cent, suspension in this way are as follows :
Streptococcus aureus and albus 10 billion
Streptococcus 8 "
Gonococcus 8 "
Pneumococcus (capsulated) 2.5 "
Bacillus typhosus 8 "
Bacillus coli. . 4 "
After the vaccines have been standardized suitable dilutions can
be made in salt solution to which 0.5 per cent, carbolic acid or some
other antiseptic has been added. The dilutions are usually so made
that from 100 to 500 million bacteria are contained in the cubic cen-
timeter, this being a suitable initial dose of most organisms. The
dilutions are placed in sterile bottles containing beads and fitted
with rubber caps. These bottles can be shaken
before use, the emulsion thoroughly distributed,
and the desired quantity can be taken out with a
sterile hypodermic syringe thrust through the
rubber cap after this has been covered with a
small amount of lysol or strong carbolic (see fig-
ure). After the dilutions have been made both
these and the stock vaccines should be sterilized.
Some workers sterilize always the stock vaccines
VACCINE STOCK and make the dilutions with aseptic proportions
EMULSION IN j ^ ^ t further sterilization is
EUBBEB TOP J. . £ ,, ,
BOTTLE. necessary. This is preferable because the less
heat that is applied the better it is for the
preservation of their antigenic properties — sterilization is usually
accomplished by heat in the water bath. Wassermann's earlier
technique called for heating to 60° C. for one hour for a
number of consecutive days. It is generally considered at the
present time that it is better not to heat above 55° C. After the
vaccine has been heated its sterility must be controlled to aerobic
and anaerobic cultivation, and possibly by animal inoculation, al-
though, except in special cases, this is unnecessary. Some workers,
OPSONIC INDEX AND VACCINE THERAPY 355
especially when the vaccine is to be extensively used, as in typhoid
immunization, inject some of the vaccine into white mice to exclude
the possibility of contamination with tetanus. In such cases also it
is not inadvisable to test out the antigenic value of the vaccine upon
animals, measuring the agglutinins, etc., which result from a number
of inoculations. In the preparation of a therapeutic vaccine where
speed is required this of course is not feasible. Moreover, it is un-
necessary in view of the fact that we wish to inject that particular
organism into the patient from whom it has been cultivated. What-
ever its antigenic value may be from animal experiments, it is pre-
ferable for the given purpose to any other strain.
Sensitized vaccines are easily made by exposing emulsions of the
bacteria to moderate amounts of a strong immune serum which has
been heated to 56° C. to destroy the complement. Bacteria will
usually agglutinate under these circumstances and can easily be
centrifugalized to the bottom. The excess serum is then washed off
and the bacteria emulsified as in the case of the preparation of vac-
cines with dead organisms.
THE TUBERCULINS
Since we shall not attempt to discuss critically tuberculin treat-
ment, as this is a subject upon which many special studies have been
made both by clinicians and by laboratory workers, and is entirely
too extensive to be reviewed in a book like this, on the other hand,
we deem it a, part of our task to discuss at least the methods by which
the antigen or tuberculin preparations are obtained. There has been
much discussion concerning the nature of the antigenic substances
obtained from the tubercle bacillus. It has been claimed by Denys
and others, for instance, that the tubercle bacillus may give rise to
small quantities of a true exotoxin with consequent endotoxin-
inducing properties. Again, most observers have believed that the
poison of the tubercle bacillus consists of substances comparable to
the endotoxin of other micro-organisms. The matter is by no means
settled, and without going into the theoretical aspects of the problem
we will confine ourselves in this place to a description of the pro-
duction of the various forms of so-called "tuberculin."
OLD TUBERCULIN (Keen)
The first tuberculin prepared by Koch is made in the following
way : Tubercle bacilli of the human type are grown for from 4 to 6
weeks upon a 5 per cent, glycerin broth. The cultures are then
sterilized in an Arnold sterilizer and are evaporated at about 80° C.
356 INFECTION AND RESISTANCE
to one-tenth the original volume. This 60 per cent, glycerin extract
of the tubercle bacilli is then filtered clear and constitutes the tuber-
culin.
The old tuberculin is a preparation which is extensively used
in the subcutaneous and intracutaneous tests upon human beings and
cattle, and forms the basis of the various preparations by von Pir-
quet, Moro, and others in the cutaneous tuberculin reactions. In
his earliest work von Pirquet used a 25 per cent, solution of the old
tuberculin. At present an undiluted old tuberculin is used for these
purposes.
The old tuberculin also is the material from which the prepara-
tion for the ophthalmotuberculin test is made. For this purpose
Calmette advises precipitating old tuberculin with double the volume
of 95 per cent, alcohol, allowing the precipitate to settle and repeat-
edly washing the sediment with 70 per cent, alcohol. The powder
which results is thoroughly dried, pulverized, and made up for use
in 0.5 per cent, solutions. Bandelier and Roepke recommend the
use of the diluted old tuberculin directly for these tests, employing
a 1 per cent, solution.
TUBERCULIN (T R AND T O)
The description of the preparation of these tuberculins we take
from Ruppell in the Lancet, March 28, 1908. Virulent cultures of
tubercle bacilli are dried in the vacuum and are then thoroughly
pulverized by specially constructed machinery, and the grinding is
continued until no intact bacilli are found in the preparation. One
gram dry weight is then shaken up in 100 c. c. of sterile distilled
water. The mixture is then centrifugalized at high speed — the
supernatant fluid is T O (tuberkulin oberschicht). This contains
the water-soluble substances of the bacillus and gives no precipitate
with glycerin. The residue — T R (tuberkulin ruckstand) — is again
dried and ground up, shaken up in water, and centrifugalized. This
is repeated 3 or 4 times, the total volume of water used for all the
repetitions not exceeding 100 c. c. At the end of several repetitions
.all the T R goes into emulsion, and the various supernatant fluids
obtained during these repeated grindings and shaking are mixed
together and constitute the final T R preparation. This preparation,
according to Koch, contains important antigenic substances, it gives
a precipitate with glycerin, and it is standardized by the determina-
tion of the solid substances contained in a cubic centimeter. This,
for a standard preparation, should be 0.002 gram to a cubic centi-
meter.
OPSONIC INDEX AND VACCINE THERAPY 357
TUBERCULIN BACILLARY EMULSION
This preparation consists of a combination of T O and T R. It
represents an emulsion of pulverized tubercle bacilli in 100 parts of
50 per cent, glycerin. The preparation as marketed contains 0.005
gram solid substance to the cubic centimeter. It is prepared simply
by mechanically grinding the bacteria as in the new tuberculin, but,
instead of centrifugalizing for the separation of T O and T R, the
bacteria are allowed to sediment after the addition of glycerin. This
is the preparation which is extensively used in many places at pres-
ent for the treatment of tuberculosis. It was adopted by Koch par-
ticularly because of experiments in which he showed that the treat-
ment of animals with such preparations greatly increased the ag-
glutinins for tubercle bacilli.
BOUILLON FILTRE (DENYS)
Denys cultivates the tubercle bacilli upon 5 per cent, glycerin
bouillon as in the preparation of old tuberculin, but does not heat,
sterilizing his cultures by filtration through porcelain. Denys be-
lieves that the application of heat in sterilization destroys exotoxins
which have valuable antigenic properties.
SENSITIZED TUBERCULIN
Following the introduction of sensitized vaccines in other dis-
eases by Besredka, Meyer 39 has introduced the sensitized tuberculin.
This tuberculin is prepared in the following way : Tubercle bacilli
of the human type are washed and dried and are mixed with a con-
siderable quantity of the serum of animals immunized with tubercle
emulsions and containing considerable quantities of tubercle-agglu-
tinins. These serum mixtures are kept at 37° C. for several days
and are then shaken in a shaking machine until intact tubercle bacilli
are no longer to be found. The tubercle bacillus fragments are then
thrown down in the centrifuge, washed in salt solution, and emulsi-
fied in 40 per cent, glycerin, 0.5 per cent, carbolic acid being used.
The emulsion contains 0.005 gram dry weight to a cubic centimeter.
We take the description of the preparation from that cited by
Bandelier and Roepke.
The above tabulation contains the most important tuberculin
preparations as they are at the present time in use. For detailed
studies of their clinical application we refer the reader to the very
valuable book of Bandelier and Roepke, "Lehrbuch der spezifischen
Diagnostik und Therapie der Tuberkulose," Curt Kabitzsch, Wiirz-
burg.
39 Meyer. Cited from Bandelier and Roepke, "Lehrbuch d. spez. Diagn.
u. Ther. d. Tuberkulose," Kabitzsch, Wiirzburg, 6th ed., 1911, p. 186.
CHAPTER XV
ANAPHYLAXIS
FUNDAMENTAL FACTS
THE fundamental principle of active immunization is the fact
that the treatment of animals with bacteria or bacterial products,
carried out according to certain empirically determined methods,
leads to increased tolerance or resistance. The limitations within
which this statement is true, and the variable factors to which it is
subject, we have considered in the foregoing discussions dealing with
the antibody-antigen reactions.
Although these reactions were studied at first purely from the
point of view of increased resistance to infection, the most extensive
studies of antibody formation have been made with such antigens as
blood cells, serum, and other substances which are in themselves en-
tirely harmless. For, in such reactions, great simplicity and ease of
experimentation could be attained. For a time, therefore, the pri-
mary problem of increased tolerance or resistance was relegated to a
secondary position, or, at least, dealt with chiefly by analogy, and the
phenomena of increased antibody formation and increased resistance
to the antigen were assumed to maintain a more or less strict paral-
lelism.
That the problem is not as simple as this has gradually become
obvious. We have come to recognize that the treatment of animals
with any antigen, bacterial or otherwise, though leading to increased
tolerance under certain conditions and within definite limits, may,
under other conditions, give rise to the very opposite, that is, to an
intolerance or increased susceptibility.
The development of this knowledge, like much else that serum
study has revealed in the last fifteen years, takes root in isolated
observations scattered throughout the early literature, but often
regarded as merely noteworthy accidents or technical errors. This
particular problem, moreover, was confused by the fact that some of
the earliest observations regarding hypersusceptibility were made in
the course of experimentation with diphtheria and tetanus toxins,
antigenic substances toxic in themselves and, therefore, as we shall
see, clouding some of the basic principles apparently involved in the
phenomenon of which we now speak as anaphylaxis. We will for the
present, therefore, limit our discussion to the development of the
658
ANAPHYLAXIS 359
knowledge of anaphylaxis merely as it concerns the hypersuscepti-
bility incited in animals and man by treatment with various antigens,
such as animal sera and other proteins, which possess but slight
native toxicity or no toxicity whatever in themselves.
The special problem of toxin hypersusceptibility ( "Gif tuber emp-
findlichkeit" of von Behring) we will deal with later in a separate
section, since it is as yet very doubtful whether these phenomena
may justly be incorporated with true anaphylaxis as we now define
it, despite the admitted fact that attention was called to the prob-
lems of acquired susceptibility largely because of these toxin in-
vestigations.
The earliest observation having direct bearing upon protein
anaphylaxis is one which Morgenroth discovered in the writings of
Magendie. Morgenroth l mentions that, in his "Vorlesungen iiber
das Blut," published in 1839, Magendie describes the sudden death
of dogs which had been repeatedly injected with egg albumen. Al-
though Morgenroth, whose paper was written before the present
facts regarding hypersusceptibility were fully developed, attributes
these results to the action of precipitins, there can be little doubt as
to the anaphylactic nature of Magendie's results.
A clear statement of the fundamental phenomena was given, also,
by Flexner,2 in 1894. In describing certain experiments he says:
" Animals that had withstood one dose of dog serum would succumb
to a second dose given after the lapse of some days or weeks, even
when this dose was sublethal for a control animal."
One of the experiments cited to justify this statement is as
follows :
"Two rabbits received % of 1 per cent, and 1 per cent, of their
body weight respectively of dog's serum, twenty-four hours old, on
January 19, 1894. With the exception of hemoglobinuria, indisposi-
tion to move, and increased respiration, no ill effects were noted.
The animals still showed hemoglobinuria on the following day.
These symptoms disappeared and apparently the rabbits entirely
recovered. On February 12, 1894, each received 1 per cent, of their
body weight of dog's serum intravenously. A control animal also
received 1 per cent, of its body weight of the same serum. The two
animals that had been previously inoculated died in two and twelve
hours respectively; the control animal showed only hemoglobinuria
which disappeared after a day or two."
The experiment here quoted is, as a matter of fact, a perfect ex-
ample of what we now know as "active sensitization."
However, the isolated observations recorded above were neither
correlated nor followed out to their logical developments, and a
1 Morgenroth. "Ehrlich Gesamme'lte Arbeiten," Transl., Wiley & Son,
N. Y., 1906; p. 332 footnote.
2 Flexner. Medical News, Vol. 65, p. 116, 1894.
360 INFECTION AND RESISTANCE
systematic and purposeful study of the problem was deferred until
Richet and Portier 3 attacked it in 1902.
Richet and Hericourt4 had observed in 1898 that dogs treated
with eel serum, which is toxic per se, could be killed by a second
injection of an amount too small to injure normal untreated animals.
Some years later Richet, in collaboration with Portier,5 determined
a similar fact in the case of a poisonous substance, "actinocongestin,"
which they isolated by extraction of the tentacles of actinia.
Some of the facts of Richet and Hericourt's observations are as
follows : Actinocongestin injected intravenously into dogs in quan-
tities of 0.05 to 0.075 gram per kilo weight may cause illness, with
vomiting, diarrhea, and respiratory distress, but does not kill. A
dose of 0.002 gram per kilo causes no symptoms in a normal dog.
If, however, 0.002 gram of the poison is injected into a dog which
has previously received a sublethal dose and recovered, the result
is violent illness and often death. It was obvious, and this was
clearly stated by Richet, that the first dose had induced a condition
of markedly greater susceptibility to the poison.
He, therefore, spoke of the phenomenon as "anaphylaxis" ("ac-
tion anaphylactique de certains venins") to express its 'antithesis to
prophylaxis or protective effects.
Although it has been disputed by a number of writers that
Richet's investigations constitute the beginnings of our modern
understanding of the anaphylactic phenomena, yet his recognition
of the distinct dependence of the hyper susceptible condition upon a
preceding inoculation with the same substance, and his conclusion
that a definite incubation time must elapse after the first injection
before susceptibility is developed, defined two of the most important
criteria of the condition and initiated purposeful investigations in
this field. It is true, on the other hand, that, like v. Behring and
most of his other predecessors, he was working with primarily toxic
substances, and the final recognition of the general biological sig-
nificance of the anaphylactic phenomenon was necessarily deferred
until a similar development of hypersusceptibility was noted in
animals injected with various antigens which of themselves were
entirely harmless. In this the history of anaphylactic investigations
is similar to that of other reactions to antigen injections, lysin, ag-
glutinin, and precipitin formation, in which the first observations
were made upon pathogenic bacteria or their products, and in which
subsequent extension of the investigations revealed that the response
to inoculation with bacterial proteins represented merely a single
phase of a general biological reaction on the part of animals to treat-
ment with the large class of substances known as antigens.
8 Richet and Portier. C. E. de la Soc. Biol, p. 170, 1902.
4 Richet and Hericourt. C. E. de la Soc. Biol, 1898.
5 Portier and Richet. C. E. de la Soc. Biol, p. 170, 1902.
ANAPHYLAXIS 361
This generalization of Richet's observations had really been
foreshadowed by the observations of Magendie and by the experi-
ments 'of Flexner quoted above, but this work had been lost sight .
of and the attention of investigators was again focused upon the
problem mainly by the publication of Arthus 6 in 1903 on the re-
peated injection of horse serum into rabbits, and some observations
made upon guinea pigs by Theobald Smith and communicated by
him in 1904 to Ehrlich.
Arthus 7 found that horse serum injected into rabbits by any of
the usual paths of entrance is entirely innocuous. It is possible to
inject 10, 20, or even 40 c. c. without harm. If, however, one re-
peatedly injects small amounts, 5 c. c. or less, subcutaneously, at
intervals of several days, eventually the later injections will give 7
rise to infiltrations, edema, sterile abscesses, and even gangrene at/
the points of injection. He recognized that this was not due to
cumulative action, and that it was not necessary to inject several
times in the same place to produce the characteristic response. For
instance, the early injections might be made into the peritoneum,
the subsequent ones into the skin, and the local reactions to the later
injections might nevertheless ensue. In other words, he recognized
the systemic nature of the phenomenon and regarded it as analogous
to the observations of Richet in that he spoke of the hypersensitive
rabbits as "anaphylactises" by a series of preparatory injections.
The "phenomenon of Theobald Smith" is closely related to that
of Arthus, and was made in the course of the standardization of
diphtheria antitoxin in guinea pigs. It was noticed that guinea
pigs which had been used for this purpose and had survived had
acquired great susceptibility to subsequent injections of normal horse
serum made several days or weeks later.
With these observations as points of departure, together with the
studies of v. Pirquet and Schick 8 upon the clinical manifestations
of antitoxin injections into human beings, a number of investigators
took up the problem, chief among them Rosenau and Anderson, of
the United States Hygienic Laboratory, and R. Otto, of the Frank-
furt Institute of Experimental Therapy.
Although the paper of Otto 9 appeared in print a little earlier
than did the first one of the American workers, the investigations
were independent and almost synchronous. Their results, moreover,
confirm each other in all essentials. Otto showed that the Theobald
Smith phenomenon was entirely independent of the toxin or anti-
6 Arthus. C. R. de la Soc. Biol., Vol. 55, p. 817, Reunion biol., Marseille,
June, 1903.
7 Arthus et Breton. C. E. de la Soc. Biol, 55, p. 1478.
8 Von Pirquet u. Schick. "Die Serumkrankheit," Deuticke, Wien, 1906.
9 Otto. "Das Theobald Smithsche Phaenomen, etc., v. Leuthold Gedenk-
schrift," Vol. 1. 1905 ; also Otto in Erganzungsband 2, "Kolle u. Wassermann
Handbuch," etc.
362 INFECTION AND RESISTANCE
toxin contents of the injected serum, but could be produced (though
somewhat less markedly) with horse serum alone. He also showed
that, while a preliminary injection of horse serum "sensitized" a
guinea pig to a subsequent dose given after an interval of 10 to 12
days, the repeated injection of considerable quantities at short inter-
vals produced a condition of "antianaphylaxis" or immunity to the
later injections. Otto, too, excluded from his results the direct rela-
tion of the anaphylactic state with the possible presence of serum
precipitins, a thought suggested by Morgenroth in his interpretation
of the observations of Magendie mentioned above.
Eosenau and Anderson 10 had attacked the problem with the pri-
mary purpose of throwing light upon the occasional accident of sud-
den death following the injection of diphtheria antitoxin into human
beings. Since the detailed description of their extensive investiga-
tions would tend to render more difficult the exposition of an already
sufficiently complicated subject, it will be best to tabulate the chief
results of this classical series of their earlier papers. Briefly, these
are as follows :
1. A single injection of horse serum into guinea pigs, harmless
in itself, renders these animals hypersusceptible to a subsequent
injection given after a definite interval or incubation time.
2. This interval, with the ordinary dosages employed (about 1
to 2 c. c.), was about 10 days. Properly carried out injections after
this period were usually fatal.
3. The known antibodies, antitoxins, hemolysins, and precipi-
tins, are not responsible for the reaction.
4. The reaction is "quantitatively" specific, injections of horse
serum sensitizing to horse serum only. (The question of specificity
will be further discussed below.)
5. The sensitive condition is transmissible from mother to off-
spring,11 the young of sensitized mothers being hypersusceptible to
a first injection of horse serum.
6. The reaction is extremely delicate. Rosenau and Anderson
succeeded in sensitizing in one case with 0.000001 c. c. (one one-
millionth) of horse serum.
7. The hypersusceptible state is not a transient condition, but
may last a long time.
8. Sensitization, or the production of the hypersusceptible con-
dition, can be carried out, not only with the various animal and vege-
table proteins employed in the first experiment, but can be brought
10Rosenau and Anderson. U. S. Pub. Health and M. H. S. Hyg. Lab.
Bull 29, 1906; 30, 1906; 36, 1907; Journ. Med. Ees., Vol. 15, 1906, Vol. 16,
1907; also Jour. Inf. Dis., Vol. 4, 1907, Vol. 5, 1908.
11 It is important practically, as Anderson points out, that a female
guinea pig may transmit to its young sensitiveness to horse serum and im-
munity to diphtheria toxin.
ANAPHYLAXIS
about by the use of extracts of various bacteria. In such cases also
the reaction is specific. The first determinations with bacterial ex-
tracts carried out by Eosenau and Anderson were made with colon,
anthrax, typhoid, and tubercle bacilli.
By these observations, then, the possibility of a direct relation
between the phenomena of anaphylaxis and infectious diseases in
animals was indicated.
This, in essence, is the harvest of the two earliest purposeful re-
searches into this problem. A large number of investigators now
took up the question, and its further elucidation, as we shall see, has
proved, not only the most directly fruitful of the phases of recent
immunological studies, but has thrown much indirect light upon
antigen-antibody reactions apart from the anaphylactic phenomena
themselves.
Before entering into the further discussion of the experimental
data, however, it will be necessary to describe briefly the clinical
manifestations which follow upon the second injection of an anaphy-
lactic antigen into a sensitized animal, manifestations which we have
heretofore summarized in the phrase "anaphylactic shock." For
there has been much controversy regarding the physiological mech-
anism which lies at the bottom of these symptoms, and the matter
has been complicated by the unquestionably different reactions oc-
curring in various species of animals in response to the anaphylactic
experiment.
Since anaphylactic studies were begun largely as the result of
Theobald Smith's observations upon guinea pigs, and subsequent
study has revealed these animals as peculiarly susceptible to the
anaphylactic poison, the large bulk of the experimental data at our
disposal was worked out upon these animals. In consequence our
understanding of the mechanism of the reaction is based largely
upon guinea pig studies.
If a properly sensitized guinea pig receives a second injection of
an antigen after a suitable incubation time a very characteristic
train of symptoms ensues. There is usually a short preliminary
period — lasting either a fraction of a minute or several minutes ac-
cording to the violence of the reaction and the mode of administra-
tion— during which the pig appears normal. At the end of this time
the animal will grow restless and uneasy, and will usually rub its
nose with its forepaws. It may sneeze and occasionally emit short
coughing sounds. At the same time an increased rapidity of res-
piration is noticeable and the fur will appear ruffled. In light cases
the animals may remain in this condition, with further irregularity
and difficulty of respiration, possible discharges of urine and f eces ;
then gradual slow recovery may set in, with complete return to
normal in from 30 minutes to several hours. In more severe cases
these preliminary stages are rapidly followed by great apparent
364 INFECTION AND RESISTANCE
weakness. The animals fall to the side, the legs and trunk muscles
twitch irregularly, and the respiration becomes slow and shallow;
the thorax never entirely contracts, but remains in a more or less
expanded condition. The very evident dyspnea is of an inspiratory
character. The excursions of the lung itself seem to grow shallower
and shallower in spite of apparent strong inspiratory efforts — the
volume of the thorax and lung remaining in the expanded condition.
At this stage evidences of motor irritation may appear, in that the
animal may arise and attempt to run. More often, however, in this
phase general convulsions set in, often several times repeated, and
in these the animals usually die.
On the other hand, after cessation of convulsions they may lie
perfectly still on the side as though paralyzed, the breathing becom-
ing gradually slower and more shallow, finally ceasing entirely. The
heart may continue to beat for a considerable time after the breath-
ing has stopped.
If such an animal is immediately autopsied a very characteristic
condition is found — to which, in the essentials, attention was first
called by Gay and Southard.12 They speak of finding "pulmonary
emphysema as a constant feature at autopsy/' and attribute the
anaphylactic death in guinea pigs to cessation of respiration in the
inspiratory phase under the influence of respiratory central intoxi-
cation.
The lungs of such guinea pigs after death are found distended
and completely filling the thorax. They are usually pale and blood-
less and do not collapse as the pleurae are opened. On microscopic
examination the alveoli are seen to be distended and small hemor-
rhages may appear upon the serous surfaces. According to Gay
and Southard, furthermore, histological study of the other organs
shows also hemorrhages in the brain, stomach, heart, cecum, and
spleen — more rarely in other organs, and there are local fatty
changes in the capillary endothelium which they regard as causa-
tively related to the hemorrhages.
That the respiratory symptoms are the most striking feature of
the clinical picture of guinea pig anaphylaxis had, as a matter of
fact, been noticed by Rosenau and Anderson. A detailed physiologi-
cal study of the mechanism of the respiratory death in these cases
was first made, however, by Auer and Lewis 13 in 1909.
These investigators showed that, during the later respiratory
symptoms, little or no air enters the lungs, although the animal
makes violent respiratory efforts. This is due, as they found, to a
tetanic contraction of the small bronchioles, which practically oc-
12 Gay and Southard. Jour. Med. Ees., Vol. 16. 1907: Vols. 18 and 19,
1908.
13 Auer and Lewis. Jour, of the A. M. A., Vol. 53, p. 458, 1909 ; Jour.
Exp. Med., Vol. 12, 1910.
ANAPHYLAXIS 365
eludes the air passages. That the origin of this contraction is not, as
previously supposed, of central origin, but is referable to peripheral
cause, they proved by showing that the same phenomena occur in
the guinea pigs even after the cord and medulla have been destroyed
and the vagi divided. In such cases, of course, with the cord and
medulla destroyed, artificial respiration had to be done, and when
the symptoms set in it was found that the lungs could no longer be
expanded by the same force of artificial respiration which before this
had been sufficient.
They showed also that the non-collapsible expansion of the lungs
after death was due to imprisonment of the air in the alveoli by the
contracted musculature of the small bronchioles, and further con-
firmed their opinion of the peripheral origin of this contraction by
the important discovery that atropin will markedly protect, often
preventing death or hastening recovery. It is noteworthy, too, that
Auer and Lewis speak of occasionally finding slight pulmonary
edema, a feature which Biedl and Kraus consider incompatible with
true anaphylaxis. *
Anderson and Schultz,14 who have (bn^rmed much of the work
of Auer and Lewis, find that not only atrojrin will prevent asphyx-
iation in these cases, but methane, chloral hydrate, adrenalin, and
pure oxygen will exert a similar effect. The animals may be saved
from suffocation in this way, but may nevertheless die, probably as
the result of lowered blood pressure.
The observations of Auer and Lewis have been further confirmed
especially by Biedl and Kraus,15 who regard it as well established
that anaphylactic death in guinea pigs is caused primarily by suffo-
cation, due to tetanic spasms of the musculature of the small bronchi.
These spasms are not of central origin, but are peripherally initiated,
possibly by direct action upon the smooth- muscle itself. The fact
that atropin is not effective in preventing death in all severe cases is
no argument against this, since such an eff ect^jKOuld naturally de-
pend upon the relation between the amount of atropin given and the
severity of the attack. In this connection the studies that have been
made upon the irritability of smooth muscle fibers in normal and in
sensitized animals are of great interest. Schultz,16 following out an
observation made by Rosenau and Anderson, studied the intestinal
muscle of normal sensitized guinea pigs excised and suspended in
HowelPs solution. In this way he showed that during the period of
hypersusceptibility the smooth muscle is abnormally sensitive to
treatment with the antigen. The contraction which normally occurs
14 Anderson and Schultz. Proc. Soc. Exp. Biol. and Med., 7, 1909, p. 32.
15 Biedl und Kraus. Zeitschr. f. Immunitatsforschung, Vol. 7, 1910;
Centralbl f. PhysioL, 1910; Wien. klin. Woch., No. 11, 1910.
16 Schultz. Jour. Pharm. and Exp. Therap., 1, 1910 ; 2, 1910.
366 INFECTION AND RESISTANCE
in smooth muscles under the influence of serum is markedly aug-
mented if the preparations are taken from sensitized animals.
In addition to these predominant features of the anaphylactic
symptomatology in guinea pigs, there are a number of secondary re-
actions which, though less prominent, are nevertheless of considerable
interest and theoretical importance. The conditions in the circula-
tion are probably, to a great extent, dependent upon the respiratory
condition, and the fall of blood pressure in guinea pigs is regarded by
some investigators as merely a secondary manifestation just preceding
death. The fall of temperature first described by H. Pfeiffer,17
however, seems to be an occurrence which, though standing in no
causative relation to the symptoms as a whole, is so constant and well
marked that it has been taken by a number of workers as one of the
necessary criteria for the characterization of the anaphylactic con-
dition.
There is, indeed, an almost regular drop of several degrees in the
rectal temperature, and a close observation of this may be of much
aid in determining the occurrence of mild reactions, when other
symptoms of shock are not strongly marked. Pfeiffer 18 himself
goes so far as to claim that by this symptom alone delicate anaphy-
lactic reactions may be determined when all other symptoms are
lacking.
Friedberger,19 20 too, has found the sudden drop of temperature
a very regular occurrence, and has employed this method of study
for the analysis of the intensity of anaphylactic shock. He calls
attention to the apparent difference between infection and anaphy-
laxis in this respect in that in the former there is fever, in the latter
there is depression of body heat ; but, at the same time, he points out
that this discrepancy is an apparent one only, and determined by
quantitative differences, for when he treated sensitized animals with
varying doses of antigen he found that quantities which produced
other anaphylactic symptoms of noticeable degree would regularly
depress the temperature as Pfeiffer had shown. It was possible,
however, to determine a minimal dose necessary for temperature
reduction. Quantities just below this left the temperature un-
changed, and still smaller quantities produced fever or even in-
creased the temperature. This fact is extremely significant in that,
as we shall see, it has an important bearing upon views which inter-
pret bacterial infection as a series of anaphylactic poisonings, the
multiplying bacteria furnishing the constant supply of minute
amounts of antigen. This thought, indeed, based also on the study of
17 H. Pfeiffer. Wien. klin. Woch., No. 1, 1909.
18 Pfeiffer u. Mita. Zeitschr. f. Immunitatsforschung, Vol. 4, 1910.
9 Friedberger. Deutsche med. Woch., No. 11, 1911.
20 Friedberger und Mita. Zeitschr. f. Immunitatsforschung, Vol. 10, 1911.
ANAPHYLAXIS 367
temperature curves in animals, was expressed by Vaughan 21 as early
as 1909, and was developed by him with Gumming and Wright22 in
an extensive study upon what he called "protein fever." It was
shown in these experiments that continued fever, not unlike that of
infectious diseases, could be produced in rabbits by repeated subcu-
taneous injections of primarily harmless substances, such as egg
white and vegetable proteins. The conditions observed and the con-
clusions drawn from them in this work, as well as in the similar in-
vestigations of other workers, were clearly foreseen by Vaughan in
his early investigations on proteid split-products studies, which we
will find occasion to discuss in a later section.
The rigidity of the diagnostic value of the temperature relations
for anaphylactic shock in particular, as advanced by Pfeiffer, was
somewhat weakened by Ranzi's 23 observations that foreign serum
may produce temperature depression when injected into -perfectly
normal animals and that, injected into sensitized animals, the same
reaction may follow if other proteins than the original antigen were
administered.
Although these objections of Ranzi are perfectly just, yet there
is such a marked quantitative difference between the reaction in nor-
mal and in sensitized animals that, in principle, Pfeiffer's claim is
not invalidated. Friedberger24 very logically remarks that, after
all, the phenomena of sensitization as well as those of immunity are
merely an exaggeration of normal physiological conditions, and in
experiment he has shown that, whereas noticeable depressions of tem-
perature will follow in the normal animal only upon quantities of
antigen exceeding 0.5 c. c., the temperature of the sensitized animal
may be depressed by amounts as small as 0.0005 c. c.
Apart from the symptoms so far discussed, there are other less
apparent characteristics of anaphylaxis in guinea pigs, all of which,
however, possess considerable importance theoretically. The most sig-
nificant of these is the reduction in the amount of alexin or comple-
ment, first noticed by Sleeswijk,25 which occurs after the injection
of the second or toxogenic dose — during the development of shock.
This phenomenon is so closely interwoven with the later theoretical
aspects of anaphylaxis that we will defer its discussion until we
have completed a more general survey of the field.
In guinea pigs, as in dogs, Friedberger and others have also seen
a lowered coagulability of the blood and a temporary diminution of
the polynuclear leukocytes (leukopenia) during shock.
21 Vaughan. Zeitschr. f. Immunitatsforschung, Vol. 1, 1909.
22 Vaughan, Gumming, and Wright. Zeitschr. f. Immunitatsforschung,
Vol. 9, 1911.
23 Ranzi. Zeitschr. f. Immunitatsforschung, Vol. 2, 1909 ; Wien. klin*
Woch., No. 40, 1909. '
24 Friedberger u. Mita. Loc. cit.
25 Sleeswijk. Zeitschr. f: Immunitatsforschung, Vol. 2, 1909.
368 INFECTION AND RESISTANCE
During the earlier periods of experimentation there was a
marked discrepancy in the ease with which guinea pigs could be
sensitized by American and German investigators, on the one hand,
and by Besredka and Steinhart in France, on the other. The mor-
tality, upon second injection, was much higher, with like quantities
of horse serum in the hands of the first-named. In attempting to
explain this, Rosenau and Anderson carried out typical experiments
with horse serum sent to them by Besredka and obtained high per-
centages of fatal results. They believe, for this reason, that the dif-
ferences cannot be accounted for by variations in the toxicity of the
horse sera, but conclude that probably there are varying grades of
susceptibility to the reaction in guinea pigs of different breeds.
Next to guinea pigs the animals most commonly employed for
anaphylactic experiment are rabbits and dogs. In both of these the
symptoms and autopsy findings differ markedly from each other and
from those observed in guinea pigs.
In sensitized rabbits the injection of a second dose of the antigen
is usually followed, after a short but definite incubation time, by
great weakness with, often, discharge of urine and feces. The ani-
mals sink down until the abdomen touches the ground, the legs are
stretched out weakly but not paralyzed, and the head may drop
forward or to one side. After this, the animal may gradually fall
upon its side and lie motionless except for labored and irregular
breathing and occasional twitching of the legs and head. Sometimes
this gradual relaxation may be interrupted by a sudden motor irri-
tation, the rabbit suddenly getting up and running a short distance
but soon falling down again apparently from a sudden return of the
muscular weakness. During these running spells it seems as though
there was no sense of direction or purpose — the animals running into
obstructions or off tables as the case may be. During this period gen-
eral convulsions and a drawing back of the head by a tetanic spasm
of the muscles of the neck are not uncommon. Death may occur
within a few minutes, or it may follow a gradually increasing weak-
ness in the course of several hours. The fall of blood pressure here
seems to be purely secondary to the general failure of all the func-
tions.26
Anaphylaxis in dogs has been very extensively studied, especially
by Bie'dl and Kraus,27 and by Pearce and Eisenbrey.28 The symp-
toms in dogs are characterized by a rapid progressive fall in the
blood pressure, followed by the symptoms of cerebral anemia. Ana-
phylactic dogs, after injection, will at first grow restless, vomit, and
26 Arthus. Arch. Internat. de Physiol., 7, 1909.
27 Biedl and Kraus. LOG. cit.; also in "Kraus u. Levaditi Handbuch,"
Erganzungsband 1.
28 Pearce and Eisenbrey. Proc. Soc. Exp. Biol and Med., 7, 1909, p. 30;
Transact. Congr. Am. Ph. and S., Vol. 8, 1910.
ANAPHYLAXIS
pass urine and feces. They then grow rapidly weak, fall to the
ground, and continue to twitch and vomit and the respiration be-
comes labored and irregular. There is general weakness of the mus-
cles, but no paralysis. The marked, constant, and characteristic
feature of the condition in these animals is the fall of blood pressure.
There is also a lessened coagulability of the blood, much more
strongly developed than in guinea pigs and rabbits.
According to Biedl and Kraus this may amount to almost a pre-
vention of the coagulation in anaphylactic dogs.
As in other animals the blood picture is changed in that there is
a falling off of the total number of leukocytes with a relative diminu-
tion of polynuclear cells.
Quantitative measurements by Calvary,29 moreover, have shown
that anaphylaxis in dogs is accompanied by a marked increase of the
lymph flow ( 7 times the amount observed in normal dogs in the same
time) and, by controlling the blood pressure with barium chlorid,
that this lymphagogue action is not directly dependent upon the low
pressure. This observation is of especial interest in connection with
the similarity of anaphylaxis to peptone poisoning in which Heiden-
hcim 30 noticed a similar increase of the lymph.
Pearce and Eisenbrey found, at autopsy of dogs dead of anaphy-
lactic shock, subserous petechial hemorrhages in the rectum and gall
bladder, hemorrhagic spots on the gastric and duodenal mucosa, and
in the colon. According to these workers, in agreement with Biedl
and Kraus, the fall of blood pressure is not due to central causes but
depends upon influences exerted upon the peripheral vasomotor sys-
tem. Biedl and Kraus believe that this action is exerted upon the
muscle cells themselves rather than on the nerve endings. They
admit the inconclusiveness of their experimental data, but take the
above standpoint because of the fact that adrenalin, which acts by
stimulation of the vasomotor nerve endings particularly, does not
raise the low pressure in dogs during anaphylaxis while barium
chlorid, which acts upon the smooth muscle fibers themselves, strongly
raises the blood pressure in such animals. Pearce and Eisenbrey are
inclined to believe that the action is chiefly upon the nerve endings,
though both factors, nerve and muscle, may be involved. They
worked with apocodein, a substance which, in large doses, paralyzes
the vasomotor nerve terminals.31
When a sensitized dog was treated with apocodein and the anti-
gen then injected, no further drop of pressure was obtained. Appar-
ently a paralysis of the vasomotor nerve endings had removed the
point of attack upon which the anaphylactic poison could act.
In addition to the symptoms already enumerated Weichhardt and
20 Calvary. Munch, med. Woch., No. 13, 1911.
30 Heidenheim. P finger's Archiv, 49, 1891.
31Brodie and Dixon. Jour, of Phys.} 30, 1904.
370 INFECTION AND RESISTANCE
Schittenhelm 32 claims that anaphylaxis in dogs is invariably accom-
panied by a severe local reaction in the gut. The intestinal rnucosa
is swollen and contains miliary hemorrhages and the lumen is often
filled with a mucus mixed with blood. In the further analysis of the
anaphylactic reaction in dogs, Manwaring 33 has recently reported
observations of great interest. He investigated the participation in
anaphylactic shock of the various organs and determined that shock
did not occur when the abdominal vessels were ligated just above
the diaphragm. In further localizing the source of shock he found
that exclusion of the spleen, stomach, kidneys, suprarenals, and
ovaries from the circulation had no effect upon the occurrence of
anaphylactic shock. However, when he operated in such a way that
the liver was thrown out of circulation, none of the seven dogs that
he used reacted with anaphylactic shock to the injection of serum.
He concludes from this that the liver is directly responsible in some
way for the production of anaphylaxis. The intestines, too, were
found, by a similar procedure, to take part, though to a less important
extent than the liver.
I Other animals than those mentioned have been little used for
^anaphylactic experiment. Observations incidental to other work,
however, have shown that horses and goats are particularly sensitive.
In goats the writer has observed both serum and bacterial anaphy-
laxis, and the symptoms here were those of general trembling, weak-
ness, labored respiration, and involuntary evacuation of urine.
The occurrence of anaphylaxis in man will be discussed in a
subsequent section.
Strange as it may seem, it is nevertheless a fact which we must
emphasize that the occurrence of anaphylactic shock is not physi-
ologically identical in different species of animals. The phenomenon
in guinea pigs is unlike that in rabbits. The isolated rabbit uterus
will not react when the guinea pig uterus does. The rabbit heart
seems to react when the guinea pig heart does not, and the character-
istic pulmonary picture of the guinea pig is entirely lacking in the
rabbit. In dogs, again, the liver seems to be the organ most definitely
involved, and intestinal symptoms are particularly severe. It is en-
tirely unsafe, therefore, to draw conclusions by analogy from one ani-
mal to another, and naturally, also, it is unwarranted to apply directly
to man conditions revealed by animal experimentation.
The manifestations of "active anaphylaxis " therefore, consist in
the profound physiological changes occurring in animals when rein-
jected after a definite interval with certain substances which, on first
injection, were practically harmless. The factors which are of funda-
mental importance in determining the development of this hypersus-
ceptible or anaphylactic state consist in the nature of the injected sub-
32 Weiehhardt and Schittenhelm. Deutsche med. Woch., 19, 1911.
33 Manwaring. Zeitschrift, f. Immun., Vol. 18, 1911.
ANAPHYLAXIS 371
stance, the quantity injected, and the interval between administra-
tions. To a great extent, too, the violence of the reaction is depend-
,ent upon the path by which the particular substance enters the body.
Each of these factors, therefore, requires detailed consideration
before we can intelligently proceed with a further analysis of the
condition.
The substances with which animals may be sensitized are, in all
particulars, identical with the class of substances which we have char-
acterized as "antigens." In fact, up to the present time, there has
not been a single authenticated exception to this, and from our pres-
ent understanding of the mechanism of anaphylaxis we may safely
predict that no such exceptions will be found. It is the large class of
proteins, therefore, whatever their source, which may act as the "ana-
phy lactic antigens." However, in this connection as well as in the
larger problem of the nature of antigens in general, it has been diffi-
cult to decide whether or not the antigenic property is entirely con-
fined to proteins or whether other substances, such as the lipoids,
must be included in the definition. The problem has been the same
here as in other serum phenomena, but much special experimenta-
tion has been done upon the question with particular reference to
anaphylaxis and the possibility of sensitizing animals with lipoids.
As in the case of similar investigations in regard to antibody for-
mation, the results obtained in this work have been somewhat confus-
ing. Pick and Yamanouchi 34 extracted beef and horse sera with alco-
hol, and evaporated and redissolved the solutions until neither con-
tained coagulable protein nor gave the Biuret reaction. With this ma-
terial they obtained a few positive anaphylactic experiments. Simi-
larly curious are the results of Bogomolez,35 who succeeded in sensi-
tizing and producing shock with the lipoids extracted from egg yolks.
Although such experiments would tend to persuade us that lipoidal
substances may actually have sensitizing, and therefore antigenic,
functions, this does not follow necessarily. As Pick and Yamanouchi
themselves point out, it is practically impossible to demonstrate with
certainty the presence of slight traces of proteins as impurities in li-
poid preparations, and we know especially from Rosenau and Ander-
son's work how minute are the quantities of antigen which still serve
to sensitize. It is possible, moreover (a thought developed particu-
larly by Pick and Schwartz36 and by Landsteiner 37), that we are
dealing in many cases with combinations of protein and lipoid — a
form of chemical substance of which very little is known analytically,
but the existence of which many biological facts lead us to assume.
That^h&cmaphAilacMr, reaction is specific we have mentioned in
34 Pick and Yamanouchi. Zeitschr. f. Immunitatsforschung, Vol. 1, 1909.
35 Bogomolez. Zeitschr. f. Immunitatsforschung, Vols. 5 and 6, 1910.
36 Pick and Schwartz. Biochem. Zeitsch., 15, 1909.
37 Landsteiner. Referat. "Weichhardt's Jahresbericht," 6, 1910.
372 INFECTION AND RESISTANCE
the brief summary we have given of Rosenau and Anderson's work.
These authors use the adjective "quantitative," by which they simply
mean to convey that the specificity here is not absolute, any more than
it is absolute in the case of any of the known serum reactions. An
animal sensitized with a certain variety of protein, animal serum, etc.,
reacts with disproportionately greater delicacy to a second injection
of the same variety than of any other substance. In fact, apiart from
a few cases mentioned by Gay and Southard, there are not many in-
stances of marked non-specific anaphylactic reactions. Still we would
expect here, as in other serum reactions, a certain limitation in the
degree of specificity, and Otto recommends the less delicate subcu-
taneous method of testing for all experiments in which questions of
specificity are involved. This point we will touch upon a little later.
An interesting addition to our knowledge of such specificity was
made by experiments of Rosenau and Anderson,38 which showed that
a guinea pig could be rendered separately sensitive at one and the
same time to blood serum, eggwhite, and milk, reacting specifically
to each on second injection.
In anaphylaxis, again analogous to antibody reactions in general,
the specificity, as a rule, is one of species. In other words, the pro-
tein of any animal is specific for the proteins of its particular spe-
cies generally, there being definitely similar characteristics in the
body proteins of animals of like species which, though chemically
indefinable, are nevertheless delicately determinable by biologic reac-
tions. In considering specificity of precipitins, however, we have
seen that there are exceptions to the specificity of species expressed
in the phenomenon of so-called organ specificity. The same thing-
has been shown for anaphylaxis. Kraus, Doerr, and Sohma 39 were
able to show that animals sensitized with protein from the crystalline
lens were hypersusceptible to lens protein generally, whether this
came from the species from which the original lens was taken, or
whether some other variety of animal had furnished it. On the other
hand, animals so sensitized, while hypersusceptible to lens protein
generally, did not react to injections of homologous blood.40 In
other words, this organ contains a characteristic variety of antigen
(protein) peculiar to this kind of organ throughout the different
animal species, but not common to other tissues and organs of the
same animal. Results similar to these were obtained by von Dun-
gern and Hirschfeld 41 in the case of testicular protein, although
here the phenomenon seemed to be less rigidly organ-specific than
in the preceding case. These writers worked not with the systemic
38 Rosenau and Anderson. Jour. Inf. Dis., Vol. 4, 1907.
39 Kraus, Doerr, and Sohma. Wien. klin. Wocli., No. 30, 1908.
40Andrejew. Arb. a. d, kais. Gesundh. Ami., Vol. 30,. 1909.
41 von Dungern and Hirschfeld. Zcitschr. f. Immunitatsforschung, 4,
1910.
ANAPHYLAXIS 373
anaphylactic reaction, but with the localized (allergic) reaction
described above as the phenomenon of Arthus. They injected ex-
tracts of the testicular materials into the ears of rabbits and inci-
dentally made the very curious observation that pregnant females
would not infrequently react to a first injection without previous
sensitization.
Of great importance also in connection with the subject of organ
specificity is the further discovery by Uhlenhuth and Haendel42
that animals can be sensitized with their own lens protein, a fact
which opens the possibility of other forms of "autosensitization" and
consequently of much opportunity for clinical speculation. Rosenau
and Anderson,43 indeed, have found that guinea pigs can be sensi-
tized by means of extracts of guinea pig placenta. They have applied
this to the possible explanation of eclampsia, and similar reasoning,
as we shall see, has been utilized in many other conditions. At-
tempts have also been made to show, by the anaphylactic reaction,
that the tissue of malignant tumors possess such ''tissue-specific" or
"organ-specific" qualities. Yamanouchi,44 indeed, claims to have
shown this, but his results were not confirmed by Apolant,45 and the
writer has carried out a series of entirely negative experiments upon
the same subject. However, in view of the great difficulty of obtain-
ing any kind of anaphylactic reaction in mice, the animals in which
the tumor experiments were carried out, there is little information to
be obtained from negative results of this kind.
The delicate quantitative method of studying problems of speci-
ficity, which the reaction of anaphylaxis supplies, has further served
to revive the unsettled question of the "organ-specific" properties of
the tissues of such organs as the liver, spleen, kidney, blood, etc.
Indeed, Pfeiffer 46 has published results which would seem to en-
courage the belief of the existence of such specificity. However,
Ranzi 47 had previously obtained entirely negative results, and
Pearce, Karsner, and Eisenbrey 48 have recently made a careful
inquiry into the same problem with similar failure to determine such
organ-specificity.
In this, then, as well as in other respects, the substances by wkich\
animals may ~be sensitized are entirely similar to antigens in general. J
The substances which sensitize, therefore, are those which have'
the property of antibody formation, a statement self-evident from
42 Uhlenhuth and Haendel. Zeitschr. f. Immunitdtsforschung, 4, 1910.
43Rosenau and Anderson. U. S. Pub. Health and M. H. S. Hyg. Lab.
Bull., 45.
44 Yamanouchi. C. E. de la Soc. Biol., Vol. 66. 1009, p. 754.
45 Apolant. Zeitschr. f. Immunitatsforschung, Vol. 3, 1909.
46 Pfeiffer. Zeitschr. f. Immunitatsforschung, Vol. 8, 1910.
47 Ranzi. Zeitschr. f. Immunitatsforschung, Vol. 2, 1909.
48Pearce, Karsner, and Eisenbrey. Jour. Exp. Med., Vol. 14, 1911.
374 INFECTION AND RESISTANCE
what has been said before, but which is again emphasized because of
its very important bearing upon later theoretical considerations.
Variations in experimental anaphylaxis are, to some extent, de-
pendent upon the manner in which the antigen is introduced into the
body. It is now well known that sensitization may be accomplished
by a first injection given subcutaneously, intravenously, intraperi-
toneally, or intrapleurally. At the second or toxogenic administra-
tion shock may be probably best induced and with the smallest
quantities by the intravenous method. Besredka and Stein-
hardt,49 50 51 52 who began their studies soon after the first publica-
tions of Eosenau and Anderson, came to the conclusion that the most
effectual and rapid method of producing the anaphylactic shock
consisted in direct injection into the brain. Curiously enough, while
Besredka and Steinhardt obtained the most violent reactions by
injection of the second or toxogenic dose into the brain, they were
unable to sensitize by this path, at least with doses of 1/4000 c. c.,
which sufficed to sensitize by the intravenous method. Rosenau and
Anderson, in repeating this work, obtained similar results with very
minute amounts, but found that intracerebral sensitization could be
accomplished by doses of .0001 c. c., or more. According to them,
animals intracerebrally sensitized become anaphylactic more rapidly
than those in which the injections were subcutaneous. In the
former the incubation time was about 7 days, while in the latter it
was never less than 9. Lewis,53 in his thorough study on the same
subject, made extensive use of the direct intracardial method of
injection. In other words, any method of introducing the foreign
protein into the blood or tissues seems to lead both to sensitization
and to toxic effect, and those methods which introduce the substance,
on reinjection, directly into the blood stream or the brain induce
the most violent symptoms with the relatively smallest dosage. Ac-
cording to Otto and others, the subcutaneous method, while followed
by less violent symptoms, is the method to be preferred when ques-
tions of specificity are involved, for, while the reaction is specific in
the ordinary sense, yet it is extremely delicate and therefore, as
Eosenau and Anderson put it, "quantitatively specific.'7 The less
violent subcutaneous method, therefore, might be said to have the
same purpose here that dilution of the antigen or immune serum has
in safeguarding against error when carrying out specific precipitin or
iigglutinin reactions.
49 Besredka and Steinhardt. Ann. de VInst. Past., p. 117, 1907; ibid.,
p. 384.
50 Besredka. Ibid., p. 777, p. 950, 1907; ibid., p. 496, 1908; p. 166, 1909;
p. 801, 1909.
51 Also: Bull, de VInst. Past., Nos. 19, 20, 21, 1908; No. 17, 1909.
52 Also: C. E. de la Soc. Biol, p. 478, 1908, Vol. 65; p. 266, 1909, Vol.
67.
53 Lewis. Jour. Exp. Med., Vol. 10, 1908.
ANAPHYLAXIS 375
Whether or not sensitization can be accomplished by introduction
of the antigen into the intestinal canal, feeding, in other words, is
still to some extent an open question and of great importance in
view of the many clinical manifestations (urticaria, albuminuria,
etc.) which are attributed to possible individual hypersusceptibility
to certain proteins taken in the diet (idiosyncrasies). Rosenau and
Anderson, in their earliest paper, report success in sensitizing guinea
pigs by the feeding of horse meat and horse serum. McClintock and
King 54 failed to confirm this, and the observations of other writers
seem to bear them out. However, when we consider that Ascoli,
Oppenheimer, and others have shown that proteins fed to animals in
large quantities may be subsequently demonstrated not only in the
circulating blood but occasionally even in the urine by means of the
precipitin reaction, there seems to be little room for doubting that
antigen may enter the circulation unchanged, though possible only
under abnormal local conditions of the intestine. This, together with
Eosenau and Anderson's demonstration of the extremely small amount
of antigen necessary to sensitize, furnishes all the conditions n(
sary for anaphylaxis by way of the intestinal canal.
A study made by Lesne and Dreyfus 55 seems to us to have ex-
plained the contradictory results of other workers on this phase of
the problem. Without being able to associate the destruction of the
sensitizing function with either the gastric or pancreatic secretions,
they were nevertheless successful in showing that sensitization could
be carried out regularly if the antigen were injected after laparotomy
into the large intestine, whereas similar injections into the stomach or
small intestine wrere negative. In these experiments we must take
into consideration that the conditions following laparotomy, such as
temporary intestinal atony and congestion, may have exerted con-
siderable influence upon the positive outcome of their large intestine
injections. Whereas they do not, therefore, permit us to assume the
possibility of sensitization through the normal alimentary canal, they
nevertheless confirm the assumption of the possibility of sensitiza-
tion by this path under the influence of slightly abnormal local con-
ditions.
In this connection Besredka's 56 experiments on the production
of anti-anaphylaxis by the intestinal administration of protein are
of interest. He found that if sensitized animals were given 5 c. c.
of the antigen (milk) by rectum, they were thereby protected from
the reaction following in controls upon a second injection. In his
later experiments with egg white it appeared that the protection
could also be conferred by mouth, but that in this case it developed
54 McClintock and King. Jour. Inf. Dis., 3, 1906. See section on normal
antibodies.
65 Lesne and Dreyfus. C. E. de la Soc. BioL, Vol. 70, p. 136, 1911.
66Besredka. C. E. de la Soc. Biol., Vol. 65, 1908; Vol. 70, 1911.
376 INFECTION AND RESISTANCE
more slowly, it being necessary to wait two days after ingestion be-
fore the anti-anaphylaxis had developed sufficiently to protect. Since
attempts by mouth were not as rapidly successful as those per rec-
tum, it is clear that these facts are in keeping with Lesne and Drey-
fus' results in showing that the antigen is probably absorbed chiefly
or solely from the large intestine. In Lesne and Dreyfus' experi-
ments the sensitizing dose was given into the intestine, the toxogenic
or second dose being administered intravenously, and since, as we
shall see, minute doses may suffice to sensitize, whereas 100 or more
times the sensitizing amount is necessary to produce intoxication, it
is easy to understand why sensitization followed in Lesne and Drey-
fus' work, but no toxic effects followed in the experiments of Bes-
redka. Furthermore, the slow absorption from the intestine in these
experiments explains the development of anti-anaphylaxis in Bes-
redka's work, in that they are, in this respect, analogous to later
experiments of Friedberger, cited below, in which it was shown that-
sensitized guinea pigs, which could (in controls) be killed by rapid
intravenous injection of 0.1 c. c. of antigen and less, would withstand
without symptoms many times this amount when it was gradually
administered by slow injection covering an hour or longer.
In referring to the quantities of antigen by which sensitization
may be accomplished, we have already called attention to the very
small amounts which have been found sufficient for this purpose.
There seems, indeed, to be a wide latitude in this regard, the re-
quired quantities ranging from as little as a millionth of a cubic
centimeter (Rosenau and Anderson) to as much as 10 c. c. or more.
rw oonnTirj ipj^tinri; however, trmc effects^are never produced by
q-nflyit.it.ipjp as -minute as thoSft TyVnVh anffW for
tual amounts exists. An important
problem, moreover, is the relation which has been said to exist be-
tween the sensitizing dose and the interval necessary for the devel-
opment of the hypersusceptible state (anaphylactic incubation time).
In their first publications, Rosenau and Anderson, Otto, and others
expressed the opinion that the length of incubation time was in-
versely proportionate to the size of the sensitizing dose; in other
words, animals sensitized with small quantities (0.01 c. c. or less)
would become hypersusceptible and react to a second injection in
from 8 to 12 days, whereas animals receiving two, three, or more
cubic centimeters of the antigen would take weeks or months to be-
come anaphylactic. The same opinion was expressed by Otto,57 and
is now generally found in the literature. Later experiments of
Rosenau and Anderson,58 however, have seemed to show that this
57 Otto. Loc. cit. See also in "Kolle u. Wassermann Handbuch," Ergan-
zungsband II, p. 241.
58 Rosenau and Anderson. U. S. Pub. Health and M. II. S. Hyn. Lab.
Bull. 45, 1908.
ANAPHYLAXIS 377
relation is not as definite as at first assumed. In the tables given
by them guinea pigs receiving 0.01 c. c. reacted severely after 14, 17,
and 155 days; others, receiving 1 c. c,, after 14, 17, and 155 days;
and, again, another series sensitized with 8 c. c. reacted severely
after similar intervals. All of these series reacted but mildly after
245 days, showing apparently that the anaphylaxis, contrary to gen-
eral belief, does not last so much longer after the larger than after
the smaller sensitizing doses.
These experiments, however, as well as similar ones by other
workers, have shown that, once sensitized, animals may remain so
for very extensive periods. In the work of Rosenau and Anderson 59
the limit of horse serum sensitization was 245 days. A few guinea
pigs, sensitized with toxin-antitoxin mixtures, gave positive reac-
tions after 732 days; more recently they have obtained a reaction
after 1,096 days.60
This seems to indicate that perhaps individuals once sensitized
may remain so for life. And, as we shall see, a hyper susceptible
subject, temporarily desensitized by antigen, will again gradually^
become sensitive.
In properly sensitized animals the result of a sufficient dose of
antigen, given at the proper time, is very often death. When the
time and quantity are so chosen that instead of death there is merely
a more or less severe anaphylactic shock, the animals are immediately
thereafter in a refractory condition. That is, they are no longer
sensitive to further injections of the antigen. This observation was
made by Otto and by Rosenau and Anderson in their pioneer inves-
tigations, was confirmed by Gay and Southard, and was subsequently
very thoroughly studied by Besredka and Steinhardt.61 The last-
named workers named this refractory or immune condition "anii-
anapliylaxis." There is obviously a great deal of both practical and
theoretical significance in this fact, and methods were sought by
which such an anti-anaphylactic state might be induced in sensitized
animals without subjecting them to the dangers of actual shock. It
was found that this could be accomplished in a number of ways.
According to Besredka and Steinhardt the injection of moderate
quantities of the antigen at a time just preceding the development of
hypersusceptibility, in the preanaphylactic period, will render them
refractory to later injections. This preventive administration, how-
ever, must be given during the later days of the anaphylactic incu-
bation time. If given too soon after the first injection it does not
prevent eventual sensitization, though it may occasionally delay its
80Kosenau and Anderson. U. S. Pub. Health and M. H. S. Hyg. Lab.
Bull, 50, 1909.
60 They express the belief from this that a guinea pig may remain sensi-
tive throughout life.
61 Besredka and Steinhardt. Loc. cit.
378 INFECTION AND RESISTANCE
development, acting then simply as though a larger dose had been
given in the first place. Thus if antigen is given by a method of
introduction and in a quantity which would justify us in expecting
hypersusceptibility to be developed at the end of 12 days, we can
render the animal "antianaphylactic" by a second administration
given, say, on the 8th, 9th, or 10th day. If we give it on the 2d, 3d,
or 4th day after the first injection, it is very likely that sensitization
will proceed nevertheless. Eosenau and Anderson have also investi-
gated the repeated injection of antigen during the incubation time,
and their results would also seem to emphasize the necessity of mak-
ing the preventive injection close to the time at which hypersuscepti-
bility may be expected. If quantities of 2 c. c. were injected 10 times
in the course of 17 days, and 15 to 17 days thereafter 6 c. c. of horse
serum were given, the animals showed symptoms proving that anti-
anaphylaxis was but partial. If amounts of 0.001 c. c. were given
5 times in a period of 8 days, and the animals were tested 23 days
later, death often ensued. It is also possible, as a number of investi-
gators have shown, to produce the antianaphy lactic state by the injec-
tion of sublethal doses, even after the time has set in at which the ani-
mals are hypersusceptible. ' This method can be carried out success-
fully according to Besredka by injecting very small amounts into the
brain (1/50 to 1/400 of a cubic centimeter). Within a few hours
after such an injection the animals may withstand an otherwise
fatal dose with slight or no symptoms; although it is generally
stated that intraperitoneal injections, carried out after hypersus-
ceptibility has set in, must be of considerable quantity (large enough
to cause symptoms) in order to induce antianaphylaxis. Besredka 62
states, in a recent resume, that 1/50 to 1/100 cubic centimeter in-
jected intraperitoneally and giving "practically no symptoms" in a
sensitized guinea pig, after the anaphylactic state has set in, may ren-
der the animal entirely refractory after 5 hours.
On the other hand, Rosenau and Anderson,63 working with sub-
cutaneous injection, obtained results which differ considerably from
those of Besredka. They sensitized a series of guinea pigs with
mixtures of toxin and antitoxin, and 48 days later, at a time when
the animals were hypersusceptible, gave 20 subcutaneous injections
of 0.001 c. c. daily. Two days after the last injection, 0.2 c. c. of
horse serum was given intracerebrally, and all of the animals showed
symptoms, and many of them died. They conclude, therefore, that
the repeated injection of small amounts of antigen into sensitized ani-
mals has no appreciable effect. The same worker has shown by ex-
periment that the introduction of large amounts of antigen into the
previously cleansed rectum of sensitive animals is entirely without
62 Besredka in "Kraus u. Levaditi Handbuch," Erganzungsband I.
63 Rosenau and Anderson. Loc. cit., U. S. Pub. Health and M. H. S.
Hyg. Lab. Bull. 45, 1908.
ANAPHYLAXIS 379
danger and will produce an antianaphylaxis, which becomes evident
after 12 hours. This is probably dependent upon the very slow pene-
tration of small amounts of antigen into the circulation from the gut,
and has, therefore, an effect similar to the repeated injection of small
amounts directly, or the very slow and gradual method of intrave-
nous injection advocated by Friedberger for the prevention of serum
sickness in man. This phase of the subject is considered in greater
detail in a subsequent discussion of serum sickness.
This observation has recently been confirmed by Coca.
As far as we can understand it at the present time, the desensiX
tization depends upon a saturation of the antibodies in the animal!
with antigen, which temporarily removes them from possibility of I.
further reaction. A gradual administration of antigen to the sensi-V
tized animal, either by fractional dosage or slow administration, may
bring this about without the fatal shock resulting from a too sudden I
and violent reaction. This is the rationale of Besredka's and Fried-
berger's methods. *
Antianaphylaxis produced in this way is specific,64 although, as
we shall see, there are other methods by which it is claimed that a
nonspecific antianaphylaxis can be produced. One of these consists
in the injection of anaphy lactic animals with peptone. The problem
of peptone poisoning and its relation to anaphylaxis will receive sep-
arate consideration.
Banzhaf and Steinhardt 65 have reported that 0.5 gram of lecithin
given to sensitized guinea pigs protects them against second injec-
tion. Rosenau and Anderson 66 have failed to confirm this.
The above methods of rendering animals antianaphylactic apart
from the bearing they may have on purely therapeutic possibilities,
serve to throw much light upon the possible mechanism of the reac-
tion within the animal body. It is of great interest for the under-
standing of the physiological conditions underlying anaphylaxis also
to consider briefly the influence upon anaphylactic shock which may
be exerted by certain drugs. The preventive influence of atropin
we have already mentioned in connection with the work of Auer and
Lewis. Besredka, who, as we shall see, attributes the major part of
anaphylactic manifestations to reactions proceeding from the central
nervous system, claims to have succeeded in injecting ordinarily
fatal doses of antigen without harm into guinea pigs previously
anesthetized with ether. Banzhaf and Famulener 67 have similarly
64 Pfeiffer has recorded an exception to this in that he claims to have
rendered a horse-serum sensitive animal refractory by an injection of swine
serum.
65 Banzhaf and Steinhardt. Proc. Soc. Exp. Biol and Med., Vol. 7,
1910.
66 Rosenau and Anderson. Hyg. Lab. Bull., 64, 1910.
6T Banzhaf and Famulener. Studies N. Y. Dep. Health Ees. Lab., 1908,
p. 107.
380 INFECTION AND RESISTANCE
prevented shock by large doses of chloral hydrate. Rosenau and
Anderson,08 carrying on similar investigations with urethane, paral-
dehyd, chloral hydrate, and magnesium sulphate, came to the con-
clusion that none of these drugs has any noticeable effect upon ana-
phylactic shock in guinea pigs.
Up to the present time AVC have confined ourselves to the descrip-
tion of the basic anaphylactic experiment, which is spoken of as
''active sen&itization" in analogy to the expression "active immuniza-
tion,'7 since, like the latter, it conveys the conception that the state
of hypersusceptibility (like the immunity in active immunization) is
here acquired by reason of physiological changes directly induced in
the treated animal in reaction to the first injection of the foreign
antigen. There is another method of inducing hypersusceptibility
which, in continviance. of the analogy to immunization, is spoken of
as "passive anaphylaxis/' since it consists in transferring the hyper-
susceptible condition to a perfectly normal animal by injecting into
it serum from an actively sensitized one. The normal animal is thus
merely the passive recipient of the reaction bodies produced in the
sensitive animal by preliminary treatment.
That such a passive transference of anaphylaxis is possible was
shown by a number of investigators almost simultaneously and M.
Xicolle,69 in February, 1907, published a study on the phenomenon
of Arthus in which he showed that, if the serum of a hypersusceptible
rabbit (sensitized with horse serum) was injected into a normal
rabbit, the recipient was rendered sensitive, so that the subcutaneous
injection of horse serum, made 24 hours later, produced typical
infiltrations. Richet 70 soon after this succeeded in transferring
hypersusceptibility toward mytilocongestin (a mussel poison) from
a sensitized to a normal dog by injecting considerable amounts of the
blood from the former into the latter. In this case, too, the hyper-
susceptibility of the second dog did not appear until one or two days
after the injection of the blood. At almost the same time Otto 71
and Friedemann 72 independently succeeded in transferring serum
anaphylaxis from hypersusceptible to normal guinea pigs in a similar
way. Experiments of Gay and Southard,73 published during the
same year, may possibly be also interpreted as instances of passive
anaphylaxis, although their experimental procedure renders this
doubtful, even in their own opinions. They injected 0.1 c. c. of
serum from both sensitive and refractory guinea pigs into normal
animals and followed this, after 10 days, with injections of antigen.
The fact that such animals reacted may be interpreted in a number
68 Rosenau and Anderson. Jour. Med. Ees., Vol. 21, N. S., 16, 1909.
69 M. Nicolle. Ann. de I'Inst. Past., Vol. 21, 1907.
70 Richet. Ann. de I'Inst. Past., Vol. 21, 1907.
71 Otto. Munch, med. Woch., No. 34, 1907.
72 Friedemann. Munch, med. Woch., No. 49, 1907.
73 Gay and Southard. Jour. Med. Ees., Vol. 16, 1907.
ANAPHYLAXIS 381
of ways. They themselves regarded the hypersusceptibility which
the injected animals developed as a "purely active one/' and it is
more than likely that this was the case, the recipient animals being
actively sensitized by traces of antigen remaining unassimilated in
the blood of the actively sensitized donors. In the following year
(1908) the facts of passive sensitization were rapidly confirmed and
extended by Besredka,74 Lewis,75 and others,76 and information of
the greatest value for the comprehension of the anaphylactic reaction
was obtained.
Otto showed that passive sensitization could be carried out with
the serum of an actively sensitized animal 8 days after the antigen
injection, at a period when this animal itself had not yet become
hypersusceptible. He also showed that the passive transfer of ana-
phylaxis need not be confined to animals of the same species, but that
guinea pigs could be Tendered passively anaphylactic with the blood
serum of sensitized rabbits. From the work of Gay and Southard,77
moreover, it appeared that not only by the blood of sensitive animals
can anaphylaxis be transferred, but that this can also be done by
injecting the blood of animals that have once been sensitive but have_
subsequently been rendered antianaphylactic or refractory. Analo-
gous to this observation is the fact observed by these authors as well
as by Friedemann that the young of antianaphylactic mothers are not _
refractory but hypersusceptible. This observation, unquestionably
correct, since it has been confirmed by several other workers, is
astonishing and contrary to expectation. It has had no inconsider-
able bearing upon our theoretical understanding of anaphylaxis.
It was soon found out, too, that hypersusceptibility was conveyed
not only by the sera of sensitive and of refractory animals, but that
it could likewise be transferred by the precipitating sera of animals
systematically immunized with a foreign proteid.
This method was later employed by Doerr and Russ 78 in their
quantitative studies on the relations between anaphylactic antigen
and antibody. We are confronted, then, with the curious facts that
animals may be passively sensitized :
(a) by the serum of a sensitized animal.
(b) by the serum of an animal not yet sensitive — in the pre-
anaphylactic period (8th day, Otto).
(c) by the serum of an antianaphylactic animal.
(d) by the precipitating serum of an "immunized" 79 animal.
74 Besredka. Ann. de I'Inst. Past., Vol. 22, 1908.
75 Lewis. Jour. Exp. Med., Vol. 10, 1908.
76 Kraus and Doerr. Wien. klin. Woch., No. 28, 1908.
77 Gay and Southard. Jour. Med. Res., Vol. 18, 1908.
78 Doerr and Russ. Zeitschr. f. Immunitcitsforschung, Vol. 3, 1909.
79 We must never forget that the term "immunized" as applied to animals
treated with harmless protein is an analogy and not absolutely correct.
Such animals, though probably capable of assimilating larger quantities of
S82 INFECTION AND RESISTANCE
Lewis further showed that normal guinea pigs could be rendered
hypersusceptible with the blood of congenitally sensitive animals.
Passive sensitization is carried by the blood serum purely, since,
in ordinary cases, as Rosenau and Anderson have shown, the blood
corpuscles and tissues of a sensitive animal do not convey the hyper-
susceptibility. An exception to this will be noted later when we
come to discuss Bail's experiments on the passive transfer of tuber-
culin sensitiveness.
Passive sensitization, once established, may persist for as long
as 3 or 4 weeks, though Rosenau and Anderson found that animals
tested 26 days after treatment reacted but weakly. In the young
of anaphylactic mothers Otto has observed positive reactions as long
as 44 days after birth, though fatal results were obtained in pigs only
a few days old.
o To summarize the matter briefly, we may state that passive sen-
sitization may be accomplished with any serum that contains anti-
/ bodies, and that the qualitative power of such a serum to convey
I passive sensitization is in direct proportion to its antibody concen-
I tration; in short the process of sensitization consists in the introduc-
tion of antibodies. This fact, which is, of course, of the greatest
I theoretical importance, will be further discussed in the succeeding
\chapter.
^ Throughout the earlier investigations upon passive sensitization
the curious fact recurs in the experiments of successive workers that
a definite period must elapse between the injection of the sensitive
blood and that of the antigen.
Both Friedemann and Otto found that when the sensitive serum
was injected subcutaneously the best results were obtained by ad-
ministration of the antigen 24 to 48 hours after this. On intra-
peritoneal injection of the sensitizing serum Doerr and Russ 80 ob-
tained the best results- by permitting an interval of 24 hours to
elapse, and the same investigators still further shortened this period
to 4 hours by injecting the sensitive serum intravenously. Beyond
this, the interval could not be shortened with success. Indeed, some
writers, notably Gay and Southard, have claimed that the maximum
hypersusceptibility in guinea pigs treated with sensitive serum is
reached only after 10 or more days, and Rosenau and Anderson,
Lewis, and others have obtained results which seemed to point in
the same direction. However, as we have already indicated, the
testing of animals so long after the injection of sensitive serum
leaves us in doubt whether we are dealing with true "passive" trans-
foreign injected protein than normal ones, and this more rapidly, may never-
theless be not a whit more tolerant of the antigen — sometimes even extremely
sensitive and vulnerable.
80 Doerr and Russ. Zeitschr. f. ImmunitdtsforscJiung, Vol. 3, p. 181f
1909.
ANAPHYLAXIS 383
ference of anaphylaxis or with active sensitization due to traces of
antigen carried over with the serum of the sensitive animal. For
the purposes of theoretical deduction, therefore, it is better to ignore
these cases and consider chiefly passive transference in which reac-
tions are obtained within 24 hours or less after the injection of the
aiiaphylactic serum — an interval so short that active sensitization
can hardly be considered as a reasonable possibility.
The important point, in this connection, is the fact that it was
found that between the administration of sensitive serum and of
antigen a definite interval, however short, was invariably necessary.81
From these observations the natural deduction was made that the
anaphylactic symptoms were the result of cellular occurrences, and
that the antigen could act only after the sensitizing substance (how-
ever conceived) had become attached to certain cells, probably to
those of the central nervous system. It was thought that a meeting
of antigen and the sensitized body in the circulation would result in
no reaction; that, in other words, the effective reaction was not a
direct, but an indirect, one after the anaphylactic "antibody" of the
sensitive serum had become bound to the cells. It will be neces-
sary to recur to this problem when we discuss the various theories
of anaphylaxis, where we will see that this point has been one of the
crucial ones in the controversy between the two main directions of
thought on anaphylaxis.
81 An exception to this, contradicting the then prevailing opinion, were
the researches of Weill-Halle and Lamaire (C. R. de la Soc. de Biol., Vol.
65, July, 1908, p. 141), who showed that, under certain conditions, guinea
pigs would react with typical, often fatal, anaphylaxis if injected simul-
taneously with the serum of sensitized rabbits and the antigen horse serum.
According to them, the success of such experiments depended entirely upon
the condition of the sensitive serum — that is, the time at which the rabbits
treated with horse serum were bled. These experiments, we shall see, were
later confirmed. We record them, though important, in a footnote, since we
wish at present to emphasize the reasoning which led to the assumption of a
cellular participation in the reaction.
CHAPTER XVI
ANAPHYLAXIS (Cont.)
FURTHER DEVELOPMENT AND THEOEETICAL
CONSIDERATIONS
WE have now briefly considered some of the fundamental facts
which the earlier investigations upon anaphylaxis have revealed and,
although there are still many important observations to record, the
material so far outlined will serve as a basis for a brief consideration
of the views that have been formulated concerning the mechanism
of aiiaphylactic phenomena.
It is clear that the chapter of anaphylaxis is hardly more than
well begun. In the earlier stages of the investigations into this prob-
lem many opinions were advanced which served the valuable func-
tions of working hypotheses, but were quickly altered, trimmed, or
expanded as new and incompatible facts were revealed in astonish-
ingly rapid succession. The final solution is probably still far be-
yond our present horizon, but the recent knowledge of the toxic
derivatives of proteins, "the anaphylatoxins," foreshadowed by the
work of Vaughan and his associates, more definitely determined by
Eriedemann and especially by Friedberger, has furnished hope that
we are not only on the right path toward understanding anaphylaxis,
but has given us a new clue to the correlation of this condition with
immunity.
It will greatly facilitate exposition of the various theories which
have been advanced if we bear in mind that, although there have
been many discrepancies on minor phases, the differences of opinion
have centered upon the cardinal points.
These are: 1. Is the anaphy lactic phenomenon a true antigen-
antibody reaction in which the sensitizing injection gives rise to the
formation of a specific antibody with which it reacts on second injec-
tion ? 2< Is sensitization the result of effects exerted upon the tissue
cells, which participate directly in the reaction, or may the reaction
take place entirely in the circulation, the tissue cells being affected
secondarily only ?
Upon these two questions we can logically classify theories of
anaphylaxis.
384
ANAPHYLAXIS 385
Among the earliest definitely stated theories is that of Gay and Southard.1 "~
These workers were emphatic in denying that anaphylaxis has the nature of an
antigen-antibody reaction. Their views are summarized in the following para-
graphs given, as nearly as space permits, in their own words:
Increased susceptibility in the sensitized animal is due to the continued
presence in the circulation of an unneutralized element of the antigen (in their
case horse serum), which they call ' ( anaphylactin, " which acts as an irritant or
stimulant to the body cells, and, in some way, causes them to assimilate over
rapidly certain other elements of horse serum. These assimilated or toxic ele-
ments are the same as those eliminated without producing intoxication during the
incubation period following the first dose. This overassimilation after anaphy-
laxis is the cause of the intoxication.
Gay and Southard find much support for their contentions in the results of
experiments done with the so-called ' ' passive ' ' transfer of hypersusceptibility.
As mentioned above, hypersusceptibility may be transferred to a normal animal
with the blood serum not only of a sensitive animal, but even more surely and
effectually with that of a refractory, or ' * antianaphylactic, ' ' animal. They believe
that such transfer is not "passive" but "active" sensitization, being accom-
plished by the transfer of "anaphylactin" to the normal animal. The re-
fractory animal has received more horse serum than the merely sensitive one,
since antianaphylaxis is produced by massive injections. Therefore its blood
contains more anaphylactin and is consequently more active in transferring sen-
sitiveness. The fact that a considerable incubation .time is necessary in active
sensitization they attribute to the gradual action of the anaphylactin.
In passive sensitization, therefore, they assumed a similar gradual irrita-
tion of the vulnerable cells by the anaphylactin and, as we have seen, obtained
their reactions in animals so treated, usually 10 to 14 days after the sensi-
tive serum had been given. This conception of the mechanism of passive
anaphylaxis was, of course, rendered unlikely by the demonstrations by Friede-
mann, Otto, and others that shock could be elicited in passively sensitized animals
within 24 hours or less after transfer of the anaphylactic serum.
To this, however, Gay and Southard2 answer by implying that this greater
speed of development of sensitiveness in the experiments of Otto is due to the
larger doses used by him. They say "if the doses are sufficient it (transmitted
sensitiveness) may be shown in a single day (Otto)." However, it is very likely
that the sensitiveness noted by them in animals two weeks after the transference
of anaphylactic serum was actually positive sensitization with antigen rests, en-
tirely comparable to the usual ' ' Theobald Smith ' ' phenomenon.
Gay & Southard's definite objections to the possibility of an antigen-antibody
reaction are found in the following arguments based on experimental observations:
1. Sensibility persists for a long time, antibodies disappear rapidly.
2. In the serum of animals sensitive to horse serum antibodies to this
serum are not demonstrable by complement fixation.
3. Although sensitiveness can be transferred to a normal animal, neverthe-
less a definite period of incubation must elapse before the recipient becomes
sensitive.
To the first of these arguments Besredka3 objects by saying that, while it
is true that sensitiveness persists for a long time, the power to transmit anaphy-
laxis passively disappears rapidly, as Otto, Eichet, and others have shown.
The second contention is contradicted by the work of Nicolle and Abt.4
But since these workers made their observations upon rabbits their experiments
do not necessarily contradict those of Gay and Southard. This point at best is
a difficult one to determine, especially as recent investigations haye shown us
that under certain circumstances antigen and antibody may be found side by
side in the same serum without uniting and without therefore fixing alexin or
complement.
*Gay and Southard. Jour. Med. Ees., Vol. 16, 1907; Vol. 18, 1908;
Vol. 19, 1908.
2 Gay and Southard. Jour. Med. Ees., p. 427, 1908.
3 Besredka. Bull, de I'Inst. Past., 6, 1908, p. 826.
4 Nicolle and Abt. Ann. de I'Inst. Past., Vol. 22, p. 132, 1908.
386 INFECTION AND RESISTANCE
The point of their third argument has been discussed above.
It is clear that Gay and Southard separate distinctly the substance in the
antigen which sensitizes from that which exerts the toxic action on second injection.
Another theory which is based on such a separation of a sensi-
tizing and shock-producing element in the original antigen is that
of Besredka.
Besredka5 assumes that in the injected antigen (serum) there are two
separate substances. One of these, the sensibilisinogen, induces, during the time
of incubation, a specific antibody (sensibilisin) . This antibody remains in part
attached to tissue cells and in part circulates freely in the blood. The other
substance in the antigen he calls " antisensibilisin." This, at the second injection,
reacts with the sensibilisin and anaphylactic shock results. The nature of the
symptoms is explained by the fact that the antibody or sensibilisin is attached to
cells of the central nervous system, and shock can result only when such attach-
ment is present. Thus, in passive transference of sensitization, the property of
hypersusceptibility is bestowed upon the normal animal by the sensibilisin or
antibody present in the circulating blood, but the significance of this body for
anaphylaxis is not in evidence until a connection with the central nervous system
has been established.
There is much in Besredka's theory which is at variance with
prevailing conceptions of biological phenomena of this category. The
fact that an antigen should give rise to an antibody which reacts
not with the substance that induced it, but with a third body, is
quite out of keeping with experience.
However, it is clear that in both theories, that of Gay and South-
ard, as well as that of Besredka, the cardinal point is this separa-
tion in the antigen of two substances, a sensitizing and a toxic or
shock-producing, and, since this forms the chief argument against an
antigen-antibody conception of anaphylaxis, it will be necessary to
examine the experimental evidence on which it is based.
Gay and Adler 6 attempted to show such a dual function of the
original antigen by chemical methods. They report that, by frac-
tional precipitation of horse serum with ammonium sulphate, the
successive protein fractions obtained, as saturation is increased, are
found to be less sensitizing and more toxic as more and more am-
monium sulphate is added. The first fraction (euglobulins) obtained
by % saturation is as sensitizing as whole serum and corresponds to
anaphylactin, but is nontoxic when injected into sensitive animals.
The last fraction, while distinctly less sensitizing than either the
whole serum or the first fraction, is at least as toxic as the whole
serum.
In these experiments, therefore, we have a strong argument in
favor of the separate presence in an anaphylactic antigen of two
bodies, the one sensitizing and the other toxogenic. However, this-
assertion has not been borne out by later work.
5 Besredka. Loc. cit.
6 Gay and Adler. Jour. Med. Res., Vol. 13, 1908.
ANAPHYLAXIS 387
Pick and Yamanouchi,7 whose extensive investigation cannot be
fully reviewed here, were unable to obtain such a separation; in
fact, they conclude that the same substances which sensitize are also
toxic, and, working^with a large variety of methods, find that both
the sensitizing ^a~nd toxogenic properties of proteins show no differ-
ences either in chemical condition or in resistance to chemical agents
or heat.
The work of Pick and Yamanouchi, however, was done with
rabbits and, therefore, as bearing on the theory of Gay and Southard,
the work of Doerr and Russ 8 is more directly to the point. These
workers using guinea pigs, and both horse and beef sera, obtained
results which are practically diametrically opposed to those of Gay
and Adler. They found that the euglobulins, obtained by % satura-
tion with ammonium sulphate, are the most strongly sensitizing and,
at the same time, the most toxic of the fractions of the sera. As
saturation with the salt is increased, the proteins which come down
decrease progressively and in parallelism, both as regards the power
to sensitize and the faculty of exerting toxic action on second injec-
tion. The albumin, which finally comes out on total saturation, is
devoid both of sensitizing and of toxic properties. Similar results
were obtained by Doerr and Russ with the precipitation of serum
proteins with CO2.
The weight of evidence, therefore, seems to point against a chem-
ical separation of the two functions in the antigen.
Besredka's contentions in favor of such a separation were based
chiefly upon a difference in resistance to heat.
His experiments showed that the sensitizing properties of serum
are not lost even if it is heated to 120° C., while the toxogenic
powers are destroyed by much lower temperatures. The results of
Besredka as to the differences in thermostability between the two
properties have found confirmation by Kraus and Yolk 9 and others,
and there can be little doubt that the sensitizing function is extremely
heat-resistant, since this has also been shown by Wells,10 Rosenau
and Anderson, and many others. However, researches by Doerr and
Russ,1 * and notably by Wells, have shown that, though not destroyed
by high temperatures, even moderate heating markedly diminishes
the sensitizing function, and that larger doses have to be given as
the temperature is increased; and since the smallest quantities of
antigen necessary for inducing shock at the second injection must be
anywhere from 100 to 1,000 times as large as the smallest sensi-
tizing doses, it is quite likely that a combination of such conditions
7 Pick and Yamanouclii. Zeitschr. f. Immunitatsforschung, 1, 1909.
8 Doerr and Russ. Zeitschr. f. Immunitatsforschung, Vol. 2, 1909.
9 Kraus and Volk. Zeitschr. f. Immunitatsforschung, Vol. 3, 1909.
10 Wells. Jour. Inf. Dis., Vol. 5, 1908.
11 Doerr and Russ. Loc. cit.
388 INFECTION AND RESISTANCE
might stimulate an actual difference in heat resistance. In fact, this
is the view expressed by Wells 12 and borne out by experiments car-
ried out by Doerr and Russ.
Wells, too, confirms the identity of sensitizing and toxic sub-
stance by his experiments on the influence of tryptic digestion upon
these properties of the antigen. He concludes that both sensitizing
and intoxicating properties are attacked and slowly decrease as the
coagulable protein disappears.
As to that aspect of Besredka's theory which deals with the
indirect participation of the central nervous system, his arguments
are based mainly on the fact that ether narcosis seemed, in his
experiments, to prevent anaphylactic shock when animals were
deeply anesthetized during the second injection, and also upon the
regularity, severity, and speed with which anaphylactic symptoms
follow injections directly into the brain. The former contention
regarding narcotics cannot, by any means, be accepted as yet, since
Rosenau and Anderson failed to confirm it and claim that ether
narcosis merely masks the symptoms but does not prevent death.
If we admit the beneficial effects of ether, moreover, it may well be
that this is accomplished by relaxation of the bronchial spasms,
known, since Auer and Lewis, to be the cause of death in guinea
pigs, and the action of ether could hardly be utilized, therefore, to
argue in favor of a central localization of the anaphylactic process.
That phase of the two theories so far mentioned, therefore, which
depends upon the assumption of two separate substances in the orig-
inal antigen does not seem established nor even sufficiently likely
to warrant the formulation of a theory upon it.
The second premise is the necessary participation of the body
cell, in that the reaction cannot take place unless the cells are ren-
dered vulnerable by preliminary alteration. In Gay and Southard's
theory this is assumed to occur by irritation exerted by the "anaphy-
lactin," in Besredka's scheme it is attributed to the antisensibilisin
which is attached to the nerve cells.
It is plain, therefore, that both Gay and Southard and Besredka
admit a preliminary preparation of the cells of the body, and this is,
as we shall see, an important factor in anaphylaxis, though not
exactly in the sense of any of the observers named.
All other views of the mechanism of anaphylaxis have held from
the beginning that, in substance, this reaction is a true antigenr
(antibody reaction. The injected antigen gives rise to a specific
antibody. This, on second injection, unites with the first antigen
and the result is anaphylactic shock. Such a point of view was held
by v. Pirquet, Rosenau and Anderson, and others, who reached this
conclusion from the nature of the anaphylactic antigens, the speci-
12 Wells. Jour Inf. Dis., Vol. 6, p. 521, 1909.
ANAPHYLAXIS 389
ficity of the reaction, the incubation time, and the phenomena of
passive sensitization.
This assumption gained much further support when Doerr and
Kuss 13 succeeded in applying quantitative methods to the study of
the anaphylactic antibody. Their methods consisted in precipitating
sera in rabbits. With these they then passively sensitized guinea
pigs, subsequently testing them with antigen 24 hours later. To
arrive at quantitative results they developed two reliable methods.
These consisted in: 1. Intraperitoneal sensitization of guinea pigs
with constant quantities of titrated precipitating serum. Twenty-
four hours later intravenous test with diminishing amounts of specific
antigen. 2. Intraperitoneal sensitization W7ith diminishing quanti-
ties of the titrated precipitating serum, and 24 hours later intrave-
nous tests with constant amounts of antigen.
In this way they showed that there was a direct relationship
tween the power of a serum to convey anaphylaxis passively and its
contents of precipitins. We may elucidate this by an example from
their work. They possessed a rabbit serum which gave precipitation
with sheep, goat, beef, pig, human, and horse sera, but not with
chicken serum. The precipitation titre of this serum for the sera
mentioned varied from 1 in 20,000 in the case of sheep and goat
sera, to 1 in 100 in the cases of the human and horse sera. When
guinea pigs were injected intraperitoneally with 1 c. c. of this serum,
and after 24 hours were intravenously injected with the various sera
mentioned above, in decreasing quantities, the sera which were pre-
cipitated in the highest dilutions gave anaphylactic shock in the
smallest quantities. Those sera in which no precipitin or little had
been present gave little or no reaction by this method even where con-
siderable quantities were used. Thus in animals prepared by 1 c. c.
of the antiserum (precipitated in dilutions of 1 in 20,000) sheep
serum caused death when injected in doses of 0.006 c. c., whereas
horse serum (which was precipitated only in concentration of 1 to
100) gave slight symptoms only when 2 c. c. were employed for re-
injection and chicken serum (non-precipitable by the antiserum)
gave no reaction in similar doses.
In this, then, we have a definite quantitative analysis which
proves that the power to sensitize passively is in direct relation to the
antibodies against the protein present in the sensitizing serum.
Whether or not this means the precipitins we consider relatively un-
important inasmuch as we have already made clear that we believe
precipitins, complement-fixing antibodies and agglutinins to be one
and the same thing.
We may therefore accept, as a definitely determined fact, that
anaphylaxis is directly or indirectly the result of the reaction within
13 Doerr and Russ. Zeitschr. f. Immunitatsforscliung, Vol. 3, pp. 181
and 706, 1909.
390 INFECTION AND RESISTANCE
the animal body of an antigen with its specific antibody. In "active"
anaphylaxis the antibodies are present as the reaction to a preceding
antigen injection. In the "passive" condition they were conveyed
with the injected antiserum.
We are now prepared to follow individually the development of
those theories in which the anaphylactic mechanism was looked upon
purely as the result o^ the union of an antigen with its antibody.
The conception which gradually grew out of the antigen-antibody
mechanism of anaphylaxis was the following: When a specific an-
tigen meets its antibody the reaction between them gives rise to a
toxic product, and this causes the characteristic symptoms. A simi-
lar idea, it will be remembered, is found in the original endotoxiii
theory of Pfeiffer. According to this, the action of the specific
lysin liberated from bacteria a preformed poison, the endotoxin. In
1902 Weichhardt, 14 bearing this conception in mind, subjected
syncytial protein of rabbit placenta to the action of specific antisera
and obtained substances toxic for normal rabbits.
This work was done long before the days of anaphylaxis studies,
and the results were interpreted in keeping with Pfeiffer's theory.
However, as Weichhardt himself now claims, it is not unlikely that
he was dealing with a phenomenon analogous to the ones we are now
discussing. A similar opinion of the production of toxic substances
by specific cytolysis was expressed by Wolff-Eisner 15 in 1904.
Probably the most important of the earlier investigations along
these lines, at least in its direct bearing on anaphylaxis, was the
work of Vaughan and Wheeler/6 published in 1907.
In its general significance this work ranks among the most im-
portant contributions to our understanding of hyper susceptibility,17
though the theoretical deductions made from it have had to be sub-
jected to considerable alteration. Their conception of anaphylaxis
takes root in the earlier investigations of Vaughan 18 and his pupils
upon the extraction of a poisonous group from the protein molecule.
Vaughan and Wheeler 19, 20 believe that the sensitizing and the
toxogenic properties of the anaphylactic antigens are in truth con-
tained within the self-same proteid molecule; but can be chemically
separated from each other. They have been able to split egg al-
bumen and other proteids by treatment with absolute alcohol (con-
14 Weichhardt. Deutsche med. Woch., 1902, p. 624.
15 Wolff- Eisner. Centralbl f. Bakt., Vol. 37, 1904.
16 Vaughan and Wheeler. Jour. Inf. Dis., Vol. 4, 1907.
17 This work also contains the germ of the more recent ideas upon the
nature of toxemia in infectious disease, advanced more particularly by
Friedberger. This will be considered in detail in the next chapter.
18 Vaughan. Transact. Ass'n Am. Phys., Vol. 16, 1901; Jour. A.M.. A.,
Vol. 36, 1901; Am. Med., 1901; Jour. A. M. A., Vol. 43, 1904.
19 V C. Vaughan, Jr. Jour. A. M. A., Vol. 44, 1905, p. 1340.
20 V. C. Vaughan, Jr. Boston Med. and Surg. Jour., Vol. 155, 1906.
ANAPHYLAXIS 391
taining 2 per cent. NaOH) into 2 fractions — a toxic alcohol-soluble
and a non-toxic alcohol-insoluble one. The former fraction gave
protein reactions, and they regard it as a true protein — while Wells,21
considering the hydrolytic nature of the cleavage resorted to, con-
siders this fraction as possibly a soluble peptone or polypeptid (the
positive protein reactions being possibly due to ammo acids). The
non-alcohol-soluble, non-toxic fraction also gives proteid reactions.
Injections into guinea pigs of the toxic fraction produce symptoms
riot unlike anaphylaxis — but do not sensitize against protein. The
alcohol-soluble portion is non-toxic and sensitizes against protein in
doses of 0.001 to 0.005 gm. Based on these results, their views of
mechanism of anaphylaxis are as follows :
At the first injection a slow lysis (cleavage) of the injected pro-
tein gradually liberates a fraction, corresponding to the alcohol-in-
soluble substance — and this by its antigenic action gives rise to the
formation, in excess, of an enzyme (lysin), which on reinjection
brings about the rapid cleavage of the injected protein — with an ex-
plosive liberation of the toxic fraction and consequent symptoms.22
Nicolle believes that the injection of a protein into an animal induces
the production in the subject of antibodies. These are preeminently two —
albuminolysins, which cause its cleavage, and albuminocoagulins or precipi-
tins, which coagulate and prevent the action of the lysin. At the time at
which an animal is hypersusceptible or anaphylactic there has been a pro-
duction of albuminolysins which cause cleavage of the protein, with the
rapid liberation of toxic substances; but the albuminocoagulins or precipi-
tins have not yet adequately developed. In a refractory animal the neu-
tralizing action of the albuminoprecipitins prevents the harm which the lytic
action might otherwise accomplish. The relative amounts of these two anti-
bodies present in the circulation of the animal at any particular time deter-
mine whether the animal is anaphylactic or refractory or immune. This
theory assumes arbitrarily the protective nature of precipitation, an idea
which has no foundation in experiment and, in fact, is rendered extremely
unlikely by more recent developments of our knowledge of the precipitating
antibodies.
Given, then, a reasonable hypothesis in which anaphylaxis is
associated with the cleavage of protein by lysis, given, in other words,
an antigen-antibody conception, it is but natural that experimenters
should ask themselves: What is the relation of the alexin to this
cleavage ? For in all known lytic reactions, of course, the union of
antigen and antibody leads to the absorption of alexin, by means of
which, then, the lysis has been assumed to take place. This problem
suggested itself to a number of the earlier investigators who attempted
to approach it by determining whether or not the sera of sensitive
21 Wells. Jour. Inf. Dis., Vol. 5, 1908.
22 For the sake of completeness it is well also to mention Nicolle's 23
iheory, which, though attractive, is not borne out by recent knowledge con-
cerning the nature of precipitins.
2:! Nicolle. Am. de I'Inst. Past., Vol. 22, 1998.
392 INFECTION AND RESISTANCE
animals, added to antigen, would fix alexin. Gay and Southard,
Sleeswijk,24 and others obtained negative results, while Nicolle and
Abt,25 and Doerr and Russ 26 obtained positive results. As far as
this particular method is concerned, therefore, no conclusions can
be drawn. Sleeswijk, however, has approached the question in an-
other way and examined whether or not there is a diminution of
alexin in the blood of an animal immediately after anaphylactic
shock. He found that this was indeed a regular occurrence, and his
results have been confirmed by Friedberger and Hartoch 27 and a
number of others.
Jt was shown by these workers that, both in active and passive
anaphylaxis in rabbits and dogs, as well as in guinea pigs, there is a
definite and considerable diminution of complement immediately
after anaphylactic shock.
The question now arises : What is the significance of this dimi-
nution of alexin? Do the animals die because of a sudden loss of
circulating, physiologically necessary alexin, or does the alexin take
an active part in producing the conditions which cause death ?
Either of these possibilities might follow from the mere fact of
alexin diminution, but the former — the possibility that complement
depletion is the cause of death — was ruled out by Friedberger and
Hartoch.28 They showed that, by supplying fresh complement to
sensitive animals at the time of rein ject ion, shock cannot be pre-
vented. They now proceeded to demonstrate the active participation
of complement in the production of anaphylaxis. They did this in an
ingenious way which depended on utilization of the fact observed by
Nolf,29 Hektoen and Ruediger,30 and others that hypertonic salt
solution (1.5-2 per cent.) will prevent the combination of comple-
ment with its sensitized cells. By slowly injecting into sensitized
guinea pigs 0.3 cubic centimeter of concentrated NaCl solution
just before the injection of antigen they were able to markedly
diminish anaphylactic shock — saving animals from injections which
invariably killed the controls. The force of this experiment we think
has been largely eliminated by work of the writer with Dwyer and
Lieb31 in which it was shown that the effect of the salt is upon the
rmooth muscle which is rendered less irritable by the salt and there-
fore less susceptible to the antigen.
24 Sleeswijk. Zeitschr. f. Immunitatsforschung, Vol. 2, 1909.
25 Nicolle and Abt. Ann. de rinst. Past., Vol. 22, 1908.
26 Doerr and Russ. Zeitschr. f. Jmmunitcitsforschung, Vol. 3, 1909.
27 Friedberger and Hartoch. Zeitschr. f. Immunitatsforschung. Vol. 3,
1909.
28 Friedberger and Hartoch. Loc. cit.
9 Nolf. Ann. de I'Inst. Past., 1900.
30 Hektoen and Ruediger. Jour. Inf. Dis., Vol. 1, 1904.
31 Zinsser, Lieb and Dwyer, Proc. Soc. for Exn. Biol. & MecL, Vol. 12,
May, 1915, p. 123.
ANAPHYLAXIS 393
An ingenious attempt to demonstrate the important role played
by complement in anaphylaxis has recently been furnished by Loef-
fler. Loeffler,32 using guinea pigs sensitized with horse serum, com-
pletely depleted their complement by injecting intraperitoneally con-
siderable quantities of sensitized sheep corpuscles. Tested by
injection of horse serum one hour later no anaphylaxis occurred,
while controls regularly succumbed.33
It seemed thus established that the complement or alexin played
an important active part in the production of anaphylaxis, and the
next logical step was to attempt to produce the anaphylactic poison
by the action of alexin upon an antigen-antibody complex in vitro.
This was first done, with direct reference to anaphylaxis, by Ulrich
Friedemann.34 Friedemann chose as his antigen-antibody complex
the sensitized red blood cell after he had demonstrated by preliminary
experiment that the basic anaphylactic experiment could be carried
out in rabbits with washed beef corpuscles. He found that if 3 c. c.
of such corpuscles were injected into rabbits and the injection re-
peated after 3 weeks anaphylactic symptoms were regularly elicited.
He then allowed alexin to act upon sensitized beef blood in vitro,
interrupted the action by cooling at a time just preceding the occur-
rence of hemolysis (to exclude the supposed toxic action of hemoglo-
bin), and injected the supernatant fluid of such mixtures into normal
rabbits. The result was marked illness resembling anaphylaxis, and
Friedemann thus had succeeded in producing the anaphylactic poison
in vitro under conditions as nearly as possible similar to those occur-
ring in the circulation of the anaphylactic rabbit. In the conclusions
drawn from his experiments he expresses the opinion that the poisons
were not preformed in the red blood cells, but were formed by the
proteolysis exerted by "amboceptor" and complement. In this state-
ment he sets down the basic conception of the production of anaphy-
lactic poisons now generally held.
Friedemann, then, in attempts to apply the same methods to the
study of serum anaphylaxis, attempted to produce similar poisons
by the action of rabbit alexin upon the washed precipitates formed
by mixtures of antigen and precipitating sera. In this he failed —
probably because of his choice of rabbits as subjects for experiment.
Where he had failed, however, Friedberger 35 succeeded by using
guinea pigs. Doerr and Russ 36 had previously shown that feeble
symptoms of shock could be produced by the injection of serum pre-
32 Loeffler. Zeitschr. f. Immunitatsforsch., 8, 1910.
33 For additional evidence pointing in the same direction see also TJhlen-
huth and Haendel, Zeitschr. f. Immunitatsforsch., Vol. 3, 1909.
34 Ulrich Friedemann. Zeitschr. f. Immunitatsforsch., Vol. 2, 1909.
35 Friedberger. Berl. klin. Woch., 32 and 42, 1910 ; also Zeitschr. f.
Immunitatsforsch., Vol. 4, 1910.
36 Doerr and Russ. Zeitschr. f. Immunitatsforsch. y Vol. 3, p. 181, 1909.
394 INFECTION AND RESISTANCE
cipitates into normal guinea pigs. With this additional evidence
in favor of his reasoning, Friedberger proceeded as follows:
One c. c. of a rabbit serum which precipitated sheep serum in a
dilution of 1 to 10,000 was mixed with 30 c. c. of a 1 to 50 sheep
serum dilution. This was kept one hour at 37.5° C. and over night
in the ice-chest, when a heavy flocculent precipitate had formed.
This precipitate was washed to remove all traces of serum, and to it
were added 2 c. c. of fresh normal guinea pig serum — as comple-
ment. This was again allowed to stand for 12 hours and then the
supernatant fluid was injected into a guinea pig intravenously. In
most cases the pigs so treated showed marked symptoms soon after
the injection and died within a few hours.
Friedberger concludes, therefore, that anaphylactic shock is a
true intoxication due to a poison produced from the products of a
precipitin-precipitinogen reaction by the action of a complement;
he speaks of the formed poison as anaphylatoxin. The experiment
just outlined, moreover, seems to show, contrary to Friedberger's
first ideas, that the entire reaction may go on under certain circum-
stances in the blood stream without intervention of sessile precipitins
upon the cells.
We have, thus, in the cited work of Friedberger the culmination
of a long series of investigations — the end result being the conclusion
that in all probability — at least as far as experimental ingenuity
has permitted us to penetrate into this very difficult problem up to
the present time — the phenomenon of anaphylaxis must be regarded
as an acute intoxication, the poison which calls it forth being the
result of the union of an antigen and its antibody, the complex being
subsequently subjected to proteolysis by the action of alexin or com-
plement. The experimental extension of this conception to the phe-
nomena of bacterial anaphylaxis has promised to exert such an im-
portant influence upon our conceptions of infectious disease that we
will take up these investigations in a separate section.
37 Here, then, we have a simple and apparently logical explanation
of anaphylaxis, entirely in accord with Vaughan's views of parenteral
digestion. An antigen is injected into an animal, specific antibodies
and enzymes against it develop in the animal ; reinjectiori of this
antigen results in relatively rapid proteolysis in the course of which
poisonous substances, the anaphylatoxins, are produced and anaphy-
laxis is the result. This hypothesis although very attractive does
not entirely meet with the facts as they have been developed since
Friedberger's first work. The premises on which it is based assume
in the first place that the poison or "anaphylatoxin' ' is formed out of
the matrix of the antigen ; further, it is definitely assumed that in the
37 Much of this discussion is adapted from Zinsser, "More Recent De-
velopments in the Study of Anaphylactic Phenomena." Harvey Lecture
Arch, of Int. Med., Aug., 1915, Vol. 16, p. 223.
ANAPHYLAXIS 395
production of the poison after the antigen and antibody have met, the
complement or alexin plays an active part. Friedberger's hypothesis
as stated by him, moreover, seems to assume that the entire process
takes place intravascularly, a matter which we will discuss at consid-
erable length in a short time. It is important to note also that Fried-
berger, with Nathan, was able to show that this anaphylatoxin produc-
tion could take place within the animal body ; that is, within the peri-
toneum of a guinea-pig into which bacteria had been injected.
The simplicity of Friedberger's explanation and the correctness
of his experimental data soon persuaded many investigators that, in
essence, his hypothesis probably contained the nucleus of the solution
of this difficult problem. However, even his own early experiments
aroused some misgivings concerning the matrix of the poisons pro-
duced, for he found that the poisons could be obtained as well when
boiled antigen was used as when the fresh, unheated substances were
employed, and the poisons were easily obtained from such organisms
as the tubercle bacillus, which is extremely insoluble and unamenable
to serum influence. It was also doubted whether one could truly
assume the participation of this specific antibody or sensitizer in the
production of Friedberger's poisons, since it soon developed that from
bacteria, at least, the poison could be produced when the organisms
were directly exposed to the action of fresh guinea-pig serum without
the presence of any immune serum.
Experiments which soon cast a definite doubt on the assumption
that the poisons were produced by a decomposition of the antigen
were reported by Keysser and Wassermann,38 These workers sub-
stituted insoluble substances like barium sulphate and kaolin for the
antigen; that is, the precipitates or bacteria used in Friedberger's
experiments. They found that if kaolin were treated with horse
serum and, then exposed to the action of guinea-pig serum or comple-
ment, poisons were produced identical in every respect to those pro-
duced by Friedberger's method. The conclusions they drew were
that the poisons were produced, not by action of the complement on
the antigen, but by its action on the horse serum absorbed by the
kaolin. In other words, they transferred the matrix of the poison
from the antigen to constituents in the serum itself, possibly the
•sensitizer or amboceptor. Bordet 39 also was able to show that poisons
similar to those of Friedberger could be produced by the action of
fresh guinea-pig serum on agar, and recently Bordet has further
shown that this is the case even when the agar has been by special
methods deprived entirely of its nitrogenous components and repre-
.sents simply a complex of carbohydrates. Agar-guinea-pig serum
mixtures of this kind show an increase in total nonprotein nitrogen
38 Keysser and Wassermann. Folia serol, 1911, vii; Ztschr. f. Hyg. u.
Infectionskrankh., 1911, Ixviii.
39 Bordet. Compt. rend. Soc. de biol, 1913, Ixxiv, 877.
396 INFECTION AND RESISTANCE
which would prove that the proteolytic action of the guinea-pig serum
must have been active against its own proteins.
An interesting further development of this work has recently
appeared in the experiments of Jobling and Petersen.40 They
showed that when bacteria are mixed with fresh active serum they
adsorb the antienzymes normally present in blood. They have shown
this experimentally and have proved that similar antienzyme re-
moval can be accomplished by the addition of kaolin or agar, and by
treatment with chloroform. Serums so treated become toxic, the
actions of the poisons formed showing great similarity to that pro-
duced by Friedberger's anaphylatoxins. According to them, the
poisons are formed because of the fact that antienzymes are adsorbed
by the antigen, thus setting the normal ferments in the fresh serum
free to act on their own serum protein.
It should be recalled that Friedemann, who was really the first
one to show that the toxic substances could result from the interaction
of fresh serum and sensitized antigens, although he succeeded only
in doing this with red blood cells, suggested rather early that the
success of such an experiment does not necessarily mean that the
antigen furnishes the matrix entirely. He had studied the metabo-
lism in anaphylactic poisoning and with Isaac has shown that the
nitrogen output following reinjection in a sensitized animal is far in
excess of that which could be derived solely from the injected antigen,
and in this he has been confirmed by many other workers, notably by
Vaughan. Moreover, recent investigations of Jobling and others have
seemed to show that proteolysis is not necessarily a function of
antibodies but is accomplished by non-specific proteoses in the blood.
It would seem to us that our present knowledge of this phase of
anaphylactic investigation permits us only to conclude that wherever
proteolytic changes take place these "proteotoxins" may be formed.
That they can be produced from a protein antigen has been shown
beyond doubt by Vaughan and his collaborators for both formed and
unformed antigens. Also this is evident from the experiments of
many workers and has been confirmed in our own experience with
poisons appearing during the autolysis of bacterial emulsions. On
the other hand, it is also clear that the antigen need not represent
the matrix which furnishes the poison, and that in the reactions as
they are generally performed both in the test tube and in the animal
body, it is more than likely that if an antigen participates at all in
furnishing the substratum for the poison, this is probably less impor-
tant than that furnished by the animal's own proteins. However,
this does not weaken the importance of the knowledge that the
antigen-antibody reaction in the presence of normal serum and cer-
tain antigens in the presence of normal serum alone, induce a reac-
tion in the course of which such poisons are formed. And the fact
40 Jobling and Petersen. Jour. Exper. Med., 1914, xix, No. 5.
ANAPHYLAXIS' 397
that they can be produced experimentally in the peritoneal cavity
of a living guinea pig renders their participation in such reactions in
the animal body a likely assumption.
Our own work 41 on these substances induces us to believe that
proteotoxins so formed are identical with Bail's aggressins, a point
to which we will later refer.
It is plain that the foregoing work has done away almost entirely
with the idea that anaphylactic phenomena were chiefly intravascular.
The brilliant experiments of Vaughan, of Friedemann, and of Fried-
berger seemed entirely convincing. They had succeeded in producing
a toxic substance apparently by cleavage of typical antigens in the
test tube by chemical and serological methods, and, on the face of it
there seemed no reason to doubt that such reactions might occur in
the circulating blood, thus explaining anaphylaxis. It has been on
the basis of the work of these pioneers, and experiments of others
based on the principles of this early work, that the intravascular
theories of anaphylaxis mechanism have been founded. For a time
these views held the center of the stage, and in their elaboration many
conditions were revealed which we believe have important bearing
on anaphylaxis in rabbits, perhaps in dogs, and in such special forms
of anaphylaxis as that produced by cellular antigens such as bac-
teria. However, from the beginning there have been irrefutable
experimental data which have prevented the complete acceptation
of the intravascular theory, indeed have caused it to seem likely that
in guinea pigs anaphylaxis was purely a cellular phenomenon and
that in anaphylaxis in general the intravascular occurrences were,
if at all important, secondary to the cellular mechanism. The most
important point which has always inclined workers to come back again
and again to the assumption of a participation of the body-cells, is
the necessary interval that must elapse (certainly in guinea pigs, and
with almost equal certainty in other animals) between the time of
injection of an antiserum, and the development of the anaphylactic
state in the animal.
This observation, made by many of the earlier writers, has been
the crucial point of controversy not yet completely settled. The work
of Weill-Halle and Lemaire, which we have cited in a preceding-
section, seemed to show that the interval was not always necessary.
Eichet, too, obtained experiments with crepitin which seemed to ex-
clude the necessity of the interval. It is interesting to note that he
spoke of his experiments as "reaction anaphylactique in vitro/' He
sensitized a dog to crepitin, then bled him during the hypersusceptible
period, mixed the serum with a harmless dose of crepitin, and in-
jected the mixture into a normal dog. Violent anaphylaxis resulted
almost immediately.
At about the same time Friedemann published his very impor-
41 Zinsser and Dwyer. . Jour. Exper. Med., 1914, xx'; No. 6.
398 INFECTION AND RESISTANCE
tant studies on the mechanism of anaphylaxis in rabbits. He found
that passive sensitization in these animals, in contrast to the work of
others upon guinea pigs, was best obtained by the simultaneous in-
travenous injection of antigen and anaphy lactic serum. If the injec-
tion of the sensitive serum preceded that of the antigen by as much
as 24 hours, the reaction became indistinct (undeutlich), and Friede-
mann concluded that here, at least, there could not be assumed the
necessity of preliminary sensitization of the body cells by the anaphy-
lactic serum, as is the case of guinea pig anaphylaxis. The anaphy-
lactic poison, whatever it may be, Friedemann concludes is, in rabbits
at least, formed in the circulating blood. In 1910 Biedl and Kraus 42
obtained immediate and severe symptoms in guinea pigs when they
injected intravenously mixtures of horse serum together with the
serum of sensitized guinea pigs. Briot 43 in the same year obtained
reactions in young rabbits into which he had injected mixtures of
horse serum and anti-horse serum. Gurd 44 in a recent publication
obtained reactions in guinea pigs when he injected intravenously
immune rabbit serum (anti-sheep serum) and immediately thereafter
sheep serum. We ourselves have been able to obtain occasional and
distinct results in rabbits and guinea pigs both by simultaneous and
immediately consecutive intravenous injections of antigen and anti-
body, though we did not succeed in attempts to duplicate exactly the
experiments of Friedemann and of Biedl and Kraus.
Recently too Manwaring and his co-workers studying the isolated
hearts of rabbits, and the isolated lungs of guinea pigs, by physio-
logical methods, have claimed that they can observe reactions due to
the meeting of antigen and antibody in the blood vessels.
There is thus a considerable amount of evidence that in rabbits,
and some other animals, the meeting of antigen and antibody in the
blood stream may lead to anaphylactic reaction. However, the evi-
dence in guinea pigs that this may occur is very slim, and to any one
who has worked in anaphylaxis consistently it is quite plain that in
these animals the simultaneous injection of the two substances leads to
either no symptoms, or such very slight ones that they play no im-
portant part in the fatal reaction. Even in other animals, further-
more, we ourselves feel that mere contact within the blood stream be-
tween antigen and antibody cannot account for the entire train of
phenomena, and one is forced to assume that any considerable degree
of hypersusceptibility must be essentially cellular and due to the
reaction of injected antigen with antibody that has previously be-
come attached in some way to tissue cells.
The idea in itself is not new. Wassermann had first suggested it in
an attempt to explain the peculiar hypersusceptibility to toxin pos-
42 Biedl and Kraus. Ztschr. f. Immunitatsforsch., 1910, iv.
43 Briot. Compt. rend. Soc. de biol, 1910. Ixviii, 402.
44 Gurd. Jour. Med. Besearch, 1914, xxxi, 205.
ANAPHYLAXIS 399
sessed by some of Behring's toxin-immunized animals. He assumed
that in such animals the formation of antitoxins may indeed have
been stimulated, but that much of it might still be attached to the
generating cells themselves, thereby rendering these proportionately
more vulnerable to the injected toxin.
Such a conception of "sessile receptors" was applied by Fried-
berger to anaphylaxis in his first attempts to formulate an hy-
pothesis. He assumed that at the first or sensitizing injection the
production of antibodies (precipitins) was stimulated. These, how-
ever, were not produced in great quantity and were not discharged
into the circulation, possibly owing to the small single dose given
for sensitization. They were present at the end of the anaphylactic
incubation time as sessile receptors or sessile antibodies (precipitins).
On the second injection a reaction occurred between the injection
antigen and these sessile precipitins and the cell was injured because
the reaction occurred on its substance, a reaction which, it is sug-
gested, might have been harmless had it taken place in the blood
stream. In passive sensitization, conversely, no injury could result
until considerable quantities of the antibody had become united to
body cells in the course of several hours. That the antibody injected
into passively sensitized animals indeed disappears from the circula-
tion with relative speed, has been shown by Doerr and again recently
by Weil.
Direct study of the cellular conception was made possible by
methods such as the transfusion method employed in anaphylactic
dogs by Pearce and Eisenbrey45 and the technique of observing
isolated tissues from anaphylactic animals as used by Schultz 4(
work which appeared as early as 1910. Pearce and Eisenbrey work-
ing with two normal and one sensitized dog, transfused the blood
of one of the normal animals into the sensitized one, transferring the
blood of the latter to the normal dog. "At the proper moment
the normal dog containing the blood of the sensitized animal and
the latter containing the blood of the normal dog, each received
intravenously the toxic dose of horse serum." The normal dog having
the sensitized blood did not react, the sensitized dog having the nor-
mal blood showed typical fall of blood pressure. Pearce and Eisen-
brey drew the conclusion "that the phenomenon of anaphylaxis is due
to a reaction in the fixed cells and not either primarily or secondarily
in the blood." Similar experiments have since been done by Weil.
In the same year Schultz began to work with what is now spoken
of as the physiological method. He determined that smooth muscle
— freshly excised from various animals — will react with contraction
when brought into contact with serum. When such muscle was
taken from anaphylactic animals after being thoroughly washed free
45 Pearce and Eisenbrey. Cong. Am. Phys. and Surg., 1910, viii.
46 Schultz. Jour. Pharmacol. and Exper. Therap., 1910, i.
400 INFECTION AND RESISTANCE
of blood, it would react more energetically and to small amounts of
the homologous serum. There are many interesting by-products of
Schultz's work, such as the differences between fresh arterial blood
and blood serum in their abilities to stimulate contraction, but this
and other points will not be discussed at present. The important and
incontrovertible fact established by Schultz is the changed reaction-
energy or, in truth, "allergie" of the smooth muscle of anaphylactic
animals to the stimulus of the sensitizing antigen. Dale 47 has con-
firmed and extended these observations of Schultz. He removed the
uteri from guinea pigs after thoroughly perfusing them with Ringer's
solution to remove all blood. He then suspended them in baths of
Rimer's solution and by the customary physiological methods meas-
ured the contractions following the addition of various amounts of
foreign protein in the form of — among other things — horse serum
and beef serum. He found that the uterus of an animal sensitized
to horse serum would react to this substance in dilutions of 1 :2,000
or 1 : 10,000, while the organ taken from a normal guinea pig reached
its limit of reactionability at dilutions often less than 1 :200. A
uterus that had reacted strongly was found to be subsequently desen-
sitized. A normal uterus could not — strangely — be passively sen-
sitized by immersion into a solution containing serum antibodies.
This method of investigation has recently, also, been taken up by
Richard Weil 48 who has fully confirmed the principles laid down
by Schultz and Dale. He has incidentally also answered an objec-
tion to the conclusions of Dale and Schultz (never indeed a very
valid objection), namely, that the reaction of the muscle tissue of a
sensitized animal might be in part due to the fact that the blood,
i.e., the antibodies, had not been entirely washed out of the tissue
spaces by perfusion. Weil performed the very simple and ingenious
experiment of injecting a normal guinea pig with large amounts of
immune serum (anti-horse serum) and, after a few minutes, killing
the animal. He then suspended the uterus in Ringer's solution in
the usual manner without washing it completely free of blood. Con-
tact with the homologous antigen produced no response. We may
accept as definitely established by these researches of Schultz, Dale,
and Weil that the fixed cells of anaphylactic animals possess an in-
creased reaction-ability toward the antigen which is in no sense sec-
ondary to processes involving the circulating antibodies. Moreover,
the work of We1"! seems to indicate that desensitization of a pas-
sively prepared guinea pig deprives the uterus of its power to re-
spond and that the gradual spontaneous diminution of hypersuscep-
tibility on the part of the guinea pig is accompanied by an entirely
parallel loss of reaction-capacity on the part of the isolated uterus.
47 Dale. Jour. Pharmacol. and Exper. Therap., 1913, iv.
48 Weil, R. Jour. Med. Research, 27, 1913; 30, 1914; Proc. Soc. Exper.
Biol. and Med., 1914, xi, 86.
ANAPHYLAXIS 401
The recent work of Coca,49 too, has further fortified the cellular
point of view by a method which in principle is similar to that
employed by Pearce and Eisenbrey. Coca succeeded in perfusing
actively and passively sensitized guinea pigs with the defibrinated
blood of normal guinea pigs in such a way that the original blood of
the sensitized animals was reduced to a necessarily slight residue.
Animals so treated could be kept alive for as long as six hours after
the transfusions and remained delicately hypersusceptible in spite
of the blood substitution.
Limiting ourselves for the present to the phenomena of anaphy-
laxis in which noncellular antigens are employed, we may safely
say that the evidence furnished by the incubation time necessary in
passive anaphylaxis, by the transfusion experiments of Pearce and
Eisenbrey and of Coca, and most conclusively by the work on iso-
lated tissue by Schultz, by Dale and by Weil, shows conclusively
that the hypersusceptible state is largely determined by a changed
reaction-capacity to the specific antigen on the part of the fixed tissue
cells — an "alergie" which is probably due to the presence of specific
antibodies in the substance of the cell protoplasm, and incidentally
accounts for such effects as the skin reactions. It is probable that
the acute symptoms and death of anaphylactic guinea pigs (and
perhaps of other animals) is in most cases of experimental ana-
phylaxis due to the reaction which takes place between the injected
antigen and these sessile receptors.
Now, as to the identification of the anaphylactic antibody with
some one of the well-known antibodies, the assumption is that ia
cellular anaphylaxis (as in the corpuscle experiments of Friede-
mann and in the bacterial experiments of Friedberger and others)
the so-called sensitizer or amboceptor is to be held responsible. This
seems reasonable, and there is much evidence in its favor, no reliable
evidence against it.
In the case of serum anaphylaxis extensive work has been done
to show a parallelism between the anaphylactic antibody and the
precipitins. This we have seen principally in the experiments of
Doerr and Russ, and those of Friedberger.
The problem becomes a complicated one when we attempt then
to define the nature of the precipitins and their relation to the anti-
bodies hypothetically advanced as "albuminolysins" by Gengou>
Without going into this point extensively at present, it may be per-
mitted to refer the reader to the chapters on alexin fixation and
precipitins, and to reiterate the writer's 50 own opinion, which is
that much reasonable evidence points to the fact that the so-called
precipitins are in truth protein-sensitizers, in structure and function
identical with the sensitizers or amboceptors of cytolytic processes.
49 Coca. Ztschr. f. Immunitatsforsch., 1914, xx.
50 Zinsser. Jour. Exp. Med., Vol. 15, 1912, and Vol. 18, 1913.
T
402 INFECTION AND RESISTANCE
The fact that precipitation occurs when these antibodies are added
to the homologous dissolved antigen is merely a secondary colloidal
phenomenon. Antigen and antibody react, forming a complex which
is then amenable to the action of alexin. Being colloidal in nature,
and mixed under suitable quantitative and other conditions which
favor flocculation, they precipitate. This point of view, then, identi-
fies the so-called precipitins with the protein-sensitizers or albumino-
lysins first hypothetically suggested by Gengou. It leads necessarily
to the conception that in cytolysis as well as proteolysis, in fact, in
all reactions in which antigen is sensitized to the action of alexin,
there is functionally but one variety of antibody — the sensitizer
, — precipitation and agglutination being incidental physical phe-
nomena not dependent upon special antibodies as heretofore sup-
posed.
In this sense, then, the "precipitins" or albuminolysins may be
regarded as identical with the anaphylactic antibody.
That animals in whose circulation antigen and antibody are
simultaneously present do not suffer from symptoms of anaphylaxis
has been referred by Zinsser and Young 51 as possibly due to the
action of a protective colloid which prevents the union of the two.
THE MECHANISM OP ANTIANAPHYLAXIS
The conditions under which these animals, previously anaphy-
lactic, may be rendered refractory or "antianaphylactic" have been
discussed in another place. This condition is not entirely comparable
to immunity since it is a purely temporary state, lasting possibly a
few weeks, but after this the animals do not return to the normal
condition, but gradually become again moderately hypersusceptible.
(Rosenau and Anderson — Otto and others.) Thus a guinea pig
which has been sensitized, then rendered antianaphylactic by a mas-
sive injection of antigen, may react with mild symptoms to an in-
jection made 20 to 30 days later. Such returning sensitiveness,
according to Rosenau and Anderson 52 is usually mild, fatal reac-
tions rarely occurring.
A satisfactory theory of antianaphylaxis has not yet been ad-
vanced.
Besredka,53 as we have seen, believes that the anaphylactic reac-
tion takes place by the union of the toxic factor in the serum (anti-
sensibilisin) with a specific antibody sessile upon the cells of the
61 Zinsser and Young. Jour. Exp. Med., Vol. 17, 1913.
52 Rosenau and Anderson. Pub. Health and M. H. S. Hyg., Bull.} 36,
1907.
53 See Besredka, "Kraus u. Levaditi Handbnch, etc.," Erganzungsband ],
p. 246.
ANAPHYLAXIS 403
central nervous system. If the antigen is injected slowly or in small
amount these sessile receptors are gradually united to antigen with-
out fatal shock, and the animal is thereby rendered insensitive.
In his own words, this "desensitization" amounts to a return of
the cells to their normal preanaphylactic or naturally unsensitive
condition. With the refutation of his theory of anaphylaxis, his
theory of anti anaphylaxis also falls to the ground, and neither of the
two can be accepted as valid at present.
If we look upon anaphylaxis as a reaction taking place entirely in
the circulation we may accept, with Rosenau and Anderson 54 Fried-
berger, and others the explanation that antianaphylaxis is due to a
saturation of the anaphylactic antibody with antigen. Hypersus-
ceptibility is then subsequently reestablished because a gradual
formation of circulating antibody continues, and eventually free
antibody will again be present in the blood. This view is only in
part satisfactory, as Friedemann 55 points out. For it does not
explain the antianaphylaxis which Biedl and Kraus 56 have noticed
after the injection of mixtures of antigen and antibody, nor the non-
specific antianaphylaxis which the same workers have observed after
peptone injections. It is clear that the nature of antianaphylaxis
remains for the present obscure, and, in view of the recent attempts
to account for certain phases of infectious disease by the anaphy-
lactic phenomena, is one of the most important problems of im-
munity.
Bearing upon this condition of antianaphylaxis is the tolerance
to the anaphylactic poison which has been observed to develop in
animals once or twice injected. Vaughan 57 has noticed this in ani-
mals injected with his toxic split products, produced by alkaline-
alcohol splitting of colon bacilli. By repeated injection of the guinea
pigs he showed that a tolerance was developed which protects the
animal from about double the fatal dose, but the animal is not pro-
tected against larger multiples, and the condition is not an immunity
in the sense in which we have used the term. Similar observations
have been made by Bessau.58 Bessau passive sensitized guinea pigs
with 1 c. c. of anti-horse serum intraperitoneally, and on the follow-
ing day injected them intravenously with 1 c. c. of horse serum. He
gauged his dose so that the animals should have severe shock but
survive. One or two days later he injected the amount of typhoid
anaphylatoxin which was fatal for normal pigs, and found that those
54 Rosenau and Anderson. U. S. Pub. Health and M. H. S. Hyg. Lab.
Bull., 64, 1910.
55 Friedemann. "Frei Vereinigung f. Mikrobiol.," Berlin, 1910. Ref.
Centralbl. Bakt. I, Vol. 47; Beiheft, p. 1, 1910.
56 Biedl and Kraus. Zeitschr. f. Imm., Vol. 4, 1910.
57 Vaughan. "Protein Split Products, etc.," Lea & Febiger, Philadelphia,
1913, p. 139.
58 Bessau. Centralbl. f. Bakt., Vol. 60, 1911.
404 INFECTION AND RESISTANCE
which had been treated as described were now able to withstand the
anaphylatoxin These experiments of Bessau would indicate that
antianaphylaxis was to a certain extent due to tolerance to the poison,
and that it was non-specific. Friedberger, together with Szymanow-
ski, Kumagai, Odaira and Lura, later studied this problem and came
to the conclusion that antianaphylaxis is strictly specific, depending,
as Friedberger had suggested, upon the diminution of specific anti-
bodies rather than upon tolerance to the poison. They claimed that
animals that had been sensitized and then had survived the "shock"
dose of homologous protein showed no tolerance for anaphylatoxin,
and that animals that had been treated with the sublethal dose of
anaphylatoxin are, for 24 hours, as sensitive to anaphylatoxin, how-
ever prepared, as are normal animals. Recent studies along the
same lines by Zinsser and Dwyer 59 have yielded results differing
from these conclusions. Working with typhoid anaphylatoxin they
found that guinea pigs treated with a sublethal dose of anaphylatoxin
develop a tolerance which enabled them to resist 1% to 2 units of
poison, the tolerance developing within three days and lasting, to a
slight degree, for as long as two months. It seemed to them that
animals treated with a second dose of anaphylatoxin within 24 hours
after the first, if the results of this first injection have been severe, as
they usually are, are still weak and generally depressed in vitality
so that a developed tolerance may be clouded by this condition. The
tolerance did not seem to be strictly specific in that typhoid anaphy-
latoxin seemed to produce a moderate tolerance to prodigiosus
anaphylatoxin.
It would seem, therefore, that in antianaphylaxis we might have
two very important elements. The one strictly specific depends upon
the depletion of antigen from the body, a true "desensitization." The
other non-specific, and probably of secondary importance since so
far it has not been shown to any very powerful degree, consists of
the development of tolerance by the body cells for the anaphy lactic
poison.
NATURE OF ANAPHYLACTIC POISON
As to the nature of the anaphylactic poison we are also to a large
extent in the dark. From the experimentation upon the production
of these poisons in vitro it appears that they are protein cleavage
products. This is indirectly indicated also by metabolism experi-
ments— such as those of Friedemann and Isaak,60 and of Weichhardt
and Schittenhelm.61 It appeared from this work that, as measured
59 Zinsser and Dwyer. Reported at the meeting of Am. Ass. of Path, and
Bact., Toronto, April, 1914.
60 Friedemann and Isaak. Zeitschr. f. exp. Path. u. Ther., Vol. 1, 1905.
61 Schittenhelm u. Weichhardt. Munch, med. Woch., 1910, No. 34, and
1911.
ANAPHYLAXIS 405
by nitrogen output, the cleavage of foreign protein injected into
specifically sensitized or immunized dogs occurred with much greater
energy and speed than occurred in normal animals after first in-
jection.
Attempts to obtain the poison by non-specific methods — that is,
by purely chemical processes without the agencies of alexin and sen-
sitizer or antibody, have been made with apparent success by
Vaughan and Wheeler,62 whose toxic, alcoholic-soluble fraction (ob-
tained by boiling egg white in absolute alcohol containing 2 per cent.
NaOH), seems to produce typical anaphylaxis in guinea pigs. This
substance Vaughan and Wheeler regard as a protein, whereas Wells 63
states that it may be this, or a "soluble peptone or polypeptid, con-
taining enough of the different aminoacids to give all the usual reac-
tions." Weichhardt,64 too, has obtained similar poisons by a method
similar in principle to that of Vaughan and Wheeler.
This substance is, according to him, pharmacologically identical
with his "keno toxin/' or fatigue toxin, obtained in the washings
from the muscles of excessively fatigued animals.
Accurate chemical definition of the anaphylactic poison has not
so far been accomplished, and it is obvious that the problem is an
extremely difficult one. Biedl and Kraus,65' 66 however, have drawn
a very close parallelism between anaphylactic intoxication and pep-
tone poisoning in dogs. They have shown that peptone (0.3 gr. to the
kilo.) injected into these animals gives rise to the same clinical
symptoms that characterize anaphylaxis. It is accompanied also by
typical fall of blood pressure, delayed coagulability of the blood, and
leukopenia. Furthermore, they claim that the injection of sublethal
doses of Witte peptone into serum-sensitized dogs leads to a non-
specific anti-anaphylaxis. They claim a physiological identity of the
Witte peptone with the anaphylactic poison.
This last observation could not be confirmed by Manwaring,67
who found that dogs that had been rendered anti-anaphylactic to-
horse serum still reacted strongly to peptone — an observation which
does not indeed weaken the contention of Biedl and Kraus as to the
similarity of peptone shock to anaphylaxis, but has significance in
contradicting the doubts their experiments have thrown on the
specificity of antianaphylaxis.
Observations similar to those of Biedl and Kraus on the toxic
action of peptone have been made by Arthus.68
62 Vaughan and Wheeler. Loc. cit.
63 Wells. Jour. Inf. Dis., Vol. 5, 1908.
64 Weichhardt. Zentralbl. /. die ges. Pliys. u. Path, des Staff wechsels,
No. 15, 1909. Ref. "Weichhardt's Jahresbericht," 1910, p. 554.
65 Biedl and Kraus. Wien. klin. Woch., No. 11, 1909.
66 Also "Kraus n. Levaditi Handhuch, etc.," Erg-anzung-sband 1, p. 26 l.
67 Manwaring. Zeitschr. f. Imrmtnitntsforschung, Vol. 8, p. 589, 1911.
68 Arthus. C. E. tie I'Acad. des Scienc., Vol. 148.
406 INFECTION AND RESISTANCE
Biedl and Kraus have found a similar parallelism in guinea pigs
in which they determined the typical bronchial spasms after peptone
administration. This is in contrast to Werbitzky,69 who found even
large doses of peptone non-toxic for guinea pigs. Nevertheless, there
is no question that the similarity between peptone shock and anaphy-
laxis is very striking and of great theoretical importance. It does
not, however, bring us much nearer to a chemical understanding of
the nature of the poisons since the "Witte" peptone used in these ex-
periments is a mixture of many different substances. Brieger,70
for instance, found toxic and non-toxic lots of Witte peptone. The
toxic ones yielded on extraction a body which he calls peptotoxin.
This variation in the constitution of different samples of so-called
"pepton" may account for some of the conflicting results obtained in
guinea pigs.71
Recently Dale has suggested that B-imidazolethylamin or his-
tamin may be the active principle concerned in anaphylactic shock.
Intravenous injection of 0.5 mg. of this substance into large guinea
pigs results in typical respiratory difficulties, convulsions with death
and distention of the lungs typical of anaphylactic shock. Treat-
ment with atropin diminishes this action, just as Auer and Lewis
found this to be the case in true anaphylaxis, and fall in blood
pressure also occurs. It would seem then that substances represent-
ing cleavage products of native proteins of highly complex nature,
the result of proteolytic cleavage not very far advanced, are prob-
ably concerned in the production of anaphylactic shock. The ana-
phylatoxins of Friedberger cannot of course be studied chemically
by the methods to which Vaughan's poisons are amenable.
PHENOMENA CLOSELY RELATED TO ANAPHYLAXIS
There are a number of well-defined phenomena of acquired
hypersusceptibility or sensitiveness which, in nature, seem closely
analogous to true anaphylaxis as we understand it to-day, but re-
garding the mechanism of which the opinions of experimenters are
still to some extent at variance.
Among the most important of these is the toxic action of normal
«era when injected into animals of another species — a phenomenon
which is now generally accepted as belonging in principle to the true
anaphylactic phenomena, though this opinion is of comparatively
recent formulation. The subject is of sufficient theoretical and
practical importance to be considered in some detail.
The older studies of phenomena belonging to this category fol-
69 Werbitzky. C. E. de la Soc. de BioL, Vol. 66, 1909.
70 Brieger. "Die Ptomaine," 1, p. 14.
71 For analysis of Witte peptone, see Hammarsten, "Physiological Chem-
istry," English translation.
ANAPHYLAXIS 407
lowed closely in the footsteps of experiments on transfusion, and as
early as 1666 a commission of the London Koyal Philosophical So-
ciety reported deaths following transfusion, alleging intravascular
coagulation as the probable cause of deatk
The cause of death following injections of foreign whole blood,
blood cells, and serum has, since that time, occupied the attention of
many workers whose studies need not be reviewed for our present
purposes. Chief among them were Magnani, Brown-Sequard, Ma-
gendie, and, more recently, Naunyn, Landois, and Ponfick.72
The work of Landois is of special interest in that he worked
with blood serum free from cells, and attempted to correlate the
occurrences after the injection of animals with the action of the
serum upon the cellular blood elements in vitro. Landois observed
both the solution of hemoglobin and hemagglutination, and was
led to regard the action of serum upon erythrocytes as the primary
cause of death after transfusion. His conception of the mechanism
is apparently twofold. On the one hand, he believed that when
small quantities of blood were transfused, a formation of fibrin
(stroma-fibrin) was initiated in the stroma of the injured erythro-
cytes which led to coagulation and thrombosis in the capillaries of
the central nervous system and lungs. In the case of the transfusion
of rabbit's blood into dogs he attributed death to embolism in the
pulmonary vessels due to "Massenhafte Verklebung der Kaninchen-
zellen im Hundeblut" — or, in other words, to hemagglutination.
Ponfick and others have disputed the validity of Landois' con-
clusions, but the basic principles of his explanations have been up-
held within recent years by workers who have gone over the same
ground with the aid of more modern methods. Two careful re-
searches have appeared during the last two years in which the prob-
lem has been approached by different routes, but in which the gen-
eral conclusions show much agreement. Coca,73 investigating the
cause of death following the injection of washed blood cells into ani-
mals of different species, concludes that in these cases death is due to
mechanical obstruction of the pulmonary circulation owing to ag-
glutination of the injected cells. It is important to note, however,
that he adds in his conclusions the following paragraph: "The
mere presence of specific agglutinins does not suffice, in the injection
of 'toxic7 erythrocytes, to occlude the pulmonary circulation. The
cooperation of another factor must be assumed — a factor probably
found in the capillary walls."
Loeb, Strickler, and Tuttle,74 investigated the cause of death fol-
lowing the injection of normal dog and beef sera into rabbits. They
72 A brief historical review of this work can be found in the paper of
€oca (1), Virchow's Arch. f. path. Anat., 1909, Vol. 196, p. 92.
73 Coca. Virchow's Archiv, Vol. 196, 1909.
74 Loeb, Strickler, and Tuttle. Virchow's Archiv, Vol. 201, 1910.
408 INFECTION AND RESISTANCE
correlated their animal experiments carefully with the action of the
sera in vitro upon the blood elements of rabbits, and utilized the
property of hirudin to inhibit the coagulation of blood, finding, in
the case of dog serum, that injections of hirudin, while not always
preventing death, at any rate prolonged life or necessitated an in-
crease in the lethal dose. The conclusions of these authors are as
follows: "Death following the injection of foreign serum is brought
about by obstruction of the pulmonary circulation either by heaps of
agglutinaled erythrocytes or by fibrinous plugs. Dog serum and beef
serum represent two different types. In the case of dog serum hem-
olysis of the blood cells of the recipient liberates substances at-
tached to the stromata, which hasten coagulation. In consequence
fibrin is formed which is carried into the pulmonary vessels and
occludes them. In the case of beef serum death is due to hemag-
glutination."
The more recent understanding of the liberation of toxic bodies
from blood cells by immune hemolytic sera, especially by the experi-
ments of Friedemann cited above, have rendered it likely that a
similar anaphylatoxin formation from the cells of the recipient may
lie at the bottom of the toxic action of normal sera. And it is a
fact, indeed, that such toxic sera are always hemolytic for the cor-
puscles of the susceptible animal.
An analysis of the toxic action of certain normal sera from this
point of view has been made by Uhlenhuth and Haendel,75 who, in
studying the necrotizing action of beef serum injected into guinea
pigs, attribute this action of the serum to a "complex process de-
pending upon the cooperation of complement," but not identical
with the hemolytic mechanism. The toxic action of such serum, how-
ever, they separate from the necrotizing action, concluding that this
is independent of complement, and more thermostable than either
the mechanism causing necrosis or that responsible for hemolysis.
Recent studies of the writer76 on the toxic action of goat serum
for rabbits have shown that, contrary to Loeb, Strickler, and Tut tie,
hemagglutination and blood coagulation can be excluded as causes
of death, and that, in agreement with Uhlenhuth and Haendel, the
toxic action is due to a proteolyHc action on the part of the serum
not necessarily identical with the hemolysins, but producing from
the protein of the recipient a poison similar to the anaphylatoxins.
Unlike Uhlenhuth and Haendel,, however, it seemed clear that the
participation of alexin was definitely necessary — the process being
probably entirely analogous to Friedemann's results with immune
hemolytic (cytolytic) sera. The poisonous action of dissolved hemo-
globin could be excluded. In principle, therefore, the toxic action
75 Uhlenhuth and Haendel. Zeitschr. f. Immunitatsforscli., Vol. 7, 1910.
76Zin,sser. Jour. Exp. Med., Vol. 14, 1911.
ANAPHYLAXIS 409
of normal sera would seem to depend upon a mechanism, similar to
that of other anaphylactic phenomena.
Toxin hyper susceptibility, which is often acquired by animals
in the course of immunization with diphtheria and tetanus toxin, is
usually classified with anaphylaxis, indeed is often cited as the
earliest observation of this phenomenon. However, it is by no
means clear that the two conditions are actually analogous, since in
the case of the toxins we are dealing with antigens which are not only
toxic in themselves, but against which neutralizing antibodies are
formed in the reacting animal. This last fact alone would separate
toxin hypersusceptibility sharply from true protein-anaphylaxis in
that entirely different reacting-mechanisms seem to be called into
play by the two varieties of antigen. It will be necessary, therefore,
to discuss toxin hypersusceptibility at some length.
Probably the earliest authentically recorded observation is that
of von Behring,77 who determined, both for diphtheria and tetanus
toxins, that animals once inoculated with these poisons were oc-
casionally more sensitive to them subsequently than were normal
animals. He spoke of "Gift Ueber empfindlichkeit" as a property
acquired by reason of a preceding injection, and the observation was
further developed by Knorr78'in 1895, and by v. Behring himself,
in collaboration with Kitashima 79 — a few years later. These writers
showed that guinea pigs which are treated repeatedly with small
doses of diphtheria toxin may, under certain circumstances, not only
fail to show immunity, but may even develop a susceptibility in-
creased to such an extent that doses far too small to injure a normal
animal will cause their death. Again, in the case of diphtheria toxin
similar observations were made upon horses by both Salomonsen and
Madsen,80 and by Krete.81 The last-named worker observed that
horses that had been immunized with diphtheria toxin would often
react to neutral mixtures of toxin and antitoxin by which normal
horses were unaffected. This so-called "paradox phenomenon" was
much discussed, and many theories advanced to explain it, a most
ingenious adaptation of the side-chain theory being applied to it by
Kretz 82 and by Wassermann.83 They assumed that the partial im-
munization in such treated animals had in truth induced the forma-
tion of excessive receptors ; that, in the stages of hypersusceptibility.
however, these receptors had not yet been cast off from the cells. In
77 Von Behring. Deutsche med. Woch., 1893.
78 Knorr. Quoted from Otto, "Dissertation," Marburg, 1895.
79 Von Behring u. Kitashima. Berlin. Jclin. Woch., 1901.
80 Salomonsen et Madsen. Ann. de I'Imt. Past., 1897.
81 Kretz. Quoted from Otto in "Kolle u. Wassermann Handbuch," Er-
ganzungsband 2, p. 232.
82 Kretz. Zeitschr. f. Heilknnde, 1902.
83 Wassermann. "Kolle u. Wassermann Handbuch,'" Vol. 4, 479. .;...
410 INFECTION AND RESISTANCE
consequence there was an excess of "sessile receptors" — by means of
which the cell was rendered more exposed to toxin action than it was
normally — it being still unprotected by the presence of freely cir-
culating "antitoxin" receptors. The difficulties arising from the
observation of similar hypersusceptibility in animals whose blood
contained free antitoxin were disposed of by Wassermann by the
convenient assumption of variations of affinity.
He assumed that the treatment with toxin, i. e., the intoxication,
may induce a condition of higher affinity for the poison on the part
of the sessile cell receptors, leading to a selective toxin-absorption by
the cells and consequent greater susceptibility to injury. With
Behring, he speaks of this as a "histogenic hypersusceptibility,"
implying an increased vulnerability of the tissue cells.
The analogy between these early observations and the phenomena
which we now classify as anaphylaxis is unquestionably a striking
one. However, it is doubtful, as Friedemann suggests, whether the
two processes depend upon similar mechanisms. For, as we have
seen in the case of the sensitiveness to toxin, we are dealing with
primarily poisonous substances against which in the reacting animal
neutralizing antibodies are found — a combination of conditions quite
different from those with which we are confronted in hypersuscepti-
bility against primarily harmless proteins. It is, of course, possible
that the toxin hypersusceptibility is a true anaphylaxis against the
toxin-protein — independent of the specifically poisonous nature of
this substance. However, this is unlikely, since Lowi and Meyer 84
have shown that with tetanus toxin, the symptoms of such hypersus-
ceptibility are not those of anaphylaxis, but of increased but charac-
teristic tetanus poisoning. The fact that toxin hypersusceptibility
cannot be passively transferred with the serum of a susceptible ani-
mal does not seem to us a good argument against its anaphylactic na-
ture, since this, as we shall see, is equally impossible in the case of
tuberculin susceptibility, which is in all probability a modified exam-
ple of true anaphylaxis. Lowi and Meyer regard tetanus toxin hyper-
susceptibility as a "'summation" — meaning thereby that it depends
upon an alteration of the cells of the spinal cord because of traces of
the poison retained in them. When the toxin was given intraneurally
no antitoxin formation occurred, but the animals developed a marked
hypersusceptibility in the course of several weeks, showing that
here, unlike true anaphylaxis, specific antibodies play no part.
Not unlike toxin hypersusceptibility is that which is noticed in
the case of certain medicinal substances. Such are the so-called
idiosyncrasies against cocain, pilocarpin, morphin, quinin, and other
drugs. These conditions have no direct relation to anaphylaxis, and,
84 Lowi and Meyer. Festschrift, Schmiedeberg Suppl., Arch. f. exp.
Path. u. Therap., 1908, p. 355.
ANAPHYLAXIS 411
according to Hans Meyer,85 depend probably upon the chemical
peculiarities of the tissues of the individual, such as calcium con-
tents, etc. Hunt 86 has also shown that poison susceptibility, in cer-
tain cases, may be influenced by the diet.
85 Meyer u. Gottlieb. "Experimentelle Pharmakologie," 2d Ed., Urban
& Schwartzenberg, pp. 520 et seq., 1911.
86 Reid Hunt. U. 8. Pub. Health and M. H. S. Hyg. Lab. Bull, 69, 1910.
CHAPTER XVII
BACTERIAL ANAPHYLAXIS AND ITS BEARING ON
THE PROBLEMS OF INFECTIOUS DISEASE
IN the case of most serum reactions the original observations
were made upon the sera of bacteria-immune animals, and later ex-
panded into generalizations applicable to antigens as a class. This
was the case with the phenomena of lysis, agglutination, and precipi-
tation. In the case of anaphylaxis the reverse was true. The fun-
damental observations were made with non-bacterial antigens, but
the thought that analogous observations could be made with bac-
terial proteins was an obvious one, and since the problem was one of
altered susceptibility there was great promise that investigation of
this subject might prove of profound significance for our knowledge
of the pathology of infectious diseases.
Accordingly Rosenau and Anderson,1 in one of their earliest
researches, carried out experiments upon the sensitizing properties
of bacterial proteins. They were successful in sensitizing guinea
pigs with extracts of colon, tubercle, anthrax, and typhoid bacilli,
with Bacillus subtilis extracts, and with those of yeast. In most
cases they used considerable quantities of bacterial extracts and
obtained but slight or moderate symptoms. However, their results
were conclusive in showing that the anaphylactic experiment could
be carried out with bacterial proteins and was, in every detail,
analogous to the similar phenomena of serum anaphylaxis.
Not only could the basic experiment of active sensitization be
carried out with these substances, but it was found that the -reaction
here, as in other cases, was specific, and that shock was followed by
a period of "antianaphylaxis" or "immunity." Rosenau and An-
derson suggested that the incubation time of many infectious dis-
eases may be represented by the period necessary for the development
of susceptibility after a first injection, and that tlie crisis of pneu-
monia might possibly find an explanation in the analogy with anaphy-
laxis.
The criteria governing the successful production of bacterial
anaphylaxis were then studied especially by Kraus and Doerr,2 Holo-
1 Rosenau and Anderson. U. S. Pub. Health and M. II. S. Hyg. Lab.
Bull. 36, 1907.
2 Kraus and Doerr. Wien. klin. Woch., No. 28, 1908.
BACTERIAL ANAPHYLAXIS 413
but,3 Delanoe,4 and others, and the essential points of Rosenau and
Anderson's experiments were confirmed. Although Kraus and
Doerr succeeded in frequently sensitizing guinea pigs with a single
injection of bacteria, this was not found to be the most favorable
method for sensitization. Braun 5 obtained entirely negative results
by such a procedure, but this may well have been because in the first
place single sensitization with bacteria is evidently irregular in
result, and because Braun carried out his intravenous test-injection
slowly, a technique by which Friedberger found later that shock
could be avoided. Delanoe, in the main, confirmed the fact that
bacterial sensitization was possible, but denied the specificity of the
resulting anaphylaxis, in that he succeeded in producing shock in
tubercle-sensitized guinea pigs with comparatively large amounts
of typhoid, paratyphoid, and other bacilli, and conversely found
typhoid-sensitized guinea pigs hypersusceptible to tubercle-injec-
tions. Other workers, however, notably Kraus and Doerr, Holobut,
and Kraus and Admiradzibi,6 agree that the reaction is specific, at
least in the same limits within which other serum reactions may be
called specific.
Holobut then developed a technique of sensitization with bac-
teria more reliable than any which had been previously employed by
other workers. He found that the most regularly successful results
were obtained when he injected small quantities of bacteria (1/100
loopful) daily for ten days, subcutaneously, and tested with fairly
large amounts (1-2 c. c. of an emulsion of the bacteria) intrave-
nously about 3 weeks after the last sensitizing injection. This is in
keeping with later experience, and in our own work with typhoid
immunization in young goats we have found that anaphylactic
reactions were not observed unless the goats had previously
received several injections. A second injection never elicited
symptoms.
It is not at all unlikely that this difference between serum sensi-
tization and bacterial sensitization is due to the comparatively larger
amounts of protein injected with very small volumes of serum than
is the case with even the thickest bacterial emulsions. When larger
sensitizing quantities of bacteria are used — (which is often difficult
because of the primarily toxic nature of some of the bacteria) — a
single sensitization gives positive results in guinea pigs more fre-
quently than when the smaller amounts are used.
Since it was objected to many of the results at first obtained with
bacterial sensitization that they might have been due to the primarily
3 Holobut. Zeitschr. f. Immunitatsforsch., Vol. 3, 1909.
* Delanoe. C. E. de la Soc. de Biol, Vol. 66, 1909, pp. 207, 252, 348, 389.
5 Braun. Quoted by Bail and Weil, Zeitschr. f. Immunitatsforsch., Vol.
4, 1910.
6 Kraus u. Admiradzibi. Zeitschr. f. Immunitatsforsch., Vol. 4, 1910.
414 INFECTION AND RESISTANCE
toxic nature of the bacteria or their extracts, it is important to note
that Kraus and Doerr and later Kraus and Admiradzibi succeeded
in well-controlled experiments in transferring bacterial anaphylaxis
"passively" with the serum of previously sensitized animals — not
only of the same, but of other species — (rabbit serum to guinea pigs).
These experiments add the final link to the chain of complete analogy
between bacterial and serum anaphylaxis.
This analogy was partly established and, in its completeness,
clearly foreseen, when Friedemann's work upon the poisons produced
from cells by hemolytic sera, and Friedberger's similar work upon
serum precipitates, turned the trend of anaphylactic experimentation
into new channels.
It will be remembered that, before this time, the toxic action of
most bacteria (exclusive of "true toxin" producers like diphtheria
and tetanus bacilli) had, since Pfeiffer, been attributed to the libera-
tion of preformed "endotoxins" from the bacterial body during the
process of lysis.
This idea is fundamental to the opinion of hypersusceptibility
expressed by Wolff-Eisner7 as early as 1904.
The underlying concept of these ideas is really a morphological
one in which the "endotoxin" is regarded as something present in the
antigen which is set free by disintegration of the cell. In applying
this to serum anaphylaxis Wolff-Eisner 8 preserves this morphological
simile in that he speaks of the dissolved protein antigen (serum,
etc.) as "nur scheinbar gelost" and "dass es erst durch die Lysine
wirklich resorbierbar wird."
Indeed the sudden liberation of endotoxins by immune sera had
been regarded by Pfeiffer and others as the cause of the rapid death
often ensuing in immunized guinea pigs when more than a definite
maximum of cholera spirilla or other organisms was injected. In
all these opinions the basic conception was that certain bacteria con-
tained a characteristic preformed poison (endotoxin) upon the
pharmacological properties of which the peculiar symptoms caused
by each organism depended.
The earliest unambiguous statements of a conception differing
from this original view of the nature of bacterial endotoxins, and
approaching the later conceptions of Friedberger, are found, we
believe, in the work of Vaughan.9 In an article by him, published in
1908, Vaughan, after describing the incubation time occurring in
man and animals after inoculation with typhoid bacilli, says : "The
sickness begins when the animal body becomes sensitized and begins
to split up the bacilli." By "splitting up" he means here, as in his
7 Wolff-Eisner. Centralbl. f. Bakt., Vol. 37, 1904.
8 Wolff- Eisner. "Handbuch der Serum Therapie," p. 24, Lehmanns,
Miinchen, 1910.
9 Vaughan. A m. Jour, of Med. Sci., Sept., 1908.
BACTERIAL ANAPHYLAXIS 415
other work 10 on protein split products, not a mere liberation of pre-
formed poisons, but a chemical (enzymotic) proteolysis by which a
poisonous group of the bacterial protein-molecule is set free.
The essential difference of this point of view from the endotoxin
theory at first sight seems a trivial one — in the one case liberation of
a preformed poison molecule — in the other liberation of a poison by
the breaking up of a molecule. The difference, however, is a funda-
mental one. For, in the earlier theory, the specific element of the
toxemia was in the nature of the different poisons — whereas in the
view of Vaughan the lysin which breaks up the protein molecule is
alone the specific element, the formed poisons being concerned as
non-specific and alike, whether produced from colon bacilli, tubercle
bacilli, or egg white.
Friedberger,1 1 finally, in 1910, repeating with bacteria his ex-
periments upon "anaphylatoxin" liberation from specific precipitates,
succeeded in obtaining such poisons in the test tube by allowing
fresh guinea pig complement to act upon sensitized bacteria.
These results were confirmed by extensive experiments carried
out soon after this by Friedberger 12 himself with a number of
collaborators.
The results of these investigations may be summarized as follows :
1. The action of alexin upon sensitized or unsensitized bacteria
yields toxic substances which, injected into normal guinea pigs,
produce the characteristic symptoms of anaphylaxis, with frequent
death and typical autopsy findings.
2. These poisons ("anaphylatoxins") may be produced from
any variety of bacteria, pathogenic and non-pathogenic.13 (The or-
ganisms used in the earlier experiments were Vibrio metchnikovi,
the bacillus of tuberculosis, the typhoid, prodigiosus, and subtilis
bacillus, and Aspergillus fumigatus.)
3. The successful production of the poisons depends intimately
upon the relative amounts of antigen (bacteria) and alexin used, and
upon the time and temperature conditions under which the ex-
posures are made.
10 Vaughan. Zeitschr. f. Immunitatsforsch., Vol. 1, 1909.
11 Friedberger. Berl. klin. Woch., Nos. 32 and 42, 1910.
12 Friedberger; Friedberger and Goldschmid; Friedberger and Szymano.w-
ski; Friedberger and Schiitze: Friedberger and Nathan. Zeitschr. f. Im-
munitatsforsch., Vol. 9, 1911.
13 Neufeld and Dold, comparing virulent and avirulent strains of pneu-
mococcus in this regard, have found that the virulence of the race has no
relation to its yield of anaphylatoxin. Indeed the anaphylatoxins from va-
rious bacteria seem to be qualitatively entirely alike.
NOTE. — We wish to note here that we are giving Friedberger's theories and
experiments at some length because they have had considerable influence on
our conceptions of infection. We do not wish to create the impression that we
accept them, however, and will discuss their fallacies later in the same chapter.
416
INFECTION AND RESISTANCE
4. The poisons can be produced from boiled as well as from na-
tive bacteria.
Although unsuccessful with none of the bacteria with which ex-
periments were carried out, different species yielded the poison with
varying degrees of intensity, though qualitatively the poisons were
similar. Bacillus prodigiosus, though non-pathogenic, seems, in
general, to be one of the most favorable microorganisms for such
experiments.
Since a clear understanding of Friedberger's basic experiments is
essential to the further development of the theoretical conceptions
which have been based upon them, it will be useful to insert here a
protocol taken from his paper with Goldschmid.
Experiment VI. 30, VI, 1910. Ten 3-day agar cultures of
typhoid bacilli washed up in salt solution — 5 c. c. to 1/2 culture.
Varying amounts of inactivated typhoid immune serum are added,
the tubes brought to 11 c. c., 24 hours in refrigerator. 1, VII —
Centrifugalized and to sediment added 4 c. c. guinea pig complement
(active or inactivated), 24 hours in refrigerator. 2, VII — Centrif-
ugalized and supernatant fluids injected into guinea pigs of 200
grams intravenously.
Exp.
N6.
Amount of
culture
Specific
immune
serum
sensitiz.
Amount of
complement
No. of
animal
Symptoms
Result
1
^ slant agar
0
4 c. c.
G61
Severe
Dead 4 min.
2
% slant agar
0
4c. c.
G64
Slight
Dead 18 hre.
3
l/2 slant agar
0
4 (heated 56° C.)
G62
No symptoms
Lives
4
J^ slant agar
0
4 (heated)
G63
No symptoms
Lives
5
H slant agar
1.0
4
G66
Severe anaph.
Lives
6
H slant agar
1.0
4
G69
Very severe
Dead 4 min.
7
J^ slant agar
1.0
4 (heated)
G65
0
Lives
8
H slant agar
1.0
4 (heated)
G68
0
Lives
9
^ slant agar
0.1
4
G67
Very severe
Dead 8 min.
10
H slant agar
0.1
4
G70
Very severe
Dead 11 min.
11
J^ slant agar
0.1
4 (heated)
G73
0
Lives
12
H slant agar
0.01
4
G71
Very severe
Dead 2 min.
13
^ slant agar
0.01
4 (heated)
G75
0
Lives
14
Y<i slant agar
0.001
4
G72
Very severe
Dead 5 min.
15
3/2 slant agar
0.001
4 (heated)
G74
0
Lives
16
1.0
o
G76
0
Lives
17
0 1
o
G77
o
Lives
18
H culture
0
0
G79
0
Lives
From Friedberger and Goldschmid, Joe. «*., p. 402.
sion of control 19.)
(Changes made only in wording and onus-
This series alone shows that, under the given conditions, 4 c. c.
of alexin will produce the poison with 1/2 slant of typhoid bacilli,
without sensitization (tubes 1 and 2), with sensitization ranging
in degree from 1 c. c. to 0.001 c. c. of the given immune serum
(tubes 5, 6, 9, 10, 12, and 14), and that inactivation of the alexin
'serum in all cases prevented the poison formation. Normal guinea
BACTERIAL ANAPHYLAXIS
417
pig serum alone, active or inactivated, the bacteria, or the immune
serum alone were without toxicity in all of numerous controls.14
The experiments of Friedberger and his associates were rapidly
confirmed by Neufeld and Bold,15 Kraus,16 Ritz and Sachs,17 and
many others,18 and, though the conditions under which the anaphy-
latoxin formation took place were defined with slight variation by
different workers, the essential features of Friedberger' s claims
were upheld.
As was to be expected, it was soon found that instead of the pro-
longed exposures at refrigerator temperature the poisons could be
obtained more rapidly by digestion for shorter periods in water 19
baths at 37° C. And with this method accurate studies on the rela-
tions between time of exposure and proportions of reagents (antigen,
sensitizer, alexin) were made, relations the importance of which
was apparent from Friedberger's first studies. The outcome of
this work was as follows: 1. There are a definite minimum and a
definite maximum quantity of bacteria with which anaphylatoxin
can be produced by a given fixed quantity of guinea pig serum.
Thus, in one of the experiments of Friedberger and Goldschmid, 4
loopsful of typhoid bacilli with 4 c. c. of complement produced a
fatal poison, 24 loopsful with the same amount produced none. (In
some of the writer's 20 experiments with typhoid bacilli a similar
principle of proportions was evident, though much larger quantities
TITBATION EXPERIMENT WITH TYPHOID-IMMUNE SERUM
Rabbit 79
Dilution of serum
Agglutination
Bactericidal titre with modified
Stern-Korte method
1:100
+ + +
480 colonies
1:200
+ + +
556 colonies
1:500
+ + +
750 colonies
1:1,000
+ +
Over 10,000 colonies
1:2,000
=fc
+ + + + +
1:5,000
—
+ + + + +
1:10,000
—
+ + + + +
14 Injury of the animals by mere volume of injection can be definitely
excluded. The writer has frequently injected 5 to 6 c. c. of salt solution
into guinea pigs of 200 to 300 grams without symptoms in any way resem-
bling anaphylaxis.
15Neufeld and Bold. Berl. Wn. Woch., No. 2, 24, 1911; Arb. a. d. kais.
Gesundheits ami., Vol. 38, 1911.
16 Kraus. Zeitschr. f. Immunitatsforsch., Vol. 8, 1911.
17 Ritz u. Sachs. Berl. klin. Woch., No. 22, 1911.
18 Izar. Zeitschr. f. Immunitatsforsch., Vol. 11, 1911.
19 Friedberger u. Mita. Zeitschr. f. Immunitatsforsch., Vol. 10, 1911.
See also Bold, "Das Bakterien Anaphylatoxin," Fischer, Jena, 1912.
20 Zinsser. Jour. Exp. Med., Vol. 17, 1913.
418
INFECTION AND RESISTANCE
of typhoid bacilli could be successfully used if the time of exposure at
37° C. was prolonged.) 2. If sensitized bacteria are used an excess
of sensitization, beyond a definite limit, weakens the formation of
anaphylatoxin. It may be permitted to illustrate this with a protocol
of one of the writer's experiments with typhoid bacilli, since, though
merely confirming the principle laid down by Friedberger, it in-
cluded a careful titration of the bactericidal contents of the anti-
typhoid serum. (See table on p. 417.)
Two-tenths c. c. of this serum added to 1 c. c. of typhoid filtrate
gave a very slight clouding in about 15 minutes.
ANAPHYLATOXIN EXPERIMENTS
Amount
Number
in
Typhoid
bacilli
of
inactive
serum
Amount
of
complement
Weight
of
animal
Result
antityphoid
1
2
| slant
| slant
5.0c. c.
3.5 c. c.
4 C. C.
4 c. c.
215 gm.
200 gm.
Very sick, recovers
Typical death 2 min.
3
£ slant
3.0c. c.
4 c. c.
198 gm.
Typical death 2 min.
4
£ slant
2.0 c. c.
4 c. c.
225 gm.
Typical death 2 min.
5
| slant
l.Oc. c.
4 c. c.
200 gm.
Sick, recovers.
The weak character of the antiserum used, and the fact that,
in this experiment, the digestion was at 0° to 5° C., explain the
failure to obtain a strong anaphylatoxin with 1 c. c. of sensitizing
serum.
The negative experiment resulting from a too vigorous sensi-
tization is practically a corollary of the next point ascertained by
Friedberger, namely, that :
3. With constant amounts of reagents a too prolonged exposure
at 37° C. will result in failure to obtain the poison.
How are we to explain these experimental results ? The first of
the three — namely, the fact that an excess of bacteria inhibits the
formation of anaphylatoxins — seems to the writer most easily ex-
plained by accepting the views of Bordet on the manner of the
union of an antigen with its antibody. For, unlike the opinion of
Ehrlich, who assumes a union of the two according to the laws of
multiple proportions, Bordet 21 believes that the distribution of
serum substances upon an antigen is such that the entire amount of
antibody is distributed equally among the antigenic elements. In
the case of an excess of bacteria, as in these experiments,' therefore,
the quantity falling to each unit is insufficient, at least in the time of
exposure here practised, to accomplish the cleavage necessary for
poison production.
21 Bordet. Ann. de I'Inst. Past., 17, p. 161, 1903.
BACTERIAL ANAPHYLAXIS 419
As regards the second and third point — the failure of producing
anaphylatoidns if, on the one hand, too intense sensitization was
employed — or, on the other, the time of exposure was too prolonged —
these seem to indicate that anaphylatoxin is not the end product of
the complement action, but rather an unstable intermediate sub-
stance which, once formed, is. rapidly further decomposed ("abge-
baut") into non-toxic derivatives.
Indeed, Neufeld and Dold,22 in experiments with the cholera
spirillum, found that whenever lysis was permitted to proceed as far
as the actual disintegration and granulation of the bacteria no pois-
onous substances were obtained. They conclude from this that rapid
lysis actually prevents the production of the poison, and that the
anaphylactic antibody has no relation to the bacteriolytic sensitizer.
They fortify this opinion by experiments in which they easily ob-
tained powerful poisons with pneumococci, organisms which are but
slightly, if at all, subject to actual lysis. They suggest identity of
the anaphylactic antibody with the opsonins, or possibly with the
"Bordetsche Antikorper" of Neufeld. This latter conclusion does
not seem valid, since the mere fact that one microorganism under-
goes lysis and another does not is not necessarily an argument for a
difference in the sensitizers produced in animals by immunization
with these bacteria. It may, and probably does, depend upon varia-
tions in the ease of disintegration of the different cell-bodies, and,
as a matter of fact, not many bacteria undergo actual complete lysis
as easily as does the cholera spirillum. Moreover, there is much
evidence in favor of the so-called "unitarian" point of view, which
holds that no fundamental structural and functional differences
between the various heat-stable antibodies — sensitizers (ambocep-
tors), precipitins, immune opsonins (bacteriotropins), and the so-
called "Bordet" alexin-fixing antibodies — have as yet been proved.
However this may be, it seems conclusively established that a too
vigorous and prolonged action does not yield poisons — and that, since
less vigorous sensitization or early interruption of the exposure will
lead to positive results, the mechanism is one of rapid poison forma-
tion with equally rapid further decomposition into a non-toxic sub-
stance. In some cases chis is jiore rapid than in others. In Neufeld
and Dold's experiments with cholera spirilla the exposure of 2 loops-
ful of the organisms sensitized with 0.02 antiserum and treated with
2 c. c. of alexin resulted in complete lysis and failure of demonstrable
anaphylatoxin in 2 hours at 37° C. In some of the writer's experi-
ments with typhoid bacilli the most regular positive results were ob-
tained when the exposures at 37° C. were prolonged to several hours
and powerful poisons were determined even after as long as 15 hours
at 37° C.
That the production of the poison can under no circumstances be
22Neufeld and Dold. Loc. cit.
420 INFECTION AND RESISTANCE
regarded merely as a giving up from the bacterial cell of preformed
endotoxins under the influence of lytic substances which produce
greater permeability of the cell membrane was shown by Neufeld
and Dold, who extracted bacteria with lecithin salt solution and pure
salt solution, and from these extracts (but moderately toxic in them-
selves) produced typical anaphylatoxins by the action of complement.
The matrix of the poison thus is shown by direct experiment to be a
soluble ingredient of the bacterial cell.
It was further shown by Friedberger and Nathan that the con-
ditions prevailing in the test tube experiment in truth represent the
processes taking place within the animal body. This they accom-
plished by injected bacterial emulsions into the peritoneal cavities of
guinea pigs, killing the animals after several hours and examining
the peritoneal exudates for their toxic properties. Centrifugalized,
cleared of bacteria, and injected intravenously into other guinea
pigs, these exudates produced the typical acute symptoms character-
istic of the poisons obtained in test-tube experiments.
It was on these premises, then, that Friedberger 23 was led to
formulate his views of the nature of bacterial infections, which give
promise of introducing a new understanding of these diseases. It
has been shown in the researches upon serum anaphylaxis that the
injection of small quantities of a foreign protein may produce reac-
tions of temperature which simulate very closely those prevailing
in infectious diseases, and variations in the quantities injected, the
path of administration, and the interval between injections may lead
to conditions, local and systemic, which may affect, more or less
profoundly, many different organs and tissues of the body. These
matters we have considered in the general discussion of anaphylactic
phenomena. Friedberger now suggests that we may regard bacterial
infection, after all, as the presence in the body of a living foreign
protein — in this case varying in distribution and quantity by reason
of the particular invasive properties of the given germ and the
balance between these and the resistance of the host. It is not neces-
sary, therefore, to assume that the character of the disease is deter-
mined by the existence of different preformed "endotoxins." He
believes that we may justly assume that the toxic substances appear
only after proteid cleavage of the bacterial bodies has been initiated
Iby the action upon them of the serum components, and that the ap-
parent specificity of the poisons, or differences between the toxemic
.manifestations of various diseases, may depend, not on differences
in the pharmacological actions of these poisons, but rather upon
variations in the invasive properties of the bacteria, both as concerns
their quantitative distribution and their accumulation and localiza-
tion in the infected body.
23 Friedberger. Loc. cit.; also Deutsche med. Woch., No. 11, 1911 ; Berl.
klin. Woch., No. 42, 1911.
BACTERIAL ANAPHYLAXIS
If we leave out of consideration bacteria which, like the diph-
theria bacillus, produce true secretory poisons, it would be the ability
to gain a foothold in the body, the degree of invasive power, the pre-
dilection in the choice of a path of entrance, and the specific local
accumulation upon which the speed and quantity of anaphylatoxin
production and absorption would depend, and which consequently
would give character to variations in the clinical pictures of different
diseases. Besides simplifying considerably our comprehension of
bacterial toxemia this point of view again brings out the great im-
portance of the work of Vaughan, and of Vaughan and Wheeler,
on the non-specific poisonous fraction obtained by hydrolysis of bac-
terial and other proteins.
To support this assumption Friedherger points out the similarity
in the clinical manifestations of several diseases in which the inciting
bacteria are biologically very different, but in which the distribution
and invasive properties are alike. For instance, lobar pneumonia
caused by the pneumococcus is clinically very similar to that caused
by the Friedlander bacillus, though the microorganisms inciting
them are extremely unlike each other. He draws a similar parallel
between true cholera and cholera nostras, and we may add another
striking example in the great similarity existing clinically between
the various forms of acute and subacute septicemia in which a defi-
nite bacteriological diagnosis can rarely be made except by blood
culture.
Conversely, the same microorganism may call forth diseases
which clinically, apart from the purely local manifestations, are very
dissimilar, according to the localization and distribution of the bacteria.
Granted that we accept this view, then the subsidence of the
disease might depend merely upon limitation of the supply of an-
tigen, as the increasing bactericidal action of the blood constituents
comes into play, and upon the consequent diminution of the anaphy-
latoxin. For, as the bacteria diminish and the sensitizer increases, a
changed proportion between them is established which, finally, as
experiment has shown, results in a failure of anaphylatoxin produc-
tion. For, although experiments have shown that, within a wide
latitude of relative proportions of bacteria and antibody, anaphyla-
toxin can be formed, beyond this range an excess of one or the other
element eventually will prevent their formation.
Infectious disease, then, according to this point of view, repre-
sents merely the reaction of the body against a foreign protein, the
bacteria. These gain a foothold in the body, and at first, during the
so-called incubation time, cause no symptoms, since the slight amount
of bacterial destruction with correspondingly slight cleavage of the
bacterial protoplasm liberates too small an amount of anaphylatoxin
to incite noticeable deviations from the normal condition. As these
slight quantities of bacterial cleavage products are absorbed, however,
INFECTION AND RESISTANCE
a reactionary formation of specific antibody occurs. Meanwhile,
also, the foreign protein increases and is distributed by bacterial
growth. In consequence of these parallel processes changes of pro-
portion between the reacting substances are created and a constantly
greater amount of anaphylatoxin is liberated and the disease pro-
gresses. This may kill the patient if the proportions become such
that the amount of poison formed exceeds the lethal dose. At any
rate, the symptoms may vary and fluctuate according to the relations
maintained between the reacting bodies, modified somewhat by the
supply of alexin or complement. If recovery is to take place the
amount of antibody (sensitizer, amboceptor) may become so great
that the bacteria are subjected to rapid destruction, the chemical
cleavage of their bodies taking place so vigorously that practically no
anaphylatoxin is distributed and vigorous phagocytosis is initiated.
Finally the antigen is completely removed. On the other hand, an
excessive increase of the bacteria or a defective supply of alexin
might also lead to a final cessation of the formation of anaphylatoxin ;
in this case, however, we would expect death by the metabolic dis-
turbance occasioned by the life processes of the great masses of bac-
teria. It is not unthinkable, moreover, that the bacterial enzymes in
such a case might produce substances comparable to the anaphyla-
toxins from the destroyed tissues of the host.
It is perfectly true, as Friedberger says, that on the basis of this
theory, rendered so likely by experimental fact, the assumption of
the existence of endotoxins to explain the various manifestations of
infectious disease is not necessary. The poisons, according to the
view just outlined, are alike and non-specific. It is the reaction
bodies, the sensitizers, induced by the bacterial protein which in each
case are these specific elements.
While it is not necessary to assume specific endotoxins, however,
it is not possible on present evidence to entirely exclude the partici-
pation of such substances in the genesis of infectious disease. The
rapid toxic action of bacterial extracts obtained in various ways
has been taken to argue in favor of this.
It is a difficult question to settle, and must undoubtedly remain
an open one until a method is found by which crucial experiments
can be formulated. Since Neufeld and Dold have succeeded in pro-
ducing anaphylatoxin from bacterial extracts, the primary toxic
action of every bacterial extract, however rapidly produced from the
bacteria, can be regarded as possibly furnishing merely an antigen
for anaphylatoxin production, and indeed such a supposition is ren-
dered more likely by the almost invariable incubation time following
upon the administration of endotoxic extracts, even when they are
introduced directly into the circulation. Pfeiifer 24 himself still
24 R. Pfeiffer. "tiber Bakterien Endotoxine, etc.," Weiehhardt's Jdhres-
bericht, Vol. 6, p. 29, 1910.
BACTERIAL ANAPHYLAXIS
believes in specific endotoxins, basing his opinion on the individually
characteristic nature of the infections caused by supposedly endo-
toxic bacteria. The differences in the degrees of toxicity, moreover,
of extracts obtained by the same technique from different micro-
organisms would certainly tend to add some weight to his argument.
We need only to recall to memory the greater toxicity of bouillon
culture extracts of B. dysenterice Shiga-Kruse as compared with
similar extracts of B. dysenterice Flexner or Hiss-Russell, or the
similar difference between typhoid and colon extracts. Altogether
the problem is an involved one, for the recent claims of Kraus,25
Doerr,26 27 and others of having discovered true (antitoxin-forming)
soluble toxins 28 in such cultures as those of cholera, dysentery
Shiga, and typhoid bacilli add another complication.
As a matter of fact, the entire problem of endotoxins is one which
calls for reexamination in that the knowledge gained of recent years
has opened a number of alternative explanations for the primary
toxicity of such bacteria as the typhoid bacillus. Briefly summarized
they are: (1) The actual intracellular existence of specific endo-
toxins in the sense of Pfeiffer 29 (toxalbumin). (2) The production
of toxic split products in the animal body from the bacterial protein
by proteolytic cleavage brought about by non-specific serum protease
(Jobling and Petersen) 29 or by the cooperation of antibody and
alexin (Friedberger). (3) The absorption of antienzymes by the
bacteria with consequent activation of the serum protease which then
splits off toxic substances from the plasma protein (Jobling and
Petersen). (4) The presence of non-specific toxic substances in the
bacterial cell body, of the nature of peptones, primary and second-
ary albumoses, etc., which are liberated by lysis from the bacterial
cell after cell death. This conception would differ from that of
Pfeiffer in that the intercellular substances are conceived as in no
sense specific toxic proteins, but rather entirely non-specific constit-
uents representing the type of poisons conceived as proteolytically
produced from the antigen by Vaughan and others. This last view,
though hitherto not particularly considered, should nevertheless in
our opinion be regarded as at least a possible explanation for a part
of the toxic manifestations resulting from the injection of bacteria
of this class. Moreover, such a possibility is suggested by the fact
that bacterial protein is relatively poor in antigenic properties.
Doerr, also, has considered this in stating that he believed the diffi-
culty of producing anaphylaxis with bacteria was in part due to
the fact that their body substances were relatively poor in coagulable
26 Kraus. MonatscJir. f. Gesundheitspflege, No. 11, 1904.
26 Kraus u. Doerr. Wien. klin. Woch., No. 42, 1905.
27 Kraus. "Kraus u. Levaditi Handbuch," Vol. 1, p. 180.
28 Exotoxins.
29 Zinsser and Parker. Journal of Exp. Med., September, 1917.
424 INFECTION AND RESISTANCE
(antigenic) proteins. We have not been able as yet to study the
matter extensively, but we have carried out a few experiments.
Typhoid bacilli from twenty-four agar cultures were weighed as
a moist mass, ground up with salt, and then taken up in distilled
water to isotonicity. After the addition of 0.2 c.c. of N sodium
hydroxid to 100 c.c. the suspension was heated to 60° C. for 30
minutes to prevent autolysis, and was then shaken for 4 to 5 hours
with a motor, at room temperature. After filtration through a
Berkefeld candle the clear solution gave a definite cloud after boiling
and adding acetic acid. The filtrate was treated with heat and acid
to remove coagulable protein. On the advice of Professor Gies the
suspension was first brought to a boil and then small amounts of
acid were added to prevent possible hydrolysis which might have oc-
curred had the acid been added first. The filtrate from this was
then half saturated with ammonium sulphate. Again a definite cloud
was obtained, and when this was filtered clear, a second turbidity
could be produced by complete saturation with the sulphate.
Although we have not yet obtained toxic reactions with these
substances after isolation, perhaps because of the difficulty of obtain-
ing them in sufficient amount, the presence of albumoses, substances
which have often been found to possess primary toxicity for animals,
suggests the possibility that their existence in the bacterial body
might indeed contribute to the injury done by injections of bacteria.
We found albumoses in extracts of typhoid bacilli not only when
the bacteria were grown on the ordinary peptone media but also on
agar made without peptone to which nutrose or sodium casemate has
been added as an enriching substance.
Moreover, we have also had the experience with the typhoid
bacillus which is referred to repeatedly by Vaughan when dealing
with the colon bacillus; namely, that bacterial suspensions sub-
jected to boiling are quite as toxic and often more so than are the
unheated and living suspensions.
Thus, if three guinea pigs are injected with equal amounts of
typhoid suspensions, one with the living bacteria, the second with
bacteria heated to 60° C. for 15 minutes, and the third with bacteria
boiled for 5 minutes, the guinea pig receiving the boiled bacilli will
often be the first one to grow sick and die several hours before the
others.
This may mean, of course, as Vaughan suggests, that the heated
protein is more promptly split by the ferments of the body. . It
also suggests, however, that in addition to this the heating has left
unchanged non-coagulable toxic constituents of the cell.
It still remains for us to consider certain experimental facts
which have had some influence upon extending and altering the con-
ceptions of anaphylatoxin formation which we have just outlined.
BACTERIAL ANAPHYLAXIS 425
In the earlier work of Friedberger, Neufeld and Dold, and others the
poisons were formed from the bacteria by the action of alexin at
low temperatures. This suggested the possibility that the alexin
fractions — "Endstiick" and "Mittelstiick" — might not both be in-
volved in the reaction, since, from previous studies, it was known that
at low temperatures the midpiece (the globulin fraction) was bound,
but that the end piece did not become active until the temperature
was increased. This point was, therefore, made the object of a
special investigation by Friedberger and Ito,30 who found that
neither fraction alone would suffice, but that bacterial anaphylatoxins
were formed only under the influence of the intact whole alexin, or
by that of the two fractions, reunited after separation.
Because of the reasoning along which the investigations of ana-
phylatoxin formation were developed, it is not surprising that it
seemed self-evident that the matrix of the poison was represented by
the bacterial protein — the antigen of the lytic complex. The only
fact which, in the earlier experiments, might have cast some doubt
upon this was the ease with which anaphylatoxins were produced
from boiled bacteria and precipitates and from such very insoluble
organisms as the tubercle bacillus.
Such vague suspicion becomes a very definite doubt, however,
in the light of the experiments of Keysser and Wassermann.31
Keysser and Wassermann utilized the fact that certain serum ele-
ments may be absorbed out of serum if this is shaken up with such
indifferent suspensions as barium sulphate or kaolin (aluminium
orthosilicate).32 They therefore substituted these insoluble sub-
stances for antigen, allowed them to absorb serum constituents, as-
sumed by them to be amboceptor, out of normal and inactivated im-
mune sera, and then allowed complement or alexin to act upon the
"sensitized" kaolin.
In this way they obtained active and powerful anaphylatoxin,
and claim, in consequence, that the matrix of the poison is not in
the bacterial antigen, but in the sensitizer or amboceptor, which is
mechanically absorbed by the bacteria (as by the kaolin), and thus
made amenable to the alexin action.
The experiments of Keysser and Wassermann have found con-
firmation in the hands of other investigators, although the results of
Neufeld and Dold, as well as our own, with this method were far
more irregular than were those of Keysser and Wassermann. !N~eu-
30 Friedberger and Ito. Zeitschr. f. Immunitatsforsch., Vol. 11, 1911.
31 Keysser and Wassermann. Folia Serologica, Vol. 7, 1911; Zeitschr.
f. Hyg., Vol. 68, 1911.
32 Kaolin emulsions will absorb amboceptor only out of diluted serum.
Out of concentrated serum complement is completely absorbed. Friedberger
u. Salecker, Zeitschr. f. Immunitatsforsch., Vol. 11, 1911; Zinsser, from
Journ. Exp. Mecl., Vol. 18, 1913.
426 INFECTION AND RESISTANCE
feld and Dold 83 and Friedberger,34 suggest that the horse serum
absorbed by the kaolin may act as an antigen itself, and is acted upon
by normal sensitizer present in the guinea pig serum. This is in
keeping with the well-known fact that small amounts of sensitizers to
many varieties of foreign proteins are present in normal serum, and
is further borne out by the fact that Neufeld and Dold, unlike Keys-
ser and Wassermann, were never able to produce anaphylatoxin by
allowing the alexin alone to act upon kaolin — without previous
absorption of horse serum.
We say "never," though the protocols of Neufeld and Dold35
show a single successful experiment. This they explain, however,
by assuming the accidental presence of some antigen in the alexic
serum. That is, the entire complex, antigen, sensitize^ and alexin,
is assumed to have been present in this particular guinea pig serum.
The same explanation may be applied to the occasional inherent
toxicity which develops in normal guinea pig sera on standing.
Whether the above complicated explanation is necessary or whether
we may assume an autolytic process in the guinea pig serum by which
anaphylatoxin-like substances are formed is an open question.
At any rate, it has been shown that, even with bacteria, the
action of alexin is not the only way in which acute poisons may be
obtained from them. And, indeed, if we look upon the action of
alexin as analogous to that of an enzyme — an assumption for which
we have much supporting evidence, we may well expect that other
methods of proteolysis will give similar toxic cleavage products.
And various methods of bacterial autolysis have actually yielded
such results. Thus Neufeld and Dold obtained poisons by digesting
typhoid bacilli, cholera spirilla, and other microorganisms for sev-
eral hours in salt solution, lecithin salt solution, and inactivated
guinea pig sera. Their extracts killed guinea pigs within several
hours. Rosenow 36 has even succeeded in obtaining acutely toxic sub-
stances which caused typical anaphylactic death in guinea pigs by
suspending pneumococci, typhoid bacilli, and other bacteria in salt
solution at 37° C. for varying periods, and the writer,37 though never
producing acute death, was able to cause typical anaphylactic shock
in isolated cases with similar salt solution extracts of typhoid bacilli.
It is not impossible that poisons obtained in this way are formed
by autolysis due to proteolytic enzymes of the bacterial cell.
In cases in which bacteria, suspended in salt solution and other
indifferent fluids, represent the only source of protein present it
must, of course, be assumed that they are the substratum or matrix of
83 Neufeld and Dold. Loc. cit.
84 Friedberger and Salecker. Zeitschr. f. Imm., Vol. 11, 1911.
85 Dold. Loc. cit.
88 Rosenow. Jour. Inf. Dis., Vol. 9, 1911; Vol. 10, 1912.
87 Zinsser. Loc. cit.
BACTERIAL ANAPHYLAXIS 427
the anaphylactic poison. They are also, of course, to be regarded as
the source of the poison in such experiments as those of Vaughan, in
which the poison was produced by chemical hydrolysis of the bac-
terial bodies. In the case of anaphylatoxin production by fresh
serum in the presence of bacteria, kaolin, precipitates, etc., the ques-
tion is much more complex.
As we have stated before, it is only natural, considering our pre-
vious knowledge of bacteriolysis in serum, that the first conclusion
arrived at should look for the source of the poisons in the bacterial
cells. The doubt which has been cast upon this assumption by the
work of Keysser and Wassermann and others, however, rests upon a
sufficiently sound experimental basis to prevent our absolute accept-
ance of this view. Jobling and Petersen 38 have recently carried out
experiments which may serve to throw much light upon anaphyla-
toxin. They believe that, by the ordinary technique of anaphylatoxin
production with bacteria and serum, most of the toxic substances orig-
inate from the serum proteins. The bacteria act merely by remov-
ing the antiferments from the serum, thereby setting free the fer-
ments normally present in the serum, and permitting them to act
upon the serum proteins. The result is cleavage and the production
of toxic split products. This would explain such results as those of
Keysser and Wassermann. Jobling and Petersen have supported
their contention by experiments in which they have obtained typical
anaphylatoxins by removing serum antiferments with chloroform,
kaolin, and agar. They have further shown that emulsions of bac-
teria actually do remove antiferments from fresh serum, and that
the bacteria used in the process become more resistant to tryptic
digestion in consequence.
This does not necessarily weaken the force of Friedberger's view
of infectious disease. For, whatever the source of the toxic sub-
stances, the result is still the same. Wherever proteolysis takes
place, and certain quantitative relations between cleavage, energy,
and substratum exist, it seems toxic bodies may be liberated.
And the result of such proteolysis, at some stage of the process,
yields apparently the same non-specific toxic substance, whatever the
particular nature of the proteolysis and whatever the variety of the
original protein matrix.
38 Jobling and Petersen. Jour, of Exp. Med., June, 1914.
CHAPTER XVIII
THE CLINICAL SIGNIFICANCE OF ANAPHYLAXIS
SERUM SICKNESS
WE have mentioned that Rosenau and Anderson attacked the
problem of hypersusceptibility primarily in the hope of casting light
upon the nature and cause of the distressing symptoms which in
human beings often ensue upon the injection of diphtheria antitoxin.
It has been one of the staple objections of lay opponents to the use
of antitoxins that the injections are apt to cause severe illness and
occasionally death, and indeed a few cases are on record in which
sudden death has followed the first injection of diphtheria antitoxin.
Since it was known by accumulated clinical experience as well as
by experiments like those of Bertin,1 of Johannesen,2 and others
that the harmful effects were not dependent upon the antitoxin con-
tents, but could be produced by injections of normal horse serum, it
was but natural to bring these ill effects into analogy with the phe-
nomena of hypersusceptibility. A large number of references to
such antitoxin illness or SERUM SICKNESS have appeared in the lit-
erature since the first beginnings of the therapeutic use of sera, yet
no careful analysis of the condition was made until von Pirquet and
Schick,3 in 1905, published their studies.
As a rule the results of serum injection have been mild and with-
out danger, though sufficiently frequent and troublesome to call for
thorough study and attempts to discover the prophylactic measures.
As stated above, a few cases are on record in which sudden death
has followed a single first injection. There are no reports in the
literature known to us, however, of fatalities after second injections,
although not infrequently such cases have taken on alarmingly seri-
ous aspects.
The percentage of incidence and the variety of symptoms have
been the subjects of many reports. The most frequent and striking
single occurrence has been an urticarial rash. Rolleston,4 in a large
1 Bertin. Gaz. Med. de Nantes, 1895. Quoted from Levaditi.
2 Johannesen. Deutsche med.Woch., No. 51, 1895.
3 Von Pirquet u. Schick. "Die Serum Krankheit," Deuticke, Leipzig,
1905. Also Munch, med. Woch., 53, p. 67, 1906.
4 Rolleston. The Practitioner, Vol. 74, 1905.
428
CLINICAL SIGNIFICANCE OF ANAPHYLAXIS 429
series of cases, found urticaria in all but 17 of 289 cases of serum
eruptions occurring between the first and tenth days after injection,
and in all but ten of ninety-four later eruptions.
Rashes occurred in from 69.4 to 81.9 per cent, of the 600 anti-
toxin cases which Rolleston reports.
Joint pains commonly accompany the appearance of the rash, and
frequently there is adenitis, involving the glands adjacent to the point
of injection, and even remotely in the submaxillary, axillary, or in-
guinal glands. Albuminuria is quite common, and with it oliguria
and relative concentration. Fever is rarely absent, though usually
slight, together with general malaise. Rolleston in his purely clini-
cal study does not classify his cases into those reacting after a first
injection and those showing symptoms after repeated treatment. He
states, however, that the serum-reaction may be extremely severe in
cases of "relapse or second attack of diphtheria" in which urticaria
with "pronounced edema surrounding the wheals, vomiting, rigors,
and collapse may ensue within a few hours of injection," and further
asserts that these severe symptoms are more apt to follow upon large
than upon small doses.
Von Pirquet and Schick have studied the condition with careful
reference to a comparison between the symptoms occurring in sub-
jects after a first injection of serum and those following upon re-
peated treatments. Their studies revealed the very important fact
that the ill effects following a second injection were not only more
severe than those occurring after the first injection, but developed
after much shorter periods of incubation. In the ordinary "first
injection" case the symptoms appear usually in from one to twelve
days. After a second injection this incubation period may be con-
siderably shortened and symptoms may appear in from five to seven
days, the local and general reactions being much more marked than
those subsequent to a first injection. Indeed, in some of the cases re-
ported they may attain very alarming degrees of severity. This is
the so-called accelerated ("beschleunigte") reaction of von Pirquet
and Schick, and is different from the "first injection" symptoms only
in its greater severity and speedier onset. In addition to this, how-
ever, the "second injection" cases may show a train of immediate
symptoms5 (sofortige Reaktion), which occur within twenty-four
hours after injection, and are characterized by marked local erythema
and edema with often urticaria and constitutional disturbance. Both
reactions may occur in the same individual, the "accelerated" reac-
tion setting in as the "immediate" reaction subsides.
Again, one reaction or the other may occur alone. The analogy
between the immediate reaction and the anaphylaxis of animal ex-
periment is obvious. The cases may be classified on the basis of
5 Rankin in the Lancet, Dec., 1911, reports a case of "immediate" reac-
tion 15 minutes after injection.
430 INFECTION AND RESISTANCE
these reactions, according to von Pirquet and Schick, the nature of
the reaction being, within certain limits, determined by the interval
ensuing between the first and the second injection. Thus, when the
interval was twenty-one days or less the immediate reaction alone
was noticed. When the interval was between two and six months
both the immediate and accelerated reactions were present, and when
the interval was still longer (seven months or more) the accelerated
reaction alone was present. Isolated exceptions to this are noted in
the series of sixty-one cases so reported.
Currie,6 7 who has made similar studies, confirms the results of
von Pirquet and Schick in all essentials, and agrees with their state-
ment that the nature of the reaction is chiefly dependent upon the
interval between injections.
That the entire train of symptoms, as well as the mere fact of
their dependence upon an injection of a foreign protein, rather than
upon the antitoxin itself, force upon us the analogy with anaphy-
laxis is clear. Moreover, this analogy becomes almost an identity
when we can show, as von Pirquet and Schick have done, that the
first injection has apparently sensitized the subject, in that the
second administrations are fraught with more violent and serious
reactions, dependent to a great extent, as in experimental anaphy-
laxis, upon the time intervening. If serum-sickness is truly an
anaphylactic phenomenon, however, it is still by no means clear why
symptoms should at all ensue after the first injection. Many ex-
planations have been offered for this; none of them, however, from
the very nature of the problem itself, can be finally accepted as
proved. Two possible explanations appear from the experimental
work of Rosenau and Anderson quoted above. These workers, we
have seen, showed among other things that the state of hypersus-
ceptibility could be transmitted from mother to offspring, and that
sensitization by way of the intestinal canal was at least possible.
Both of these factors may have determinative significance in the
present case.8 There may be, because of such conditions, a pre-
existent sensitization which, especially in cases of accidental injec-
tion of the antitoxin directly into a small vein (an accident prob-
ably not infrequent in deep muscular injections), may possibly ex-
plain the few instances of sudden death following the first antitoxin
injection and the isolated instances of "immediate" reaction follow-
ing "first" injections. Rosenau has also suggested recently that sensi-
tization may be unconsciously acquired against various forms of
protein by absorption through the lungs of the organic matter car-
ried in the expired breath of animals. In this way possibly hyper-
6 Currie. Jour, of Hyg., Vol. 7, 1907.
7 See also Goodall, Jour, of Hyg., 7, 1907.
8 Regarding intestinal sensitization see also Richet, C. R. de la Soc. cU
Biol., Vol. 70, 1911 ; Lesne et Dreyfus, C. E. de la Soc. de Biol, Vol. 70, 1911
CLINICAL SIGNIFICANCE OF ANAPHYLAXIS 431
susceptibility against horse protein may be acquired and subsequently
be expressed by a reaction to the first injection of antitoxin.9 While
this must be considered a possibility, however, not all investigators
are ready to accept it and its significance is at present very uncertain.
In the ordinary case of serum-sickness after first injection, how-
ever, the long incubation time elapsing between the injection of the
serum and the onset of symptoms, often more than 10 days, would,
it seems to us, tend to argue against a previous hypersusceptible
state of the patient. On the other hand, we have learned since 10
the earlier studies of Eisenberg, von Dungern, and others that for-
eign proteins injected into rabjbits may be excreted very slowly, and
that even after the formation of antibodies (precipitins) the antigen
may still be demonstrable in the blood serum of the rabbit. Thus
at periods eight to twelve days after the injection of comparatively
large amounts there is often present in the same individual both an-
tigen and its specific antibody side by side, and the essential condi-
tions for the production of anaphylatoxin are thus established. That
the two bodies do not, as a rule, unite in such serum in quantities
sufficient to be demonstrated by alexin fixation has been discussed
in another place, but this by no means excludes a gradual, slow union
of small amounts of antigen with antibody, consequent fixation of
alexin, and the liberation of anaphylatoxic products. In fact, al-
though there does not seem at present to be any way to bring experi-
mental proof to support it, it seems very likely that a slow splitting
of the antigen begins by virtue of the normal antibody, and as, in
the course of eight to ten days, the antibody appears in relatively
larger amounts, the toxic products of the reaction are sufficient to
give rise to symptoms. Such a point of view is supported only by the
experimental knowledge that antibody may appear in considerable
concentration before the antigen has disappeared from the circula-
tion, and upon the facts we know concerning the toxic substances
which arise from the union of two such reagents subjected to the in-
fluence of alexin.
In fact, it seems likely that this process of antibody formation
may represent merely an emergency mechanism for the purpose of
ridding the body of foreign dissolved proteins which have penetrated
into the circulation, cannot diffuse unchanged through the healthy
excretory channels, and must remain in the blood stream until sub-
jected to proteolysis by the enzymes of the blood. In the course of
ordinary life the quantities of such substances gaining entrance into
the circulation are necessarily small, and would call forth but slight
reactions. The sudden injection of large amounts of serum, not
9 Weichhardt, Arch. f. Hyg.,Vo\. 74, 1911, has made similar studies and
claims to have found toxic protein cleavage products similar to his kenotoxin
in exposed air.
10 See Zinsser and Young, Jour. Exp. Med., Vol. 17, 1913.
432 INFECTION AND RESISTANCE
easily disposed of, would, on the basis of the preceding assumptions,
result in a very gradual antigen destruction with consequent antibody
formation, so that, at the end of eight to ten days, there would be
present side by side remnants of unchanged antigen and newly
formed specific antibody. The destroyed antigen fraction, in other
words, gradually sensitizes the body to the fraction which persists and
has not yet been assimilated or excreted at the end of this time. Such
a point of view would explain, not only the reaction after a first in-
jection, but would account for the incubation time in such cases, and
for the differences between these reactions and both the "immediate"
and the "accelerated" reactions of cases twice injected. The bearing
which this point of view would have on the problems of incubation
time in general is obvious.
The recognition of the anaphylactic nature of serum sickness
has led to many attempts to develop methods of antitoxin adminis-
tration by which these reactions could be avoided. Since it was de-
termined that the degree of reaction was directly dependent upon
the amount of the foreign serum injected, it was an obviously logical
procedure to attempt in antitoxin production to concentrate as high
a potency of antitoxin into as small as possible an amount of serum.
Attempts have also been made to alter the serum itself in such a way
that it would lose its properties of acting as an anaphylactic antigen
without suffering materially in antitoxin value. Bujwid11 found
that serum sickness was less frequent after the use of sera which
had been allowed to stand for prolonged periods, and we have seen
that Besredka and others have claimed a reduction of toxic property
in sera heated repeatedly to 60° C. It was hoped, moreover, that
the so-called concentration methods — such as those of Gibson, Banz-
haf, and others — would yield an antitoxin that would be devoid of
anaphylactic properties. None of these methods of altering the
serum can, however, be said to have been satisfactory in that the
antitoxic property seems to be closely associated with the globulins,12
which we have seen are at the same time closely associated with the
production of anaphylaxis.
The conclusions of Rosenau and Anderson 13 regarding this are
based on direct experimentation with concentrated antitoxin made
at the New York Department of Health by the Gibson method.
They found the refined antitoxin, volume for volume, quite as toxic as
the unrefined, but since the same amount of antitoxin is by this and
other methods concentrated in a considerably smaller amount of
11 Bujwid. Quoted from Friedberger and Mita, Deutsche med. Woch.,
No. 5, 1912.
12 Among others previously mentioned see also Turro and Gonzales, C. R.
de la Soc. de Biol., Vol. 69, 1910.
13 Rosenau and Anderson. U. S. Pub. Health and M. H. S. Hyg. Lab.
Bull. 36, April, 1907.
CLINICAL SIGNIFICANCE OF ANAPHYLAXIS 433
protein solution there is a distinct gain for safety in the use of
such preparations. Endeavors to produce potent antitoxic sera by
chemical or physical methods without any sensitizing properties have
thus been unsuccessful.
On the other hand, the knowledge gained by animal experimen-
tation regarding the influence upon the anaphylactic manifestations
exerted by various methods of administering the antigen has led to
results which have proven of much value, both in the immuniza-
tion of experimental animals and in human serum therapy. Prob-
ably the most carefully studied of these methods is the one which
Besredka 14 has recommended on the basis of his work on antianaphy-
laxis in animals. He found that sensitized guinea pigs could be
injected with quantities of serum amounting to about one half or
less of the fatal dose without showing symptoms, and subsequently,
at intervals of 2 to 5 minutes, further injections of the serum could
be given, the total amount five to twenty times exceeding the lethal
dose without causing symptoms of any kind. From these experi-
ments he has developed a method of serum injections the principle
of which is very simply a division of dose. In lieu of injecting into
an animal or man the entire quantity of serum at once, small, gradu-
ally increasing amounts are administered in two, three, or more
doses, the intervals varying from five minutes to several hours, ac-
cording to the necessities of speed indicated by clinical considera-
tions. The process as applied to man consists, then, in preceding
the injection of the larger quantity of the serum by one or two sub-
cutaneous injections of smaller amounts. With this principle well
defined it would be quite unwise to lay down definite rules of quan-
tity or interval at present, since in no instance will it be possible
to estimate the exact condition of susceptibility of the particular
case. It goes without saying that the precautions should be par-
ticularly respected in children in whom the relation of 5 or 10 c. c.
of serum volumes to the body weight approaches the dangerous pro-
portions dealt with in animal experiments.
Besredka has also shown that if the rectum of a sensitive animal
is cleaned out by enema, and a relatively large amount of the an-
tigen then introduced, an injection may be given in within twelve to
twenty-four hours later without danger, however delicate the hyper-
susceptibility of the animal has been. This method apparently
must depend upon a slow, gradual absorption of antigen, and would
seem to furnish a most convenient and advisable method to apply in
man.
14 Besredka. Ann. tie I'Inst. Past., Vol. 24, 1910; C. E. de la Soc. de
Biol, 65, 1908, p. 478; C. E. de la Soc. de Biol., Vol. 66, 1909, p. 125; ibid.,
67, 1909, p. 266; C. E. de I'Acad. des Sc., Vol. 150, 1910, p. 1456; ref. Bull,
de I'Inst. Past.} Vol. 8, 1910, p. 735.
434 INFECTION AND RESISTANCE
Friedberger and Mita 15 have suggested another method which
also depends upon very slow administration rather than division of
dose. In experiments upon guinea pigs they had found that sensi-
tized animals which, as tested by controls, would succumb to intrave-
nous injections of 0.01 c. c. of sheep serum per 100 grams weight
when the entire quantity was injected within one minute, would sur-
vive a similar administration of. as much as 0.1 c. c. if, by means of a
specially constructed apparatus, the injection was made gradually,
extending over a period of 100 minutes. While this method offers
many practical difficulties to ordinary bedside application, it does
show that the intervals of injections by the Besredka method do
not need to exceed fractions of an hour — or, at most, a few hours —
in order to add materially to the safety of injection.
There is another phase of specific therapy in which the question
of possible anaphylaxis must be taken into consideration, and that
is the treatment of patients with bacterial vaccines. As a matter of
fact the danger of anaphylaxis in such cases is probably very remote
— both because of the shortness of the intervals at which these injec-
tions are usually made and because of the extremely small amounts
of protein represented by the usual dose of 100 or 200 millions of
bacteria. However, the possibility cannot be disregarded, especially
in children, and two cases were verbally described to the writer by
Dr. Philip Van Ingen, in which gonococcus vaccines caused immedi-
ate symptoms of such a character that anaphylaxis could not be ex-
cluded.
Ohlmacher 16 also has described localized reactions at the place
of inoculation as well as swelling and tenderness at points of former
inoculations following bacterial vaccine injections. He has oc-
casionally seen slight systemic symptoms (dizziness, nausea, etc.)
which he explains on the basis of anaphylaxis.
Moreover, it must be remembered that active sensitization with
bacterial antigens has been most regularly successful in the hands
of Kraus and Doerr,17 Holobut,18 and Kraus and Admiradzibi,19 as
well as in confirmatory experiments carried out in the Stanford
laboratory, when repeated injections at short intervals were made,
rather than when, as in serum anaphylaxis, a single injection only
was given. This would lend an even closer analogy to the procedures
carried out during vaccine treatment. For instance, in the successful
experiments of the last-named writers ten daily injections of 1/100
of a slant culture of dead colon bacilli were made for the purpose of
15 Friedberger and Mita. Deutsche med. Woch., No. 5, 1912; figures
taken from Versuch., 3.
16 Ohlmacher. Jour. Med. Res., Vol. 19, 1908, p. 113.
17 Kraus u. Doerr. Wien. klin. Woch., No. 28, 1908.
18 Holobut. Zeitschr. f. Immunitatsforsch., Vol. 3, 1909.
19 Kraus u. Admiradzibi. Zeitschr. f. Immunitatsforsch., Vol. 4, 1910.
CLINICAL SIGNIFICANCE OF ANAPHYLAXIS 435
sensitization — the toxic dose of 1/2 slant being given fifteen days
after the last of these.
In order to obtain some opinion regarding the possible dangers
of vaccine therapy in this regard the writer a few years ago ob-
served carefully a pair of young goats, animals extremely favorable
for anaphylactic experiment (and of approximately the weight of a
child of three) in the course of frequent and irregularly spaced
intravenous injections of typhoid bacilli. In both cases marked
anaphylactic symptoms were observed after the animals had attained
a considerable agglutinative and bactericidal power (1 to 5,000 to
1 to 20,000), but in each case only after intravenous injections of
large quantities of bacteria, 1/2 to 2 slant cultures. While, of
course, such experiments are not conclusive in any way, from these,
as well as from a number of laboratory accidents in the course of
animal immunization, it is the writer's impression that the intrave-
nous injection of bacteria or bacterial products in human beings
would be a procedure involving some risk, unless more thorough ex-
perimental data than we at present possess were available to guide
us as to dosage and intervals. The ordinary subcutaneous treatment
of patients, however, with bacteria in the amounts customarily em-
ployed in vaccines would seem to be practically without risk as far
as acute anaphylaxis is concerned.
In the treatment of animals with vaccines of various kinds Le-
clainche2021 has repeatedly called attention to the fact that inocu-
lation with a vaccine may lead to a condition of hypersusceptibility,
serving to light up a latent lesion which 'might have been held in
check if the normal resistance had not been interfered with. This
objection, we have seen, has been made on numerous occasions against
tuberculin therapy, and is one of the factors which have led to the
great caution in dosage and control of all therapy based on active
immunization. These considerations, even more than the rather
remote dangers of serious active anaphylaxis, require that all forms
of specific therapy should be carried out only under the safeguards
of thorough familiarity with the experimental phases of such work.
Our own recent studies on anaphylatoxins, moreover, have in-
clined us to believe that hypersusceptibility to bacterial protein may
well be a strong predisposing factor in infection.
Serum sickness, occurring as a direct consequence of the injection
of a foreign protein into a human being, forces itself upon us as
manifestly related to anaphylaxis. There are a number of other
clinical conditions which are less obviously anaphylactic in nature,
but in which we have many good reasons for attributing an important
part of the etiology to a state of hypersusceptibility. Thus the pe-
20 Leclainche and Vallee. Ann. de I'Inst. Past., 1902.
21 Leclainehe. Eevue Gen. Med. Vet., Sept., 1911; Bull, de I'Inst. Past.,
9, 1911, p. 1089.
436 INFECTION AND RESISTANCE
culiar so-called "idiosyncrasies" observed in many people who suffer
from urticarial skin rashes, gastro-intestinal difficulties, and even
severe systemic illnesses after certain varieties of food seem to de-
pend upon an acquired or possibly inherited hypersusceptibility to
the particular proteins involved, which, at certain times of abnormal
gastro-enteric conditions, can get into the circulation in small quan-
tity. It is not impossible, furthermore, that such unfortunate cases
as the severe forms of angioneurotic edema, which seem, at least in
part, to be associated with gastro-intestinal disturbance, and which
may be transmitted from mother to child, have their root in anaphy-
laxis. For this, however, we have only inference based on clinical
observation.
ASTHMA AND HAY FEVER
Conditions in which there seems to be more definite ground for
association with anaphylaxis are ASTHMA and HAY FEVER. In
asthma the analogy has been clearly set forth by Meltzer.22 He
points out that in both asthma and anaphylaxis the symptoms consist
in a tonic stenosis of the small bronchioli of peripheral origin, and
that both conditions are favorably affected by the administration of
atropin. It is, of course, not certain, but it seems extremely likely
that so-called "nervous asthma" is nothing else than an anaphylactic
attack in a hypersusceptible individual when the particular protein
to which he is sensitive gains access either by the alimentary or
respiratory path.
Very closely related to asthma is the condition known as "hay
fever." This disease has been of recent years most thoroughly stud-
ied by Dunbar.23 Dunbar has ascertained that the hay fever preva-
lent in Europe is dependent chiefly upon a protein substance found
in the pollen of most grasses, while that of America, which occurs
chiefly in the autumn, is caused by the proteins of the pollen cells of
the ambrosiacese and solidaginese — plants which are generally dis-
tributed on the North American continent and bloom in August and
September. The disease occurring in China is caused by another
plant, the Ligustrum vulgare. The suggestion that the disease was
due to anaphylactic action of these pollen proteins upon hypersus-
ceptible individuals was first made by Wolff-Eisner.24 Dunbar has
gone into the question with great thoroughness, and has come to the
conclusion that the disease has much in common with anaphylaxis —
though he believes that, in addition to a hypersusceptibility to the
pollen "toxin," there must be present in the patients, at the same
22 Meltzer. Jour, of the A. M. A., Vol. 55, 1910, p. 1021.
23 Dunbar. Berl. klin. Woch., Nos. 26, 28, 30, 1905; Zeitschr. f. Immuni-
tatsforsch., Vol. 7, 1907; Deutsche r.ied. Woch., Vol. 37, 1911, p. 578.
24 Wolff- Eisner. "Das Heufieber sein .wesen u. seine Behandlung," 1906.
CLINICAL SIGNIFICANCE OF ANAPHYLAXIS 437
time, an abnormal "Durchlassigkeit" or penetrability of the cutis
and mucosa for the pollen substances. He claims to have shown that
a solution of pollen protein instilled into the eye — or even dropped
upon the skin of a hay-fever patient — gives rise to a prompt .and
severe reaction, while it produces no effect upon normal persons.
Unlike experimental serum anaphylaxis, the repeated instillation of
the pollen substances rather increases than diminishes the suscepti-
bility even when these are carried out daily. Furthermore, unlike
serum anaphylaxis, against the manifestations of which no direct
passive immunization has so far been possible, Dunbar claims to
have produced a curative immune serum by the treatment of horses
with the pollen extracts ("Pollantin"). Dunbar, therefore, while
admitting an anaphylaxis-like hypersusceptibility of the patients,
still believes that the antigen in this case is a true "toxin" against
which an antitoxin can be produced — the condition being more
directly comparable to the sensitization against diphtheria and
tetanus toxins observed during the earlier phases of these investiga-
tions by v. Behring and his associates rather than to the phenomena
of serum anaphylaxis themselves.
Schittenhelm and Weichhardt,25 on the other hand, regard hay
fever as truly anaphylactic in every sense. They speak of it as
epithelial anaphylaxis (hay fever being specifically designated as
"conjunctivitis and rhinitis anaphylactica," in distinction from
other forms of cellular anaphylaxis, i. e., enteritis anaphylactica).
They believe that the manifestations of the disease result from a
local hypersusceptibility in which a toxic substance (Abbau Produkt
— similar to anaphylatoxin) is produced. The so-called "antitoxin"
of Dunbar acts favorably only when locally applied, and not on sub-
cutaneous administration. For this reason they do not regard it as
a true antitoxin, but think it acts as a local antiferment which pre-
vents or delays the cleavage of the pollen-substance into its toxic
split-product — thereby preventing or ameliorating the attacks.
Similar to hay fever are the sudden attacks of catarrhal naso-
pharyngitis and conjunctivitis — often of asthma-like respiratory dif-
ficulty, with itching of the nose and eyes and sneezing which many
individuals experience when coming close to horses, cats, or other
animals. In the Stanford University laboratory the writer had an.
assistant who invariably had such attacks, sudden, violent, and of
several hours' duration, when handling guinea pigs for experiment.
The character of such attacks has long aroused the suspicion that
the reaction was anaphylactic in nature, especially since it was known
that extremely slight amounts of antigen could give rise to symptoms
in susceptible subjects. The difficulty in these cases was the ques-
tion of the nature of the antigen which emanated from the animal
to excite an attack. Recently, however, observations having impor-
25 Schittenhelm and Weichhardt. Deutsche med. Woch., 37, No. 19, 1911.
438 INFECTION AND RESISTANCE
tant bearing upon this problem have been made by Weichhardt 26
and by Rosenau,27 who have demonstrated the presence of organic
matter in expired breath. Rosenau condensed the moisture of the
expired breath of man and injected the liquid so obtained into
guinea pigs. After two weeks these animals were injected with nor-
mal human serum, and out of 99 test animals 26 responded with
symptoms of anaphylaxis. This demonstrated not only the presence
of organic matter in the breath, but showed at the same time that
such organic matter was probably protein in nature or at least surely
capable of acting as anaphylactic antigen. Rosenau surmises, there-
fore, that such protein may be slightly volatile under the given con-
ditions. He suggests that sensitization in this manner may explain
the harmful effects resulting from a first injection of horse serum
into patients, previous sensitization having occurred by close associa-
tion with horses. Surely it would explain logically the "cellular"
or epithelial anaphylaxis experienced by certain people in the pres-
ence of animals. In our opinion this is rendered more likely, since,
in the case mentioned as occurring at Stanford University, the in-
halation of washings (both aqueous and alcohol soluble) obtained
from the hair and skin of guinea pigs, and dried in Petri dishes,
produced absolutely no effects in the susceptible individual, whereas
continued handling of a living pig almost invariably caused such
marked effects that the person in question often became useless as
an assistant because of violent attacks of sneezing. It must not be
omitted, however, that not all observers have confirmed Rosenau's
work, and his explanation must therefore be regarded as merely tenta-
tive.
An interesting train of suggestions connecting human pathology
with anaphylaxis has followed the discovery of "organ-specificity"
in the case of hypersusceptibility similar to that noted by Uhlenhuth
in connection with the precipitin formation and described in another
.chapter.
It was shown by Kraus, Doerr, and Sohma,28 we have seen, that
animals sensitized with the crystalline lens protein of one animal
species would react to lens protein in general, and not necessarily
to the tissue protein of the animal species from which it was taken.
In other words, the ordinary "species" specificity did not hold good.
Specificity was determined in this case by the character of the organ
rather than by that of the species. The same thing was shown for
testicular protein by v. Dungern and Hirschfeld.29 The proteins of
these organs from various animals have therefore a certain common
antigenic property which is independent of the antigenic element
26 Weichhardt. Arch. f. Hyg., Vol. 74, 1911.
27 Rosenau and Amoss. Jour. Med. Res., Vol. 25, Sept., 1911.
28 Kraus, Doerr, and Sohma. Wien. klin. Woch., Vol. 21, 1908, p. 1084.
2* Von Dungern u. Hirschfeld. Zeitschr. f. Immunitatsforsch., 4, 1910.
CLINICAL SIGNIFICANCE OF ANAPHYLAXIS 439
common to the particular species. Further than this, Andre jew 3d
claims to have shown that it is possible to sensitize an animal with
its own lens protein. A few guinea pigs injected by him with their
own lens proteins, and reinjected with the same substances after a
suitable interval, reacted with definite anaphylactic symptoms. The
possibility is thus given than an animal or human being could become
sensitized by its own organ proteins if these were traumatically or
otherwise destroyed and absorbed. The train of reasoning is similar
to that which has given much hope of enlightenment to pathologists
when the earlier work upon the cytotoxins was done. Rosenau and
Anderson,31 for instance, injected guinea pigs with guinea pig pla-
centa, and found that, after the usual period of incubation, the ani-
mals reacted to a second injection with marked symptoms of anaphy-
laxis. On the basis of these experiments Rosenau and Anderson
suggest that certain of the toxemias of pregnancy are of anaphylactic
origin. They believe that it is possible that a mother may become
sensitized by the "autolytic products of her own placenta," the result
being eclampsia.
By a similar process of reasoning Elschnig32 has attempted to
explain sympathetic ophthalmia. He claims to have shown that the
laws of organ specificity apply to the proteins (especially the pig-
ment) of the uveal tract. The destruction and absorption of injured
uveal tissue, according to him, induce the formation of organ-
specific antibodies by which the remaining uveal structures of the
same, as well as of the opposite, eye are sensitized. The consequence
is a "sympathetic" inflammation which "is to be regarded purely as
an anaphylactic reaction."
These and other similar suggestions less well founded experi-
mentally illustrate the possibilities for clinical reasoning furnished
by a knowledge of the anaphylactic phenomena. In no cases of this
sort, however, can the association with anaphylaxis be as yet re-
garded as more than an extremely interesting suggestion.
From all that has gone before it is quite evident that most of
the positive facts which may be regarded as determined concerning
the phenomena of anaphylaxis have been obtained in experiments
with small and very sensitive animals, comparatively large and
measured quantities of antigen, and often by the violent method of
intravenous injection in which the entire mass of antigen comes
rapidly into contact with the available antibodies and the vulnerable
tissues. We cannot, therefore, draw rigid parallels between these
experiments and clinical manifestations in human beings in whom
30 Andrejew. Arb. a. d. kais. Gesundh., Vol. 30, 1909.
31 Rosenau and Anderson. U. S. Pub. Health and M. H. S. Hyg. Lab.
Bull 45, 1908.
32 Elschnig. Von Graefe's Archiv f. Ophthal, Vol. 75, p. 459; Vol. 76,
p. 509; Vol. 78, p. 549.
440 INFECTION AND RESISTANCE
the localization, quantitative discharge of antigen, and consequent
production of antibodies is of necessity irregular and different in
each individual case. We have learned, as a general conception,
however, that the introduction into the animal body of antigenic sub-
stances of all varieties leads, under certain conditions, to increased
tolerance or resistance — under other circumstances to a state of
greater susceptibility — these diametrically opposed physiological
consequences being, in all probability, determined by relative concen-
trations of antigen and antibody, their speed of contact, and their
quantitative relationship to available alexin. The problems of clini-
cal medicine in which the possibility of anaphylaxis can be at all
considered, therefore, are extremely complicated, and few of them
can be approached by direct experiment.
In the case of serum sickness the analogy has been so clear and
the experience with human beings so extensive that practically no
doubt can exist as to the common mechanism of this condition with
that of experimental anaphylaxis. In the other conditions men-
tioned the connection is one of great likelihood, but after all is in-
ferential, and calls for much further investigation. For this reason
it is best to abstain from a further enumeration of many other
maladies in which the condition of hypersusceptibility has been sug-
gested as a vaguely possible etiological factor.
ANAPHYLAXIS AND THE TUBERCULIN REACTION
There is one class of phenomena, however, which calls for
further discussion in this connection, since its dependence upon
anaphylaxis, while generally assumed, is still opposed by many
authorities. This consists of the various DIAGNOSTIC REACTIONS in
which extracts of micro-organisms are injected, or brought into con-
tact with the skin or conjunctiva of infected subjects. Such are the
various forms of the tuberculin reaction, the typhoid reaction of
Chantemesse, the one of Gay, and the luetin reaction of Noguchi. In
the tuberculin reaction the conditions have been thoroughly studied,
and we may make a detailed consideration of this example serve to
bring out the general principles involved.
In all forms of the TUBERCULIN REACTION there is a very evident
hypersusceptibility to various forms of antigen derived from the
bacillus. When the tuberculin is injected subcutaneously the reac-
tion is systemic and also localized to a certain extent in any tubercu-
lous foci which may be present. When the v. Pirquet or Moro skin
reactions are carried out, or the Calmette ophthalmic test is made, the
reactions are almost purely local. In all cases reactions are induced
by quantities of antigen which cause no effect whatever in normal
individuals.
CLINICAL SIGNIFICANCE OF ANAPHYLAXIS 441
The -basic observation leading to the diagnostic use of tuberculin
was made by Koch 33 upon guinea pigs. He describes his observa-
tion as follows:
" Tuberculin may be injected into normal guinea pigs in consid-
erable quantities without causing noticeable symptoms. Tuberculous
guinea pigs, on the other hand, react to comparatively small doses
in a very characteristic manner."
Since, in Koch's experiments upon tuberculin, it was desirable
for his particular purposes at the time, to obtain very sharp reac-
tions, he did not content himself with the production of moderate
symptoms by the injection of slight amounts of tuberculin into in-
fected animals, but increased his dosage until the guinea pigs were
killed. He showed that guinea pigs having a moderately advanced
infection — 4 to 5 weeks after inoculation — could be killed by doses
of 0.2 to 0.3 gram, while animals in very advanced stages would suc-
cumb within 6 to 30 hours to quantities as small as 0.1 gram sub-
cutaneously. In the animals so studied he determined not only a
systemic effect, but a very marked local reaction as well in the skin,
areolar tissues, and adjacent lymph nodes.
Koch's observations upon guinea pigs were applied by him, Gutt-
stadt,84 Beck,35 and others to man, and the result was the develop-
ment of the present important diagnostic test. The fundamental
fact in this as well as in other tests of this kind, then, is the
appearance of local and systemic reactions in infected subjects to con-
tact with specific antigenic material which, at least in the same
quantities, produces no effects in normal individuals. The analogy
with the phenomena of anaphyJaxis is thus indicated.
Koch's original interpretation of the phenomenon was of course
unaided by any of the later observations on anaphylaxis. According
to him the tuberculin contained substances which caused tissue
necrosis. The necrotizing action was particularly powerful upon
tissues which were tuberculous, and therefore already saturated with
the toxic material. The destruction of such tissues resulted in sys-
temic symptoms.
Very similar to this view is the one later expressed by Babes and
Broca,36 who attribute the systemic symptoms to a sudden lighting
up of the existing lesions by the small amount of extra tuberculin
added to that already present in these foci.
The first suggestion of the possible association of the tuberculin
reaction with the union of an antigen and its antibody was made
by Wassermann and Bruck.37 They accepted Ehrlich's assumption
33 Koch. Deutsche med. Woch., No. 43, 1891.
84 Guttstadt. "Klin. Jahrbuch" Erganzung-sband, 1891.
35 Beck. Deutsche med. Woch., No. 9, 1899.
36 Babes u. Broca, Zeitschr. f. Hyg., Vol. 23, 1896.
37 Wassermann and Bruck. Deutsche med. Woch., No. 12, 1906.
442 INFECTION AND RESISTANCE
that certain cells of the tuberculous foci (those situated just below
the periphery and already affected by the tubercle toxin, though still
resistant) were possessed of an increased receptor apparatus for the
tubercle antigen. For this reason the injected tuberculin was con-
centrated in these foci, attracted out of the circulation by the in-
creased avidity of these cells, the consequence being increased ac-
tivity of the lesions and systemic symptoms. The tuberculin reac-
tion, according to these writers, therefore, would be caused by the
union of the tuberculin with the "sessile receptors" upon the diseased
tissues — a point of view which would specify the diseased tissues and
their products as the sources from which emanated the toxic factors
inciting the systemic symptoms.
The theories of Koch and of Babes do not, as Meyer points out,
explain the frequent absence of the tuberculin reaction in very ad-
vanced cases of human tuberculosis, as contrasted with its frequency
and regularity in the earlier cases. For, according to both of these
views, the more severe the existing lesions the more actively would
the injected tuberculin initiate tissue necrosis and consequent symp-
toms. The theory of Wassermann and Bruck avoids this objection
since it presupposes the acceptance of Ehrlich's view that the in-
creased receptor apparatus is present and free only in those cells in
which necrotic destruction has not yet set in. In the necrotic areas
the receptor apparatus is already saturated or satisfied as to its
affinities, and extensive areas of necrosis, therefore, are unaffected
by contact with further quantities of tuberculin.
All of these theories, however, inasmuch as they refer the tubercu-
lin reaction to alterations taking place in more or less active lesions,
are unable to account for the occurrence of the reaction in persons
in whom healed foci only are present, and are entirely inconsistent
with the facts we now possess regarding the cutaneous and ophthal-
mic tests in which the reactions occur in previously healthy tissues.
These facts practically exclude the acceptation of any theories
which regard the tuberculous focus as the sole source of the reaction.
We may still accept the Koch or Wassermann views to explain local
swellings and other changes in infected lymph-nodes or other lesions,
but we must assume in addition to this a generalized hypersuscepti-
bility at least analogous to the phenomena of anaphylaxis.
This ability of previously healthy tissues, remote from any center
of tuberculous infection, to react to the application of tuberculin
was discovered by von Pirquet38 in the development of his skin
reaction, and by Calmette 39 and Wolff-Eisner 40 in their work upon
88 v. Pirquet. Berl. Urn. Woch., No. 20, p. 644, and No. 22, p. 699, 1907;
also "Klinische Studien iiber Vaccination," Deuticke, Wien, 1907.
39 Calmette. C. E. de I'Acad. des Sciences, June, 1907.
10 Wolff-Eisner. Berl. klin. Woch., 1907, p. 1052. Discussion of paper
by Citron.
CLINICAL SIGNIFICANCE OF ANAPHYLAXIS 443
the ophthalmoreaction. The principle involved in these reactions
was then further emphasized by the introduction of the Moro tuber-
culin-ointment method and the intracutaneous test of Romec. The
mere observation that the infection with tuberculosis results in a
general tissue hypersusceptibility immediately suggests the interpre-
tation of the tuberculin reaction as a manifestation of anaphylaxis.
Von Pirquet, accordingly, on the basis of his previous studies upon
serum sickness includes the reaction in the category of what he calls,
"allergic."
He assumes that the reaction depends upon the presence in the
system of antibodies, which form a union with the applied tubercu-
lin, the result being the formation of poisons and a reaction. This
assumption, according to the anaphylatoxin theory of Friedberger,
would imply the participation of alexin in the reaction — acting upon
the united tuberculin-antituberculin complex, though v. Pirquet
does not express himself positive as to this.
This, moreover, is the clearly expressed opinion of Friedberger41
himself. Consistently with his general theory of anaphylaxis he
assumes that the injected tuberculin comes into relation with specific
antibodies with which it unites, the alexin then splitting off anaphy-
latoxin from the complex. He bases this view upon his experimental
demonstration, mentioned above, of the production of anaphylatoxin
from tubercle bacilli by in vitro digestion with guinea pig comple-
ment.
In principle the view of v. Pirquet is similar to that previously
expressed by Wolff-Eisner42 that the union of tuberculin with its
lytic antibody, present in the tuberculous animal, gave rise to pois-
ons as the result of lysis. Both of these theories simply apply to the
special case of the tuberculin reaction theories of mechanism applied
to anaphylactic reactions in general.
We must admit that the facts of the "allergic" reactions as a
class seem to force upon us the acceptation of von Pirquet's views.
Apart from the purely clinical observations made in carrying out the
routine tests we have the additional evidence that the instillation of
tuberculin into the eye of normal individuals gives rise to no reac-
tion, but a repetition of the instillation into the same eye after ten
or more days results in a marked and typically positive test. Further-
more, von Pirquet 43 states that individuals showing no clinical tu-
berculosis and negative to a first test will often react ("sekundare
Reaktion") to a second test carried out a few days after the first.
These facts all seem to indicate acquired hypersusceptibility more
analogous to true serum-anaphylaxis than to the toxin hyper suscepti-
41 Friedberger. Munch, med. Woch., Nos. 50 and 51, 1910.
42 Wolff-Eisner. Berl. klin. Woch., Nos. 42 and 44, 1904.
43 Cited from Lowenstein in "Kraus u. Levaditi Handbuch," Vol. 1, p.
1039.
444 INFECTION AND RESISTANCE
bility of von Behring, in that the tuberculin is but slightly toxic in
itself.
If the analogy is such a close one, therefore, it should be easy to
formulate experiments by which the phenomena now ascertained
regarding serum-anaphylaxis could be demonstrated for tuberculin
hypersusceptibility. The obvious procedure, therefore, would be to
attempt to passively transfer tuberculin sensitiveness to a normal
animal with the serum of a tuberculous one. This has indeed been
attempted by Friedemann,44 later by Bauer 45 and a number of
others — usually with negative result. The writer, hoping to develop
a diagnostic method for tuberculosis, has also attempted this by the
transference of human tuberculous blood to guinea pigs, but invari-
ably obtained negative results. Yamanouchi46 alone has reported
positive experiments by a similar procedure with rabbits, but so far
his results, according to Friedemann, have completely failed of con-
firmation. Austrian succeeded by sensitizing guinea pigs with 5
c. c. of titrated whole blood, using for the second injection a tuber-
culo-protein prepared by the method of Baldwin.47 In this particu-
lar, therefore, the analogy between anaphylaxis and the tuberculin
reaction, though not easily worked out, has nevertheless been estab-
lished. Another objection which has been made by a number of ob-
servers is the fact that anaphylaxis is accompanied by temperature
depression while tuberculin reactions are followed by a rise. This
objection may be regarded as invalid, however, in the light of Fried-
berger's 48 experiments which showed that temperature depression
follows only when large doses of the antigen are injected into the
sensitized animals, smaller doses often giving rise to increased tem-
perature.
We gain a certain amount of insight into the conditions here
prevailing by considering the information which has been obtained
from the study of the antibodies formed in animals in tuberculosis.
It appears from the work of Christian and Rosenblatt 49 that anti-
bodies to the tubercle bacillus are formed by tuberculous animals
only. Normal animals form these to a very slight degree only, if at
all, when immunization with tuberculin is attempted. In other
words, as Friedemann (aWeichhardt's Jahresbericht," 6, 1910)
points out, the specific reaction of antibody formation in tuberculosis
seems to be closely associated with the tuberculous tissues themselves.
44 Friedemann. "tiber anaphylaxie," "Weichhardt's Jahresbericht," Vol. 6,
1910.
45 Bauer. Cited ibid.; also Munch, med. Woch., 1909, p. 1218.
[Q Yamanouchi. Wien. klin. Woch., 1908, p. 1263.
47 Austrian Bull, of the Johns Hop. Hosp., Vol. 24, 1913 ; Baldwin, Journ.
Med. Res., Vol. 17, 1910.
48 Friedberger. Deutsche med. Woch., No. 11, 1911.
49 Christian and Rosenblatt. Munch, med. Woch., 1908.
CLINICAL SIGNIFICANCE OF ANAPHYLAXIS 445
The same inference can be made from Bail's 50 experiments on
passive sensitization. For, although passive sensitization of guinea
pigs with the serum of tuberculous animals has been unsuccessful,
Bail succeeded in obtaining lethal anaphylactic reactions by injected
macerated tuberculous tissues, following these on the next day by
injections of tuberculin. It is plain from this, as Friedemann cor-
rectly argues, that we must assume that the antibodies (receptors)
formed against tubercle bacilli are closely bound up with the tissue
cells, the reaction of tuberculin being largely with "sessile receptors."
Indeed, it seems as though the antibodies formed against tubercle
bacilli undoubtedly remain in close relation to the cells of the tis-
sues except in cases of active tuberculosis in which localized areas
of cells are under the influence of a very intense action of the poi-
sons, and a consequent overproduction and discharge of receptors
(using the Ehrlich nomenclature) may occur. This would corre-
spond with considerable accuracy, moreover, to the histological facts,
for in this infection, similar to leprosy and a number of other con-
ditions, but unlike most acute infections, the battle against the micro-
organisms is carried out chiefly by the adjacent tissue cells.
We might assume, therefore, that, in tuberculous individuals,
there is indeed a reaction, at first local, then generalized to a slight
degree, in which antibodies are actually formed. These antibodies,
however, remain to a preponderant extent sessile, or incorporated in
the reacting cells. Upon the injection or application of tuberculin
the reaction takes place in or upon the cells. Whether or not the
further cooperation of complement or alexin is then necessary for the
lysis and poison production from the antigen, as in the similar reac-
tions taking place in the circulation, or whether the intracellular
ferments themselves suffice for this, cannot be decided at present.
It is certainly not unlikely that the circulation of tubercle-antigen —
even in small quantities — throughout the body may produce such
hypersusceptibility of cells (represented graphically by the concep-
tion of sessile receptors in many parts of the body remote from the
lesion — a quality remaining constant for prolonged periods and ex-
plaining the subsequent skin and ophthalmic reactions obtained upon
test). Certain clinical observations cited by v. Pirquet 51 would seem
to support such a view. For instance, he states that, having em-
ployed his left forearm repeatedly for tests, he was able to obtain
positive reactions in this area with tuberculin diluted 1 to 1,000,
whereas his right forearm was negative to tuberculin ten times as
concentrated. Furthermore, as Kohn 52 has shown that, while the
first injection of tuberculin into the eye of a normal person produces
no reaction, this eye will not only react to a second instillation,
50 Bail. Zeitschr. f. Immunitatsforsch., Vol. 4.
51 V. Pirquet, "Kraus u. Levaditi Handbuch," Vol. 1, p. 1050.
52 Kohn. Quoted from Lowenstein, Kraus and Levaditi, Vol. 1, p. 1033.
446 INFECTION AND RESISTANCE
but will show a reaction when the second application is made sub-
cutaneously. Negative evidence pointing in the same direction is
the observation that absolutely no influence is exerted upon the out-
come of tuberculin reaction if the tuberculin is previously mixed
with blood serum from either positively or negatively reacting
cases.53 This would tend to show that, whether reacting or not, the
factors which determined this are certainly not present in the cir-
culating plasma.
That the circulation of tubercle antibodies in the blood may even
interfere with the localized tuberculin reaction is rendered likely by
the fact that skin reactions are often negative in cases of advanced
tuberculosis, and that, as we are told by Dr. Blair, of the N". Y.
Zoological Park, such reactions are usually negative in tuberculous
monkeys in which the disease is invariably very rapid and acute.
PRACTICAL DIAGNOSTIC USES OF ANAPHYLAXIS
The specificity of the anaphylactic reaction has led to extensive
attempt to utilize it for forensic protein determinations in the same
way in which the precipitin test is used. Uhlenhuth,54 Thomsen,55
Pfeiffer, and others have carried on extensive experimentation in
this problem, the technique, in general, consisting in sensitizing
guinea pigs with solutions of the unknown protein (dissolved blood
spots, etc.) and testing them by a second injection of the suspected
protein after the usual anaphylactic incubation time. The results of
such work have shown that indeed positive reaction may be obtained
and diagnosis made in this way. However, the reactions are not ordi-
narily very striking, and this method is not as reliable as the precipi-
tin method. Uhlenhuth 56 believes that the anaphylactic reaction
has value only in cases in which the amount of unknown protein is so
small or so changed by preservation or decomposition that its pre-
cipitable qualities have been lost.
Yamanouchi's 57 attempt to utilize anaphylaxis for the diagnosis
of tuberculosis, by passively sensitizing guinea pigs with the serum
of tuberculous patients and testing subsequently with tuberculin, has
been mentioned before. Although he claims positive experiments,
our own experience with a similar technique has given us results
which were so irregular that we feel that this technique has very
slight practical value, if any.
Pfeiffer 58 has attempted to apply anaphylaxis to the diagnosis
53 V. Pirquet. Loc. cit.
54 Uhlenhuth. Deutsche milit. Zeitschr. No. 2, 1909. Cited from same
author, Zeitschr. f. Immunitatsforsch., Vol. 1, 1909.
55 Thomsen. Zeitschr. f. Immunitatsforsch., Vol. 1, 1909.
56 Uhlenhuth. Zeitschr. f. Immunitatsforsch., Ref. Vol. 1, 1909, p. 525.
57 Yamanouchi. Wien. klin. Woch., No. 47, 1908.
58 Pfeiffer. Zeitschr. f. Imm., Vol. 4. 1910.
CLINICAL SIGNIFICANCE OF ANAPHYLAXIS 447
of malignant tumors. ' Together with Finsterer 59 he sensitized
guinea pigs with the serum of carcinomatous patients following the
injection 48 hours later with press-juices of tumors. His conclu-
sions were drawn from the anaphylactic temperature reaction, and
he claims that animals so sensitized are hypersusceptible to the
juices obtained from carcinomata, whereas animals sensitized with
normal serum or the serum of sarcoma patients show no hypersus-
ceptibility. Ranzi 60 has not been able to confirm this. The signifi-
cance of such experiments, if correct, apart from their practical
value for the diagnosis of carcinoma, would be considerable in that
they tend to show that cancer tissues contain a specific protein
which is antigenically distinct from the other tissue-proteins of
the afflicted individual. However, we cannot yet accept these facts
as absolutely established.
59 Pfeiffer and Finsterer. Wien. klin. Woch., No. 28, 1909.
60 Ranzi. Zeitschr. f. Imm.} Vol. 2, 1909.
CHAPTER XIX
THEEAPEUTIC IMMUNIZATION IN MAN
FACTS CONCERNING ANTITOXIN TREATMENT IN MAN
THERAPEUTIC USE OF DIPHTHERIA ANTITOXIC
IT is not consistent with the purpose of this brief treatise to dis-
cuss extensively the therapeutic benefits obtained by serum therapy in
diphtheria. We can convey briefly an adequate idea of this by citing
some of the tables given by Northrup in Nothnagel's "Encyclopedia
of Practical Medicine," American Edition, Volume on Diphtheria,
etc., p. 143. These figures are taken from the statistics of the New
York Board of Health, which began treatment of diphtheria with
antitoxin in January, 1895. Dr. Northrup states, however, that
serum treatment cannot be considered to have been in general use
until some time later.
Without Antitoxin
Year
Cases reported
Deaths
Mortality, per cent.
1891
5,364
1,970
36.7
1892
5,184
2,106
40 0
1893
7,021
2,558
36.4
1894
9,641
2,870
29 7
Total
27210
9,504
Avg. 34 9
With Antitoxin
1895
10353
1 976
19 0
1896
11,399
1,763
15 5
1897
10896
1 590
14 5
1898
7 173
919
12 8
1899
8,240
1,085
13.1
1900
8364
1 176
14 0
Total
56425
8509
Avg. 15 0
Table taken directly from Northrup, loc. cit.
THERAPEUTIC IMMUNIZATION IN MAN 449
From this table there appears a reduction of 58 per cent, in
mortality and a similar drop is evident from the German statistics
of Dieudonne,1 from those of Welch, and many others.
It should be considered, moreover, in reading such statistics that
they are made on gross mortality reports without elimination of the
many cases that have not come under observation until too severely
diseased to react to any form of treatment. The reason for the fail-
ure to obtain results with antitoxin when the cases have proceeded
beyond a certain stage of intoxication will become evident when we
consider the manner of absorption of the poison in a succeeding para-
graph. The mortality sinks to between 8 and 9 per cent., when such
cases are omitted, as is shown by the collective investigations of the
American Pediatric Society in 1896 — figures which we take also
from Northrup's comprehensive study. This purely statistical evi-
dence, however good, is further reenforced by the unquestionable
and considerable diminution of emergency operations,2 such as intu-
bation and tracheotomy, since introduction of the antitoxin. More-
over, there is the manifold clinical evidence of benefit, after the
serum treatment, familiar to every practicing physician.
Although the injection of antitoxin is of benefit by whatever
route and in whatever quantity it may be given, nevertheless recent
experimental investigations have taught us much regarding the
proper use of this therapeutic agent. Especially interesting are the
investigations of Meyer,3 who showed the extreme importance of an
early use of the antitoxin. Apparently, as we have mentioned in
another place, like tetanus antitoxin, the diphtheria poison may be
in part absorbed directly by the nerves.4
There is apparently a great difference in therapeutic efficiency,
according to the method in which the serum is administered, a differ-
ence probably depending upon speed of absorption. Berghaus5
showed that intravenous injection is 500 times more potent therapeu-
tically than the subcutaneous, and 80 to 90 times. more so than the
intraperitoneal injection. Schick, in discussing this problem from
the clinical point of view, for this reason lays special stress upon the
speed of administration. He says: "Not only days but hours are
of great importance." He bases this opinion largely upon the fact
that the toxin which has already united with the nerve substance can
probably no longer be influenced by the injection of the serum.
According to the experiments of Meyer and Ranson diphtheritic
1 Dieudonne. Arb. aus dem kais. Gesund., XIII, 1897.
2 Siegert. "Jahrbuch f. Kinderheilkunde," Vol. 52, cited after Wernicke.
3 Meyer. Berl. kl Woch., 25, 26, 1909 ; Arch. f. exp. Path. u. Ther., Vol.
60, 1909, and Berl. kl. Woch., No. 45, 1911.
4 For a thorough discussion of these conditions see Schick, Centralbl. f.
Bakt., Rev. Vol. 57, 1933, "Report of 7th Meeting of the Mikrobiol. Gesell.,"
Berlin, 1913.
5 Berghaus. Cited from Schick, loc. cit.
450
INFECTION AND RESISTANCE
paralysis may follow even when vigorous serum treatment has been
employed. For, according to them, only the toxin which has reached
the central nervous system through the circulation can be influenced
by the serum, but no effect is possible upon the fraction which has
been absorbed from the nerve endings directly.
Schick,6 on the basis of extensive experiments, comes to the con-
clusion that the subcutaneous injection of 1,000 to 2,000 units in
diphtheritic cases has an immunizing value only, which protects the
tissues from further injury and leads to cure if, at the time of injec-
tion, the lethal dose has not yet united with the sensitive cells. "If,"
he states, awe wish to obtain antitoxic action upon toxin which has
already gone into action before the injection of the serum, then re-
sults can be obtained both in man and in animals only if a great deal
of antitoxin is injected intramuscularly or intravenously." 7
Interesting also from a clinical point of view are the studies of
Schick,8 Hahn,9 and others10 upon the presence of antitoxin in the
blood of normal, untreated individuals at different ages. These
investigations were carried out by the intracutaneous method of
toxin and antitoxin determination described in greater detail in a
later section. The following table, taken from the article of Hahn,
illustrates the experience, in such investigations, both of Schick and
of Hahn himself. The determinations were carried out upon indi-
viduals who had never had diphtheria, as far as could be learned.
Age
Cases with
antitoxin serum
Cases without
antitoxin serum
Highest antitoxin
value in 1 c. c.11
Schick -
r Newborn ....
L0-l year
11
1
0
3
under 1.5 units
0.11 unit
:- 2-10 years..
7
5
1.0 unit
11-20 years. .
8
9
0.75 unit
Hahn
21-30 years
9
5
2 5 units
3 1-40 years..
5
1
0.25 unit
L 4 1-65 years. .
2
8
2 . 5 units
The table shows that in newborn children there is almost regu-
larly a definite and sufficient protective value in the serum which
diminishes up to the first year, so that at the end of the first year
three out of four individuals had no antitoxin in their serum. In
subsequent years up to the age of 40 an increasing percentage of
6 Schick. Loc. cit.
J Schick. Loc. cit., p. 32.
8 Schick. ' "tiber Diphtherimmunitat," Wiesbaden, 1910.
9 Hahn. Deutsche med. Woch., Vol. 38, No. 29, p. 1366, 1912.
10 Karasawa and Schick. Zeitschr. f. Kinderkrankheiten, 1910, and " Jahr-
buch f. Kinderheilkunde," 1910.
11 Table taken directly from Hahn, loc. cit.
THERAPEUTIC IMMUNIZATION IN MAN 451
people have sufficient amounts of diphtheria antitoxin in their blood.
After the age of 40 an increasing percentage is without such protec-
tion. The first observation, that newborn children usually possess
considerable amounts of antitoxin, is very probably due to passive
immunization by the blood of the mother, a fact which we have men-
tioned in another place. The exact method by which such measure-
ments are made is described in a subsequent section on the intra-
cutaneous method of determining toxin and antitoxin.
The work of Schiek, that of J. Henderson Smith, and recent
studies by Park and Biggs promise to alter considerably the methods
of antitoxin therapy as at present in use in diphtheria. Smith meas-
ured the speed of absorption of antitoxin injected subcutaneously into
the abdominal wall of a healthy man. His results are shown in the
following table, which we take from his communication (page 213) :
TABLE V
One c. c. of the patient's serum contained:
Before injection No demonstrable antitoxin
5 hours after injection 0.1 unit antitoxin
14 hours after injection 0 . 225 unit antitoxin
32 hours after injection 0.68 unit antitoxin
44 hours after injection 1.0 unit antitoxin
3 days after injection 1.3 units antitoxin
4 days after injection 1.3 units antitoxin
6 days after injection 0 . 68 unit antitoxin
13 days after injection 0 . 17 unit antitoxin
15 days after injection 0 . 14 unit antitoxin
20 days after injection 0 .08 unit antitoxin
27 days after injection No demonstrable antitoxin12
Park and Biggs 13 have made similar studies and have contrasted
the speed of absorption after subcutaneous administration with that
after intravenous injection, basing their curves upon careful measure-
ments of the sera of the treated patients. We reproduce their
charts as given in their recent publication.
It is apparent from these charts, as well as from the work of Hen-
derson Smith, that antitoxin, subcutaneously given, is slowly ab-
sorbed, and does not reach its maximum concentration in the blood
stream until forty-eight hours or more after the injection. It fol-
lows that, as Park and Biggs point out, it is more rational to inject
a single adequate dose than to divide the dosage and inject at inter-
vals. They have obtained results in animal experiment which
graphically illustrate this principle. A rabbit which had received
ten fatal doses of toxin intravenously was given a total of 500 anti-
toxin units in divided doses as follows : 100 units after twenty min-
12 J. Henderson Smith. Journal of Hygiene, Vol. 7, 1907, p. 205.
13 Park and Biggs. Collected Studies from the N. Y. Department of
Health, Bureau of Laboratories, Vol. 7, 1912-1913, p. 27.
452
INFECTION AND RESISTANCE
6.0
CHART I. — Showing the extent and rapidity of absorption of 10,000 units of
antitoxin given subcutaneously. Each line represents the antitoxin content
of 1 c. c. of blood at different intervals of time. (From Park and Biggs,
loc. cit.)
THERAPEUTIC IMMUNIZATION IN MAN
453
CHART II. — The antitoxic power of human blood after an intravenous injection
of 10,000 antitoxic units. (From Park and Biggs, loc. cit.)
utes, 100 after 40 minutes, and 150 units each after 60 and 80 min-
utes. This rabbit died. Another animal given the same dose of
toxin received 200 units of antitoxin twenty minutes later and lived.
The amount necessary to save life in rabbits receiving ten fatal doses
intravenously was as follows :
Given after 10 minutes 5 units antitoxin
Given after 20 minutes 200 units antitoxin
Given after 30 minutes 2,000 units antitoxin
Given after 45 minutes 4,000 units antitoxin
Given after 60 minutes 5,000 units antitoxin
Given after 90 minutes . . No amount
These extremely important experiments of Park and Biggs bear
out the opinion of Schick and show beyond question that the proper
way to give antitoxin is to give a single adequate dose as early as
possible. They emphasize the fact that probably the most important
single point in the specific therapy of diphtheria is the speed with
which the diagnosis can be made and the antitoxin given. At the De-
partment of Health the dosage now employed, as given by Park and
Biggs, is the following :
UNITS ii
i CASES
Mild
Moderate
Severe
Very Severe
Infants under 1 year
2,000
3,000
10,000
10000
Children 1 to 5 years
3,000
5,000
10,000
10,000
Children 5 to 9 years
Persons over 10 years
4,000
5,000
5,000
10,000
10,000
10,000
15,000
20,000
454 INFECTION AND RESISTANCE
PRACTICAL CONSIDERATIONS CONNECTED WITH DIPHTHERIA ANTI-
TOXIN PRODUCTION AND STANDARDIZATION
The conditions which govern the active production of toxins by
bacteria in culture media are not only of great theoretical interest
but possess unusual practical value in that the most important factor
for the successful production of a strong antitoxin consists in the
preliminary preparation of a potent toxin. The bacterial true toxins
are all "exotoxins" in that they are soluble, moderately diffusible
substances which pass readily from the bacterial bodies to the en-
vironment, and for this reason can be obtained most readily by the
cultivation of the bacteria upon fluid media and subsequent filtration
of the cultures through earth or porcelain filters.
The choice of culture or strain is an important element in this
procedure, since within the same species of toxin-producing micro-
organisms there is much variation in the speed and energy of toxin
production. Thus for unknown reasons some strains of diphtheria
bacilli will far outstrip others in this respect. An excellent illustra-
tion of this is the experience of Park and Williams 14 with two diph-
theria cultures — a very virulent and a very weak one. Of the
former, 0.002 c. c. of a forty-hour bouillon culture killed a guinea
pig, while of the latter 0.1 c. c. of a similar culture was necessary
for the same result.
In the case of tetanus, cultural differences do not seem to be as
common. Individual strains also may gain or lose in toxin-pro-
ducing powers, according to the method of handling them which is
practiced. It is stated,15 for instance, that a diphtheria culture will
lose in energy of toxin production if permitted to grow without suffi-
ciently frequent transplantation. However, transplanted on solid
media with reasonable frequency, these bacteria show a remarkably
constant toxin production. A well-known strain, the Park-Williams
No. 8, now in use in many antitoxin laboratories throughout the
world, has persisted for over 15 years in producing a strong toxin.
There are occasional strains among toxin-forming species which are
entirely devoid of this property. Diphtheria bacilli which were
virulent while possessing all the other cultural characteristics of the
group have been described, but appear, from the experience of the
writer, to be rather uncommon.16 Of tetanus bacilli little is known
in this respect.
Given a powerfully toxic strain of. the proper bacteria the
method of cultivation is also of great importance in influencing the
eventual yield of poison. These relations have naturally been stud-
ied with the greatest care in the case of diphtheria and tetanus
14 Park and Williams. "Pathogen. Micro-organ.," N. Y., 1910.
15 Park and Williams. Loc. cit.
16 Zinsser. Jour. Med. Ees., N. S., Vol. 12, 1907.
THERAPEUTIC IMMUNIZATION IN MAN 455
bacilli, since in these cases there has been the greatest practical appli-
cation for such knowledge.
In the case of diphtheria, though toxin will be produced on all
media on which the bacillus grows easily, the most favorable medium
for this purpose is a slightly alkaline broth made of lean beef or
veal infusion and containing peptone. Since acid formation hinders
the production of toxin, Martin 17 has suggested fermentation of the
APPARATUS ARRANGED FOR THE STERILE FILTRATION OF DIPHTHERIA CULTURES
IN TOXIN PRODUCTION.
(After Eosenau, U. S. Hyg. Lab. Bull. 21, 1905, p. 38.)
muscle sugar with yeast, while Theobald Smith 18 recommends pre-
liminary fermentation with Bacillus coli.
Park and Williams 19 regard this as unnecessary. They recom-
mend a 2 per cent, peptone broth made of veal. This is neutralized
to litmus and 7 to 9 c. c. of normal NaOH solution to the liter are
added. In such a medium at 37.5° C. the production of toxin begins
within 24 hours and reaches its highest point in from five to ten
days. When at its height the process must be stopped and the cul-
tures exposed to a lower temperature, otherwise rapid deterioration
takes place because of the instability of the toxin. Even when kept
cold and in the dark this deterioration proceeds steadily though
slowly. At first, however, even under these conditions a compara-
tively extensive loss of toxin goes on — a process sometimes spoken
of as "maturing of the toxin" — after which the poison strikes a
17 Martin. Ann. Past., 1896.
18 Th. Smith. Journ. Exp. Med., IV, 1899, p. 373.
19 Park and Williams. Journ. Exp. Med., Vol. 1, 1896.
456 INFECTION AND RESISTANCE
fairly constant and very gradual rate of weakening, and is, com-
paratively speaking, stable.
In the United States Hygienic Laboratory in Washington, ac-
cording to Rosenau, the recommendations of Theobald Smith are
largely followed in the production of toxin. The procedure is as
follows :
The culture medium, "Smith's Bouillon," is prepared from
chopped beef from which fat and tendon have been cut out. This is
adjusted by phenolphthalein titration to 0.5 per cent, acidity. It is
then placed into Fernbach flasks and inoculated on the surface with
a Park-Williams bacillus ~No. 8. The flasks are incubated for 7 days
at 37.5° C. The reaction of the medium after such incubation is
determined, and flasks showing an acidity of 1.5 or over are dis-
carded. The usual reaction at the end of incubation is 0.6 to 0.8
per cent, acidity. This broth is filtered through Berkefeld filters or
porcelain candles.
Toxin so prepared is now tested and its L0 and L+ doses deter-
mined by the methods described above. Rosenau20 states that poi-
sons are discarded as containing too large a proportion of toxon if
the difference between L0 and L+ is greater than 15 M L D. The
toxin is now set aside in flasks for the process which Rosenau
calls "seasoning." At intervals of about a month it is retested
and finally it is found that the rate of toxoid formation decreases
and the poison reaches a period of equilibrium. It can now be used
for accurate determination of the L+ dose, and this is done from
careful measurements on a large number of guinea pigs.
Examples 21 of such measurements, abbreviated for the sake of
simplicity, are given in the following tables :
Toxin Determinations of M L D or "T"
Dose in c. c. Result
0.03 = death in 1^ days
0.02 = death in \y% days
0.01 = death in 2 days
0.008 = death in 3 days
0.006 = death in 3}^ days
0.005 = death in 4 days M L D
0.004 = death in 6 days
0.003 = death in 8 days
0 . 002 = late paralysis
0.001 = well in 16 days.
Toxin Determination of L+ Dose
1 Antitoxin unit -f- 0 . 2 c. c. = 0
1 Antitoxin unit -j- 0.21 c. c. = 0 = L0
1 Antitoxin unit + 0.22 c. c. = local infiltration
20 Rosenau. Hyg. Lab. Bull No. 21, April, 1905.
21 Examples are taken from measurements reported by Rosenau, loc. cit.
THERAPEUTIC IMMUNIZATION IN MAN 457
1 Antitoxin unit + 0.23 c. c. = fatal in 17 days
1 Antitoxin unit -j- 0.24 c. c. = fatal in 14 days
1 Antitoxin unit -f 0 . 26 c. c. = fatal in 9 days
1 Antitoxin unit -j- 0 . 28 c. c. = fatal in 6 days
i Antitoxin unit -f 0.29 c. c. = fatal in 4 days = L+
1 Antitoxin unit + 0.3 c. c. = fatal in 3 days
The production of antitoxin is carried out by the graded injec-
tion of antitoxin into horses. Young, healthy horses are chosen,
tested for freedom from, glanders, and the first injections are made
either with toxin attenuated by the addition of Lugol's solution or
terchlorid of iodin, or, as in the New York Health Department, the
first injections consist of mixtures of toxin and antitoxin. We take
our description largely from the account given by Park.22 The first
injection consists of 12 c. c. of toxin (M L D 1/400 c. c.), together
with 100 units of antitoxin. After the reaction from such an injec-
tion has completely subsided — after 3 to 5 days- — a second injection
is given of toxin without antitoxin; then 15 c. c., 45 c. c., 55 c. c.,
65 c. c., 80 c. c., 95 c. c., 115 c. c., 140 c. c., etc., the intervals be-
tween injections being about three days and depending upon the
reaction of the horse and the speed with which it entirely recovers
from the preceding injection. In a particular case cited by Park
675 c. c. of toxin could be given by the 60th day; in this case by the
28th day the horse was yielding 225 units to the c. c. ; on the 40th
day, 850 units; on the 60th day, 1,000 units.
The determination of the antitoxin unit, carried out from time
to time on the serum of such a horse against the L+ dose described
in our preceding table, would be carried out as follows :
In all such standardization great care must be taken in employ-
ing accurately standardized glassware. Rosenau recommends em-
ploying "capacity instruments" rather than "outflow instruments."
Dilutions of unknown antitoxin are made in 0.85 per cent, sterile
salt solution. As a basic dilution one part of the antitoxic serum to
nine of the salt solution gives 1/10 c. c. to each cubic centimeter,
and from this initial dilution further dilutions may be easily made
as follows : 1 c. c. of dilution I. + 9 c. c. salt solution = 1-100, etc.
A series of mixtures is then made in each of which the quantity of
toxin equals the L+ dose, and in which the quantity of antitoxin
varies within a wide margin of the limits of strength to be expected.
This is illustrated in the following table :
L+ (0.29 c. c.) + 1/500 c. c. of antitoxic serum = lives
L+ (0.29 c. c.) -j- 1/600 c. c. of antitoxic serum = lives
L+ (0.29 c. c.) -j- 1/700 c. c. of antitoxic serum = lives
L+ (0.29 c. c.) -j" 1/800 c. c. of antitoxic serum = dies in 8 days
L+ (0.29 c. c.) 4" 1/900 c. c. of antitoxic serum = dies in 4 days
L+ (0.29 c. c.) + 1/1,000 c. c. of antitoxic serum = dies in 2 days
22 Park and Williams. "Pathogenic Bacteria," p. 213.
458
INFECTION AND RESISTANCE
In the above tables, according to our previous definition of the
antitoxin unit, the serum would contain 900 units to the cubic centi-
meter, since 1/900 c. c., injected together with the L+ dose of the
standard toxin, resulted in the death of the guinea pig in four days.
In order to allow a margin of safety Rosenau and others have sug-
gested that the unit should be determined, not by the quantity of
antitoxin, which delays death by the L+ dose for four days, but
rather by the quantity which, with the L+ dose, results in saving
the life of the guinea pig. According to this latter standard the
serum employed in the table would be spoken of as containing 700
units to the cubic centimeter. Of course the tabulated measurements
fnMMMMT
BATTERY OF EOSENAU SYRINGES PREPARED FOR ANTITOXIN STANDARDIZATION.
(Taken from Eosenau, U. S. Hygienic Bulletin 21, 1905.)
are rough, leaving an undetermined zone of 100 units. The exact
number of units to the cubic centimeter could, of course, be deter-
mined with greater accuracy by now carrying out another series of
tests in which the amount of serum varied between 1/700 and 1/900
of a cubic centimeter.
In carrying out such a standardization the toxin is diluted so
that the L-j. dose is contained in 2 c. c. This can easily be done.
For instance, in the above the L+ dose being 0.29, it merely necessi-
tates adding to each 0.29 c. c. of toxin 1.71 c. c. of salt solution,
to each 2.9 c. c., 17.1 c. c. The antitoxin also is made up in such a
way that the required dilution is contained in two cubic centimeters,
since a total volume of 4 c. c. has been agreed upon as standard for
these tests, the injected volume having much influence upon the speed
of absorption. In using the so-called Rosenau syringe, shown in the
figure for the standardization of antitoxin, the antitoxin is made up
to 1 c. c. in each case, so that 1 c. c. of salt solution may be added to
wash out the syringe after injection of the mixture. The mixtures
THERAPEUTIC IMMUNIZATION IN MAN 459
can be made directly in these syringes or in test tubes, and are
allowed to stand one hour at room temperature, so that there
may be time for complete union. If the mixtures are made di-
rectly in the syringes the needles are dipped into sterile vaselin,
which closes them and prevents leakage while standing. The
mixture is then forced out of the syringe with a rubber bulb, thus
ensuring complete injection of all the fluid. As Rosenau states,
much depends on the guinea pigs. They must be of standard
weight, about 250 grammes, well fed and cared for, and must not
be descendants of pigs that have shown marked or unusual resist-
ance to diphtheria toxin. This, as Theobald Smith has shown,
occasionally happens.
The antitoxic serum as obtained from the horse directly may
be concentrated in a number of ways, representative of which is the
method developed at the New York Department of Health by Gib-
son,23 Banzhaff, and others.24 The original method consisted in
heating horse serum to 56° C. for 12 hours, by which some of the
pseudoglobulin was converted into euglobulin, the antitoxin remain-
ing in the pseudoglobulin fraction. After this an equal volume of
saturated ammonium sulphate solution is added and the globulin
precipitated. After several hours the precipitate is filtered off and
again taken up in water corresponding in amount to the original
volume of serum. After filtration this solution is precipitated with
ammonium sulphate and this precipitate is treated with saturated
solution of E"aCl in quantity twice that of the original serum. After
standing for 12 hours the supernatant fluid containing the antitoxin
is decanted, and this is precipitated with 0.25 per cent, acetic acid.
The resulting precipitate is dried by pressing it between filter papers
and is placed in a parchment dialyzing bag, after neutralization with
sodium carbonate. At the end of seven or more days of dialyzation
against running water, the globulin solution remaining in the dial-
yzer is filtered and made isotonic.
More recently the method as modified by Banzhaff is as follows :
The serum, as obtained from the horse, is diluted by one-half the
volume of water, and to this a saturated solution of ammonium
sulphate is added up to 30 per cent, saturation. This is heated to
61° C. for two hours. It is then filtered and the residue on the filter
paper, which contains the antitoxin, is thoroughly dried by pressing
between filter papers and is directly dialyzed.
Observations by Park and Throne 25 have shown that this con-
centrated antitoxin which, according to Gibson, represents a yield of
about TO per cent, original antitoxic power of the serum, is equally
efficient for therapeutic purposes as an unconcentrated preparation
23 Gibson. Journ. of Biol Chem., Vol. 1, 1906.
24 Gibson and Collins. Journ. of Biol. Chem., Vol. 3, 1907.
25 Park and Throne. Amer. Journ. of Medical Science, Vol. 132, 1906.
460 INFECTION AND RESISTANCE
and has the advantage of introducing less foreign protein into the
human body. It retains its potency, according to Park and Throne,
as long as does the whole serum.
ACTIVE IMMUNIZATION IN DIPHTHERIA WITH MIXTURE OF TOXIN
AND ANTITOXIN
Recently Behring26 has advocated the immunization of human
beings with mixtures of diphtheria toxin and antitoxin. This
method represents essentially active immunization with toxin ren-
dered harmless by neutralization with antitoxin. The use of such
mixtures had previously been studied with considerable care, in the
case of the toxin of symptomatic anthrax, by Schattenfroh and
Grassberger,27 and the procedure had been used in the New York
Department of Health for some years in the initial treatment of
antitoxin horses. Theoretically considered on the basis of Ehrlich's
opinions, one would be inclined to wonder at the fact that relatively
neutral mixtures of toxin and antitoxin should possess any antitoxin-
inciting properties. Behring explains the immunizing value of such
mixtures by the reversible nature of toxin-antitoxin union in the
animal body. He calls attention to the fact that our analyses of
diphtheria toxin-antitoxin mixtures have been made entirely with
guinea pigs as indicators. In studying such mixtures in other ani-
mals Behring has come to the conclusion that complete detoxication
of the poison in vitro does not occur. He found, for instance, that a
toxin-antitoxin mixture that wyas entirely innocuous for guinea pigs
produced an active febrile reaction in an ass. In monkeys (Maca-
cus rhesus) he finally found an animal in which he obtained evidence
satisfactory to him that toxin may be powerfully active in the animal
body, even if it has been previously mixed with antitoxin. If, for
instance, he gave a monkey a mixture in which as much as 20 to 40
antitoxin units were mixed with one toxin unit, and repeated the
injection two or three times, the animal died of subacute diphtheria
toxin poisoning. The mixture ceased to be poisonous for monkeys
only when the relation of antitoxin to toxin became one of 80 to 100
antitoxin units to one toxin unit. This final detoxication when suffi-
cient amounts of antitoxin were used, it seems to us, may be taken
as sufficient evidence that Behring' s monkeys did not die of ana-
phylaxis.
We gather from Behring' s writings that he attributes these dif-
ferences in susceptibility to toxin-antitoxin mixtures in various ani-
mals to differences in the reversibility of the toxin-antitoxin com-
plex in the bodies of the individual species.
26 Behring. Deutsche med. Woch., Vol. 39, No. 19, 1913.
27 Schattenfroh and Grassberger. Deuticke, Wien, 1904 ; see also Schat-
tenfroh, Wien. kl Woch., No. 39, Sept., 1913.
THERAPEUTIC IMMUNIZATION IN MAN 461
Human beings are less susceptible to such mixtures than are
monkeys, but nevertheless more so than guinea pigs. It also appears
that diphtheria bacillus carriers or such persons who, because of a
previous infection, have antitoxin in their blood are much more sus-
ceptible to these mixtures than are others. Newborn children are
less susceptible than are children from 4 to 15 years. Mixtures
which are entirely neutral for the newborn may incite febrile reac-
tion in older children. In all cases the injection of such mixtures is
followed by a more or less active production of antitoxin.
The mixtures which von Behring advocates at present are so
prepared that the toxin action upon guinea pigs is practically nil;
in other words, the mixture is completely neutralized.
The method represents in purpose, and apparently in achieve-
ment, a safe process of actively immunizing against diphtheria.
Heretofore the method of protecting human beings prophylactically
against diphtheria has consisted in the injection of antitoxic serum.
This, unquestionably a wise procedure, has nevertheless the disad-
vantage of bringing about an immunity of short duration only.
Within 20 to 30 days the antitoxin injected may have completely or
almost completely disappeared from the blood stream. Prophylactic
immunization with the toxin-antitoxin mixtures, however, repre-
senting as it does an active immunization, is likely to be more pro-
longed in its effects. According to Behring a human being possess-
ing 0.01 antitoxin unit in 1 c. c. of blood may be regarded as still
moderately protected against diphtheria. According to his estima-
tion a decline to this amount, in a person actively immunized by the
mixtures (an estimation based upon curve measurements of treated
cases), would take about two years. He has observed that horses
that had been actively immunized by him, and subsequently used in
agricultural work, retained measurable antitoxin values in their
blood after five years without treatment.
Schreiber28 and others state, also, that this method of active
immunization with mixtures of toxin and antitoxin has the advan-
tage of avoiding the anaphylactic dangers incident to the injection
of antitoxin alone. Their opinion is probably erroneous, since it is
most likely that whatever anaphylactic dangers there are result from
the injection of horse serum rather than from the antitoxin con-
tained in the injected substance. Moreover, the recent studies of
Park have shown satisfactorily that the danger of anaphylaxis in the
injection of antidiphtheritic sera is practically nil. Among 330,000
cases on record there were but five deaths.
The chief value of this new method of immunization is that it
represents a safe technique for the prophylactic treatment of indi-
viduals exposed to the disease and possibly for the general prophy-
lactic immunization of school children, nurses, physicians, etc. In
28 Schreiber. Deutsche med. Woch., Vol. 39, No. 20, 1913.
462 INFECTION AND RESISTANCE
the case of children during the ages at which they are most suscep-
tible to the disease, the prolonged immunity resulting from the treat-
ment should strongly recommend it as a method of promise for the
gradual eradication of epidemics. Behring also suggests it as a
hopeful method of treatment in the case of bacillus carriers.
Schreiber and others have reported upon the effects of treatment
when carried out with Behring's mixtures. In the earlier experi-
ments of Hahn, mixtures were used in which there was a slight excess
of toxin. The later experiments were made with mixtures which
were completely neutralized for guinea pigs. In Schreiber's cases
from two to six injections were made at intervals of three to five
days, most of them subcutaneously, and some of them intramuscu-
larly. In no case were there serious reactions, although occasionally
there were slight swelling of regional lymph nodes and a little fever.
The effects of immunization were noticeable about 23 to 25 days
later. When two injections only had been made, at least 0.075 of an
antitoxin unit to the cubic centimeter was present. The highest value
obtained after two injections was one unit to one cubic centimeter.
In nine patients who had been treated by four to seven injections
with gradually increasing doses, as much as 10 to 75 antitoxin units
to the cubic centimeter resulted. It appears, therefore, that in med-
ical practice this method is safe, and that with as little as two injec-
tions antitoxin values may be obtained which entirely suffice for the
protection of human beings against the ordinary dangers of diph-
theria infection, an immunity which, as far as we can judge at
present, may last about two years.
Another advantage which Behring claims for his method is the
production of homologous antitoxin in human beings for the passive
immunization of other human beings. Mathes has tried this in
children with the idea of thereby avoiding the dangers of anaphy-
laxis. Incidentally it was claimed in this case that the passive im-
munization, when carried out with homologous serum, lasted longer
than did that conferred by horse serum. However, one case is
hardly enough to establish such a fact.
THE INTRACUTANEOUS METHOD OF DETERMINING TOXIN AND
ANTITOXIN VALUES
Marks 29 was the first to utilize the prevention of local edema or
injury for the determination of antitoxin values. He mixed diph-
theria antitoxin and toxin and injected them subcutaneously into
guinea pigs, claiming that this method was considerably more deli-
cate than the Ehrlich method, since the amount of toxin capable of
causing localized edema amounted to as little as one-twentieth of a
29 Marks. Centralbl. f. Bakt., Orig. Vol. 36.
THERAPEUTIC IMMUNIZATION IN MAN 465
minimal lethal dose. This method has many points in its favor, and
has been recently utilized and improved upon by Romer.
Romer 30 31 32 has developed a method of diphtheria antitoxin
standardization which depends upon intracutaneous injections into
guinea pigs. The principle of this test consists in the observation
that, when very slight amounts of diphtheria toxin are injected intra-
cutaneously into the abdominal skin of guinea pigs, small areas of
local necrosis result within about 48 hours. When such injections
are made with mixtures of toxin and antitoxin the presence of free
toxin is indicated by the appearance of such necrosis.
Before proceeding to the standardization by this method it is
necessary to determine the "limes-necrosis" (just as Ehrlich deter-
mines his L+ dose), that is, the amount of toxin which, together
with a given amount of toxin (1/50, 1/200, or 1/2,000), will still
produce a minimal amount of necrosis after intracutaneous injection
into guinea pigs. It is necessary, therefore, arbitrarily to choose a
certain definite fraction of an antitoxin unit and mix this with vary-
ing amounts of toxin and inject the mixtures into guinea pigs intra-
cutaneously. Those mixtures in which the toxin is fully neutralized
will give rise to absolutely no lesion further than, possibly, a slight
local edema. Those in which there is a large excess of toxin will
cause extensive necrosis. Between the two, in the series, there will
be a mixture in which slight local necrosis results from the injection.
In this mixture the amount of toxin, just sufficient to cause notice-
able necrosis in spite of admixture with the antitoxin, contains the
L-n (limes necrosis) dose.
When this has been determined, then unknown antitoxin can be
similarly measured against this L-n dose of the standard toxin. The
method has the advantage of permitting one to work with very small
quantities, since only a small fraction of a cubic centimeter need be
used for intracutaneous injections; also it permits great economy of
animal material, since four or five tests can be simultaneously car-
ried out upon the abdominal wall of the same guinea pig.
The technique is not easy. We have found in studying this
method in connection with some work carried on in our laboratory by
Dr. M. C. Terry, that a considerable amount of practice and experi-
ence is necessary, both in carrying out the procedure accurately and
in judging the lesions. However, when carefully and consistently
done by an experienced worker, this method gives results which cor-
respond with fair accuracy to measurements made of the same anti-
toxin by the Ehrlich method. This has been the experience of
Lewin,33 and also of Terry in the few experiments carried out by him.
30 Romer. Zeitschr. f. 7mm., Vol. 3, p. 208, 1909.
31 Romer and Sarnes. Ibid., p. 344.
32 Romer and Somogyi. Ibid., p. 433.
33 Lewin. Centralbl. f. Bakt., Orig. Vol. 67, 1913.
464 INFECTION AND RESISTANCE
The Homer method has been recently used by clinicians for the
determination of the presence of free toxin or antitoxin in the circu-
lating blood of patients suffering or convalescent from diphtheria.
Homer himself suggested this, since his method is adapted to the
determination of extremely slight amounts of either substance. A
recent study by Harriehausen and Wirth 34 illustrates the results
obtained in such tests, formal human serum injected intracuta-
neously into guinea pigs never caused necrosis. Neither did the
similar injection of the sera of children suffering from varicella and
other diseases. Of twelve children suffering from diphtheria, how-
ever, serum taken before the administration of antitoxin caused
necrosis upon intracutaneous injection into guinea pigs, in every
case. In spite of the administration of antitoxin, toxin was demon-
strable in the blood in five cases as long as the 35th day. Of ten
cases of post-diphtheritic paralysis, toxin was demonstrated in the
blood of five.
Since this method of determining antitoxin values in the blood
of human beings is of considerable importance and may have much
practical value, it may be useful to insert an example of such an
application of this method as used by Hahn 35 in a series of investi-
gations mentioned elsewhere.
The standard toxin was obtained from Marburg. In a series
of guinea pigs a determination was made of the smallest quantity of
this standard poison which would produce just noticeable necrosis
of the skin if injected into the pig intracutaneously, together with
1/2, 000th of a unit of a standard antitoxin. The toxin and antitoxin
were left together for 24 hours before injection, 3 hours in the incu-
bator, and 21 hours in the refrigerator.
When this quantity of the antitoxin had been determined, it
could be used in similar experiments and similarly mixed with vary-
ing amounts of the patient's serum. The amount of antitoxin pres-
ent in such serum could then be easily computed. For, let us sup-
pose that this amount of toxin, together with 1/5 00th of a c. c. of
the serum injected intracutaneously into the guinea pig, gave the
same amount of necrosis in the same time as the identical quantity
of the toxin, together with 1/2, 000th of a standard unit. Then
l/500th of a patient's serum was equivalent to 1/2, 000th of a stand-
ard unit, and the patient's serum would contain 0.25 of a unit
per cubic centimeter.
Michiels and Schick 36 have carried out intracutaneous reactions
with diphtheria toxin directly upon the human body to determine
whether or not diphtheria immunity was present. They injected
0.1 c. c. of a 1 to 1,000 dilution of toxin and claim that a positive
34 Harriehausen and Wirth. Zeitschr. f. Kinderlieilkunde, Vol. 7, 1913.
35 Hahn. Deutsche med. Woch., Vol. 38, No. 29, 1912.
36 Michiels and Schick. Zeitschr. f. Kinderheilkunde, Vol. 5, 1912.
THERAPEUTIC IMMUNIZATION IN MAN 465
intracutaneous reaction with this amount indicates an absence of
antitoxin from the blood, or at least an insufficient protection. The
Schick reaction is at present carried out at the New York Depart-
ment of Health, under the direction of Park, with 1/5 Oth M L D
intracutaneously injected. The dilutions are so made that this
quantity is contained in a total volume of 0.1 c. c.
TETANUS ANTITOXIN AND ITS STANDARDIZATION
The methods employed in the production and standardization of
tetanus toxin are in every way analogous to those used in the case
of diphtheria antitoxin. A strong toxin is obtained by growing the
organisms under anaerobic conditions on suitable media. Accord-
ing to Vaillard and Vincent37 it is essential that the media upon
which the tetanus bacilli are grown should be freshly made and
sterilized. Apparently this precaution, which has been similarly
recommended by Wladimiroff, ISTovy, and others, is made necessary
by the gradual absorption of oxygen which takes place if the media
are allowed to stand for a long time without heating. It is further
necessary in preparing tetanus toxin that the culture medium should
not be acid, and a weakly alkaline initial titre is advised. For the
same reason, also, most workers have advised against the use of
glucose or other carbohydrates in the media, since the acid formed
by the fermentation of these substances inhibits growth and toxin
production. Recently Hall 38 has advised the use of a simple meat
extract broth to which have been added 1 per cent, of dextrose and
0.5 per cent, of finely powdered magnesium carbonate. The last-
named substance, by neutralizing any acid that is formed from the
glucose, prevents the harmful acidity. Anaerobic conditions are ob-
tained by growing the organisms under a layer of oil in tightly stop-
pered flasks.
Although mice were formerly used in the standardization of
tetanus toxin and antitoxin, the more recent usage has been to sub-
stitute guinea pigs as in diphtheria standardization. According to
the recent directions of Rosenau and Anderson 39 the purposes of the
standardization are carried out as follows:
The unit of antitoxin is arbitrarily designated as 10 times the
smallest amount of serum necessary to preserve the life of a guinea
pig weighing 350 grams for 96 hours, when given together with an
official test dose of toxin. The test dose of toxin contains 100 min-
imal lethal doses. And the minimal lethal dose is measured against
a 350-gram guinea pig.
37 Yaillard and Vincent. Ann. Past., 1891.
38 Hall. "Univ. of Cal., Publ. in Path.," Vol. 2, No. 11, 1913.
39 Rosenau and Anderson. U. S. P. H. Service Hyg. Lab. Bull. 43, 1908.
466
INFECTION AND RESISTANCE
In carrying out the standardization the L+ dose of 'the toxin is
used, but, unlike diphtheria standardization, in this case the L+
dose means an amount of toxin which will kill a guinea pig of 350
grams in four days, although united with 0.1 unit of antitoxin (it
must be noted that the L+ dose in this case is measured against one-
tenth unit of antitoxin rather than against 1 unit, as in the case of
diphtheria.
In determining the value of an unknown antitoxin, mixtures are
made, each containing the L_j_ dose of the toxin and varying quan-
tities of antitoxin. As in diphtheria measurements, the various in-
jection volumes are brought to 4 c. c. with salt solution, and are then
injected subcutaneously into guinea pigs of about 350 grams. The
table given below is taken from the Bulletin of Rosenau and Ander-
son.
Subcutaneous injection of
a mixture of
No. of
Weight of
guinea
guinea pig
Time of death
pig
(grams)
Toxin
(Test dose)
Antitoxin
(gram)
(C. C.)
1
360
0.0006
0.001
2 days 4 hours
2
350
0.0006
0.0015
4 days 1 hour
3
350
0.0006
0.002
Symptoms
4
360
0.0006
0.0025
Slight symptoms
5
350
0.0006
0.003
No symptoms
In this experiment 0.0015 equals 0.10 antitoxin unit.
ANTITOXINS AGAINST SNAKE POISONS
(Antivenin)
Antitoxins against snake poisons have been produced by a num-
ber of different workers, but the subject has been most extensively
studied by Calmette. As early as 1887 Sewall 40 succeeded in in-
creasing the resistance of pigeons to snake poison. Later Calmette
and Physalix and Bertrand independently succeeded in producing
immunity in rabbits and guinea pigs with the poison of the cobra.
The serum of animals treated with snake poisons gradually acquires
antitoxin properties, but the process of immunization is not a simple
one, and considerable time is needed for the immunizations.
Snake poisons, as we have seen, have attracted considerable atten-
tion because of their peculiarities in being antigenic and yet differ-
ing in heat resistance and a number of other properties from the
40 Sewall. Cited from Calmette.
THERAPEUTIC IMMUNIZATION IN MAN 467
bacterial toxins. It was with snake poisons that Calmette definitely
showed that the union of toxin and antitoxin is a true neutralization
and is not accompanied by the destruction of the toxin. These ex-
periments, as we have seen, were elaborated later by Morgenroth,
who succeeded in producing the snake poison HC1 combination. It
is these poisons also that have been the subject of extensive study by
Flexner and Noguchi, by Kyes, and later by von Dungern and Coca.
This work has been sufficiently discussed in other places and need
not occupy us here. The important poisonous snakes may be divided
into the colubridse, to which class the cobra belongs, and the viper-
.idse, which includes the ordinary European vipers, the rattlesnake,
and most of the poisonous snakes of North and South America. Ac-
cording to Calmette the poison of the cobra is much more heat-stable
than that of the rattlesnake. Pharmacologically the poisons of these
two main classes of snakes show considerable difference. In the case
of the cobra there is very little local disturbance and the systemic
symptoms dominate the clinical picture. Calmette describes the
cobra bite as being followed only by a feeling of stiffness at the site
of the bite, followed very soon by great general weakness, difficulty
in respiration, slow heart action, and finally death with unconscious-
ness. In the case of the vipers the local symptoms are very much
more marked, there being great pain and swelling and apparent clot-
ting of the blood about the point of the bite, with a rather slower
onset of systemic symptoms. In a description by Sparr 41 of a case
of bite by Russell's viper there was almost immediate swelling of the
limb with a faint bluish tint around the pin-point puncture, and
within 15 minutes great weakness, restlessness, and retching. In
spite of very active local treatment, within a short time after the
bite, the patient died within 24 hours of asphyxia and heart failure.
According to Calmette 0.0002 gm. of cobra poison will kill a
guinea pig; Noguchi states that 0.0005 gm. of rattlesnake venom will
kill a guinea pig of 250 gr. within 24-30 hours when injected intra-
peritoneally. The snake poisons apparently contain substances which
are especially active upon nerve cells (neurotoxins), and hemolysins
which act particularly upon the red blood cells. Flexner and
Noguchi 42 also speak of another poison which acts particularly upon
the endothelium of the blood vessels producing hemorrhages.
According to Calmette the antisera which are produced by im-
munization with cobra poison are most strongly potent against neuro-
toxic poisons of the colubridse and, to a certain extent, against some
of the poisons of the vipers. However, the action of the cobra anti-
toxin against viper poison seems at best to be weak. On the other
hand, antitoxins produced with rattlesnake poison are not potent
against the cobra venom since, as Calmette states, the rattlesnake
41 Sparr. Biochem. Bull., Dec., 1911, No. 2.
42 Flexner and Noguchi. Univ. of Pa. Bull., Vol. 15, 1902.
468 INFECTION AND RESISTANCE
poison contains hardly any neurotoxin. Antitoxins may be produced
by the gradual immunization of horses, and have been produced in
this way by Calmette in the Pasteur Institute of Lille for some
years. Calmette standardizes his antitoxin by determining the
amount of serum which completely neutralizes in vitro 0,0001 gm.
of the poison as tested upon white light. He also determines the
protective power by injecting a rabbit with 2 c. c. of the serum and
two hours later gives 1 gm. of the poison.
JSToguchi has studied rattlesnake poison particularly and suc-
ceeded in preparing a strong antitoxin by the gradual immunization
of a goat. Great difficulty has always been experienced in attempts
at immunization with rattlesnake poison because of the very violent
local injury produced by injections of the venom. The potency of
the serum produced by him was such that 2% c. c. of goat serum
protected guinea pigs against 12 times the fatal dose of rattlesnake
poison if given at the same time. If the antivenin was given one
hour later, 5 times the amount of serum had to be given.
PASSIVE IMMUNIZATION IN DISEASES CAUSED BY BAC-
TERIA WHICH DO NOT FORM SOLUBLE TOXINS
As we have stated in another place the greatest therapeutic suc-
cesses with passive immunization have been achieved in bacterial
diseases in which the malady is essentially a toxemia due to a soluble
toxin. In such cases the serum of actively immunized animals con-
tains specific antitoxins by virtue of which the toxins circulating in
the blood of the patient are directly neutralized, quantity for quan-
tity, with consequent therapeutic benefit. In the case of bacteria in
which no toxins are formed, the immunization of an animal is not
followed by the formation of any poison-neutralizing principle.
Here the injection of bacteria, dead or alive, or the invasion of the
bacteria in the course of spontaneous disease, is followed by the
formation of specific antibacterial substances, lytic, opsonic, agglu-
tinating, or precipitating bodies, the nature of which we have dis-
cussed in other chapters. The toxemia which occurs in such cases
is due as we have seen to derivatives of the bacterial protein which
by some observers are regarded as preformed endocellular poisons
liberated by the lytic action of the serum, and by others as split
products of the bacterial protein, non-existent until the bacterial cell
has been acted upon by the serum components and destroyed. How-
ever this may be, the recovery from diseases of this nature is accom-
plished by bacterial destruction ; this may be direct, by the bacteri-
cidal action of the serum, or indirect by opsonic properties which
induce phagocytosis. The poisons which are liberated from the bac-
terial bodies, if free, can do their injury, and no neutralizing sub-
THERAPEUTIC IMMUNIZATION IN MAN 469
stance is formed in the body fluids to prevent their action as far as
we know. Immunity in such cases, then, is not an antitoxic immu-
nity in any sense of the word ; it is rather an antibacterial immunity
in which the disease is prevented or cured only when complete de-
struction of the bacteria has taken place. If an animal or a human
being is prophylactically immunized against diseases of this kind
(typhoid, cholera, etc.), it is easy to see that an increased presence
of antibacterial substances, bactericidal or opsonic, in the circulation
would serve efficiently and rapidly in disposing of the small numbers
of invading micro-organisms which ordinarily enter the body in spon-
taneous infections. And, indeed, experience has shown that prophy-
lactic immunization can be successfully carried out in the case of
cholera, typhoid fever, plague, and other diseases which are suffi-
ciently prevalent endemically or epidemically to justify prophylaxis
on an extensive scale.
However, when in diseases of this kind the body is already exten-
sively infected and has begun, as is usually the case, to respond spon-
taneously with the formation of specific antibodies, it has been a
matter of doubt whether or not passive immunization, that is, the
introduction of specific antibodies in the form of the serum of a
highly immunized animal, is therapeutically of value. Indeed, it
has been feared that the use of such sera may even be harmful in
that the sudden introduction of large amounts of bactericidal sub-
stances might lead to a sudden liberation of large quantities of poi-
sonous products and consequent rapid toxemia.
The conditions in such cases are exceedingly complex and many
gaps exist in our knowledge concerning them. The bacteria when
invading the body, immediately enter into conflict with the protective
forces, as we have stated in the chapter on Infection. If a consider-
able degree of resistance exists, let us say as the result of preceding
immunization or a recent attack of the disease, there is a rapid
destruction of the bacteria, probably by active phagocytosis. It has
been shown by Bordet in the case of cholera and more recently by
Gay with typhoid, that injection of the organisms into immunized
animals is followed by prompt and high leukocytosis, whereas sim-
ilar injections into normal animals usually induce a temporary leuko-
penia. When the invaded animal is not particularly resistant the
bacteria may accumulate and, as in the case of pneuraococci and
streptococci, develop phagocytosis-resisting properties (capsule forma-
tion, etc.) ; or, as in the case of typhoid bacilli, there may be an
immediate liberation of toxic substances (anaphylatoxins) by reac-
tion between bacterial cell and blood plasma, which can induce leuko-
penia, and by this means the organisms may be protected from
phagocytic destruction. Experience with curative sera in all of the
conditions of this class has yielded promising results only when the
cases have been treated with the sera at early stages of the disease,
470 INFECTION AND RESISTANCE
either when the invading germ was still localized or, at least, wnen
the septicemic condition was not yet thoroughly established. It may
be that the doses heretofore given have been insufficient, and indeed
recent experiences with pneumonia seem to indicate that this may
have been, in part, the cause of earlier failures. Yet in pneumonia
the septicemia probably does not represent the firm establishment of a
foothold by the pneumococcus in the circulation but rather a con-
tinuous discharge of new organisms into the blood from the localized
lesion in the lung.
It is our own opinion moreover that septicemia as usually ob-
served clinically represents in most cases exactly this condition, that
is", a more or less continuous discharge of the bacteria into the blood
from some active focus with a continuous destruction of the organ-
isms after they have entered the blood stream. It is only when the
resistance of the body is overwhelmed, in the later stages of the
disease, that the bacteria can continue to grow and develop in the
circulation, and this stage probably does not occur until death is
imminent. In such septicemic diseases as streptococcus infection,
typhoid fever, plague, anthrax, and many others the presence of the
bacteria in the blood at the time when the patient is still in a condi-
tion of powerful resistance probably means that the bacteria are
being supplied to the blood from the local lesions. There is prob-
ably just such a continuous discharge of bacteria from the focus into
the blood with active destruction after the bacteria have entered the
circulation. This seems especially probable from the fact that in
many of these diseases the protective antibodies, bactericidal and
opsonic, can often be demonstrated in the blood serum in quantities
higher than normal at the very time when blood culture yields posi-
tive results. In typhoid fever, of course, it is well known that bac-
tericidal titres of over 1-50,000 are often present while the patient
may still be very sick, and in the more chronic streptococcus condi-
tions with malignant endocarditis we have often seen that opsonic
properties on the part of the patient's serum against the very organ-
ism invading him are considerably higher than normal. We take
this to mean that the injection of immune sera would simply aid in
more rapidly freeing the blood stream of the bacteria, the cure of'
the disease, however, involving a destruction of the focus. This, of
course, is not possible merely by the injection of the serum. When,
as in some cases of streptococcus infection, the focus can be surgically
reached, the septicemia will often disappear and cure result, as we
have ourselves had the opportunity to observe. When the focus
•cannot be reached surgically, it may nevertheless be a wise procedure
to inject considerable amounts of immune serum, for, by keeping the
Wood stream free of bacteria, the case may be influenced favorably.
Pneumonia is an example of this. Former failures have recently
been turned into partial success by the work of Neufeld and of Cole
THERAPEUTIC IMMUNIZATION IN MAN 471
merely by the use of larger quantities of immune sera essentially
similar to sera used at previous times, and Cole attributes the appar-
ently favorable results to the fact that the blood stream can be cleared
of bacteria although the focus cannot itself be affected.
Cure of such diseases, therefore, by serum treatment can hardly
be expected. Favorable influence of the disease by energetic serum
treatment may, however, be hoped for.
In discussing this subject it must not be forgotten, however, that
in most of the diseases which we have classified, on the basis of pre-
vailing opinions, as caused by bacteria that do not form true toxins,
the formation of such poisons has been claimed by a number of care-
ful and eminent observers. In the case of the typhoid bacillus, espe-
cially, Chantemesse, Kraus and Stenitzer, and others have claimed
the existence of a true toxin and a consequent antitoxin in immune
sera. Similar claims have been made for the cholera spirillum by
Kraus and Doerr, for the streptococcus by Marmorek, and for the
plague bacillus by Markl and Rowland. Since these claims have
been made on the basis of extensive experimentation by competent
men the question must be left open, and the possibility of antitoxic
properties on the part of the sera cannot be completely ignored.
Since in most cases, however, the poison-neutralizing properties of
the immune sera in this disease have not exceeded more than 1 to 2
multiples of the M L D of the bacterial poisons, it does not seem
impossible that the apparent antitoxic properties may have repre-
sented merely an acquired tolerance to anaphylatoxic poisons of
which we have spoken in another place.
SERUM TREATMENT IN EPIDEMIC CEREBROSPINAL, MENINGITIS
Serious attempts to produce curative sera -against the epidemic
form of cerebrospinal meningitis were not made until 1906 and
1907, when this disease appeared epidemically chiefly in Europe,
where it appeared most severely in Eastern Germany, and in the
Eastern United States.
In 1906 Kolle and Wassermann immunized three horses with
meningococci, using for immunization purposes the dead organisms
followed by living cultures and cultures shaken up in distilled water,
the so-called artificial aggressins of Wassermann and Citron. They
obtained sera of considerable potency when measured against menin-
gococcus cultures, and suggested standardizing the sera by comple-
ment fixation. They did not at this time treat human beings, but sug-
gested the use of the serum subcutaneously and intravenously in
meningitis cases. Very soon after the publication of the work of
Kolle and Wassermann Jochmann 43 also produced an antimeningo-
43 Jochmann. Deutsche med. Woch., 1906, Vol. 32, p. 788.
472 INFECTION AND RESISTANCE
coccus serum by immunizing horses with proved meningococcus cul-
tures, in his cases making a polyvalent serum by the use of many
different strains of the organism. The sera which he obtained were
highly agglutinating, somewhat bactericidal, and, according to him,
not antitoxic. He first succeeded in immunizing guinea pigs against
meningococci by injecting the serum 20 hours before infecting the
animals. He also treated 40 cases of meningitis in man and ob-
tained encouraging results in cases treated before the development of
hydrocephalus. Believing that possibly intraspinous injection of the
serum might offer advantages, he first determined by experiments
upon the dead body that the injection of methylene-blue intra-
spinously passed from the point of injection in the lumbar regions as
far up as the olfactory nerves. After having determined this he
treated 17 cases by tapping the spinal canal, taking out 30 to 50 c. c.
of spinal fluid and then injecting about 20 c. c. of the serum. Of
these 17 cases only 5 died, and Jochmann expresses himself opti-
mistically in consequence.
Meanwhile Flexner 44 had been working upon the same subject,
laying a rather more thorough basis for therapy in careful animal
experimentation. He produced the typical disease in monkeys by
intraspinous inoculation of the meningococci and then saved the
animals from death by following the infection with the injection of
serum intraspinously six hours later. In his earlier articles he ex-
presses himself with much conservatism, but his studies were con-
tinued and extensive opportunity for testing the serum which he then
produced, together with Jobling,45 was offered by the continuance of
the epidemic throughout the United States.
The results with the serum produced at the Rockefeller Institute
have since then proved to be uniformly favorable. The method of
intraspinous inoculation of the serum after the removal of some of
the spinal fluid was the method finally adopted by Flexner as most
favorable, and this is the method in current use to-day. In 1908
Flexner and Jobling reported upon 47 cases treated with the anti-
serum of which 34 recovered. Of 12 additional cases reported in an
addendum only 4 died. In the most recent summary by Flex-
ner 46 records of 1,294 cases that have been treated with the serum
prepared at the Rockefeller Institute are analyzed. Of this num-
ber, unselected and treated in many different parts of the world,
69.1 per cent, recovered. It is of course very difficult to obtain
exact comparative data on the efficiency of any method of treatment
in a disease as irregular in its clinical manifestations as epidemic
meningitis, especially since the mortality attending upon different
44 Flexner. J. Exp. Med., Vol. 9, 1907, and /. A. M. A., 1906, Vol. 47, p.
560.
45 Flexner and Jobling. J. Exp. Med., Vol. 10, 1908.
46 Flexner. J. Exp. Med., Vol. 17, 1913.
THERAPEUTIC IMMUNIZATION IN MAN 473
epidemics is subject to great variations. For this reason we can
draw conclusions only from a large statistical material. However,
we know that the average mortality of epidemic meningitis before
the introduction of specific therapy ranged certainly higher than 65
per cent., and in carefully studied epidemics usually between TO and
80 per cent. The statistics of Flexner showing a mortality hardly
exceeding 30 per cent, in unselected cases unquestionably marks a
wonderful therapeutic triumph. It must be remembered in consider-
ing the benefits of this serum that in unselected cases there must be
many in whom the disease has produced marked anatomical changes
in the central nervous system before the serum is used. It is well
known, of course, that the later manifestations of this disease, which
often lead to death with hydrocephalus, asthenia, and malnutrition,
are the remote results of the anatomical injuries produced by the in-
flammatory reactions accompanying the earlier manifestations of the
acute infection. These conditions of course cannot be expected to
yield to. serum treatment. It must be assumed, therefore, that were
we able to obtain statistics of cases diagnosed and treated soon after
the onset the figures would be even more favorable than those stated
above.
The action of the serum seems very largely to be an opsonic one,
in that, under the influences of serum, a powerful phagocytosis of
the meningococci takes place. It is also possible that to a certain
extent bactericidal action participates, in that the injection of the
serum into the closed space may give rise to a sort of intraspinous
Pfeiffer reaction with energetic ingestion of the bacteria by leu-
kocytes.
The standardization of the antimeningococcus serum has been
worked out particularly by Jobling.47 After attempting to stand-
ardize the sera by their protective power against meningococcus in-
fection in animals and by complement fixation, as suggested by
Kolle and Wassermann, Jobling has come to the conclusion that
neither of these methods is sufficiently regular, and that the most
suitable procedure is a standardization by opsonin determination.
The method as worked out by him depends upon determining the
highest dilution of the immune serum at which opsonic action
against the meningococcus is still evident. He suggests as a definite
and suitable standard of strength opsonic activity at a dilution of
1-5,000 of the antiserum.
SERUM TREATMENT IN STREPTOCOCCUS INFECTIONS
The attempts to produce powerful immune sera against strepto-
cocci date back to the earliest days of immunology. That the sub-
ject is a particularly difficult one follows from the great confusion
47 Jobling. /. Exp. Med.} Vol. 11, 1909.
474 INFECTION AND RESISTANCE
which has prevailed, and, to a great extent, still prevails, regarding
the classification of the streptococci and their interrelationship.
There are apparently a large number of different strains of strepto-
cocci which vary from each other, not only culturally, but also in
regard to agglutination and bactericidal reactions. For this reason
it is not at all a foregone conclusion that a serum prepared by the
immunization of an animal with a streptococcus of one type will
have any protective action against other strains. The subject has
been still more complicated recently by the discovery of Rosenow 48
that the various types of streptococci (viridans, hemolyticus, etc.)
are not constant in their properties, but may be artificially trans-
formed one into the other, and that even mutation of true pneumo-
cocci into true streptococci may take place. Most important in this
connection is the observation that a pneumococcus sent to Rosenow
was altered by him by special methods of cultivation in such a way
that not only its morphological and cultural properties were changed,
but also its agglutination reactions. These observations are of the
utmost importance in connection with attempts at producing specific
sera which can be utilized therapeutically. In all cases, therefore,
in which streptococcus immune serum is at all used it must be re-
membered that the disease produced in human beings by organisms
classified among the streptococci are by no means necessarily closely
related in biological reactions, and the same immune serum may be
extremely potent in one case and entirely useless in another.
That animals could be successfully immunized against strepto-
cocci was shown early in the history of investigations in immunity
by a number of workers, notably Roger, Behring, von Lingelsheim,
and Mironoff. The first extensive attempts to produce a curative
serum for use in passively immunizing human beings were made by
Marmorek 49 at the Pasteur Institute in 1895. The basic idea from
which Marmorek worked was the similarity of all the streptococci
producing disease in human beings. He also believed that the most
powerful serum could be produced with cultures whose virulence
had been greatly enhanced by animal passages. When such cultures
were grown on mixtures of human serum and broth he asserted
furthermore that soluble poisons were produced which could be ob-
tained by filtration of the culture fluids. For these reasons he im-
munized horses with cultures rendered highly virulent by very
gradual injections first of dead then of living organisms, finally
injecting also considerable quantities of culture filtrates.
Testing these sera upon animals, he was successful in protecting
against streptococcus infection when the serum was administered 12
to 18 hours before the bacteria were injected. He expressed the
opinion that the serum was antitoxic as well as antibacterial, In
48 Rosenow. Journ. A. M. A., Feb., 1914.
49 Marmorek. Ann. Past., Vol. 9, 1895.
THERAPEUTIC IMMUNIZATION IN MAN 475
his earliest reports the results of the treatment of 413 cases of ery-
sipelas leave one very much in doubt as to the value of the serum since
the difference in mortality between the treated and the untreated
cases is less than 2 per cent. However, an analysis of the individual
cases makes the serum treatment appear more favorable. He re-
ported good results also in 7 cases of puerperal septicemia and in
scarlatinal angina. Later observers, notably Lenhartz,50 Baginsky,51
and many others, have not been able to confirm the favorable results
reported by Marmorek, and it may be stated that at the present day
the value of Marmorek's serum is very much in question. Anti-
streptococcus sera have also been produced by Aronson 52 and Tavel,
Van de Velde, Menzer,53 Moser, 54 and some others. Aronson at
first worked from the idea which Marmorek also had used that there
was a close relationship between the various streptococci pathogenic
for man. He adopted the opinion first developed by Denys 55 and
Van de Velde that many different strains should be used for im-
munization in order to allow for possible difference in the character-
istics of the pathogenic streptococci. This principle of the necessity
for the production of polyvalent sera was also emphasized strongly by
Tavel, who based his opinion on careful agglutination tests, and by
Menzer and Moser.
That the action of the antistreptococcus sera, however produced,
is very largely due to its opsonic properties has been shown by
Bordet,56 by Meier and Michaelis, and a number of other workers.
If there is any bactericidal power it is probably relatively slight.
It would be quite impossible to attempt in this place to analyze
the large number of streptococcus infections of man which have been
treated with one or the other antistreptococcus sera. Those men-
tioned, moreover, do not by any means include all the sera which
have been produced and marketed for this purpose. In general we
may say that here, as well as in the cases of other sera in which no
antitoxic action is evident, beneficial results have been obtained
chiefly in cases in which the streptococcus infection has been localized
and treated early after its inception. In generalized or advanced
cases it cannot be said that the results are encouraging. Even in
animals, in which experimental conditions can be so much more
thoroughly controlled, the protective action of even the strongest
sera is evident only if the serum is administered either before in-
50 Lenhartz. "Die Septischen Erkrankunsren Holder," Wien, 1903.
51 Baginsky. Berl kl. Woch., 1896, p. 340.
52 Aronson. Berl. kl. Woch., Vol. 39, 1902; Deutsche med. Woch., 29,
1903.
53 Menzer. Berl kl. Woch., 1902, and Munch, med. Woch., 1903.
54 Moser. Wien. kl Woch., 1902.
55 Denys. Bull de I'Acad. Beige, 1896, cited from Schwoner K. and
L. H., Vol. 2.
56 Bordet. Ann. de Vlnst. Past., 1897.
476 INFECTION AND RESISTANCE
fection or within a very definite period after inoculation. The
standardization of streptococcus sera may be accomplished by de-
termining its protective value for animals when injected 18 to 20
hours before infection. When the sera are produced by immuniza-
tion with streptococci obtained from the human body and without
pathogenicity for animals the standardization is of course unsatis-
factory.
SERUM TREATMENT IN PNEUMONIA
Attempts to work out a therapeutically valid method of passive
immunization in pneumonia have been many and date from the very
beginning of the discovery that pneumonia was a bacterial infection.
Sera have even been marketed and used, but until recently no very
encouraging results were obtained. Recent studies have revealed
that in pneumonia the serum of convalescents contains practically no
bactericidal properties for the pneumococcus, and that the protective
powers of such serum depend upon the presence of immune opsonins
or bacteriotropins, by means of which the pneumococci are ren-
dered amenable to phagocytosis. Virulent pneumococci are not as a
rule phagocytable in the presence of normal serum. However, in
the presence of immune serum powerful phagocytic action can be
observed.
Neufeld has studied the conditions of pneumococcus immunity
most thoroughly in recent years. The most important advance from
a practical point of view was a discovery made by him, with Han-
del,57 in 1909. They determined that there was a definite difference
between various pneumococci in their reactions to immune serum ; in
other words, pneumococci could be grouped into various serological
types. The serum produced with organisms of one type did not pro-
tect against infection with other strains. In consequence they called
attention to the importance of determining the type of pneumococcus
which causes the individual pneumonia so. that the corresponding
immune serum might be used. They produced a highly potent anti-
pneumococcus serum by the injection of horses and donkeys with
highly virulent pneumococci grown on fluid cultures, then deter-
mined the high protective power of this serum upon animals and
used it in the treatment of patients by intravenous injection. Their
results were exceedingly encouraging. In reporting their results
Neufeld and -Handel state that considerable doses must be given.
They call attention to the fact, revealed by their animal experiments,
that moderate amounts do not, as in the case of diphtheria serum,
exert a correspondingly slight amount of beneficial action, but that
in the case of the pneumonia serum amounts smaller than a certain
57Neufeld and Handel. Zeitschr. f. 1mm., Vol. 3, 1909, and Arb. am
(Jem kais. Gesundh. Amt.3 Vol. 34, 1910.
THERAPEUTIC IMMUNIZATION IN MAN 477
active minimum seem to exert absolutely no beneficial action. This
is a fact which later was also determined by Dochez.
In confirmation of the work of Neufeld and Handel 58 Dochez
and Gillespie 59 have also been able to determine that there are at
least two distinctive groups of pneumococci which differ from each
other as far as agglutination and serum protection experiments are
concerned. In addition to these two fixed types they separate as a
third group the streptococcus or Pneumo coccus mucosus and a fourth
heterogeneous group which seems to fit in with none of the others as
far as serum reactions can determine.
Cole,60 therefore, adopts the reasoning of ISTeufeld in that he
advises the determination of the type of pneumococcus present in
cases of pneumonia as a guide to the type of antiserum to be used.
The type of organism is determined as soon as a case comes under
observation and 50 to 100 c. c. of the homologous antiserum is in-
jected intravenously. The result in the few cases so far treated by
Cole and his associates has been encouraging.
Protective substances, according to Dochez, appear in the serum
of treated cases very shortly after the administration of the serum,
whereas in untreated lobar pneumonia such protective substances
usually do not appear until after the crisis. Apparently Cole believes
that the great value of passive immunization of this kind in pneu-
monia lies in the fact that the bacteriemia shown to prevail in prob-
ably all cases of lobar pneumonia is either cured or improved by the
treatment, converting the disease, which is by nature, at least for a
time, a septicemia, into a localized pulmonary infection. Experi-
ence with antipneumococcus serum so far has been too limited to
warrant final judgment as to its permanent place among therapeutic
agencies.
THE SEBUM TREATMENT OF TYPHOID FEVER
The first extensive attempts to treat typhoid fever by passive im-
munization with the serum of treated animals were made by Chante-
messe, who immunized horses with filtrates of typhoid cultures sub-
cutaneously, and with emulsions of virulent bacilli intravenously.
Chantemesse believed that the serum of horses which had been
treated in this way for very long periods possessed, not only bacteri-
cidal action, but stimulated phagocytosis, and possessed a certain
limited amount of neutralizing power against the toxic properties of
the typhoid filtrates. At the International Congress of Hygiene in
Berlin in 1907 Chantemesse 61 reported upon a thousand cases
58 Neufeld and Handel. Arb. aus dem kais. Gesundh. Ami., Vol. 34, 1910.
59 Dochez and Gillespie. Jour. A. M. A., Vol. 61, p. 727, 1913.
60 Cole. Jour. A. M. A., Vol. 61, p. 663, 1913.
61 Chamtemesse. International Congress of Hygiene, Berl., Sept., 1907 ;
Ref. Bull, de I'Inst. Pasteur, Vol. 5, 1907, p. 931.
478 INFECTION AND RESISTANCE
treated with his serum. Of this number 43 only died, whereas the
average mortality during the same six years at the Paris hospitals
was 17 per cent. The injection of the serum he claimed very mark-
edly improved the condition of patients in that, after a preliminary
period of no apparent change lasting from several hours to 5 or 6
days, the temperature goes down and the general condition of the
patient changes considerably for the better. He noticed very few
complications in these cases, and intestinal hemorrhage occurred
four times only.
A remarkable feature of Chantemesse's treatment is that he in-
jected into the patients a few drops only of the serum, and rarely
made a second injection, facts which alone tend to persuade one that
his apparent therapeutic success was a fortunate accident.
The opinion originally expressed by Chantemesse that the serum
of horses vigorously treated with typhoid bacilli possesses in addition
to its bactericidal and opsonic powers definite antitoxic properties
recurs again in the work of a number of investigators. Besredka 62
prepared a serum by the intravenous injection of typhoid cultures
heated to 60° C., continuing the immunization for 6 months. He
claims that this serum possesses what he designates as "anti-endo-
toxic" properties. A dry extract of typhoid bacilli which in dose of
0.01 gram killed a guinea pig of 300 grams regularly became innocu-
ous when mixed with small quantities of this horse serum. One c. c.
of the horse serum neutralized often as much as two fatal doses of the
serum, but it is important theoretically to recognize that Besredka
states particularly that even an increase of the quantity of serum
never neutralized more than two fatal doses. This is particularly
important in connection with the more recent studies on toxic split
proteins by Vaughan, and on anaphylatoxins by Bessau and by Zins-
ser and Dwyer, in which it has been shown that an animal acquires
a tolerance against the toxic substances produced from bacterial and
other proteins which, however, never exceeds one or two multiples
of the minimum lethal dose. This fact alone would militate against
considering the serum of Besredka in any way antitoxic in the sense
in which the word is used concerning diphtheria and tetanus anti-
toxins where neutralization of poison follows roughly the law of
multiples. Besredka's anti-endotoxic sera has recently been very
thoroughly investigated by Pfeiffer and Bessau.63 These investi-
gators have found that Besredka's serum exerted a very definite
beneficial influence upon typhoid infection in guinea pigs if injected
at the same time with the bacilli. In their experiments it also pro-
tected somewhat against the toxic properties of substances derived
from the typhoid bacillus, and Pfeiffer and Bessau did not believe
that this was due to a true antitoxic action, nor that the serum was
62 Besredka. Ann. de I'Inst. Pasteur, 19, 1905, and 20, 1906.
63 Pfeiffer and Bessau. Centralbl f. Bakt., Vol. 56, 1910.
THERAPEUTIC IMMUNIZATION IN MAN 479
superior in this respect to the ordinary bactericidal sera prepared by
inoculating animals with typhoid bacilli. Kraus and Stenitzer 64
have also taken up the study of typhoid immunization from the
point of view that the typhoid bacillus produces a true toxin, and that
therefore a true antitoxic action could be expected from the sera pro-
duced by immunization with typhoid filtrates. It should be noted
that, in spite of the most common opinions against this at present,
a similar point of view was advanced by MacFadyen,65 and more
recently by Arima.66 Kraus and Stenitzer 67 immunized horses and
goats very highly with extracts of agar cultures and with broth fil-
trates by intravenous injection. The serum which they produced in
this way not only possessed the ordinary bactericidal action, but,
they claimed, neutralized also toxic broth filtrates, not only of the
typhoid, but of the paratyphoid bacilli. The serum of Kraus and
Stenitzer has been used by a number of observers, among whom are
Herz,68 Forssmann, linger, Russ, and others, and the results are
said to be encouraging in early cases.
Rodet and Lagrifoul69 immunized horses with living typhoid
cultures, and also claim favorable results.
Mathes,70 continuing the work of Gottstein after the death of the
latter, employed the method of immunizing with the product ob-
tained by digesting typhoid bacilli with trypsin. The poison so pro-
duced he speaks of as "fermotoxin." Liidke 71 slightly modified the
Gottstein-Mathes method by digesting the typhoid bacilli with pepsin
and hydrochloric acid, and with the poison so produced immunized 8
goats, reenforcing the immunization by the subsequent injection of
the bacilli themselves. He claims that 0.05 to 0.1 c. c. of the serum
so produced protected animals against five times the lethal dose
of the poison. In a small series of human cases treated by this
method he reports good results.
Garbat and Meyer 72 immunized animals with sensitized typhoid
bacilli, and claim that the most potent sera for typhoid immunization
can be obtained by the combination of sera produced by the injection
of sensitized and of unsensitized bacteria. They assert that the
typhoid bacillus contains two definite antigens, one particularly as-
64 Kraus and Stenitzer. Wien. kl. Woch., Vol. 20, 1907, pp. 344 and 753,
and Vol. 21, 1908, p. 645.
65 MacFadyen. Cited from Stenitzer in "Kraus und Levaditi Handbuch,"
Vol. 2.
66 Arima. CentralU. f. Bakt., 65, 1912, p. 183. Orig.
67 Kraus and Stenitzer. Wien. kl. Woch., Vol. 22, 1909, p. 1395; Deutsche
med. Woch., March, 1911.
68 Herz. Wien. kl. Woch., Vol. 22, 1909, p. 1746.
69 Rodet and Lagrifoul. C. R. de la Soc. de Biol., April, 1910.
70 Mathes. D. Archiv f. kl. Med., Vol. 95, 1909.
71 Liidke. D. Archiv f. kl. Med., 98, 1910.
72 Garbat and Meyer. Zeitschr. f. exp. Path. u. Ther., Vol. 8, 1911.
480 INFECTION AND RESISTANCE
sociated with the bacterial ectoplasm, which becomes active when the
bacteria enter the animal body, and a truly endocellular poison which
does not become active until the surrounding ectoplasm is dissolved.
They believe that sensitizing bacteria is a method for the production
of endotoxin, and think that therefore the ideal serum for the treat-
ment of typhoid consists of a mixture of two sera produced each
with one of the antigens, that is, with sensitized and unsensitized
bacteria. Rommel and Herman 73 did not obtain encouraging results
with this serum.
From a study of the literature it seems to us that in spite of the
many different methods of production employed by various observers
in their studies on typhoid sera it is quite likely that all these sera
are essentially alike, containing, quantitatively, according to the de-
gree of immunization, bactericidal, agglutinating, and opsonic prop-
erties, with possibly a limited amount of neutralizing power for the
poisons liberated from the typhoid bacilli in the body. As far as we
can judge from clinical reports the therapeutic value of the sera
so far produced is not very great. It seems that cases treated early
in the disease may be benefited, and possibly an early cessation of
the bacteriemia can in this way be attained. However, it does not
seem either theoretically or from the study of clinical publications
that any very marked effects have followed the use of any of the sera
in advanced cases.
THE SEBUM TREATMENT OF PLAGUE
That the serum of animals immunized with killed plague cultures
may actively protect normal animals from experimental infection
was first shown by Yersin, Calmette, and Borrel.74 The serum which
they produced possessed apparently powerful bactericidal action,
but no antitoxic properties were demonstrated. They determined
its protective powers by injecting measured quantities into mice and
infecting them with fatal doses of virulent plague bacilli 24 hours
later. The Yersin serum which was produced for the treatment of
plague as a result of these experiments was made, then, by the
gradual immunization of horses with first dead plague bacilli, finally
with virulent living organisms. The serum has been extensively
used by many observers with results that leave one much in doubt
as to its efficacy. Yersin 75 himself, reporting on an epidemic in
Nhatrang, reports a general mortality of 73 per cent, for the whole
epidemic, a mortality of 100 per cent, in untreated cases, and of 42
per cent, among those treated with his serum. Good results were
also reported from the epidemics in Amoy and Canton in 1896.
73 Rommel and Herman. Centralbl. f. Bakt. Ref. Vol. 53, 1912.
74 Yersin, Calmette, and Borrel. Ann. de I'Inst. Past., 1895.
76 Yersin. Ann. de I'Inst. Past., 1899.
THERAPEUTIC IMMUNIZATION IN MAN 481
However, these results apparently were not accepted by all observers
as proving the efficiency of the serum, since the number of cases
observed were few, and the irregularity in the gravity of the disease
in different individuals makes statistical evidence unreliable unless
large material can be studied. Kolle and Martini 76 announce that
Dr. Choksy reported very poor success with the Yersin serum, and
cite a number of later writers whose results with this serum were
also unsatisfactory when used on human beings. That the serum
unquestionably contains antibodies against the plague bacillus is
testified to, not only by the French observers themselves, but also by
the German Plague Commission of 1899, and by Kolle and Mar-
tini 77 themselves. The Commission experimented with this serum
upon monkeys, and showed that it possessed unquestionable protec-
tive powers in rodents and in monkeys when given 24 hours before
the plague infection, and in monkeys possessed fair curative prop-
erties when injected 24 hours later than inoculation with the plague
bacilli. Because of the doubtful success in the treatment of human
beings with this serum Yersin and Roux at the Pasteur Institute
later altered their methods of serum production by injecting, not
only dead and living plague cultures, but considerable quantities of
culture filtrates after the horses had attained a high degree of im-
munity. Later observations on the Yersin 78 serum have been pub-
lished by the British. Plague Commission in 1908 and 1911. In this
investigation the cases, were controlled as to their severity by blood
culture, since it had been claimed by a number of earlier investi-
gators that the Yersin serum was efficient in mild cases, but failed
entirely in the severe ones. It seems from the report of this Commis-
sion that ordinarily 70 per cent, of cases of plague without bacilli in
the blood survive while three-quarters of those with mild septicemia
die, and all of those with a marked septicemia succumb. In the
summary given of 146 cases treated with Yersin's serum by the
British Commission 65.1 per cent, died, whereas of 146 untreated
controls 71.90 per cent. died. These figures, together with an
analysis of the percentages, classified according to the severity of the
infections, do not show a very marked curative action on the part
of the serum.
Markl,79 who claims that the plague bacillus produces a soluble
toxin, has produced a plague serum by immunization of animals by
filtrates of broth cultures. He claims that 0.1 c. c. of his serum, as
produced at Vienna, will protect various animals against lethal doses
of plague bacilli if given at the same time. He attributes much of
76 Kolle and Martini. Deutsche med. Woch., 1902, p. 29.
77 German Plague Commission. Arb. a. d. kais. Amt., Vol. 16, 1899.
78 British Plague Commission. Journ. of Hyg., Vol. 12, Sup., 1912, p. 326.
79 Markl. Centralbl. f. Bakt., 24, 1898; Zeitschr. f. Hyg., 37, 1901;
Zeitschr. f. Hyg., 42, 1903.
482 INFECTION AND RESISTANCE
its curative action to the fact that in the presence of this serum active
phagocytosis takes place.
Dean,80 in 1906, also claimed to have produced strongly antitoxic
plague sera by treating horses with nitrates from 8 to 10 weeks old
bouillon cultures. He claims that 1 c. c. of his serum will neutralize
150 or 450 minimal lethal doses of the plague poison according to
whether one measures the M L D by death in 48 hours or in 4 days.
Rowland 81 also has produced a serum by the immunization of ani-
mals with the "toxins" produced by his sulphate process. Rowland
has apparently utilized the idea previously advanced by Lustig of
immunizing with "nucleoproteins" derived from the plague bacillus
instead of with the whole bacteria. Lustig's 82 method consisted of
washing up agar cultures of plague bacilli in 1 per cent, sodium
hydrate solution, precipitating with ascitic acid, taking up the pre-
cipitate in an indifferent fluid and injecting it into horses. The
serum produced by Lustig's method was used in Bombay, and is re-
ported by Hahn as effective in milder cases, but without action in
the more severe ones. There was but slight difference in the latter
type between the treated and the untreated cases.
Rowland's 83 method consisted in the treatment of the moist
bacteria with enough anhydrous sulphate of soda to combine with
all the water present, freezing and thawing the mixture and filtering
off the bacterial deposit at 37° C. Subsequently he extracted this
bacterial mass with water. The extract so obtained was fatal to
rats in quantities of 0.05 to 0.1 mg., killing them in 18 hours. In
his experiments doses of 0.001 to 0.01 afforded protection, the last-
named quantity reducing the mortality after inoculation of fatal
doses from 80 per cent, to 10 per cent.
The sera produced by the immunization of horses with these
supposed nucleoproteids are taken to be antitoxic in nature by Row-
land himself and by MacKonky. They were used in the treatment of
plague cases in the epidemics of 1908 and 1911 by the Maratha Hos-
pital in Bombay, and reported upon by the Indian Plague Commis-
sion on the basis of observations made by Dr. Choksy. The cases in
this series were controlled, as were those treated by the Yersin serum,
by blood culture. Here the results were not striking — 68.40 per
cent, of the serum-treated cases died, while 77.60 per cent, of the
controls died.
' Altogether we cannot draw any definite conclusions as to the
value of the serum treatment in plague. On the whole it does appear
80 Dean. Cited from MacKonky, Journ. of Hyg-, Vol. 12, Plague Suppl.
II, 1912, p. 402.
, 81 Rowland. Journ. of Hyg., Vol. 11, Plague Suppl. I, pp. 11-19.
82 Lustig. "Monograph Sierotrapia e Vaccin Prev. Control la Peste,*
Turin, 1899 ; cited from Kolle and Martini, loc. cit.
83 Rowland. Journ. of Hyg., Vol. 10, p. 536.
THERAPEUTIC IMMUNIZATION IN MAN 483
that the milder cases are materially benefited by the treatment, and it
is not at all impossible that in such cases aggravation of a milder
case into fatal septicemia may be prevented by the timely adminis-
tration of the plague serum. Animal experimentation also seems to
indicate that the administration of the serum may be of great value
as a prophylactic measure. It seems, on the other hand, as far as we
can judge from the evidence of statistics, that when a case of plague
has developed into the condition of active septicemia the administra-
tion of even the strongest plague sera at present available is of little
use. And this is indeed unfortunately true of all passive immuniza-
tion where the activity of the serum seems to depend chiefly upon
bactericidal and opsonic properties. For we cannot definitely accept
at the present day the claims that a true antitoxic serum, in the sense
of those produced against diphtheria and tetanus poisons, can be
really produced in the case of plague. The toxic substances derived
from plague bacilli by a number of observers do not correspond in
many particulars to true toxins.
FACTS CONCERNING ACTIVE PROPHYLACTIC IMMUNIZATION
IN MAN
In a previous chapter we have dealt with the treatment of in-
fectious disease with emulsions of dead bacteria or vaccines. The
discussion there was confined to the use of these substances in the
case of developed disease in which the infectious agent had already
gained a foothold in the body. Concerning this form of therapy
much difference of opinion exists, and we have seen that careful
judgment must be applied to the selection of cases to which treatment
with vaccines is adapted.
Concerning the prophylactic immunization of human beings with
bacteria there can be little difference of opinion ; this procedure finds
its justification in prolonged laboratory experience in the hands of
many men since the days of Pasteur.
The principle of specifically increasing the resistance of an in-
dividual by treatment with an altered form of the disease, either
with the attenuated bacteria, with dead bacteria, or with bacterial
extracts, has been sufficiently discussed in Chapter IV. It is indeed
surprising that this phenomenon of prophylactic protection, though
discovered by Jenner in small-pox, and developed by Pasteur in
rabies, did not find more general application to the diseases of man
until recejit years. At present such methods are in extensive use in
typhoid fever, in which they have had unquestionably excellent re-
sults. In the cases of cholera and plague numerous attempts have
been made, but the results here are not as clear-cut as they have been
in the case of typhoid. In the succeeding paragraphs we have set
484 INFECTION AND RESISTANCE
forth as briefly as possible the methods employed in prophylactically
immunizing man against this disease in which this procedure has
been most commonly attempted.
PBOPHYLACTIC IMMUNIZATION IN TYPHOID AND PARATYPHOID
FEVER
The first attempt to inoculate human beings with typhoid bacilli
with the intention of producing prophylactic active immunization
was probably that made by Pfeiffer and Kolle 84 in 1896. During
the same year also Wright 85 made similar studies in England, and
soon after this he reported upon the development of antibodies in the
blood of 17 people inoculated with typhoid. By these studies it was
shown that human beings could be inoculated with dead typhoid
bacilli without danger, and this logically led to the attempt to vac-
cinate human beings on a large scale.
It is hardly necessary to dwell upon the desirability of such a
procedure. From tables recently published by Russell 86 we take the
information that, in our own Spanish- American war, 20,738 cases
of typhoid with 1,580 deaths occurred in a total enlistment of 107,-
973. In this entire war 243 men were killed in action or died of
their wounds, while almost 7 times as many died of typhoid fever.
In the British army during the Boer war there were over 75,000
cases of typhoid in 380,000 men, and in the Eussian army during the
Russo-Japanese war over 17,000 cases of typhoid occurred, over half
as many as the number of men killed in action. Such appalling fig-
ures leave no possible doubt as to the desirability of prophylactic im-
munization in armies, and there can be little question that typhoid
fever is sufficiently prevalent in many parts of the civilized world to
encourage prophylactic immunization of individuals, even when not
living under the especially dangerous conditions of camps.
Following the preliminary studies of Pfeiffer and Kolle and of
Wright extensive practical studies of vaccination were made in the
German colonial army during the Herrero war, and by British
bacteriologists during the Boer war. Leishmann 87 also studied care-
fully the results of vaccination among regiments of the British army
in India.
The vaccine employed by Wright and his associates in England
•consisted of broth cultures of a typhoid bacillus killed by exposure to
53° C., and by the further addition of 0.4 per cent, of lysol. The
German vaccine consisted of emulsified agar cultures similarly killed.
The results obtained with these vaccines, although encouraging,
84 Pfeiffer and Kolle. Deutsche med. Woch., 22, 1896, p. 735.
85 Wright. Brit. Med. J., Jan., 1897, p. 256.
86 Russell. Amer. J. of Med. Sciences, Dec., 1913, Vol. 146.
87 Leishmann. Glasgow Med. Journ., 1912, Vol. 77, p. 408, cited from
Russell.
THERAPEUTIC IMMUNIZATION IN MAN
485
were not as striking as had been hoped. Russell 88 summarizes the
earlier attempts by stating that the morbidity was reduced to about
one half among vaccinated persons with a slightly greater reduction
of mortality. The last-named writer also attributes the early fail-
ures to the overheating of the vaccines with a consequent reduction
of their antigenic properties, and to timidity in their administration
resulting from Wright's fear of a negative phase. Russell, of the
United States Army Medical Corps, made a most extensive study of
typhoid vaccination in this country. After careful consideration of
the methods of others he produces his vaccines as follows : A single
strain of typhoid bacilli is used (culture Rawlings obtained from
England), and this is grown on agar in Kolle flasks for 18 hours.
The purity of the culture is tested out both morphologically and by
transplantation upon the double sugar media devised by Russell.
Agglutination tests are also made. After 18 hours the growth is
washed off in small quantities of salt solution and the emulsion heated
in a water bath for one hour at 53° C. ; it is then diluted with sterile
salt solution to a concentration of one billion bacteria to a cubic
centimeter. Then 0.25 per cent, of tricresol is added. Before use
the safety of the vaccine is ascertained both by aerobic and anaerobic
cultivation and by the injection into mice and guinea pigs of consid-
erable quantities for the exclusion of possible tetanus contamination.
The efficiency of the vaccine is then tested by injecting rabbits with
three doses at 10-day intervals, and determining the resulting ag-
glutinating power.
With these vaccines under the direction of the United States
Army Medical Corps the troops mobilized in Texas, California, and
along the Mexican border were treated. Compulsory vaccination was
established in March, 1911, and the results as reported by Russell
have fully justified the measure. The following table taken from
Russell's paper will illustrate the results obtained :
Typhoid Fever. Officers and Enlisted Men, United States Army
Totals
Yr.
Jan.
Feb.
Mar.
Apr.
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
for
9 months
Volun-
f 1908
5
6
4
2
3
11
14
31
25
26
12
8
101
tary
i 1909
4
10
6
4
11
15
26
14
16
45
20
6
106
T1910
8
11
1
4
2
6
12
27
21
16
20
11
92
Com-
I 1911
3
3
3
7
4
4
4
7
4
4
1
0
39
pulsory
1 1912
1
2
2
0
0
3
1
3
1
4
0
1
13
I 1913
0
0
0
0
0
0
0
0
0
0
0
0
0
Paratyphoid fever included in figures for 1908, but excluded in other years. Cases paratyphoid
1909, 3; 1910, 3; 1911, 2; 1912, 3; 1913, 0.
88 Russell. Am. Journ. of Med. Sc., Vol. 146, Dec., 1913.
486 INFECTION AND RESISTANCE
We have mentioned in another place that Metchnikoff and Bes-
redka in their studies on typhoid vaccination in the chimpanzee
have concluded that very little protective value resided in vaccina-
tion with dead typhoid vaccines, whereas animals vaccinated with
small amounts of living cultures were very efficiently protected.
Metchnikoff and Besredka adopted finally the method of immunizing
with living sensitized vaccines. By this is meant typhoid bacilli that
have been exposed to the action of heated immune serum, or, in
other words, typhoid bacilli that have absorbed specific antibodies.
There is no question as to the efficiency of this form of vaccination.
The method of employing sensitized bacteria for these purposes
utilized by Besredka in the case of plague has unquestionably won
an important place in active immunization. However, the results
of Russell and others seem to indicate that in human beings the use
of dead vaccines is certainly of considerable value, and there are
certain practical objections to the use of living vaccines in immuniza-
tion of large numbers of people as in armies to which Russell calls
attention. In the first place, living vaccines cannot be stored for
any considerable period, and may become a source of possible infec-
tion by mouth if carelessly handled; furthermore, contamination is
not so easily ruled out in the case of living vaccines when used on
a large scale, and it is not possible at present to require compulsory
vaccination with living bacteria. The choice of strains is important.
For vaccines in the British and American armies the Rawlings strains
usually used by Wright are employed at present. The work of Weiss,
Hooker and others recently has shown that perhaps we may have to
Ude for best results a polyvalent typhoid injection, and that in im-
munizing with paratyphoid B, a polyvalent vaccine will be necessary
eventually. The paratyphoid A strains are so uniform antigenically
that this will be unnecessary in their case.
Gay has recently recommended the use of sensitized killed vac-
cines. He controls the efficiency of his vaccines by testing them out
on rabbits in which typhoid septicemia has been produced by inocu-
lation with cultures grown on rabbit blood agar. These vaccines have
not yet been used upon sufficient numbers to justify conclusions. It
would seem, however, that any one of the methods mentioned must
possess considerable value, since they all represent merely slight
variations of the same procedure. The method at present used in
the German, British, and American armies, namely, vaccination
with dead cultures, seems certainly, according to Russell's statistical
studies, to have yielded excellent results and recommends itself by
its extreme simplicity and safety.
THERAPEUTIC VACCINE TREATMENT IN TYPHOID FEVER
A number of workers have attempted to treat typhoid fever after
it occurred with typhoid vaccine sensitized and unsensitized.
THERAPEUTIC IMMUNIZATION IN MAN 487
Frankel did this as early as 1893. He was followed by Petruschky 89
and better and in 1912 Ichikawa began the intravenous injection
of typhoid vaccine in patients with the rather astonishing results of
obtaining in many cases rapid falling of temperature and often ap-
parent shortening of the disease. Boinet in 1912 obtained similar
results and recently the same thing has been done by Gay.89a
Of these writers the only one who recognized immediately that
perhaps the intravenous injection of. such vaccines was not due to
purely specific activity was Ichikawa, who thought the paratyphoid
bacilli injected into typhoid cases often produced similar results,
and soon after this Kraus obtained similar degrees of temperature
and favorable results by injecting colon bacilli into typhoid patients,
into a few cases of pyocyaneus infection, and into streptococcus septi-
cemias.
With Jobling and others we have believed from the beginning
that at least an important factor in the activity of these vaccines
was non-specific, due perhaps to a rapid mobilization of leucocytes
and of ferments. This is discussed in another section on page 522.
At any rate the therapeutic results obtained justify the cautious
continuance of such intensive treatment in various diseases.
ACTIVE PROPHYLACTIC IMMUNIZATION IN CHOLERA
Attempts to protect human beings against cholera by prophylactic
vaccination were made as early as 1885 by Ferran,89b a pupil of
Pasteur. At the time at which Ferran's experiments were done little
was known regarding the production of immunity with killed cul-
tures or with bacterial extracts, and Ferran, under the influence of
the French school and its endeavors to immunize with living attenu-
ated organisms, applied similar methods to cholera. First experi-
menting with guinea pigs, he soon applied his method to human
beings, inoculating them with small quantities of living broth cul-
tures of cholera spirilla. In many of his experiments he gave, at the
first injection, 8 drops of a fresh broth culture, following this after
8 days with 0.5 c. c. of a similar culture. There is no reason why
Ferran's method should not have yielded excellent results. How-
ever, it is stated that he worked with impure cultures, and other
observers, notably Nikati and Rietsch, van Ermengen, da Lara, and
others, failed to obtain encouragement in their subsequent investiga-
tion of this method of vaccination.
The method which Haffkine 90 worked out some years after Fer-
89 Petruschky, Cent. f. Bakt., I, XIX, 1896.
89a Gay. Arch, of Int. Med., 1914.
89b Ferran. C. E. de VAcad. des Sc., 1885.
90Haffldne. The Lancet, February, 1893; Brit. Med. Journ., December,
1895.
488 INFECTION AND RESISTANCE
tan's experiments also depended upon the injection of living cul-
tures, but Haffkine attempted, by a rather elaborate technique, to
produce two separate vaccines, one attenuated, the other enhanced in
virulence. Attenuation was accomplished by growing the cholera
spirilla at a temperature of 39° C0 in broth over the surface of which
a constant stream of sterile air was passed. Under these conditions
the first crop of cholera organisms died rapidly, but Haffkine prac-
ticed reinoculation into new broth flasks before complete death of the
original culture had taken place; after a series of generations of
cultivation in this way he obtained cultures which produced merely
temporary and slight edema when injected under the skin of guinea
pigs. This weakened virus was used for the first inoculation.
He enhanced the virulence of cholera cultures with the purpose
of producing a strain of maximum potency comparable to virus fixe.
His procedure was as follows:
a. Giving an animal an intraperitoneal injection of cholera
spirilla larger than the fatal dose.
b. Taking out the peritoneal exudate and exposing it for a few
hours to the air.
c. Injecting this exudate into another animal and treating the
resulting peritoneal exudate in the same way.
After a series of such animal passages he claims to have obtained
a virus of great virulence, and this is his second and stronger vaccine.
In applying the method to human beings Haffkine planted the
cholera spirilla upon agar slants of the standard size, emulsified the
growths in sterile water, and injected 1/5 to 1/20 c. c. of such a
culture, using first the weak vaccine and five days later a more
virulent culture.
Beginning his work as early as 1893, Haffkine and others vac-
cinated as many as 40,000 people in India. On the whole, the results
obtained were very encouraging. It is a question, however, whether
or not his method is unnecessarily complicated. In the light of our
more recent knowledge concerning cholera immunity it seems likely
that the importance which Haffkine attached to the virulence of the
cholera culture used for injection was exaggerated, and we have
reason to believe that simple immunization with killed cultures may
produce results fully as efficacious. After all, we could not expect,
at least at present, to produce by active artificial immunization an
immunity as permanent as that which results from an attack of the
disease. Concerning the reasons for the acquisition of such perma-
nent immunity we have as yet little knowledge. Even Haffkine's
method of inoculation with living virus does not, by his own estima-
tion, last longer than possibly two years. It is therefore likely that
prophylactic immunization in cholera is efficacious by reason of the
appearance in the blood serum of the specific bactericidal and opsonic
substances by which the small numbers of cholera organisms enter-
THERAPEUTIC IMMUNIZATION IN MAN 489
ing during spontaneous infection can be disposed of before a foot-
hold in the body is gained.
Tamancheff later used Haffkine's method, but killed the cultures
by the addition of a 0.5 per cent, solution of carbolic acid.
Kolle 91 later recommended the injection of dead cholera or-
ganisms, maintaining that a single injection of about 2 milligrams
of a culture killed by exposure to 50° C. for a few minutes, and
by the addLl.n of 0.5 per cent, of phenol, is sufficient to immunize
successfully. Good results with Kolle's method have been reported
from Japan.
Strong,92 also proceeding from the idea that the immunizing
antigen is present, as such, within the cell body of the cholera
spirilla, recommends the injection of autolytic products obtained by
digesting cholera spirilla in aqueous suspension and filtering. He
prepared his "prophylactic" by growing the organisms upon agar,
then suspending the growth in sterile water and keeping it at 60° C.
for from one to twenty-four hours. The mixture was then exposed
to 37° C. for from two to five days and filtered through Reichel
filters. One to 5 c. c. of this was used in his experiments upon
human beings.
PKOPHYLACTIC IMMUNIZATION AGAINST PLAGUE
The first attempts to immunize human beings prophylactically
against plague were those of Haffkine.93 The first vaccinations were
carried out with broth cultures killed at 65° C. He tested out his
vaccines on a large scale in Bombay, and obtained apparently prom-
ising results. In a plague epidemic occurring in a Bombay prison
only 2 of 151 vaccinated persons became ill, and neither of these
died; whereas, of 177 unvaccinated persons 12 became ill and 6
died. In large series of vaccinated people only 1.8 per cent, were
infected with plague, with a mortality of 0.4 per cent, for the total,
whereas of unvaccinated individuals in the same epidemic 7.7 per
cent, fell victim to the disease, with a mortality of 4.7 per cent.
The German Plague Commission, consisting of Gaffky, Pfeiffer,
and Dieudonne, recommended a vaccine of killed agar cultures.
Kolle and Otto,94 basing their earlier results upon experiments car-
ried out with monkeys, mice, guinea pigs, and rats, have come to the
conclusion that vaccination with dead plague cultures is much in-
ferior to that obtained when attenuated living cultures are used.
The same conclusion has been reached by Kolle and Strong.95 Kolle
91 Kolle. Deutsche med. Woch., 1897, No. 1.
92 Strong. Journ. Inf. Dis., Vol. 2, 1905.
93 Haffkine. Bull, de I'Inst. Past., Vol. 4, 1906, No. 20, p. 825.
94 Kolle and Otto. Deutsche med. Woch., 1903, p. 493, and Zeitschr. f.
Hyg., Vol. 45, 1903.
95 Kolle and Strong. Deutsche med. Woch., XXXII, 1906, p. 413.
490 INFECTION AND RESISTANCE
and Otto found that the immunization of animals with large doses
of killed agar cultures of plague bacilli and with Haffkine's prophy-
lactic did not protect them against subsequent inoculation with viru-
lent cultures.
Strong 96 subsequently made a very careful comparative study of
the various methods of plague vaccination, and concluded that the
most efficient method is immunization with attenuated living cul-
tures. He showed that when carefully done this method can be
safely employed in human beings, but admits that his work must be
as yet considered as experimental and further studied before it can
be universally employed.
Besredka97 has advised the use of sensitized dead plague cul-
tures, claiming, from animal experimentation, that such vaccines
produce an efficient and relatively durable immunity.
Rowland98 confirms the immunizing properties of Besredka's
vaccines in plague, and believes that the antigenic properties of the
plague bacillus are attached to the bacterial nucleoproteins, and can
be extracted with these. Rowland prepares a vaccine by the treat-
ment of the moist bacteria with enough anhydrous sodium sulphate
to combine with all the water present, freezing and thawing the
mixtures, then filtering off the bacterial deposits at 37° C., and ex-
tracting them with water. The solution so obtained was fatal to
rats in small quantities and afforded substantial protection, reducing
the mortality on subsequent inoculation of a standard culture from
80 to 10 per cent.
PROPHYLAXIS AGAINST SMALL-POX
In the case of small-pox the general method of active prophylactic
immunization is in principle identical with that devised by Jenner
in the 18th century. The original observation from which Jenner
worked was that dairy maids and other individuals who had been
infected with cow pox were thereafter spared when a small-pox epi-
demic appeared in the region in which they lived. It is now agreed
by most observers who have studied the problem that the virus of cow
pox and that of small-pox are identical in nature ; the former repre-
senting a strain attenuated by passage through the animal body. This
is based chiefly upon the observation that true variola can be trans-
mitted to cattle, and that it can be thus carried from animal to ani-
mal, during this process becoming attenuated for human beings to
such a degree that reinoculated into man a simple vaccinia is pro-
duced."
96 Strong. Journ. of Med. Res., N. S., 13, 1908.
87 Besredka. Bull, de I'Inst. Past., Vol. 8, 1910.
98 Rowland. Journ. of Hyg., Vol. 12, 1912, p. 344.
99 Haecius. Cited from Paul, Kraus and Levaditi, Vol. 1, p. 593.
THERAPEUTIC IMMUNIZATION IN MAN 491
Small-pox, therefore, represents in principle active immunization
by means of attenuated virus. When vaccination was first introduced
the virus was taken from preceding pustules produced in other human
beings. This has been given up in most countries to-day largely be-
cause of the dangers of transferring syphilis and other diseases in
this way. At present the method of obtaining virus for vaccination
purposes is carried out as follows: The initial material consists of
what is known as "seed" virus, which can be obtained from spon-
taneous cow pox or from vaccination pustules in children, or again
from pustules obtained in calves after several passages of small-pox
virus through these calves. From such seed virus calves may be in-
oculated for vaccine production or else the calves may be inoculated
from the material obtained from other calves in the usual way.
Healthy young . animals are used; they are washed along the
abdomen, strapped down upon specially prepared tables, and the
abdominal skin thoroughly cleansed with soap and water. The
exact procedure varies in different places; often the skin is thor-
oughly cleansed with carbolic solution, and this is thoroughly re-
moved with sterile water before inoculation,' or else cleansing is re-
lied upon without the use of germicides. Over the clean area
longitudinal scratches 1 to 2 c. c. apart are made, and into these
the seed virus is rubbed. The animals are then kept in a clean
stall, preferably over asphalt floors, and rigid cleanliness is observed
during the period of development of the pustules. After the 6th or
7th day, when the vesicles are beginning to appear, the abdomen is
well washed and cleansed of superficial dirt without the use of an
antiseptic, and the pulp removed from the lesions with a curette.
The pulp so removed is placed into 60 per cent, glycerin and thor-
oughly ground up in a specially constructed mill. According to
Eosenau, the animal should always be killed before the vesicles are
removed, not only for humane reasons, since the same object might
be attained with anesthesia, but because a thorough autopsy can then
be performed to determine the health of the calf.
Vaccines so obtained always contain bacteria, the glycerin there-
fore serving a double purpose : one, the preservation of the virus, the
other a gradual destruction of the bacteria. Rosenau has shown that
the addition of 2 to 4 parts of 60 per cent, glycerin to one part
weight of the pulp prevents the growth of bacteria and probably
destroys them by dehydration. Most of the bacteria are destroyed
within one month at 20° C. During this period, then, from 4 to 6
weeks, the glycerinated virus should not be used, and should from
time to time be controlled by cultivation. At the end of this time
the lymph is ready for use.
Formerly the material for the vaccination of human beings was
obtained very simply by dipping ivory splinters into the fluid of pus-
tules, allowing this to dry, and rubbing these ivory or bone points
492 INFECTION AND RESISTANCE
into the exudate obtained by scratching the skin of the individual to
be vaccinated. This method has practically gone out of use, and
to-day the ripened glycerin pulp prepared as above is taken up in
small capillary glass tubes and from these blown upon the vaccina-
tion scratch. The efficiency of vaccine virus can be tested for po-
tency by the inoculation of the ears of rabbits before use.
ACTIVE PROPHYLACTIC IMMUNIZATION IN EABIES (HYDROPHOBIA)
Although many modifications have been suggested and actually
used in different parts of the world, the most common method of
immunizing against rabies still remains that originally devised by
Pasteur. The Pasteur treatment takes advantage of the prolonged
incubation period of rabies and is planned to confer immunity be-
tween the time of inoculation and the time at which the disease
would naturally appear. Since this period in ordinary street infec-
tion by dog bite is usually 40 days or more, a considerable interval
for active immunization is available. Formerly much of this time
was lost in that the diagnosis of hydrophobia in the dog or other
animal that had caused the injury could not be made with certainty
until the results of rabbit inoculations had been obtained. Nowadays
the ease with which a diagnosis of hydrophobia can be made within
a few minutes by finding negri bodies in the hippocampal and cere-
bellar cells has added considerably to safety in that it has made pos-
sible a gain of almost two weeks in determining whether treatment
should be instituted or not.
Here again, although the infectious agent of rabies is not known
with certainty even at the present day, the method of Pasteur de-
pends upon active immunization by means of an attenuated virus.
In standardizing the virus for the purpose of treatment Pasteur
first produced what he calls the "virus fixe." This consists of the
ordinary street virus as obtained from rabid animals passed through
a considerable series of rabbits (20-30) until its virulence for these
animals has reached a maximum. After a sufficient number of such
rabbit passages the incubation time after intracerebral inoculation is-
reduced to 7 or 8 days, but can no longer be shortened by further
passage. The brain and cord material of rabbits dead of rabies after
such repeated passages constitutes virus fixe. This can be preserved
•for considerable periods in 60 per cent, glycerin, and this is the
initial material from which the attenuated preparations for treat-
ment are produced.
In preparing the material for treatment a small amount of virus
fixe is injected subdurally into rabbits, about 0.2 c. c. of a salt solu-
tion emulsion being given. The inoculation is very easily made
through a small trephine opening in the skull, and contamination
is very easily avoided. Just before the rabbit dies when completely
THERAPEUTIC IMMUNIZATION IN MAN 493
paralyzed it is killed by chloroform and the cord is removed best by
the method of Oschida.100 The rabbit is nailed to a board, back up-
permost, and washed with a weak antiseptic ; a longitudinal incision
is then made along the backbone from the occiput to the lumbar
region, and the vertebral column laid bare.
After searing the tissues around the back of the head the spine is
cut across just behind the occiput, and again in the same way just
above the sacrum. The neck and lumbar regions are dissected loose
from the skin and gauze is inserted under it to avoid contamination.
The assistant grasps the end of the spinal cord as it appears in the
cervical region and pulls on it very slightly while the operator with
a glass rod or a piece of wire pushes against it from below. If this
is carefully done the spinal nerves are torn and the cord can be
gradually pulled out of place. This procedure is by far the best,
although it requires a certain amount of practice.
The cords so removed are hung up by a thread in bottles contain-
ing sticks of caustic potash and exposed in a dark place to 22° to 23°
C. Under these conditions of drying and temperature the virus is
gradually attenuated until at the end of 13 days or more the viru-
lence is practically nil. If removed from the drying bottles at any
time during the process and kept in a refrigerator in sterile glycerin
the virulence, whatever it may be at the time of placing into the
glycerin, remains fairly constant for long periods. When any of
this material is used for treatment little pieces of the cord % cm- in
length are cut off and emulsified in 2.5 c. c. of salt solution, and this
emulsion is used for injection.101
When patients are to be treated the principle of the treatment is
to inoculate them first with cords that have been dried for consider-
able periods, gradually proceeding toward those that have been dried
for less prolonged times and are therefore more virulent. The treat-
ment is varied in the individual case according to the severity of the
injury. Formerly treatment was begun with cords dried as long as
16 days. More recently it has been found that cords dried for longer
than 8 days are practically non-virulent and correspondingly lack
in antigenic value. They are no longer employed, therefore, since
their use is regarded as a waste of time. The tables on p. 494, taken
from Stimson's article in Bulletin 65 of the Hygienic Laboratory of
the U. S. Public Health Service, give the standard methods of treat-
ment as recommended by the United States Public Health Service.
This is the standard treatment used almost everywhere in the
100 Osehida. Centralbl f. Bakt., Vol. 29, 1901.
101 In our description of the methods of drying rabies, for the sake of
adhering to a standard, we follow closely the directions laid down by A. M.
Stimson, in the U. S. P. H. S. Bull 65, 1910. There are various modifica-
tions used in different countries, in many cases unimportant, and it seems
well to adhere to the U. S. regulations as a standard for this country.
494
INFECTION AND RESISTANCE
Scheme for Mild Treatment
Amount injected
Amount injected
T\- _„
Cord
Cord
uay
(injections)
Adult
5-10
1-5
Day
(injections)
Adult
5-10
1-5
(c. c.)
yrs.
(c. c.)
yrs.
(c. c.)
(c. c.)
yrs.
(c. c.)
yra.
(c. c.)
i
8-7-6 = 3
2.5
2.5
2.0
12
4=1
2.5
2.5
2.5
2
5-4 = 2
2.5
2.5
1.5
13
4 =
2.5
2.5
2.5
3
4-3 = 2
2.5
2.5
2.0
14
3 =
2.5
2.5
2.0
4
5 = 1
2.5
2.5
2.5
15
3 =
2.5
2.5
2.0
5
4 = 1
2.5
2.5
2.5
16
2 =
2.5
2.0
1.5
6
3 = 1
2.5
2.5
2.0
17
2 =
2.5
2.0
1.5
7
3 = 1
2.5
2.5
2.0
18
4 =
2.5
2.5
2.5
8
2=1
2.5
1.5
1.0
19
3 =
2.5
2.5
2.5
9
2 = 1
2.5
2.0
1.5
20
2 = 1
2.5
2.5
2.0
10
5 = 1
2.5
2.5
2.5
21
2=1
2.5
2.5
2.0
11
5 = 1
2.5
2.5
2.5
Scheme for Intensive Treatment
Amount injected
Amount injected
Cord
Cord
Day
(injections)
5-10
1-5
Day
(injections)
5-10
1-5
Adult
yrs.
yrs.
Adult
yrs.
yrs.
(c. c.)
(c. c.)
(c. c.)
(c. c.)
(c. c.)
(c.c.)
1
8-7-6 = 3
2.5
2.5
2.5
12
3 = 1
2.5
2.5
2.0
2
4-3 = 2
2.5
2.5
2.0
13
3 = 1
2.5
2.5
2.0
3
5-4=2
2.5
2.5
2.5
14
2=1
2.5
2.5
2.0
4
3 = 1
2.5
2.5
2.0
15
2 = 1
2.5
2.5
2.0
5
3 = 1
2.5
2.5
2.0
16
4 = 1
2.5
2.5
2.5
6
2 = 1
2.5
2.0
1.5
17
3 = 1
2.5
2.5
2.5
7
2 = 1
2.5
2.5
2.0
18
2 = 1
2.5
2.5
2.0
8
1 = 1
2.5
1.5
1.0
19
3 = 1
2.5
2.5
2.0
9
5=1
2.5
2.5
2.5
20
2=1
2.5
2.5
2.5
10
4=1
2.5
2.5
2.5
21
1 = 1
2.5
2.5
2.0
11
4 = 1
2.5
2.5
2.5
world at present. Other methods have been recommended. One of
these is that of Hogyes, in which virus fixe unattenuated is used in
dilution. Hogyes begins by injecting 3 c. c. of a 1 to 10,000 dilu-
tion of virus fixe, gradually proceeding within 14 days to 1 c. c.
of a 1 to 100 dilution.
Fixed virus attenuated by the addition of antirabic serum and
chemical disinfectants (carbolic acid) and by partial digestion in
gastric juice has also been used, but none of these methods has
attained widespread application.
THERAPEUTIC IMMUNIZATION IN MAN 495
INFECTION AND IMMUNITY IN POLIOMYELITIS
The active experimental investigation of poliomyelitis started
with the discovery by Landsteiner and Popper,102 that monkeys
could be inoculated with this disease by the intracerebral or the
intraperitoneal injection of saline emulsions of brain or spinal cord
of individuals dead of the disease. The same observations by Flex-
ner and Lewis 103 a little later, and then by a large number of work-
ers throughout the world, gave an opportunity for the careful study
of immunological conditions, and of the nature of the virus. It is
our own opinion that the disease is caused by the globoid bodies of
j^oguchi and Flexner 104 and not by streptococci. However, the work
on this controversy may be regarded as unfinished at the present
writing, and we may abstain from a prolongued discussion of the
etiological factor. The mere fact that the virus of poliomyelitis has
been found to keep as long as four to six years in glycerinated ner-
vous tissue, and the analogy which this offers to such virus as that
of rabies, small-pox, etc., make it alone seem likely that true bac-
teria are not concerned in the cause of the disease. Our own ex-
periments, and those of Dr. Tsen of this laboratory, on the isolation
of streptococci from poliomyelitis animals, incline us to think that
animals afflicted with this disease are readily subject to secondary
tissue infection with organisms of various kinds, but chiefly with
streptococci of the viridans type, which are so universally distributed
throughout the body.
It has long been suspected that one attack of poliomyelitis pro-
tected from subsequent infection. It is plain therefore that some
form of active immunization takes place in the course of the disease.
Experimentation on monkeys subsequently confirmed this, in that it
was shown that monkeys that had contracted the disease, and re-
covered, were thereafter resistant to inoculation. Strangely enough,
however, monkeys that have been unsuccessfully inoculated are just
as susceptible as they were before, showing that the immunity in
this disease is closely analogous to that existing in syphilis and some
other diseases where inoculation with attenuated virus, dead virus,
or sub-infectious doses of living virus, is entirely incapable of pro-
ducing immunity.
A brief survey of the epidemiology of the disease will be of value
because it throws some light on questions of immunity. An excel-
lent and extensive summary of this may be found in the Mono-
graph on the disease published by the New York Department of
102 Landsteiner and Popper. Zeit. f. Immunitats., 11, 1909.
103 Flexner and Lewis. Journ. Amer. Med. Assn., 1910, liv. 45.
104 Flexner and Noguchi. Jour. Exp. Med., 1913, XVIII, 461.
496 INFECTION AND RESISTANCE
Health, in 1917, on the basis of the work done during the great
epidemic of 1916.
Seasonably the disease has almost invariably been one of hot
weather. There are a few exceptions to this, one of them a great
epidemic in Sweden in 1911, which was at its height from October
to December. Some others have begun as early as March.
Formerly the disease was regarded as chiefly rural, but of late
years it has also extensively prevailed in cities.
The age incidence is important, in that the majority of cases
(about 90 per cent) have occurred in infants under ten, although a
few epidemics among adults have been observed and a few adult cases
always accompany epidemics. In other people the disease often
takes the form of what was formerly known as Landry 's paralysis.
Indeed, in a little epidemic which we studied in Palo Alto, Cal.,
the first cases which died were adults, with typical Landry symp-
toms; and successful monkey inoculation was made from a woman
of 26 whose symptoms were those of a typical Landry paralysis.
Relation to domestic animals has often been suggested, and
Frost (U. 8. Hygienic Laboratory Bulletin 90, 1916), who studied
this extensively during the last epidemic, made careful observations
on this factor, finding that there is no demonstrable relationship.
The diseases of dogs and chickens investigated by others, which
simulated the human disease, have been found to be due to distinct
pathological conditions without any relationship to human poliomye-
litis. A paralytic disease of cats occurring in New York during the
last poliomyelitis epidemic was studied in our laboratory by Dr.
Messing and found to have been caused by a Gram negative bacillus.
Because of the seasonable occurrence insects were long suspected
as possible carriers, and M. J. Rosenau a few years ago carried out
experiments in a few of which he succeeded in infecting monkeys
by the bites of stable flies (Stomoxys) which had been fed on the
virus. There has been but a single confirmation of this by Ander-
son, and all other attempts to confirm Rosenau' s observation have
been negative. Investigations of other flies, or bed-bugs, fleas, mos-
quitoes, etc., have invariably been negative. While therefore stable
fly transmission cannot be accepted as a common method of infec-
tion, and probably plays no practical role whatever, we must bear
in mind that there have been two successful experiments in the work
of two reliable and skilled experimenters, and should be alert for
further evidence in this direction.
Important positive knowledge concerning the transmission of
the disease is that which has been gained by the investigations of so-
called "carriers." It has been found both in monkeys and human
beings that the virus of the disease can be carried in the upper
respiratory passages of convalescents for as long as months after
recovery. It has also been found in a single instance that a monkey
THERAPEUTIC IMMUNIZATION IN MAN 497
could be successfully inoculated with dust from a room in which a
patient lay ; and Josef son states that infection may be acquired from
handkerchiefs and other recently handled articles for a short time.
Under these conditions it would seem rather likely that the greatest
factor in transmission would be found in direct and indirect trans-
mission by means of carriers, themselves immune or recently re-
covered, harboring the virus and transmitting it to susceptible in-
dividuals. And it is possible that many individuals in a community
may be immune, owing to mild forms of the disease sustained in
childhood which were not recognized as poliomyelitis; for evidence
is constantly accumulating that poliomyelitis is not essentially a
paralytic disease, that paralysis is merely one of the severe symp-
toms, and that mild but immunizing attacks may occur in the form
of very slight indispositions, attacks of mild grippe, headache, etc.,
which under present conditions are not recognized as poliomyelitis.
On this basis it would be comprehensible why so often one child of a
family is afflicted while others equally exposed are spared. And it
would become plain why epidemics sweep through a community se-
lecting cases irregularly here and there, and then die out, leaving
behind an endemic smoulder of sporadic cases. It would seem as
though for the time being all the susceptible children had been
afflicted and no further fuel for spread was available.
Under these conditions the study of contact infection becomes
very important, and Frost concludes that none of the other factors
usually considered in the spread of an infection, other than personal
contact, seems to offer adequate explanation for the spread of epi-
demics. To some extent discouraging as far as the contact theory
is concerned, the epidemic in New York showed that in 96 per cent of
the families afflicted, that is, in 8,287 families, there was one case
only to the family, although in these families there were a total of
24,883 children. In only 3.6 per cent were there two cases to the
family. Just why this should be it is difficult to surmise. However,
to some extent this and other radical phenomena in the transmission
of this disease may depend upon differences in susceptibility, due per-
haps to a light unrecognized attack previously sustained, or to some
anatomic or physiological predisposition to infection of the meningeal
choroid plexus. In this connection recent studies by Flexner and
Amoss 105 are of unusual interest. They find that among the
mechanisms that defend the central nervous system from infection is
this plexus, which, if in any way injured, becomes unable to function-
ate, and in consequence permits infection to pass. The injection of
normal monkey or horse serum of isotonic salt solution, or of Ringer's
and Locke's solutions, when injected into the meninges, though in
itself innocuous, causes a sufficient injury of the membrane to
render a subsequently injected lethal dose of virus infectious.
105 Flexner and Amoss. Jour. Exp. Med., 1914, XX, 249.
498 INFECTION AND RESISTANCE
Early in the study of poliomyelitis it was found that one at-
tack protected and that the serum of a human being or monkey
that had recovered from the disease was capable of neutralizing the
virus. In consequence extensive studies have been made by Flex-
ner, Amoss, and Chesney, Draper, Zingher106 and others in wlrich
the serum taken from recently recovered cases was injected intra-
spinously into children suffering from the disease for therapeutic
purposes. The work on this problem is not yet complete, but it
seems, especially from the studies of Amoss and Chesney,107 that
when injected early in the disease a very definite favorable influence
can be expected.
The serum should be taken from cases which have recovered as
recently as possible and collected with ordinary precautions of
asepsis. It should then be injected — best in quantities not less than
50 c.c., part of which is given intraspinously, and the rest intra-
venously, to neutralize any circulating virus.
IMMUNITY IN SYPHILIS
Until relatively recent years there seems to have been little ques-
tion in the mind of clinicians regarding the existence of true im-
munity in syphilis. It was stated uncompromisingly by Ricord 108
that "an individual who had once acquired syphilis was thereafter
protected against reinfection.7' This opinion acquired wide accep-
tance and was shared by most of his contemporaries. Baumler 109 in
1875 summarizing the authoritative opinions of that period stated
that "One who has once had small-pox, scarlet fever, typhus, etc.,
is, as a rule, not liable to these diseases again for the rest of his life.
The same is true of syphilis." However, even at the time Baumler
wrote this, exceptions to the supposed rule were accumulating and
he appends, to the positive statement given above, references to
observed instances of second infection reported by Bidenkap,110 H.
Lee,111 Diday,112 Kober,113 Zeissl,114 and others.
The conception of the existence of a true acquired immunity was
also expressed and widely accepted in the so-called "laws" of Colles
and Prof eta. The former, first enunciated by Beaumes and later,
106 Zingher. Jour. Amer. Med. Assn., March, 1917.
107 Amoss and Chesney. Jour. Exp. Med., April, 1917.
108 Ricord. Eecherches sur le Chancre, Paris, 1858, cited from Baumler.
109 Baumler. Ziemssen's Cyclop, of Pract. Med., Amer. Ed., Wm. Wood
and Co., N. Y., 1875, iii.
110 Bidenkap. Wien. med. Wchnschr., 1865, cited from Baumler,
111 Lee, H. Lecture on Syphilis, London, 1863.
112 Diday. Arch, gener., 1862, ii, cited from Baumler.
113 Kober. Berl. klin. Wchnschr., 1872, No. 46.
114 Zeissl. Ztschr. d. K. K. Gesell. d. Aertze in Wien., 1858.
THERAPEUTIC IMMUNIZATION IN MAN 499
in 1837, stated by Abraham Colles,115 of Dublin, is the well-known
generalization based on the observation that mothers who have borne
syphilitic infants were not infected by their children while suckling
them, although such children might often infect wet nurses. Pro-
f eta's 116 observation was the converse of this, namely, that children
born of mothers who suffered from active syphilis during the period
of conception did not acquire the disease from their mothers.
Both of these phenomena appear too well founded in clinical
observation to be questioned, at least as frequent occurrences. More-
over, considering the intimate contact between mother and infant
during the first months after birth, they acquire unusual importance.
However, as we shall see, they have been deprived of much of their
bearing as proofs of true acquired or inherited immunity by serologi-
cal investigations such as those of Bauer,117 Knopf elmacher,118 and
others, which have shown that mothers of syphilitic children usually
give positive Wassermann reactions, a fact which makes it seem
likely that such women are suffering from syphilis in a latent form
and are not immune in the ordinary sense. This indeed was the
interpretation given to the "laws" of Colles and Profeta by
Fournier 119 and by Matzenauer 12° at a time prior to that at which
serological data were available. It is still unclear why such mothers
should so frequently exhibit the disease in a latent form. However,
this is a matter for the intelligent discussion of which we are not
at the present time in the possession of sufficient information.
In the years just preceding the period of experimental investi-
gation of syphilis upon animals much purely clinical research was
carried out on the problem of reinoculation and what is commonly
known as "superinfection" of syphilitic human beings. Much of this
work is unavailable as scientific evidence owing to the difficulties of
distinguishing, at that time, between the true chancre and the
chancroid, but a considerable number of the observations then made
have been of much value in pointing out directions of later research
upon animals. The work has been so thoroughly analyzed both by
Neisser and by Levaditi that it would be needless repetition to do so
again. The records include both accidentally occurring superinfec-
tions, and purposeful experimental reinoculations. Without, there-
fore, going into details concerning the individual cases we may sum-
marize the conclusions justified from a study of these observations.
115 Colles. Cited from Osier & Churchman, "Syphilis," Osier's System of
Medicine.
116 Profeta. Cited from Osier and Churchmann, "Syphilis." Trattato
Practico delle Malattie Veneree, Palermo, 1888.
117 Bauer. Wien. klin. Wchnschr., 1908, No. 28.
118 Knopf elmacher and Lehndorff. Med. klin., 1909, No. 40; Wien. med,
Wchnschr., 1909, No. 38.
119 Fournier. L'Heredite Syphilitique, Paris, 1890.
120 Matzenauer. Vererbung d. Syphilis-Wien., 1905, cited from Bruck.
500 INFECTION AND RESISTANCE
1. The reports of Lasch,121 Jadassohn,122 Sabeareanu,123
Queyrat,124 Taylor/25 H. Lee, Knowles,126 and many others, have
shown that patients are susceptible to a second inoculation during
the first incubation time, that is, during the period elapsing between
the first infection with syphilis and the appearance of the chancre.
Second positive inoculations have also been successful at periods
shortly subsequent to the appearance of the primary sore.
Autoinoculations and reinoculations undertaken after the chancre
has become well developed have been, in the main, negative, though
Queyrat reports a case successfully inoculated daily up to the
eleventh, and Taylor one inoculated on the fourteenth day after the
appearance of the primary induration. In contrast to these and other
successful attempts, many observers record failure. However, the
point is not one of particular importance, since, after all, the ap-
pearance of the chancre does not mark off any fundamental change
in the progressive pathological development of the disease, and indi-
cates only the completed reaction at the point of entrance. The fact
remains that analysis has revealed that reinoculation, with the ap-
pearance of the second initial lesion, is possible up to about the twen-
tieth to the thirtieth day after the first infection with the trepone-
mata (Mauriac 127 and Sfeisser state the twenty-second day, Queyrat
cites a case eleven days, Linderman one twenty-four days after the
appearance of the chancre). It is claimed by some of the observers,
however, that even when reinoculation during this period is success-
ful, the incubation of the second and subsequent lesions is shorter,
that the induration itself is less severe, may not ulcerate, and heals
more readily than the first.
In judging of the success or failure of reinoculation practised
during the later days of the period above referred to, the possibility
must be borne in mind that the trauma produced at the inoculation
might have served to favor the development of a localized focus, the
treponemata at this time being very probably well distributed
throughout the body. Unsuccessful control inoculations with non-
syphilitic materials in Quey rat's cases would tend to eliminate this
possibility, whereas lesions resulting at the sites of such control
abrasions in the experiments of Neuman and Cehak,128 and of
121Lascli. Arch. f. Dermat. u. Syph., 1891, 61.
122 Jadassohn. Arch. f. Dermat. u. Syph., 1907, 86 ; Festschr. f. Neisser,
1907.
123 Sabeareanu. These de Paris, 1905, Chancres Syph. Successif.
124 Queyrat. Bull, de la Soc. Med. des. Hop., 1904, No. 28; 905.
125 Taylor. Jour. Cutan. Dis., Dec., 1890.
126 Knowles. New York Med. Jour., Dec., 1906.
127 Mauriac. Cited from Neisser, loc. cit.
128Neuman and Cehac. Wien. med. BL, 1890, cited from Neisser.
THERAPEUTIC IMMUNIZATION IN MAN 501
Levaditi,129 appear to support it. However, this is not of great im-
portance inasmuch as it determines merely a relative lengthening
or shortening of the period during which reinoculation or superin-
fection is possible.
2. After the disease is well established as a systemic infection,
that is, from the time of development of the chancre throughout the
so-called active "secondary" period, reinoculation is either impos-
sible or, at any rate, extremely difficult. Neisser cites Rollet as
follows :
"Although I and my predecessors have a thousand times attempted
to reinoculate luetic subjects, we have never observed a successful
case. I know no single fact more thoroughly proven than the insus-
ceptibility of a syphilitic to the action of a new virus, and, more-
over, these experiments are so harmless that they may be performed
without scruple."
The same opinion was held by Mauriac and is in a general way
assented to by Neisser in his summary of this place of his studies.
(Neisser, loc. cit., pp. 180-181.)
That the patient with well-developed lues has acquired a con-
siderable degree of resistance to fresh inoculations is pretty gener-
ally accepted, therefore, by all who have studied the question — but
there are investigators, notably Finger and Landsteiner, who believe
this resistance to be less absolute than stated by earlier writers.
Their observations on patients with active syphilis seem to indicate
that superinfection is possible "under certain circumstances in all
stages of the disease," but "the positive effect can be obtained only
with considerable quantities of the virus." Furthermore, Land-
steiner 13° states that, in general, lesions so obtained are relatively
slight in severity, do not appear as primary indurations, but have
the tendency to simulate the particular variety of lesion spontane-
ously manifest in the individual at the time. Were it not for Finger
and Landsteiner 7s monkey experiments, to which we will refer di-
rectly, and which seem to bear them out in their interpretation of
these inoculations as "positive" results, the simulation of the spon-
taneously occurring lesions by the inoculation-products would again
justify suspicion that these experimental results represented merely
traumata in which, as points of less resistance, the patient's pre-
existent disease had found a favorable spot for localization.
Observations in this respect, also, are corroborated by the older
literature. A report which has direct bearing on this point, is one
oif Queyrat and Pinard who inoculated a tertiary patient with chan-
cre material, obtaining not a primary sore but an ulcerated lesion
having the clinical characteristics typical of the late skin manifesta-
129Levaditi. De I'lmmunite acquise dans la Syphilis. Ztschr. f. Imm.
Ref., 191, 277-318.
130 Landsteiner. Centralbl /. Bakteriol, Ref. 190S. xli, 785.
502 INFECTION AND RESISTANCE
tions of the disease. The autoinoculation experiments of Ehr-
mann 131 also would tend to show that the resistance of the luetic
subject is a relative one only. Ehrmann succeeded in producing
positive autoinoculation-products in 45 syphilitics with papular erup-
tions. Control inoculations with sterile water were negative. Al-
though his observations and those of Finger and Landsteiner, with
other similar ones, teach us that superinfection during the disease
is possible the nature of the lesions, their short incubation time, and
their exceptional character when averaged with the total of such at-
tempts, prevent them from invalidating the conclusion that there is
a resistance at this stage higher than that of the normal subject.
We may conclude, therefore, concerning the secondary period of
the disease, that the luetic individual has acquired a resistance which
while not absolute is at least very high, and protects him from fresh
external inoculation, although at the time his disease may still be
progressing and in no sense overcome.
3. During the late stage of syphilis, the stage at which, according
to the more or less arbitrary divisions of Ricord, we are accustomed
to speak of it as "tertiary," the resistance is still manifest, though
apparently not so regularly potent as during the preceding "secon-
dary" period. JSTeisser expresses himself with great caution and
accepts few, if any, of the observed reinoculations of tertiary cases
as surely representing unquestionable freshly acquired infections.
Nevertheless, taking into consideration a detailed study of individual
reports, he concludes that resistance during the late stages is pro-
nounced but already beginning to wane. He himself in the Fest-
schrift to F. Joseph Pick in 1898 cites a case of the development of
a chancre in a tertiary case which, however, was not followed by
constitutional symptoms.
r- 4. In the preceding paragraphs we have concerned ourselves
entirely with questions of "superinfection," that is, the implantation
of the syphilitic virus into subjects still suffering from manifesta-
tions of the disease. A problem of equal theoretical, and of much
greater practical importance, is that dealing with true "reinfection."
.By "immunity" in the ordinary sense, we mean an increased resist-
ance to specific infection which persists for a more or less prolonged
period after the active disease has been overcome and the causative
agents removed from the body. It is not easy to draw conclusions
-concerning this point from observations on human beings, since it
is most difficult to decide, even with the aid of serological methods,
whether or not a given case is cured in the bacteriological sense;
for syphilis is preeminently the disease in which there occur fre-
quent and prolonged latent periods terminated, often after lapses of
years, by the reappearance of foci of a grave nature.
Moreover, our own experiments both with syphilis in rabbits and
131 Ehrmann. Cited from Neisser, loc. cit.
THERAPEUTIC IMMUNIZATION IN MAN 503
the treponemata in culture have convinced us that these micro-
organisms may assume for long periods a condition of metabolic
latency, a sort of resting period, during which they incite no reactions
of any kind on the part of the tissues, apparently do not multiply
to any great extent and yet remain alive and capable of development
when conditions favor this.
In spite of these difficulties, however, careful clinical studies by
Jonathan Hutchinson, Taylor, Hudelo,132 Neisser, and many others,
have furnished data which warrant an opinion upon this problem.
Of especial value is the painstaking analysis of reported cases of re-
infection in syphilis made in 1909 by Felix John.133
In contrast with some others who have attempted similar analyses
John takes the utmost care to separate these cases into the ones in
which the evidence of true reinfection is absolute, and those in which
the reported details are insufficient to exclude the possibility of
recrudescence. In agreement with Taylor he insists upon a symptom-
free interval of five years between the last manifestations of the
first attack and the appearance of the second. As an example of
what he calls an "Ideal Fall" we may cite the following:
X. J., April 1, 1872, primary sore.
Roseola, polyadenitis, mucous patches. September, 1872, papular rash.
March, 1873, palmar and plantar spots, iritis. 1875, gumma of tibia and
serpiginous syphilide of right thigh.
Four courses of inunctions and KI.
1876, married — two healthy children.
No symptoms till 1887 when he acquired a second chancre followed by
typical roseola. In 1888 wife had still-birth.
John has analyzed in this way 356 cases of supposed reinfection,
in 34 of which the first attack was of congenital origin. Of the
remaining 322, fourteen were cases which seemed unquestionable
instances of reinfection and in 16 more there was practically no
doubt of this. Of the 34 hereditary cases there were three, one of
Emery, one of Taylor, and one of Hochsinger, in which there was
practically no question of their nature as valid reinfections. In all
of the others there was one point or another which rendered them
doubtful as evidence.
John concludes that true reinfection can unquestionably occur
in syphilis but that it is relatively rare.
To John's cases Neisser has added others reported between 1909
and 1911. Yet even with these, the total number is not a large one.
Nevertheless, we should not be tempted to conclude from this that
the relative infrequency of such cases is evidence in favor of the
132 Hudelo. Ann. de dermat. et de syph., 1890, 353.
133 John, Felix. Eeinfectio Syphilitica, Samml. klin. Vortr., Volkmann,
1907-1909, 559 et seq.
504 INFECTION AND RESISTANCE
existence of a true immunity analogous to that following typhoid,
plague, etc. For we have seen that a very definite insusceptibility
is coincident with the persistence of actual disease in the subject
and, as Neisser points out, the more recent investigations carried out
with the aid of serological tests have shown that the number of cases
uncured though long without symptoms, is much larger than for-
merly supposed. The scarcity of true reinfection, therefore, may
well be due to the relative scarcity of completely cured cases. More-
over, it must be remembered that, in this disease, even when final
recovery results, it is usually achieved only at an age when the in-
dividual is less exposed to reinfection because of changed economic
and domestic conditions, or by reason of the virtue which comes with
arteriosclerosis and the wisdom of the burnt child that fears the fire.
Granted, then, that true reinfection is possible, is there any evi-
dence that when it does occur, the second attack is less severe and
more easily cured than the first, a fact which would also tend to
support the opinion that a certain degree of immunity persists.
Jonathan Hutchinson 134 expressed this view in a clinical lecture in
which he states that "second chancres are far more common than
second attacks of constitutional syphilis." However, John in his
summary, in which Hutchinson' s cases are included, finds that in
general the disease has run a second course very similar as to severity
to that incident to the first infection.
If we gather together, then, the facts revealed by clinical study we
may conclude with Levaditi, Neisser, and most others, that :
1. The syphilitic subject acquires definite resistance to rein-
oculation which becomes manifest soon after the appearance of the
primary sore, at a time when the virus may be regarded as having
gained universal systemic distribution.
2. This resistance, high though not absolute, persists through-
out the secondary or most active period of the disease and into the
tertiary stage. During the latter, however, it appears somewhat to
decline, reinoculation or superinfection being more frequently pos-
sible at this period.
3. When syphilis is entirely cured, susceptibility may in all
probability be regarded as returning, possibly, though not certainly,
to the same degree as it exists in the normal subject. The reasons
for this last belief will become more clear when we study the evidence
contributed by animal experimentation.
Notwithstanding the admirable thoroughness with which clinical
data had been collected and analyzed in the study of syphilis, prog-
ress beyond the points indicated in the preceding sections was quite
impossible without the aid of animal experimentation and a knowl-
edge of the causative agent. Fortunately these two deficiencies in
our methods were removed when in 1903 Metchnikoff and Roux
134 Hutchinson, J. Brit. Med. Jour., 1882, i, 699.
THERAPEUTIC IMMUNIZATION IN MAN 505
succeeded in transmitting the disease to a chimpanzee, and in March,
1905, Schaudinn135 discovered the treponema pallidum.
As a matter of fact, probable transmission of syphilis to lower
monkeys had been accomplished as early as 1879 by Klebs 136 and
subsequently by Neumann,137 Martineau,138 and Charles Nicolle,139
but in none of these experiments had it been possible to prove beyond
question the syphilitic nature of the inoculation-products. In the
chimpanzees inoculated by Metchnikoff and Roux, the animals de-
veloped not only primary sores but also secondary eruptions, poly-
adenitis, and enlarged spleens in such characteristic manner that the
identity of the inoculation disease with human syphilis could no
longer be doubted.
Successful transmission to other anthropoids and to lower mon-
keys were then announced, in rapid succession, by Metchnikoff and
Roux, Ch. Nicolle, Neisser, Baermann and Halberstaedter, Einger
and Landsteiner, jloffman, and others. The susceptibility of mon-
keys was tabulated by Neisser in the following series : Chimpanzee,
Gibbon, Orang-Outang, Cynocephalus babuin, Cynocephalus sphinx,
Cynocephalus hamadryas, Cercopithecus fulginosus, Macacus niger,
M. nemestrinus, M. cynomolgus, M. sinicus, M. speciosus, M. rhesus.
In 1906 Bertarelli 14° produced syphilitic keratitis in rabbits, dem-
onstrating the treponema pallidum in sections of the cornea and in
1907 Parodi 141 first produced syphilitic orchitis in the same animals.
Apart from monkeys and rabbits, no animal species have so far
been shown sufficiently susceptible to be available for systematic
study. It is true that the production of keratitis is claimed in dogs
and sheep (Bertarelli, Hoffman and Briinig), in guinea pigs (Ber-
tarelli), in cats (Levaditi and Yamanouchi), and in goats (Ber-
tarelli). However, these experiments have been isolated and too
uncertain with present methods to offer material for experimentation.
Our own attempts on cats, pigs, guinea pigs,- rats, mice, and a few
birds, have yielded negative results only. There are many observa-
tions of great scientific interest which might be discussed in con-
nection with the problem of susceptibility of various animal species,
however, we will confine ourselves at present to those phases of the
work only which have bearing on the questions of immunity.
In the fundamental premises, the work on monkeys has pretty
accurately confirmed the observations concerning reinoculation and
superinfection previously made on human beings.
135 Schaudinn. Deutsch. med. Wchnschr., 1905, No. 42; Arb. a. d. k.
Gsndhtsamte., 1907, xxvi.
136 Klebs. Arch. f. exper. Path. u. Pharmakol., 1879.
137 Neumann. Cited from Miihlens, loc. cit.
138 Martineau. Arch. f. Dermat. u. Syph., 1884, No. 16.
139 Nicolle, Ch. Cited from Miihlens, loc. cit.
140 Bertarelli. Centralbl. f. Bakteriol, Orig. xli, 1906, and xliii, 1907.
141 Parodi. Centralbl. f. Bakteriol., Orig. xliv, 1907.
506 INFECTION AND RESISTANCE
The most extensive studies in this respect are those of Neisser 142
and his associates in the Javan expedition. Briefly stated, Neisser
reinoculated 135 animals 165 times with negative result 'on the second
and subsequent inoculations. The second inoculations were made at
periods ranging from 21 days to two years after the first. In but 27
animals did the reinoculation show positive results, and in ten of
these cases only does Neisser recognize the experiments as valid.
Although these ten positive reinoculations, it is true, add an element
of irregularity to the series, they constitute but 6.8 per cent, of the
entire number, a proportion which in no way invalidates the experi-
ments when we consider that the work was done entirely on the
lower monkeys, animals that are far less susceptible to syphilis than
are human beings and in many of which, therefore, systemic dis-
tribution of the virus (a generalization apparently necessary for the
development of resistance) may not have taken place. Neisser's
conclusions, therefore, that monkeys, like human beings, are not
reinoculable while suffering from systemic syphilis, seem entirely
justified.
His work, as well as that of Finger and Landsteiner, and of
Kraus and Volk143 on lower monkeys, has shown that resistance
does not develop until the twelfth to the twentieth or twenty-first day
after the first inoculation; that is, again as in man, when the virus
has become generally distributed. Finger and Landsteiner, further-
more, noticed that reinoculation-products, obtained by reinfection
during the first incubation period, that is, before the development of
the primary lesion, were less severe and developed in a shorter time
than did the first lesion. This phenomenon which would tend to
mark another analogy to the conditions prevailing in human beings,
was not observed in the experiments of Neisser and of Kraus and
Yolk. However, like the similar observation in human beings, it
seems to indicate the gradual acquisition of resistance as the virus
begins to exert its influence upon the tissues.
Again, with monkeys as in man, the question arises whether the
resistance so unquestionably proven is a condition merely coexistent
with active disease, or whether it may be interpreted as a true im-
munity which persists after the microorganisms have been com-
pletely removed. The most directly pertinent experiments are those
of Neisser. Neisser reinoculated monkeys at periods ranging from
27 to 645 days after the first infection. After waiting a time suffi-
ciently long to insure the negative result of the reinoculation, he used
organ-substance from these animals to inoculate other monkeys. In
22 experiments of this kind he obtained positive results — showing
that the organs of the apparently immune animals still harbored
142 Neisser. Beitr. z. Pathol. u. Ther. d. Syphilis, Springer, Berlin, 1911.
143 Kraus and Volk. Wien. klin. Wchnschr., 1906, No. 21, Ref. IX. Kon-
gress d. Dermatol. Gsell. in Bern.
THERAPEUTIC IMMUNIZATION IN MAN 507
virulent treponemata. In seven animals only did the inoculations
with organ-substance fail to produce lesions, but of these all but one
died before the 30th day after inoculation.
In contrast to these results Neisser found that animals which had
been "cured" by various treponemacidal agents such as atoxyl,
arsacetin, etc., were almost regularly reinoculable. In fact, these
experiments were so uniform that Neisser later utilized the reinocu-
lation method as an index of cure or persistence of the disease.
The results obtained in monkeys, therefore, are very similar to
those determined by clinical observations in man, and the following
statements may be taken as summarizing the conditions revealed by
monkey experiment.
1. That the body develops resistance, progressively increasing
as the virus becomes generally distributed.
2. That the resistance probably reaches its highest development
during the early tertiary period.
3. That complete cure is probably synonymous with gradual
return of susceptibility.
The only other animals on which systematic experimentation has
been possible up to the present time have been rabbits. Since
Bertarelli's successful production of keratitis and Parodi's inocula-
tion of the testes in these animals, they have been studied carefully
by a number of workers, chiefly Uhlenhuth and Mulzer,144 and
Noguchi 145 ; and in our own laboratory, with Hopkins and Mc-
Burney, the writer has observed rabbit syphilis for a number of
years. From a large mass of observations it appears that the condi-
tions in rabbits are not identical with those observed in man and
monkeys. So far, the testes and the eyes are the only organs in,
these animals in which syphilitic lesions can regularly be produced,
and although treponema pallidum may apparently be distributed
generally to the organs after inoculation, it does not easily arouse
pathological reactions except in the organs named, and the lesions
produced are not accompanied by a generalized resistance compar-
able to that discussed in connection with the higher animals in pre-
ceding sections of this paper.
Bertarelli found that he could reinoculate the cornea in a rabbit
that had previously been inoculated with syphilis. Uhlenhuth and
Mulzer, and Neisser and Piirckhauer, showed that infections of the
eye could be produced while the opposite eye was still syphilitic.
Tomazewski found that scrotal infection did not protect against in-
fection of the cornea, and vice versa. Furthermore, Uhlenhuth and
Mulzer, on the basis of a very thorough study, believe that such rein-
fections are neither less extensive nor more rapid in healing than
144 Uhlenhuth and Mulzer. Arb. a. d. k. Gsndhtsamte., 1913, xliv.
145 Noguchi, H. Jour. Am. Med. Assn., 1912, Iviii, 1163.
508 INFECTION AND RESISTANCE
was the first lesion. The same authors noticed resistance to reinocu-
lation in two animals only, and these were young rabbits in the first
weeks of life, which, they believed, had been generally syphilized by
intracardial injections. Interesting in this connection are claims by
Ossola and Truffi who believe that successful skin inoculation in
rabbits confers a certain amount of skin resistance and this is in
harmony with the belief of Kraus and Yolk, that a specific skin
immunity in syphilis is possible. Of great interest to us and of a
possible theoretical bearing which we will discuss below, are observa-
tions made by the writer with Hopkins and McBurney 146 on 20
rabbits which were reinoculated into the testes after apparent healing
of lesions in these organs. It appeared that in rabbits the opposite
testis can be successfully inoculated before, during, or after, the
existence of a testicular lesion on one side, but that reinoculation of
the same testis which had apparently returned to normal, at periods
ranging from 6 weeks to one year, was not often successful. There
is a certain amount of evidence here of a purely local immunity, a
matter which we will discuss at greater length presently. Perhaps
the difference between rabbits and the higher animals lies chiefly in
the fact that syphilis is not generalized in the same sense that it is
in man and monkeys, and even though inoculations from the organs
of syphilitic rabbits may often result positively, this might signify
only that the treponemata have been generally distributed by the
blood stream and latently lodged in the organs, without, however,
arousing in the tissues any sort of pathological response. This idea
would seem to be borne out by the two experiments of Uhlenhuth
and Mulzer cited above in very young rabbits since in such animals
true generalization seems to be more common. The study of very
young rabbits will be continued with this point in view.
It is apparent from the preceding considerations that resistance
to syphilis differs from that acquired in many bacterial infections
chiefly in the fact that it does not persist after the disease is over,
but probably coexists only with the presence of the living incitants
in the body. In this respect it seems to be similar to the conditions
prevailing in many protozoan diseases. Thus Schilling147 cites
a case of Trypanosoma Brucei in a steer, experimentally infected
by Koch, in which reinoculation was repeatedly negative, this result
being at first falsely interpreted as immunity; but six years later
Kleine found that the same steer still harbored the trypanosomes
in his blood — showing that the resistance to reinoculation was not,
in this case, an evidence of true immunity, but rather represented a
condition of insusceptibility to "superinfection" analogous in every
146 Zinsser, H., Hopkins, J. Gv and McBurney, M. Jour. Exper. Med.,
1916, xxiii, 329, 341.
147 Schilling. In Kolle u. Wassermann Handb. d. Path. Mikroorg., xvii,
600.
THERAPEUTIC IMMUNIZATION IN MAN 509
respect to that existing in syphilis. Similar conditions have been
shown to prevail in Texas fever and Schilling believes that they may
be regarded as also existing to a certain extent in Malaria and Sleep-
ing Sickness. In both of these conditions complete spontaneous
sterilization of the body (without medicinal aid) is probably rare,
possibly does not occur at all, and the apparent immunity to rein-
fection is, as in syphilis, an evidence of persistence of the disease in a
latent form.
In weighing the analogy of syphilis to such protozoan diseases,
one is inclined to wonder whether syphilis in man may be regarded
as at all spontaneously curable. So few are the cases left untreated
and so rarely does the reinfection occur even in the face of specific
remedies, that it seems to us more than likely that in syphilis, as in
Sleeping Sickness, a spontaneous "sterilizing" immunity does not
occur. This is a point, however, regarding which it is impossible
to gather data.
In order to distinguish the conditions outlined above tersely from
the ordinary conception of immunity, Neisser 148 speaks of the altera-
tions which govern the reactions of the luetic body to freshly intro-
duced virus as "Anergie;" and "Umstimmung" or "Allergie." By
"Anergie" (a term first used by v. Pirquet and subsequently intro-
duced by Siebert in working out the analogy between pigeon-epitheli-
oma and syphilis), Neisser designates a condition, of inability to
react by cellular change to contact with the virus. As he uses it, it
implies a passivity on the part of the invaded tissues (in which "die
Zellen auf die Spirochaeten schwer oder garnicht reagiren), by
which there is not necessarily a destruction of the invading trepone-
mata, and which, therefore, cannot in any sense be interpreted as a
"Schutz wirkung." By the "Umstimmung" of Neisser, the "change-
ment dans la mode de reaction" of Levaditi, is meant the changed
reaction capacity of the syphilized tissues which determines the char-
acters of the lesions at various stages of the disease. Thus it is
obvious that cellular reactions which result in the primary induration
are quite different from those which produce the tertiary gumma,
and that the histological changes of the roseola are distinct from
those of the serpiginous syphilide of the late stages. And since, as
we shall see, there is no valid reason to assume that the incitant has
been modified in virulence or vitality, we are forced to believe that
the reaction capacity of the body cells has been altered.
Since it is a fact, then, that syphilitic infection so changes the
body tissues of man and monkeys that, during its course, resistance
to reinfection is produced, it should be possible to analyze this re-
sistance into its responsible factors and perhaps utilize the knowledge
so gained for practical therapeutic purposes. Before we proceed
148 Neisser. Baermann u. Hall) erst aedfer d. med. Wchnschr., 1906, Nos.
1 and 3.
510 INFECTION AND RESISTANCE
to do this, it will be of advantage to review briefly the attempts at
active and passive immunization which have been made in animals
and man.
Metchnikoff and Roux 149 followed their first animal inoculation
studies by extensive vaccination experiments. Some of their first
work along these lines is perhaps marred to some extent by an in-
sufficient recognition of the resistance which depends upon per-
sistence of the disease rather than upon true immunity, but a good
many of their observations are of fundamental importance. It will
be convenient to classify their work and that of others into experi-
ments dealing with "active" and those dealing with "passive" im-
munization.
Metchnikoff and Roux first worked with filtered virus and virus
killed at 51° C. All their attempts with such material were negative
in that the monkeys treated with it could not be regarded in any
sense as immunized. In their reports of these experiments they
remark that they believed this to be due to an absolute loss of power
to incite reaction on the part of the vaccine-material. We emphasize
this point here because our own subsequent work inclines us to be-
lieve, with them, that the production of a reaction is necessary for
the development of any considerable degree of resistance.
Neisser carried out a large number of attempts at vaccination in
which he used extracts of syphilitic primary lesions and of the organs
of congenitally syphilitic children, killed by the addition of carbolic
acid. Unfortunately he assumed that his extract contained syphilitic
antigen because it gave positive reactions by the complement fixation
technique of Wassermann, an assumption which we of course know
now to be unfounded as far as any relation to the body substance of
the treponemata is concerned. This, to our mind, deprives these
particular experiments entirely of their negative importance.
In rabbits a large amount of work has been done by Uhlenhuth
and Mulzer. They injected living material from rabbit lesions in-
travenously and subcutaneously, without ever observing any evidence
of protection against subsequent inoculations.
Of perhaps the greatest importance in connection with active im-
munization are the attempts made upon human beings by different
investigators.
Casagrandi and de Luca 15° tried prophylactic immunization on
six human beings by injection of filtrates obtained from primary
lesions. Two of these people later contracted syphilis in the ordinary
way.
Possibly the most hopeful results are those obtained by Spitzer ir>1
149 Metchnikoff and Roux. Ann. de. I'Inst. Pasteur, 1903, xvii, 809; 1904,
xviii, 1 and 657; 1905, xix, 673; 1906, xx, 785.
150 Casagrandi and De Luca. Gior. ital. de mal. ven., 1905.
151 Spitzer. Wien. klin. Wchnschr., 1905, 45, and 1906, 38.
THERAPEUTIC IMMUNIZATION IN MAN 511
by a method suggested by Kraus. Kraus,152 reasoning from the fact
that syphilis like hydrophobia was a disease with Ipng incubation
time, expressed the hope that perhaps the method of Pasteur in
hydrophobia, that is, active immunization during the period of in-
cubation might, in syphilis also, tend to abort the disease. Accord-
ingly, Spitzer treated 15 cases of early syphilis immediately after
the appearance of the chancre by subcutaneous injections of emul-
sions made from human chancre material in dilutions of 1 :200 to
1:20. The cases received from 11 to 20 injections and in seven of
them the disease was uninfluenced. In the others, however, subse-
quent symptoms were delayed, and in four, no generalized symptoms
occurred. In a later communication, Spitzer reported 23 further
cases similarly treated, 10 of which failed to develop generalized
symptoms and in 9 of these the Wassermann test remained negative.
One of them, a fact which is of great importance, was spontaneously
reinfected with syphilis two and one-half years later. These results,
if accurate in every way, are of the greatest importance, but are
diametrically opposed to the experience of all other investigators.
Monkey experiments along the same lines by Neisser gave entirely
negative results and Brandweiner 153 as well as Kreibich 154 were
unable to confirm Spitzer's results in man. Further comment on
the Kraus and Spitzer method is valueless without more experi-
mental data. It is one of the few rays of hope, but so isolated that
one is forced to skepticism. Metchnikoff and Roux did almost the
identical thing in an orang-outang and obtained lesions both at the
point of the original inoculation as well as that at which the subse-
quent "protective" injection was made.
The only experiments in which an attempt at vaccination with
attenuated virus was made with some indication of efficacy, is one of
Metchnikoff and Roux, the outcome of an accidental laboratory in-
fection. It appears that a laboratory assistant who had been attend-
ing to the animals, noticed a small lesion on his lip which did not
look like a typical syphilitic chancre. In order to allay the patient's
fears, however, Metchnikoff and Roux did inoculations from the
patient to monkeys and these were positive. Nevertheless, Fournier
after examining the original lesion declared it so unlike the ordinary
primary sore that he did not advise treatment. No secondaries de-
veloped in the patient nor in the three chimpanzees inoculated with
the material. From this occurrence Metchnikoff and Roux con-
cluded that the patient had probably been infected in handling the
monkeys, and that the virus had become attenuated by passage
through these animals. On the basis of this observation they later
inoculated a willing subject 79 years old with virus carried for five
152 Kraus. Wien. klin. Wchnschr., 1905, No. 41, and 1906, No. 21.
153 Brandweiner. Wien. med. Wchnschr., 1905, No. 45.
154 Kreibich. Wien. klin. Wchnschr., 1906, No. 8.
512 INFECTION AND RESISTANCE
generations in lower monkeys. The lesion which developed was very
slight, consisting only of a local induration, and no generalized symp-
toms developed. A previous attack of syphilis Metchnikoff believes
could be reliably excluded in the subject. The experimenters suggest
that the passage through lower monkeys may attenuate the virus for
man, this leading to relative immunity to subsequent inoculations,
and furnish a possible means of protection. Their experiments are
too few to permit conclusions as yet, but even should they hold good,
the method would seem to imply a considerable degree of risk and
our experience with superinfections and reinfections in syphilis does
not encourage the hope that a method so drastic would be justified
when the benefit to the patient is apt to be of such duration only.
The history of passive immunization is an extensive one and
hardly worth going into with any degree of detail since many times
extravagant claims have been made only to be refuted by accurate
study. There is a great similarity in respect to this between syphilis
studies and those in tuberculosis and cancer. A brief examination of
the bibliography in Neisser's book is sufficient to convince one of the
many attempts that have been made in this direction and often by
methods as ludicrous as the claims of success for which they formed
the basis. The most careful and skillful workers have uniformly
reported failure. Metchnikoff and Roux treated various monkeys
with blood from syphilitic patients and used the serum of these ani-
mals for protective experiments. There are a few instances in which
mixtures of such serum with syphilitic virus rendered this inactive
on inoculation. A powder made of this serum was supposed to have
some protective effect when applied to fresh inoculation spots within
the first hour after inoculation. However, injection of the serum
had no effect whatever. Casagrandi and de Luca using serum of a
dog treated with syphilitic virus, obtained entirely negative results,
and Finger and Landsteiner 155 report negative results with monkey
blood in man. The most extensive experiments were again those
carried out by Neisser and his associates. They were done by the
treatment of animals with dead and living syphilis virus, organ
extracts, with the blood of syphilitic man and monkey, and horses,
sheep, and monkeys were used for the production of "immune"
serum. In no case was there the slightest protective effect on the
part of the serum either in vitro or in vivo, and the results of the
experiment were unqualifiedly negative.
When the method of complement fixation was successfully ap-
plied to the diagnosis of syphilis first by Detre, and then by Wasser-
mann, Neisser and Bruck,156 it was generally assumed that this
reaction incidentally demonstrated that specific antibodies were
155 Finger and Landsteiner. Centralbl. /. Bakteriol., Ref. xxxviii, 1906.
156 Wassermann, A., Neisser, A., and Bruck, C. Deutsch. meet. Wchnschr.,
1906, xxxii, 755.
THERAPEUTIC IMMUNIZATION IN MAN 513
formed in syphilis. As matters have developed, however, this point
of view can no longer be maintained. It was soon discovered that
the antigen used for these reactions by Wassermann and his associates
derived its "fixing" constituent not from the body substances of the
treponemata contained in the syphilitic organs, but from certain
tissue extractives chiefly of lipoidal nature which could be obtained
readily from normal as well as syphilitic tissues. Although Bruck 157
and others who have occupied themselves with the theoretical basis
of the Wassermann reaction, still maintain that a specific antibody
may be incidentally involved, they admit that this is the less im-
portant factor in the reaction which depends chiefly upon the ex-
istence of lipotropic substances which appear in the course of the
disease as metabolic products either of the body or possibly of the
treponemata. This is a subject which we will deal with in extenso
in another paper. It is sufficient for our present purposes to point
out that, although to a slight extent specific antibodies may play a
part in the Wassermann reaction, this is certainly not the chief
element or even a very important factor involved. The truth of this
conception has been further confirmed by the work of Noguchi,158
of Craig and Nichols,159 of Kolmer,160 and ourselves, in which it
has been found that antigens made with pure cultures of treponemata
produced a complement fixation with syphilitic sera to a very limited
extent only, and in our work in which we have been able to duplicate
the Wassermann reaction in a large series of antigens made from the
treponema cultures, we have found that similar results could be
obtained with cultures of colon and typhoid bacilli identically pre-
pared, the fixing power to a large degree depending upon the lipoidal
constituents of the bacteria. Whatever the ultimate explanation of
the Wassermann reaction may turn out to be (and this is a subject
which it would not be profitable to discuss at length at this place), it
cannot be maintained that as it exists at present, it can be interpreted
as demonstrating the existence of true circulating antibodies, analo-
gous to those found in bacterial diseases in the blood of syphilitic
patients.
Early in the history of such investigations, Fornet,161 with his
collaborators Schereschewsky, Eisenzimmer, and Rosenfeld,162 found
that when the sera of syphilitics were mixed with clear extracts of
syphilitic livers similar to those used in the Wassermann reactions,
precipitates formed which were not seen in similar experiments done
157 Bruck. Imm. bei Sypn. in Kolle u. Wassermann Handb. d. Pathog.
Mikroorg., vii, 1045 et seq.
158Noguchi, H. Jour. Exper. Med., 1911, xvi, 99.
159 Craig, C. F., and Nichols, H. J. Jour. Exper. Med., 1912, xvi, 336.
160 Kolmer, J. A. Jour. Exper. Med., 1913, xviii, 18.
lei Fornet. XIV Internal. Kongress f. Hyg. und Derm., Sept. 1, 1907.
162 Fornet, Schereschewsky, Eisenzimmer and Rosenfeld. Deutsch. med.
Wcbnschr., 1907, No. 41.
514 INFECTION AND RESISTANCE
with normal sera. This Fornet163 in a number of communications
interpreted as showing the formation of precipitins in the course of
syphilis. At almost the same time, L. Michaelis 164 made similar
observations giving them the same interpretations. The experiments
of Fornet have not found universal confirmation, but even if such
precipitin reactions can be occasionally observed, they do not indicate
true precipitins for the same reasons that the Wassermann reaction
does not demonstrate true complement fixation of antibodies. The
antigens did not represent strong treponema antigens and Jacobs-
thai 165 and others have shown that with the ordinary Wassermann
antigens syphilitic sera can produce precipitations visible under the
dark field, the basis of the complement fixation being, therefore, one
of probable colloidal precipitation. The formation of true precipi-
tins, therefore, has not been shown for syphilis.
Investigations of the effect produced upon virulent treponemata
by the sera of syphilitic individuals have likewise been unsatisfac-
tory. Hoffmann and Prowazek 166 reported in 1906 that the serum
of syphilitics in the later stages produced immobilization of the
virulent treponema pallidum, an observation which was confirmed
by Zabolotny.167 Landsteiner and Mucha 168 were not able to observe
such immobilization, nor did they see any evidence of agglutination
in such experiments. In a limited observation of our own, we have
also failed to see any regular or distinct influences of this kind. It is
not impossible that a slight difference may exist in this respect
between syphilitic and normal sera, but even if it occurs, the action
is feeble, irregular, and entirely insufficient to be interpreted as
having much practical importance in the acquisition of syphilis
immunity.
Treponemacidal experiments have been done by some workers
also with negative results. As we have stated in another part of
this paper, the attempt to treat syphilitic animals and man with the
serum of syphilitics has led to no reliable results, and no evidence
whatever that can be accepted has been adduced to show that virulent
treponemata may be killed by active luetic serum.
As far as any opsonic action is concerned, Levaditi 169 in histo-
163 Fornet, Schereschewsky. Munchen med. Wchnschr., 1907, No. 30 ;
Berl klin. Wchnschr., 1908, 85.
164 Michaelis. Berl. klin. Wchnschr., 1907, No. 46.
165 Jacobsthal. Munchen. med. Wchnschr., 1910, Ivii, 214.
166 Hoffmann, E., and Prowazek, S. Centralbl f. Bakteriol, I. 0., 1906,
xli, 741 and 817. (Balanitis and montu spirochaetae.)
167 Zabolotny, D., and Maslakowitz. Centralbl f. Bakteriol, I. O.,
xliv, 532.
168 Landsteiner and Mucha. Wien klin. Wchnschr., 1906, xix, 1349; Cen-
tralbl f. Bakteriol, 1907, I. R., xxxix, 540.
169 Levaditi. Compt. rend. Soc. de biol, 1906, Ix, 134; Ann. de I'Inst.
Pasteur, 1906, xx, 41.
THERAPEUTIC IMMUNIZATION IN MAN 515
logical studies in congenitally syphilitic children has seen phagocy-
tosis or at least intracellular localization of treponemata in the alveo-
lar cells of the lung and in the parenchyma cells of the liver and
kidneys. For reasons not entirely clear to us, he interprets the
former as true phagocytosis and the latter as a penetration of the
t,reponemata into the liver cells to the detriment of the latter. We
ourselves have occasionally seen phagocytosis in sections of syphilitic
rabbit testes, but in all cases the process was not a very active one
and not much can be said about its importance at present. As a
matter of fact, Hopkins and the writer,170 in studying the mechanism
of the natural resistance of mice against syphilis, injected virulent
organisms into the peritoneal cavities and observed the treponemata
in peritoneal puncture fluid, alive, actively motile, and unphagocyted,
though surrounded by masses of leucocytes, as long as three days
after their injection. It seemed almost as though the natural im-
munity of such animals might be similar to the "atreptic" cancer
resistance spoken of by Ehrlich. The treponemata did not multiply
in the mice but though, naturally, diminishing in number, were ap-
parently neither killed nor even inhibited in motility by the perito-
neal exudates and, for several days, swam in and out among the
accumulating leucocytes, often adhering to them peripherally but
not taken up by them. Lack of entirely satisfactory methods of
staining cells containing treponemata make it difficult to speak
with certainty of the actual occurrences. But we gained the distinct
impression that the treponemata were not actively injured or de-
stroyed until they had spontaneously died out owing to lack of
suitable environment, i. e., nutrition.
In the case of natural immunity at any rate, we do not think that
phagocytosis by the mobile leucocytes plays a primarily important
role. However, these experiments will need further elaboration.
The search for antibodies gained new vigor when the efforts of
Schereschewsky,171 Miihlens,172 Hoffmann 173 and especially Nogu-
chi, had resulted in successful cultivation of treponemata from syphil-
itic lesions in man and animals. It was hoped that with the causative
organisms isolated, immunization and a clear understanding of the
antibodies in syphilis might yield practical results. Kolmer 174 ob-
served that cultivated treponemata were agglutinated in the sera of
rabbits treated with culture material, and such agglutinins were pro-
170 Zinsser, H., and Hopkins, J. G. Jour. Am. Med. Assn., 1914, Ixii,
1802 ; Jour. Exper. Med., 1915, xxi, 576.
171 Schereschewsky, J. Deutsch. med. Wchnschr., 1911, xxxvii, Nos. 20
and 39.
172 Miihlens. Treponema Pallidum in v. Prowazek Handbvch der Paihog.
Protozoen, i, Barth, Leipzig, 1912 ; Klin. Jahrh., 1910, xxiii, 339.
173 Hoffmann. Berl. klin. Wchnschr., 1905, No. 46.
174 Kolmer, J. A., Williams, W. W., and Laughbaugh, E. E. Jour. Med.
Research, 1913, xxviii, 345.
516 INFECTION AND RESISTANCE
duced in high potency by the writer with Hopkins. Subsequently
in our own laboratory with Hopkins and McBurney, extensive ex-
perimentation on the production of antibodies with culture pallida
was carried out.
We were able to show that not only were agglutinins formed by
the immunization of rabbits with such cultures, but also that tre-
ponemacidal antibodies which were analogous to the ordinary bac-
tericidal substances in such sera were present. Also, it was shown
by cross agglutination and absorption experiments, that treponemata
cultivated from various sources were related in group reactions.
Indeed, experiments done with cultivated treponemata (much
facilitated by the discovery of a simple method of obtaining mass
cultures of old strains) seemed at first very encouraging in that
animals immunized with the cultures responded by powerful anti-
body formation. It was a perfectly justified hope, therefore, that
antibodies produced with these "attenuated" or rather "avirulent"
strains might have some action on the virulent treponemata in luetic
lesions. However, subsequent work in this direction disappointed
such expectations. We may briefly review this work as follows :
The serum of rabbits immunized with "culture" pallida although
potent against "culture" pallida, had no effect either in agglutinating
the virulent organism from rabbit lesions, nor did it exert any pro-
tective influence when the virulent organisms were subjected to its
action before injection. 4
Conversely, the serum of syphilitic rabbits showed but a very
slightly increased agglutinating power for the "culture" pallida.
This increase of potency in a few experiments was definite but very
slight, a few of the syphilitic animals agglutinating as highly as 1 :25
and 1 :50, whereas most of the normal rabbits agglutinated in 1 :10
and some of them 1 :25.
Although Kissmeyer has recently reported that diagnostic use
might be made of the fact that sera of syphilitic individuals ag-
glutinated the "culture" pallida, we have tried this with a consider-
able number of cases and found that although the sera of tertiary
syphilitics will sometimes agglutinate a little more highly than will
the sera of normal individuals, yet many patients suffering from
non-syphilitic diseases agglutinated as highly and almost as regularly
.as did the syphilitics.
We have come to the conclusion, therefore, as far as our work
lias gone, that in the syphilitic human being there is as little ag-
glutinin formation against the "culture" treponema as there seems
to be against the virulent organisms. If the slightly greater ag-
glutinating power found in some of the tertiary syphilitics can be
considered at all, the reaction is so feeble that it is negligible from
the points of view either of diagnostic value or protective importance.
Furthermore, vaccination either intravenously or locally into the
THERAPEUTIC IMMUNIZATION IN MAN 517
testis with cultures has thus far failed to protect rabbits against
subsequent inoculation with virulent material, and passive immuniza-
tion with sera produced with "culture" pallida has been without
effect.
From all this it appears that the "culture" treponema has.im-
munologically no relation to the virulent organism. It has lost its
virulence completely, as six and more successive inoculations into
rabbit testes have sufficiently demonstrated to us.
The sera produced by many injections of dead and living culture
organisms have no effect whatever on the virulent organisms in vitro,
and vaccination with it does not protect against subsequent infection.
The luetin reac ion is the only method by which the relationship
between the two is demonstrable at all, and there too, we have to
reckon with what is generally spoken of as the non-specific increased
sensitiveness of the syphilitic skin.
Were it not for the production of lesions with cultures in their
early test tube generations by Hoffmann175 and by Noguchi in a few
experiments, one would be almost in doubt as to the identity of the
virulent with the culture organisms.
The reactions in syphilis between the invading microorganism
and the invaded subject thus differ in certain fundamental premises
from those prevailing in diseases caused by most bacteria. The
treponema pallidum is an organism which, unlike many bacteria, is
rarely subjected to the necessity of adapting itself to extra-corporeal
existence during the interval between its passage from one host to
another. It practically always infects directly, being inoculated from
one human being to the next and has in consequence developed a very
delicate parasitism not unlike that seen in certain trypanosome
diseases of rats and that which we ourselves have observed in the
well-known spirochete infections of white mice. A considerable
percentage of laboratory white mice have been found to harbor
actively motile spirochetes, often in considerable numbers in the
blood and peritoneal fluid without there being any objective signs of
illness in the animals. It is an instance of what has been spoken of
by Bail as "infection without disease," and approaches what biolo-
gists speak of as symbiosis, except that the host in this case does not
benefit in any way by the invasion. Even in the case of the mice,
a certain amount of gradual injury, perhaps only metabolic, by the
slow removal of nutritive material, is taking place. In syphilis the
mutual adaptation may perhaps be less complete, a sufficient accumu-
lation of the invaders and especially a mechanical injury of tissue
cells, of closing of tissue spaces together with a certain amount of
toxic action, leading eventually to pathological changes.
The virulent treponemata apparently do not arouse true antibody
formation in any marked degree. When they have been cultivated
175 Hoffmann, W. H. Deutsch. med. Wchnschr., 1911, xxxvii, 1546.
518 INFECTION AND RESISTANCE
and have become accustomed to the test tube conditions they entirely
lose their virulence, are easily attacked by the active constituents of
animal serum, and are probably amenable to phagocytosis. When
such cultures, living or dead, are injected into animals they act like
other specific protein antigens and incite the formation of antibodies.
However, these antibodies have no effect whatever upon the virulent
organisms.
In consequence, it cannot astonish us that all efforts at passive
immunization with the sera of syphilitic man and animals, or with
those of animals systematically treated with dead virulent materials,
have been unqualified failures.
However, this does not preclude the theoretical possibility of
active immunization or vaccination with such materials since, in this
case, the antigen distributed from the points of injection might act
upon tissue cells throughout the body. However, with the exception
of the unconfirmed reports of Spitzer, all attempts to vaccinate
either with dead virulent material, or with living and dead culture
material, have been disappointing. The few experiments of Metch-
nikoff with virus "attenuated" by passage through monkeys have
indeed seemed to indicate some possibility of approaching the subject
from this direction, but these isolated observations have been very
logically criticized by Neisser and should not bear too much weight.
The observations were made on two cases only, both of them well
along in life, and the validity of the important conclusions drawn
rests entirely on the always problematical fulcrum of complete ex-
clusion of previous syphilitic infection in the two subjects. More-
over, attempts in this direction would be fraught with a considerable
amount of danger and it is therefore questionable whether experi-
mentation along these lines is sufficiently promising to be justified.
There is certainly no attenuation for man by passage through rabbits
as has been sufficiently proven by a number of accidental infections,
an instance of which in a laboratory attendant has been reported by
Graetz and Delbanco.176
We may state, therefore, as safely summarizing our knowledge
of the conditions in syphilis, that the resistance which undoubtedly
develops during the course of the disease is one which depends upon
reaction to the living virus only, cannot so far be produced in animals
by systemic treatment with dead treponemata, and does not express
itself in the formation of significant amounts of circulating anti-
bodies analogous to those observed in bacterial diseases. Moreover,
it is a well-known fact that the treponemata can continue to do injury
to many organs and tissues at a time when reinfection by the paths
of skin and mucous membranes is no longer possible.
How, then, are we to explain this peculiar state of affairs? A
clue to the problem we think is found in the 20 rabbits which Hop-
176 Graetz and Delbanco. Med. klin., 1914, 375 and 420.
THERAPEUTIC IMMUNIZATION IN MAN 519
kins, McBurney and the writer 177 inoculated into the testes after
apparent recovery from a previous lesion. Ordinarily in rabbits,
as we have stated before, no generalized resistance is developed dur-
ing the disease, and the opposite testis can be successfully inoculated
before, during, and after the existence of a lesion on the other side.
In these rabbits it was found that testes that had apparently recov-
ered from a previous lesion, were not subsequently as easily infected
as were normal testes. It has seemed to us from this as well as from
a careful study of the observations of other investigators, that re-
sistance in syphilis was probably a matter of localized reaction.
Tissues which have sustained active invasion with the living virus
react and gain thereby a certain degree of resistance which expresses
itself in a failure to react to subsequent inoculation. This would
explain why in syphilis of the human being reinoculation is unsuc-
cessful and reinfection of the skin and mucosse does not occur spon-
taneously at a period later than the early secondary stages when the
virus has become systemically distributed. It would furthermore
explain why in this disease organ after organ may be pathologically
involved when skin infection is no longer possible. This point of
view is entirely in harmony, though perhaps from a slightly different
aspect, with the skin immunity suggested by Kraus and Volk, where
the resistance is attributed to the tissue as a whole rather than to a
local cell group. The cells which have once reacted to the living
virus no longer respond, i. e., can no longer be injured for an in-
definite period after recovery. "However, the factors which lend
them this resistance, whatever they may be, are not distributed to
the blood stream in a way analogous to that in which antibodies are
mobilized in bacterial diseases, and the effect of the resistance of the
local area is not distributed to remote parts of the body. It is of
course likely that a certain amount of phagocytosis of treponemata
by the now resistant fixed cells may account for the absence of local
injury. This, however, we have not yet been able to prove suffi-
ciently, and further studies on this point are necessary. As far as
any positive evidence can be adduced at the present time, the newly
entering treponemata may not be entirely destroyed. It may well
be that the tissues do not react and are in the state which Neisser
calls "Anergie." The treponemata that enter during this period
may nevertheless remain uninjured and be as capable of subsequently
causing lesions in other locations as are those already present in the
patient. This phase is being studied by comparative histological
observation on the fate of treponemata which have been injected into
normal tissues and into tissues rendered resistant by previous in-
fection.
In animals like monkeys and man where generalization is rapid
177 Zinsser, H., Hopkins, J. G., and Gilbert, R. Jour. Exper. Med., 1915,
xxi, 213.
520 INFECTION AND RESISTANCE
and apparently complete, the resistance becomes a general one. In
animals like the rabbit in which the lesion — or in other words
pathological response, occurs in a few organs only, the resistance is
limited to the particular organ or organs that have previously de-
veloped a lesion.
It must not be forgotten that such a .resistance probably persists
for a limited time only, and does not imply the sterilization of the
body and the complete destruction of the microorganisms. These
may, and probably do, remain alive and potent in various parts of the
body, capable of again setting up new lesions in parts hitherto unin-
volved or again susceptible after a diminution of their local, acquired
resistance. That the virulent treponemata may remain thus latent,
alive and virulent has been sufficiently shown in animal experimen-
tation by successful inoculation with tertiary lesions and by the
frequent late accidents, especially of the nervous system, in indi-
viduals apparently cured or for a long time without symptoms.
It would seem when we analyze the conditions in syphilis, that
complete sterilizing immunity or, in other words, complete cure,
occurs but rarely without specific medicinal aid, and that the un-
treated syphilitic (if such an unfortunate individual exists in a
civilized country) might go on to apparent cure, in that a general
syphilization of his body would bring about a general resistance,
but would always harbor virulent treponemata which could cause
recrudescences in parts in which resistance was diminished, and
eventually kill by degenerative processes in the central nervous sys-
tem where many injuries cannot be compensated for as is possible in
other organs.
The resistance which develops is apparently a new attribute only
of the cell groups which have undergone direct reaction with the
treponemata. This resistance may consist merely in the complete
failure of the tissue cells to react to the virus, a sort of "tissue indif-
ference" or "Anergie." It may be, however, and probably is, ac-
companied by a certain amount of active defense in the form of local
phagocytosis of the treponemata by the fixed tissue cells.
THE INFLUENCE OP INJECTIONS OF NON-SPECIFIC SUB-
STANCES UPON INFECTIOUS DISEASES
Therapy of infectious diseases has been very logically dominated
in the past by attempts to increase resistance, either passively, by
the injection of specific anti-sera, or actively, by treatment with bac~
terial antigens or their derivatives. The idea of specificity, in other
words, has dominated such attempts almost exclusively. In spite of
this, however, there has gradually grown in the minds of bacteriolo-
gists an impression that not all the effects of the injection of bacterial
THERAPEUTIC IMMUNIZATION IN MAN 521
protein were purely specific. Jobling, who has recently summarized
most of this work, calls attention to the fact that Matthes178 as
early as 1895 showed that the effect of tuberculin injection could
be obtained equally well with deuteroproteose. From that time on,
many observations have been made to show that often profound
physiological effects could be produced in human beings and animals
suffering from infectious diseases, if they were treated not with
specific antigen but with proteins and protein derivatives of many
sources.
The work that the writer did with Hiss on the injection of
leucocytic extracts is a case in point. Similar in significance prob-
ably are the results obtained by the injection of substances such as
the filtrate of bacterial cultures, commercially sold as phylacogens,
the intramuscular milk injection practiced by Schmidt179 and the
injection of various ferments, reported by a number of writers
throughout the literature of the past ten or fifteen years. Into the
same category fall the favorable reports supposed to have been ob-
tained in syphilitics in tuberculin exhibits by Blach.180 Perhaps
the clearest example of this principle has been obtained in connection
with attempts at vaccine therapy in typhoid fever. Attempts to
cure typhoid fever by the injection of typhoid bacilli date back to
Fraenkel, who treated it as early as 1893, and since then, as we have
seen, extensive study has been made on this subject by Petruschky,
Isletter, and others. Ichikawa in 1912 and Boinet in 1914 obtained
astonishing results by the use of sensitized vaccine, the former being
the first to introduce the intravenous method of injection. The
results of the injection intravenously into a patient suffering from
typhoid fever have consisted in rapid falling of temperature, often
followed by a chill, with occasionally rapid general improvement of
the patient.
Gay, too, has recently made such observations by the methods
of Ichikawa, and most of these writers were inclined to believe that
their opinion was effective by some sort of specific reaction. Doubt
has been cast upon this point of view, however, by observations such
as those of Kraus. Ichikawa alone had some doubt of this, as is
indicated by the fact that he injected typhoid bacilli into some of his
paratyphoid patients with like results. Kraus subsequently ob-
tained similar effects by injecting colon bacilli into typhoid patients
and used non-specific bacterial virus with good results on cases of
pyocyaneus infection and upon a single streptococcus puerperal septi-
cemia. Leudke injected typhoid patients with non-bacterial proteose
and Jobling and Petersen have obtained very marked reactions in
178 Matthes, M. Deutsch. Arch. f. klin. Med., 1895, Vol. 45.
179 Schmidt, R. Med. Klin., 1910, No. 43.
180 Blach, M. Wien. klin. Wchnschr., 1915, No. 49.
INFECTION AND RESISTANCE
typhoid patients by injecting them intravenously with quantities of
1 to 2 c. c. of 2 per cent, solution of proteoses.
There seems to be little doubt of the fact that the injection of
bacterial proteins intravenously produces a profound therapeutic ef-
fect without being in any but a minor degree specific. Boinet is still
an adherent of specific reaction, believing that his typhoid patient was
benefited by a rapid mobilization of antibodies and Gay had the idea
that a specific hyperleucocytosis was the basis of improvement. This
last phenomenon has been discussed in another place. We, ourselves,
are inclined to believe that the specific reactions have little to do
with the improvement in such cases and that the results are due to
the injection of foreign protein and perhaps proteoses, the specific
nature of the sources of these have little significance. We believe
that one of the most important factors of such procedure is the rapid
mobilization of leucocytosis which is not specific, in the sense of
Gay and Claypole. In addition to this, as Jobling and Petersen
suggest, the hyperpyrexia which is often observed may have some-
thing to do with it as well as the mobilization of ferments which
the last named writers have observed. They have shown that the
intravenous injection of bacteria and of other protein substances
is followed by a marked mobilization of serum protease and lipase.
They suggest that these factors may have considerable influence on
the result.
The subject is hardly begun but it is being made the motive of
many researches both in the clinic and in the laboratory and un-
questionably the "ext few years will bring much further under-
standing.
CHAPTER XX
SERUM ENZYMES; LEUCOCYTIC ENZYMES; ABDER-
HALDEIST REACTION; PHYSICAL PRINCIPLES IN
SERUM REACTION; MEIOSTAGMIN AND EPIPHA-
NIN REACTIONS; COLLOIDAL GOLD REACTION
IN our section upon the nature of the precipitins, especially in
the discussion of Gengou's conception of "albuminolysins," we have
called attention to the probable significance of protein antibodies as
a mechanism for the disposal of such foreign substances. In the
bodies of the higher animals in which a special alimentary system,
with its many digestive ferments, is well developed, it is most prob-
able that the normal condition of digestion is one in which the
foreign substances utilized for nutrition are completely split into
their simpler components before they gain entrance to the circulation.
Nevertheless, abnormal conditions or accidents, such as gastro-enteric
diseases, digestive disturbances, and bacterial infections, may lead to
a condition, probably frequent enough in ordinary life, during which
such foreign substances may get into the blood stream without pre-
vious cleavage. The problem is to determine where and how such
substances, protein or otherwise, are broken up so that they may be
either assimilated or eliminated. We have referred in another place
to the fact that foreign proteins may occasionally pass through the
kidneys and be eliminated unchanged. This has been shown actually
to occur by Oppenheimer, Ascoli, and others, but probably represents
a very unusual state of affairs produced by special experimental
conditions. As a rule these substances are disposed of within the
body by chemical cleavage or by assimilation. It is more than likely,
therefore, as has been emphasized by such workers as Jacoby, Sal-
kowski, Wells, and others, that every fluid and cell in the body con-
tains enzymes which, by their action upon proteins, fats, and carbo-
hydrates, play an important part in the metabolism of the body.
It has been known for a long time that the organs of animals re-
moved from the body will undergo self-digestion by a process spoken
of as autolysis. The action of bacteria as causing such autolysis can
be excluded by covering organ emulsions with toluol, chloroform,
and some other substances which have the peculiar property of pre-
venting the growth of bacteria without in any way interfering with
the action of enzymes. Thus, any organ of the body so treated will
show rapid changes consisting in the splitting of its own proteins,
523
524 INFECTION AND RESISTANCE
fats, and carbohydrates. That such processes also go on in the living
body is probable, although, as we have stated in an earlier chapter,
it is a well-known fact that living cells oppose a more or less mysteri-
ous resistance to enzymic digestion, just as they oppose a similar
resistance to bacterial invasion. The phenomenon is more fully
discussed by Wells in his "Chemical Pathology," to which the reader
is referred. It is not improbable that this resistance to invasion and
enzyme attacks, spoken of vaguely as "vital resistance," can be anal-
yzed into more exact factors, one of which seems to be the question
of reaction, digestion depending to some ex:ent upon the reduction
of alkalinity by the formation of acid in the tissue cells.
However, the physiological importance of tissue enzymes is a
subject which we cannot go into in this connection, since the rela-
tionship of these enzymes to normal body metabolism is a subject
more fully dealt with in text books of physiology and is too far
removed from the subject of our present discussion to warrant ex-
tensive analysis. It must, of course, be self-evident that the existence
of cellular ferments of this nature must play an important role
wherever and whenever cell death occurs, and since such cell death
is an accompaniment of many phases of bacterial infection it may
well be that the enzymes which destroy dead cells, whether they be
bacterial or those of the body itself, contribute to the general patho-
logical picture characterizing such diseases. Moreover, the impor-
tance of the enzymes is particularly enhanced by the knowledge we
have gained from the work of Vaughan and others, who have shown
that in the course of proteolytic cleavage toxic substances are lib-
erated. It is such proteolysis which is the probable basis of the
formation of the "anaphylatoxin" of Friedberger, the "proteotoxin"
about which we ourselves have written, and the "sereotoxin" of
J obling and Petersen ; and similar processes are involved in the toxic
substances, studied by Whipple, which form in the intestine as a
consequence of ligature of the gut. It is conceivable that wherever
and whenever a proteolytic ferment reacts in the body with its
substrate, toxic cleavage products result, perhaps in the form of
albumoses, peptones, etc., which are rapidly absorbed and cause
symptoms of varying intensity. In the chapter on anaphylaxis we
have seen that many writers have attributed the injury occurring in
this phenomenon to the formation of such split products and there is
much evidence in favor of such a view, although the rapidity and
vehemence of anaphylactic shock indicate that probably there are
purely physical elements also involved which we do not as yet com-
prehend.
Of the cells in the body with which enzyme study has been most
assiduously followed, the most important are the leucocytes. We
have already had much to say in the chapters on phagocytosis of the
digestive functions of the white blood cells. However, we dealt with
SERUM ENZYMES 525
them there chiefly from the point of view of their phagocytie and
antibacterial functions. It will be useful to consider at a little
greater length the importance of these cells from the point of view
of purely enzymotic activity.1
As early as 1885 Hammarsten called attention to the fact that
leucocytes aid in dissolving fibrin. Leber 2 subsequently studied pus
and found that it possessed a powerful digestive action for gelatin,
fibrin, and other protein substances. He correlated his studies par-
ticularly with the processes of inflammation, and his work, as well
as that of later investigators, brought into the foreground the fact
that in the resolution of abscesses, especially of the staphylococcus
variety, the leucocytes might play an important role in liquefying
the necrotic tissue cells and bringing about the breaking down of
the center of the abscess. Any one who has carefully studied the
histology of a staphylococcus abscess can easily see the halo of disin-
tegrating tissue lying just inside the ring of aggregated polymorpho-
nuclear leucocytes. That these cells carry on an important part in
this liquefaction is certain, although as yet we are not sure whether
or not the proteolytic bacterial enzymes participate. Friedrich Mill-
ler,3 in 1902, also studied these processes and gave clinical signifi-
cance to the proteolytic activities of the white blood cells in the
resolution of the pneumonic lung. The study of inflammatory exu-
dates by Opie 4 confirmed much of the preceding work and revealed
that the ferments of all white blood cells were not the same. He
found a leucoprotease which was contained in the polynuclear leuco-
cytes which was active chiefly in a slightly alkaline medium, whereas
the large mononuclear cells of uncertain origin which appear toward
the end of an inflammatory process contained a protease which was
active in weak acid. This differentiation of a leucoprotease and a
lymphoprotease was more or less confirmatory of assumptions made
earlier by MetchnikofL It seemed probable that the acid protease
noticed by Opie is specific to the large mononuclear cells of such
exudates and is not common to lymphocytes in general, for the con-
sensus of opinion of other workers seems to be that the small lympho-
cytes contain no proteolytic enzyme whatever. The study of tubercu-
lous exudates particularly has failed to reveal proteolytic properties
when only the characteristic small lymphocytes were present, al-
though some writers, Bergell 5 especially, have shown that these cells
contain a fat-splitting enzyme. Such lipase could also be determined
in the press-juice of lymphoid organs, such as the spleen and lymph
nodes. It is not improbable that the characteristic dry abscess or
1 See Wiens. Ergebnisse d. Allg. Pathol. Lubarsch & Ostertag, Vol. 15,
1911.
2 Leber. Entstehung der Entzundung. Leipzig, 1891.
3 Fr. Miiller. Verh. d. 20 Kongress f. innere Mediz., Wiesbaden, 1902.
4 Opie. Journal of Exp. Med., Vols. 7 and 8, 1905 and 1906.
5 Bergell. Munch, med. Woch., 1909.
526 INFECTION AND RESISTANCE
"caseous" abscess accompanying typical tuberculous lesions owes its
peculiar histological characteristics to the lack of proteolytic
enzymes.
Jochmann 6 and his collaborators have extensively studied leuco-
cytic extracts in this regard and, curiously enough, found that the
proteolytic enzymes of leucocytes of which we have been speaking
can be found only in the cells of man and monkeys and, to a slight
degree, in dogs. In these species only, according to Pappenheim,7
do the leucocytes contain true neutrophile granules and it is there-
fore a possibility that the neutrophile granules and the enzyme action
are related to each other. The leucocytes of rabbits do not ap-
parently contain protease, but recently in our laboratory Mrs. Parker
and Miss Francke have shown that rabbit leucocytes contain erepsin.
The proteolytic enzymes of leucocytes curiously enough continue
their activity at temperatures as high as 55° C., a fact which makes
it possible to investigate their activity at temperatures at which most
bacteria will no longer grow and functionate, and this incidentally
facilitates sterile experimentation. Their activity is astonishing in
that Jochmann found instances where dilution even as high as 500-
fold with salt solution did not completely eliminate the proteolytic
activity of pus. The simplest method of demonstrating such action
is to place pus or washed leucocytes, in droplets, upon the surface of
plates of Loeffler's coagulated blood serum, such as that used in the
cultivation of diphtheria bacilli. On such plates, small indentations
rapidly give evidence of the liquefaction of the coagulated protein.
Casein also can be used as an indicator of proteolytic digestion,
a casein solution being made by dissolving a gram of casein in 100
c. c. of N/10 NaOH solution and neutralizing this to litmus with
deci-normal HC1. A very curious property of the leucocytic ferment
has been described by Jochmann and Ziegler, who reported that pres-
ervation in 10 per cent, formalin solution will long preserve the fer-
mentative activities of the cells.
It was formerly supposed that the proteolytic properties of
leucocytes were more or less pathological. But we have since learned
that these powers are common to leucocytes both in health and in
disease. The older idea was due to the fact that the earliest investi-
gations of the ferments were made in connection with myelogenous
leukemia. Jochmann and Ziegler studied the organs of patients
dead of this disease and found that the bone-marrow, spleen, and
lymph nodes of such cases had powerful proteolytic properties in
contrast to the very weak activities in this regard of normal organs.
The proteolytic activity was more or less in direct proportion to the
degree of myelogenous infiltration. In lymphatic leukemia no such
6 Jochmann. For summary of his work and literature see Kolle u*
Wassermann, 2nd ed., Vol. II, 2.
7 Pappenheim. Cited from Wiens, loc. cit.
SERUM ENZYMES 527
activity was discovered. In consequence there has been a tendency
to draw a fundamental physiological distinction between these types
of cells, a distinction which would have considerable importance
in discussions of their origin and significance.
Jochmann has an idea that the activities of these ferments in
connection with tissue destruction and other pathological conditions
where they become active, may be an important phase in the produc-
tion of fever.
An important question which immediately arises is whether these
ferments can be identified with the bactericidal substances described
in a preceding chapter as existing within leucocytes. According to
Jochmann the enzyme extracts have no bactericidal properties. In-
deed, it is a curious fact that living bacteria oppose a very powerful
resistance to digestion by these and other ferments. Kantorowicz,8
who has studied this particularly, has shown that living bacteria
contain a very powerful antiferment which is similar to the anti-
ferment presently to be described for blood serum. He has shown
that living bacteria cannot be digested by trypsin, a resistance which
is lost when the bacteria are heated to 80° C., and an extract of
living bacteria will prevent the tryptic digestion of the heated bac-
teria. It will appear, therefore, that in the process of phagocytosis
the bacteria are first killed by the bactericidal substances contained
in the cells and are later digested by the leucoprotease.
That the lymphocytes contain lipase has been mentioned above
and since these cells are specifically accumulated about tuberculous
foci it has been many times suggested that their function is particu-
larly directed against these acid-fast bacteria in whose constitution
the presence of waxes and fats plays such an important role. It is
a fascinating thought, though entirely conjectural, that perhaps the
specific benefit of feeding fats in tuberculosis may have some basis in
the possible increase of lipolytic ferments which appear in response
to the stimulation of introducing larger quantities of fats.
It is hardly necessary to reiterate here the possible importance
of these cellular enzymes in the many different phases of the ab-
sorption of larval organs which occurs in the lower animals in the
course of normal development, or of dead tissue in mammalia in
disease and in the processes of senescence.
In addition to proteolytic enzymes, it has been long known that
the leucocytes also contain an oxidase. This enzyme can be demon-
strated by the well-known guaiac test, in which tincture of guaiac is
added to leucocytic exudates or pus, and a blue color results as a
consequence of oxidation of the guaiac. The oxidizing ferment is
apparently limited to the polynuclear leucocytes. Exactly what its
significance is we do not know but it is highly probable that it plays
an important part in the intracellular metabolism of the leucocyte.
8 Kantorowicz. Munch, med. Woch., p. 897, Vol. 56, 1909.
528 INFECTION AND RESISTANCE
It has only recently become clearly apparent that the function of
intracellular digestion of foreign substances, such as invading bac-
teria, is not limited to the circulating leucocytes but is to a very
large extent also carried on by the fixed cells of organs. Studies
by Wyssakowitz,9 by Kyes,10 by Bartlett,11 and, recently, by Hopkins
and Parker,12 have shown that in the course of many bacterial infec-
tions the greater proportion of invading bacteria is taken care of
by cells of the liver, spleen, and especially the lung. Streptococci
injected into rabbits rapidly disappear from the circulation, largely
because they are taken up very actively by the (probably) endothelial
cells of the lung where apparently they are killed, their digestion
taking place within the cells very slowly. The endocellular enzymes
bringing this about can, of course, not be separately studied as can
those of leucocytes, for obvious technical reasons.
While the cellular enzymes have been very carefully studied, it
is only of relatively recent years that much attention has been paid
to the enzymes present in the circulating blood. This is to some ex-
tent due to the fact that German writers especially have taken for
granted that enzymes in the blood were nothing more or less than
liberated leucocytic enzymes, and also because the activity of the
circulating enzymes has been largely masked by the existence of a
powerful antiferment which must needs be there for physiological
reasons to prevent injurious autodigestion. We have already men-
tioned the fact that the study of pathological conditions, such as
myelogenous leukemia, was the point of departure for work upon the
leucocytic ferments, and this, to a certain extent, was also the be-
ginning of the studies on serum ferments. When blood serum is
incubated with various substrates, it is not unlikely, as we shall see
in our subsequent discussion of the Abderhalden reaction, that the
antiferment is absorbed and thereby the proteolytic and other fer-
ments of the blood are liberated. There is normally an excess of
antiferment in the blood, and normal human serum usually does not
contain strong proteolytic enzyme. These two factors together there-
fore prevent powerful proteolysis by normal serum. There are con-
ditions, however, under which the enzyme contents of the blood are
increased. Thus, the sera of pregnant women have been shown
to be relatively rich in proteolytic properties, and such an increase
is also present during starvation, as Schultz,13 and Heilner and
Poensgen 14 have shown. The same thing has been noted in cases
9 Wyssakowitz. Zeitschr. f. Hyg., Vol. 1, p. 1.
10 Kyes. Journal of Inf. bis., 1916, Vol. 18, p. 277.
11 Bartlett. Journ. Mcd. Res., Nov. 5, 1916, p. 465.
12 Hopkins and Parker. In press at present writing.
13 Schultz. Deutsch med. Woch., No. 30, 1908; Munch, med. Wocn.. Vol.
60, 1913.
14 Heilner and Poensg-en. Miinch. med. Woch., Vol. 61, 1914.
SERUM ENZYMES 529
of pneumonia, in which condition Falls 15 has shown a sudden dimi-
nution of this enzyme at the time of crisis, and the same author has
made analogous observations in the fluctuation of serum protease in
malaria, typhoid fever, and some other diseases.
The lipolytic activity of the serum has been found increased in
syphilis, in diseases involving liver function, such as chloroform
poisoning, and in a number of other conditions. In syphilis, indeed,
the increased lipase contents have been held responsible for the Was-
sermann reaction by a number of writers. They believed that the
lipolytic ferment acting upon the lipoid antigen produced fatty acid
which by increasing the H-ion contents rendered the complement
inactive. This to a certain extent was strengthened by the knowledge
that in the Wassermann reaction the end-piece of complement re-
mained free. However, we ourselves in hitherto unpublished experi-
ments have been able to convince ourselves that there is no relation-
ship between lipase contents and Wassermann positiveness.
Activity of intravascular enzymes of any kind may perhaps be
going on constantly in the normal course of life as a part of general
body metabolism ; yet it is probable that any great extent, especially
of proteolytic action, would be incompatible with health or even
life. The work of Vaughan, of Friedberger, and others has shown
us that in the course of proteolysis toxic cleavage products, probably
of the albumose variety, are liberated. This we have seen has been
made the basis of some theories of anaphylaxis, and the "split prod-
ucts" of Vaughan, the "anaphylatoxin" of Friedberger, the "proteo-
toxins," as we ourselves have chosen to call them, and the "sero-
toxins" of Jobling and Petersen, all are probably the results of such
proteolysis. Indeed, as we have seen, the incubation of almost any
tissue substrate with fresh serum will result in this proteolytic-
change and even in the formation of toxic substances. It was at first
thought that the toxic split products were derived from the substrate,
but we have seen in another chapter that the mechanism is probably
one in which antiferments are removed by the substrate, and the
serum-protease is liberated to act upon the serum protein itself.
This was undoubtedly the case in such instances as those in which
anaphylatoxin was obtained by digestion of serum with kaolin,
barium sulphate, tubercle bacilli, etc. Now it appears that the
injection of almost any substance, and especially of bacterial pro-
teins, produces a sharp rise of enzymes in the blood. This seems to
be true not only of proteolytic enzymes but of lipolytic and amylolytic
ferments as well. Ferments so stimulated are entirely, non-specific.
This greater serum activity after protein injection may be due both
to a direct increase of production and a simultaneous diminution of
antiferments, by means of which activity hitherto inhibited is al-
lowed to run riot. The result may be in some respects beneficial
15 Falls. Journ. of Inf. Dis., Vol. 16, p. 466.
530 INFECTION AND RESISTANCE
and in other respects injurious. The marked increase of the enzymes
may aid in rapidly disposing of the unchanged foreign substance,
and perhaps the beneficial effects following the injection of typhoid
vaccines into patients suffering from this disease may be due to the
mobilization of ferments as suggested by Jobling and Petersen.16
On the other hand, the proteolytic activity may result in protein
cleavage, by means of which albumoses, etc., are liberated in greater
or smaller quantities which act injuriously and lead to clinical
symptoms of various kinds. That fever may perhaps be explained
by poisoning with albumoses so produced has already been suggested
by Jochmann, and that albumoses are present in various products
of suppurative inflammation has also been shown. It has been shown'
by Pfeiffer and Jarisch,17 and others, that fluctuations in proteolytic
enzymes accompany anaphylaxis. Jobling and Petersen18 have
formulated a theory of anaphylaxis based upon studies of serum
protease. They observed that during the course of sensitization
there occurred a gradual mobilization of non-specific protease which
increased in intensity up to the time of maximum sensitization.
They attributed acute shock to the fact that on the second injection
there occurs an instantaneous mobilization of large amounts of non-
specific protease together with a decrease in antiferment and an
increase in non-coagulable nitrogen and amino-acids, and they be-
lieve that the cause of the acute intoxication is the rapid cleavage of
the serum protease. The specific element they believe consists in the
rapid mobilization of the ferments and the colloidal serum changes
which bring about the change in antiferment titre. Fascinating as
this theory is, and although it has a number of things in common with
our own ideas concerning the colloidal balance in the plasma which
ordinarily prevents the rapid union of antigen with antibody, we
feel that recent knowledge concerning the essentially intracellular
nature of anaphylaxis in the guinea pig prevents its adoption.
Furthermore, we believe that the specific element of anaphylaxis is
insufficiently explained by Jobling and Petersen' s conclusions.
Again the question arises: Are these serum proteases in any
way to be identified with bacteriolytic antibodies or with alexin or
complement ? It is more than likely that no such relationship exists.
In the first place, the seroproteases are non-specific, and it has been
shown that in the course of bacteriolysis no increase in non-coagulable
nitrogen occurs. Moreover, Jobling and Petersen,19 to whom we
owe much of our recent knowledge on this subject, have shown that
the treatment of bacteria with complement alone or with complement
together with human serum renders them more resistant to proteoly-
16 Jobling and Petersen. Journal of the A. M. A., 1915, Vol. 65, p. 515.
17 Pfeiffer and Jarisch. Zeitschr. f. Imm., Vol. 16, 1912.
18 Jobling and Petersen. Journ. of Exp. Med., 1914, Vol. 20.
19 Jobling and Petersen. Journ. of Exp. Med., Vol. 20, 1914, p. 321.
SERUM ENZYMES 531
sis, probably owing to the absorption of antiferments from the blood
serum. The identification of complement with lipase has been sug-
gested. The idea gains likelihood from the fact that both complement
and lipase are destroyed by lipoid solvents and that many of the
substances upon which complement acts, such as red blood cells,
contain lipoid constituents. On the other hand, no definite knowl-
edge is available in this regard, and it has been so far impossible
to prove even the enzymotic nature of complement, though this, at
least, seems likely. The peculiar resistance exhibited by bacteria to
digestion by ferments has already been alluded to.
It develops therefore that the antiferments in the blood are ex-
tremely important factors in maintaining the balance of enzyme
activity, and in consequence the antienzymatic properties of serum
have been investigated by many workers. For, depending upon their
fluctuation, cleavage processes are permitted or prevented from tak-
ing place in the circulating blood. The earliest workers concerned
themselves largely with the phenomenon that blood serum would
prevent the proteolytic activity of leucocytes. They utilized the ob-
servation therapeutically by injecting serum into suppurating ab-
scesses for the purpose of preventing tissue destruction by leuco-
protease and increasing the ability of the body to limit the processes.
Indeed it may be that this form of therapy in chosen cases may still
have possibilities.
Later investigators studied antiferment fluctuations in disease.
In septic conditions, Wiens 20 claims to have noticed a diminution
of serum antiferments. Similar observations have been made after
shock in anaphylaxis by Pfeiffer and Jarisch. In patients with can-
cer, Landois21 claimed to have noticed a marked increase in anti-
ferments, which for a time was believed to be sufficiently regular to
have diagnostic significance. Brenner 22 noticed the same phe-
nomenon in severe anemias. In cachexia the same phenomenon has
been noticed and fluctuations in this constituent have been noticed
in a great many different diseases, such as diabetes, pneumonia, etc.,
without our being at present in possession of any definite or corre-
lated knowledge of the principles upon which these fluctuations de-
pend.
Earlier observers believed that the anti-ferment in the blood was
a function of the albumin fraction of the serum. More recent work-
ers have attributed it to lipoid constituents. Schwarz,23 Sugimoto,24
and Kirchheim suggested the lipoid idea because of the fact that
20 Wiens. Munch, med. Woch., No. 53, 1907.
21 Landois. Berl. klin. Woch., No. 10, 1909.
22 Brenner. Deutsch. med. Woch., No. 9, 1909, and Mediz. Klinik, No. 28,
1909.
23 Schwarz. Wien. klin. Woch., 22, 1909.
24 Sugimoto. Arch. f. Ex^. Path. u. Pharmakol, Vol. 74, 1913.
532 INFECTION AND RESISTANCE
the antiferment properties were destroyed by lipoid solvents. They
did not settle the question satisfactorily because their results were in-
consistent, but Jobling and Petersen25 found that antiproteolytic
activity of blood serum was almost wholly inhibited when sera were
extracted for several days with chloroform or ether at room tempera-
ture or with chloroform in the incubator for an hour. The anti-
ferment could be recovered from the ether and chloroform solution,
and its action could be destroyed by incubation with iodin or po-
tassium iodid. They believed this to be due to the saturation of
the free carbon atoms. Furthermore, the same authors showed that
toxic substances analogous to anaphylatoxins could be produced in
serum when incubated with chloroform. The lipoidal nature of the
antiferment s is a likely one although it does not account for the fact
noticed uniformly by all workers that the antitryptic activity can
be entirely eliminated by heating the serum to 70° C. for a half hour.
This was the observation which at first gave rise to the idea that the
antiferment was in itself an enzyme. For this phenomenon of heat
lability no satisfactory explanation has as yet been advanced.
The recent researches of Abderhalden 2G upon the intravascular
digestion of foreign substances introduced into animal bodies prom-
ised to have considerable bearing upon problems of immunity.
Abderhalden, whose work we cite chiefly from his monograph, "Die
Schutzfermente des tierischen Organismus," took as his point of de-
parture the conception that the animal body must necessarily preside
over a mechanism whereby it can assimilate foreign substances which
obtain entrance unchanged into the circulation. Abderhalden be-
lieves that this process depends upon the mobilization of "protec-
tive ferments/' a term which he borrows from Heilner,27 and sug-
gests the possibility that these ferments may originate in the leuko-
cytes.
Experimentally Abderhalden approaches his problem by deter-
mining the presence of specific ferments in the blood of animals into
which various foreign substances have been introduced by paths
other than the alimentary canal. For this purpose he has developed
a number of methods, the most important of which are his optical
method and his dialysis method. The optical method used for the
determination of the proteolytic properties of the serum depends
upon the fact that many of the amino-acids are optically active.
Moreover, most of these substances are chemically known and their
optical activity determined, so that it is possible to take blood serum
which is to be examined for its contents of particular ferments, mix
25 Jobling and Petersen. Journ. Exp. Med., 1914, Vol. 19, p. 480.
26 Abderhalden. "Schiitzfermente des tierischen Organismus," Springer,
Berlin, 1912.
27 Heilner. Cited from Abderhalden Zeitschr. f. Biol, Vol. 50, 1907.
SERUM ENZYMES 533
them with a suitable protein, or preferably a polypeptid, and de-
termine with a polariscope the rotation which takes place. We will
not go into the technique of this method more extensively because
we have no personal experience with it, and the method is one of
such delicacy that it is best obtained from Abderhalden's original
publications directly.28 His dialysis methods depend upon placing
the blood serum and fermentable substance into dialyzing bags, sus-
pending them into distilled water, and determining the presence of
peptone, amino-acids, or total nitrogen in the liquid outside of the
bag after definite intervals of time.
By these and other methods Abderhalden 29 has carried out tests
with a large number of different substances. Experimenting first
with proteins, he injected egg albumen, horse serum, silk peptone,
gelatin, edestin, casein, etc., into dogs and rabbits, then, several days
later, bled the animals and mixed 0.5 c. c. of the serum with 0.5 c. c.
of a solution of the respective substances which had been injected.
He found in such cases that definite proteolytic action was exerted
upon the injected substances by the active serum of a treated animal,
whereas, in the case of most of the substances used, the normal serum
possessed no proteolytic action whatever. These results were con-
sistently obtained both by the dialysis and by the optical methods.
It should be especially noted that the ferments studied by Abder-
halden were not as specific as are the antibodies which we have dis-
cussed in another place. For Abderhalden found that the serum of
an animal treated with proteins developed enzymes which were active,
not only against the particular protein used for injection, but rather
against proteins in general. They were specific only in that, when
produced with proteins, they were not active against fats or carbo-
hydrates. This is especially important in connection with the recent
discussion concerning the identity of Abderhalden's protective fer-
ments and the specific protein antibodies.
In later experiments Abderhalden claimed to show further that
similar ferments could be induced in animals by treatment with car-
bohydrates and with fats. The serum of normal dogs is not capable
of splitting cane sugar. However, the blood serum or plasma of a
dog that has been treated with cane sugar develops the property of
inverting the cane sugar into dextrose and fructose within fifteen
minutes after injection. This could easily be determined both by
putting together the serum with cane sugar and determining the
increase of reducing powers, and by means of subjecting such active
plasma or serum, together with saccharose, to polariscopic examina-
tion.
28 See especially Abderhalden, Hoppe-Seyler, Zeitschr. f. pTiysiol. Cliemie,
Vols. 60, 65, and 66; also "Handbuch der Biochem. Arbeitsmethoden," Vol.
5, p. 575, 1911.
29 Abderhalden. "Schiitzfermente," p. 49.
534 INFECTION AND RESISTANCE
The earlier experiments with fats were negative because the
simple method of titration for fatty acids proved insufficient as an
indicator of activity. However, Abderhalden succeeded in determin-
ing fat splitting properties in the blood of treated dogs by using the
method of Michaelis and Rona.30 The presence of fats largely in-
creases the surface tension of mixtures, and their cleavage in such
mixtures consequently leads to reduction of this tension. Utilizing
this principle, Abderhalden claims to have determined that the paren-
teral introduction of fats into dogs is followed by a reactionary in-
crease of lipases.
The general significance of Abderhalden's researches is this:
When any foreign substances, protein, carbohydrate, or fats, gain
entrance to the circulation of an animal, the animal body reacts by
the mobilization of ferments or enzymes specifically capable of re-
ducing these substances to assimilable form. It is likely that these
ferments represent a mobilization of substances normally present but
not concentrated in the blood stream under ordinary conditions, since
they appear with a speed out of all proportion to that obtaining in
the case of the antibodies discussed in another place. In one case
cited by him a dog injected on November 25th, 29th, and December
4th showed powerful peptolytic serum properties on December 6th.
Apparently the injection of homologous proteins into animals (i. e.,
rabbit serum into rabbits, etc.) does not incite reaction.
These enzymes seemed to differ from specific antibodies in that
they did not react solely with the substance injected, but also with
other substances belonging to the same chemical group. Other dif-
ferences from antibodies are the rapid appearance of the ferments
after treatment and their rapid disappearance after the inciting
stimulus is removed. Thus Abderhalden reports that the enzymes
found in a case of pregnancy disappeared within eight days after
abortion or child birth.
It is plain that these researches of Abderhalden offer many op-
portunities for diagnostic utilization, and he has applied them to the
diagnosis of pregnancy. In this condition substances from the
chorionic villi get into the blood. These, according to Abderhalden,
may be looked upon as in a certain sense foreign in nature, and must
be chemically disintegrated by the body. In consequence it is likely
that the ferments which accomplish this would appear in the sera of
pregnant individuals and could be determined by his methods.
"When he prepared peptone from the placental substances of human
beings and allowed the blood plasma of normal individuals to act
upon it, observing it both by the dialysis and the optical method, no
peptolytic action could be observed. However, when the plasma of
pregnant women was used proteolytic action was determined. In
these cases the ferment seemed to be specific for peptones produced
30 Michaelis and Rona. Cited from Abderhalden, loc. cit.
SERUM ENZYMES 535
from placental tissue both in animals and human beings, but did not
act upon casein, gelatin, or other proteins. There are certain techni-
cal difficulties connected with the production of a test material from
the placental tissue which render this method difficult. For their
more detailed description we refer the reader to the original articles.
The Abderhalden reaction and its various logical ramifications
promise for some time to bring an entirely new and important factor
into the field of immunology, namely, that specific ferment mobiliza-
tion in definite response to the introduction of various substances,
and yet distinct from the specific antibody heretofore known. The
reaction rapidly became the subject of active research mostly of a
clinical nature, and carried out with insufficient critical judgment.
This literature is extensive and, because almost entirely refuted at
the present day, is hardly worth citing. In this country the re-
action has been carefully studied by a great many workers, more
especially by Bronfenbrenner, and Jobling, Petersen and Eggstein.
From all these studies, we may say for the sake of brevity, it has
become clear that the enzymes involved in the Abderhalden reac-
tion are probably not specific as supposed by this writer, that during
the reaction the tissues employed (that is placenta, etc.), act by
absorbing antienzymes from the serum. In consequence the fer-
ments in the serum act upon the serum protein itself which there-
fore becomes the substrate of the cleavage product determined as a
result of the reaction. This robs the reaction of any claim to the
specific and diagnostic value attributed to it by Abderhalden and his
collaborators, and as a final result of these extensive and intricate
researches we have merely a better understanding of the non-specific
enzyme activity of normal serum.
SOME OF THE PHYSICAL FACTORS WHICH ENTER, INTO
SERUM REACTIONS
In a number of discussions in preceding chapters we have em-
phasized the fact that serum reactions are gradually being recognized
as having more in common with so-called colloidal phenomena than
with those taking place during the union of crystalloids. The funda-
mental physical principles underlying colloidal reactions are as yet
pretty vague and the serologist especially (ourselves included), when
he speaks of colloidal reactions is often groping in the dark. Never-
theless, even without such fundamental understanding we can recog-
nize the close analogies which exist between the various serum re-
actions and those taking place between substances recognized by
their attributes as colloids. We have sufficiently discussed this in
the chapters on agglutination and precipitation, and need not en-
large upon it here except in indicating how the gradual tendency to
pay attention to the physical factors involved has led to the actual
536 INFECTION AND RESISTANCE
working out of serum experiments in which the reasoning from the
beginning and the technique finally developed were physical rather
than chemical. Perhaps the earliest phases of such work were those
in which complement fixation was shown to occur when complement
was brought together, with non-specific inorganic and organic
colloids. It is due chiefly to the work of Landsteiner and his
associates that attention was called to such phenomena. Landsteiner
with Stankovic showed that complement may be fixed by silicic acid,
and Seligmann soon afterwards found that in the precipitation of
mastic suspensions complement may be fixed. Much work has been
done by others, as well as by ourselves, that shows that a variety of
colloidal suspensions of known chemical constitution fix complement,
probably by adsorption. Perhaps the most striking instance of such
complement adsorption is that occurring in the Wassermann re-
action. Here, as we know, complement is fixed by a combination of
syphilitic serum and various lipoidal suspensions, which may be en-
tirely non-specific in origin. When the reaction occurs, as Jacob-
stahl and, later, Bruck have shown, a precipitation occurs which
can be seen in the ultramicroscope. Furthermore, this precipitation
takes place more rapidly in the ice-chest than in the incubator, a
principle upon which is based the so-called refrigerator method of
performing the test, and which strikingly suggests that the reaction
is an adsorption rather than a true chemical union. It is this pre-
cipitate which fixes the complement ; whether or not this is due to
quantitatively increased globulin or to purely physical change in the
syphilitic serum, is a matter which we cannot discuss at present.
Whatever it is, it is unquestionable that the availability of the anti-
gen for the Wassermann reaction depends not only upon its lipoidal
nature but also on its state of dispersion. Since it is not possible,
as we have found, to make available antigens for this reaction with
non-lipoidal substances, like mastic, gelatin, gum arabic, salicic acid,
albumins, and a number of other substances, even when in dispersion
more or less similar to that of the Wassermann antigen, it seems that
the secret of the Wassermann antigen must lie in the fact that sub-
stances of the chemical and physical constitution of lipoids when
brought into a definite state of dispersion offer surface tension con-
ditions not easily obtained with colloids of another nature. It is,
therefore, at least in our opinion at present, the physical condition
of the Wassermann antigen which makes it available for the test, a
physical condition which is secondarily dependent upon the chemical
nature of the dispersed substance. The importance of the state of
dispersion and therefore the surface tension properties is quite ap-
parent from the fact noted by many Wassermann workers that a con-
siderable difference in the fixing power of the antigen may be ob-
tained by in one case adding the salt solution to the alcoholic ex-
tract quickly, and in another case adding it very slowly, the two
SERUM ENZYMES 537
separate preparations showing, one a very transparent, the other a
very turbid, condition.
The Bordet-Danysz phenomenon has already been discussed, but
is another case in point. It consists of the fact that if to a definite
amount of antitoxin a definite quantity of toxin is added, the re-
sult is one of greater toxicity in the mixture if the poison is added
to the antitoxin fractionally than when the entire amount is added
at once. An interesting difficulty of this phenomenon is the fact that
such a reaction seems to force upon us the assumption that the toxin-
antitoxin union is not reversible, whereas later work on immuniza-
tion with neutralized mixtures of these substances seems to necessi-
tate the assumption that within the body they are reversible.
It is a consideration of these and many other apparently physi-
cal factors of immune reactions which has led experimenters to seek
to utilize physical changes for practical serological purposes. One
of the reactions resulting from such studies is that known as Weich-
hardt's Epiphanin reaction. Weichhardt noticed that when two solu-
tions of toxin and antitoxin are brought in contact with each other on
exactly horizontal glass plates, diffusion takes place between them
much more rapidly than in controls in which one or the other, antigen
or antibody, was lacking. He tested this at first on the horizontal
glass plates by adding coloring matter to the two solutions and ob-
serving its diffusion currents. Later he developed a method in which
he tried to determine such increased diffusion by means of a delicate
chemical balance. His method, briefly described in his own words,
was as follows : "To the two arms of scales, to the right and left, a
little bell-shaped jar is attached into which dilutions of serum are
placed, on the one side specific immune serum, on the other normal
serum. These little bell-jars are closed at the bottom with
'Schleicher-SchulP filter paper and are immersed into solutions con-
taining antigen in salt solution. Through the filter-paper mem-
brane diffusion takes place, and in observations carried on for from
several hours to several days it can be determined that the scales
sink on that side in which specific immune serum had been placed
into the little bell-jar. In other words, the antigen solution which
is contained in the salt solution diffuses more rapidly on the side on
which the specific immune serum was present. In consequence, the
weight of the little bell increases and a definite reading can be made."
Similar observations have been made by other observers, Kraus and
Amiradzibi doing the experiment as follows : they employed a U-tube
in the connecting horizontal part of which there was a glass stop-
cock. Into one side they put diphtheria toxin with a little aqueous
methylene blue, and into the other side antitoxin. For each experi-
ment two controls were set up, one with salt solution and the other
with normal horse serum, and they noticed that after a definite period
of time after the stopcock connecting the two had been opened,
538 INFECTION AND RESISTANCE
methylene blue could be observed to diffuse across in the tube in
which antigen and antibody had been present and not in the con-
trols. Many modifications of technique to demonstrate this principle
have been made by Weichhardt himself, and since the reaction is not
likely to find much immediate diagnostic application because of its
great delicacy, we need not describe them at length but refer the
reader to Friedemann's excellent description in the Kolle and Wasser-
mann, second edition, Vol. III.
Similar in principle but by no means identical is the so-called
Meiostagmin reaction, chiefly developed by Ascoli.
THE MEIOSTAGMIN REACTION
Ascoli and Izar 31 have attempted to work out a diagnostic reac-
tion which depends upon an alteration of surface tension of a fluid
when an antigen unites with its specific antibody. Ascoli in his first
experiments worked with typhoid bacillus extracts and the sera of
typhoid patients, and found that when the two suspensions were
mixed a reduction of surface tension resulted after time for union
between the two had been allowed.
They determined the reduction of surface tension by Traube's 32
method by the use of apparatus spoken of as the "stalagmometer."
The principle of this method depends upon the fact that as surface
tension is reduced the number of drops to a given quantity of fluid
is increased.
Diluted serum of patients was mixed with diluted antigen, and
the number of drops contained in one cubic centimeter of the mix-
ture was immediately determined and again measured after the mix-
ture had remained for two hours in the incubator at 37° C. An
example of one of Ascoli' s early measurements is given in the protocol
cited below.
Of course a certain amount of reduction of surface tension results
when various antigens are brought together with normal sera, but
this can be easily controlled by suitable dilution, and must be care-
fully taken into consideration in each individual case. Ascoli and
Izar have applied this method to the diagnosis of tuberculosis, ty-
phoid, and various other diseases, and have reported what seemed to
them reliable results. So far experience with the meiostagmin reac-
tion has not been very extensive ; not all observers have been able to
obtain results as apparently reliable as those of Ascoli and his col-
laborators. It is not possible therefore to express a final opinion
regarding this method of investigation ; it contains, however, an in-
teresting principle which with more exact methods of measurement
may well become very important in serum diagnosis.
31 Aseoli and Izar. Munch, med. Woch., Nos. 2, 7, 18, 22, 41, 1910.
32 Traube. Pfliigcr's Archiv, Vol. 123, 419.
SERUM ENZYMES
1 c. c. of serum of typhoid patient diluted to 1-10.
Number
of drops
Immedi-
ately
After
2 hours in
incubator
' 1 0/00
57.8
58.1
59.7
59.9
1 c. c. alcoholic typhoid extract diluted to — ,
1 0/000
57.5
57.5
59.4
59.6
1 0/000
57.0
57.0
59.3
59.2
" 1 0/00
58.1
57.7
59.7
59.6
1 c. c. alcoholic typhoid extract diluted to .
1 0/000
57.4
57.6
59.4
59.2
1 0/000
57.0
56.9
59.2
59.4
" 1 0/00
56.5
56.5
58.0
57.8
1 c. c. alcoholic precipitate taken up hi distilled ^
1 0/000
56.5
56.6
57.5
57.4
1 0/0000
56.7
56.5
57.4
57.5
1 c. c. in 1 0/00 alcohol in 1 c. c. 0.85 per cent.
NaCl solution
56.6
57.5
56.7
57.6
COLLOIDAL GOLD REACTION
An important reaction, depending entirely upon the principle of
colloidal precipitation, is the so-called colloidal gold reaction exten-
sively used for the diagnosis of syphilis of the central nervous sys-
tem. In 1901, Zsigmondi, knowing that protein substances will pro-
tect various colloids from precipitation by electrolytes, attempted to
work out a method by means of which he could quantitatively esti-
mate protein in solution by its protective effect. He worked with
colloidal metals, using especially colloidal gold. He determined what
he called the "Goldzahl" for various proteins, by which he meant the
amount in weight of a protein which was just enough to protect from
precipitation 10 c.c. of colloidal gold, of a percentage of 0.0053, in
the presence of 1 c.c. of 10 per cent. NaCl. It, of course, had long
540 INFECTION AND RESISTANCE
been known that in various pathological conditions of the central
nervous system the protein contents of the fluid varied. Thus, the re-
actions of Nonne, of Noguchi, etc., were all aimed at revealing an ab-
normal globulin content of the spinal fluid. Lange, attempting to ap-
ply Zsigmondi's method to the quantitative detection of proteins in the
spinal fluid, observed a curious reaction, quite the opposite of what
he had set out to find. He observed that spinal fluid, especially of
syphilitic cases, in which there was protein material beyond the nor-
mal, precipitated rather than protected colloidal gold. He also ob-
served that the dilution at which such precipitation took place was
more or less constant in syphilitic cases and might therefore be
utilized for the diagnosis of such diseases as paresis, locomotor ataxia,
etc., etc. The reaction was taken up and carefully studied by a
large number of observers, among whom were Zaloziecki, Jaeger and
Goldstein, Flesch, Kafka, Lee and Hinton, Miller and Levy, and
many others. One of the most practical discussions and clear descrip-
tions available for American and English readers is that of Miller,
Brush, Hammers, and Felton. The chief difficulty of the test con-
sists in the preparation of a proper colloidal gold solution.
Lange adopted the method used by Zsigmondi with slight modi-
fications. He added 10 c.c. of a 1 per cent, solution of gold colloid and
10 c.c. of a 2 per cent, solution of potassium carbonate to a liter of
carefully and freshly distilled water. The solution was rapidly
warmed almost to a boil and just before boiling, while actively stir-
ring, 10 c.c. of 1 per cent, formalin solution is added. The solution has
remained colorless up to this point but upon this addition should al-
most instantaneously become a deep, transparent red. There should
be no iridescent or smoky "Schimmer." The utmost care in getting
pure distilled water must be exercised and Jena glass must be used
throughout.
Other methods are those of Eicke, who adds dextrose, and many
workers, such as Flesch, Kaffge, Eskuchen, and others, have found
great difficulty in making up the solution by Lange's technique or by
any of the other methods described. Miller and Levy state that
with Lange's method good solutions can always be obtained if the
water is absolutely pure and the glassware is satisfactorily cleaned ;
later, Miller, with the collaborators mentioned above, found that this
was not universally true. Also, these writers, as well as Eskuchen,
found that occasionally solutions which appeared all right did not
react, and special studies followed dealing with the technique of the
preparation of the gold sol. Miller and his collaborators find that
the reduction method of producing this solution is the best one, and
proceed as follows. They use the following reagents :33
33 The descriptions are taken direct from the thorough papers of Millr-
and Levy.
SERUM ENZYMES 541
1. The Gold Solution:
AuCl3 — Merck's yellow crystals hermetically sealed in brown
glass ampules 1 gram
Water triply distilled, up to 100 c. c.
This stock solution is kept well stoppered in dark glass bottles
away from any bright light.
2. The Alkaline Solution:
Merck's Blue Label Potassium Carbonate (desiccated) 2 grams
Water triply distilled, up to 100 c. c.
3. The Reducing Agents :
a. Formaldehyde, Merck's 40 per cent stock solution, highest
purity 1 c. c.
Water triply distilled, up to 40 c. c.
b. Oxalic acid, Merck's Blue Label, Crystals 1 gram
Water triply distilled, up to 100 c. c.
Solutions No. 2 and No. 3 must be made up immediately prior
to use.
4. Bichromate cleaner for glassware:
Potassium Bichromate, powdered 200 grams
Water, distilled, up to 1500 c. c.
Sulphuric Acid cone 500 c. c.
The potassium bichromate should be well dissolved before the
sulphuric acid is added. If this solution is reserved for cleaning
glassware only, it can be used repeatedly.
Great attention must be paid to the cleaning of glassware. They
boil their Jena beakers in Ivory soap solution, then brush under hot
tap water. After being rinsed for five minutes, hot bichromate and
sulphuric acid is added for half an hour. The beaker is then
emptied and washed in running water for five minutes, rinsed with
distilled water and then in triply distilled water. Similar methods
are used with other glassware. The chief errors are insufficient
brushing, failure to get all bichromate out, and allowing beakers to
dry in air before using.
In obtaining the distilled water, they first take ordinary distilled
water which they then re-distil from Jena flasks. They wash out
the collecting flasks with the first 200 c. c. of the second distillate,
then collect and re-distil in the same way, again rinsing with the
first 100 c. c. of the second distillate. After all these steps have
been performed the rest is more or less simple. A beaker rinsed out
with triply distilled water is filled to the liter mark and the tem-
perature raised to 50° C. gradually, then the gas is turned on full
and when the temperature has reached 60° C. 10 c.c. of the 1 per cent.
542 INFECTION AND RESISTANCE
gold solution and 7 c.c. of the 2 per cent, potassium carbonate solu-
tion are added. The solution should remain perfectly clear. At
80° C. while stirring with a clean thermometer, 10 drops of oxalic
acid are slowly added. The solution may now turn a delicate bluish-
pink, often due to an excess of alkali. Otherwise the solution re-
mains colorless until 90° C. has been reached. When 90° C. has been
reached, the gas is turned out and, while stirring, 5 c. c. of 1 per cent,
formaldehyde is added, drop by drop. If a pink color makes its
appearance before all the reducing agent has been added, Miller
advises to stop at once. He also states that the best solutions are
those in which the color change is slow. For further technical points,
the reader is referred to the paper of Miller and his collaborator in
the Johns Hopkins Hospital Bulletin, 1915, xxvi, 391, from which
this description is almost bodily taken.
The test is done as follows: into a clean test tube 1.8 c. c. of
fresh sterile 0.4 per cent. NaCl solution is placed. Into ten further
tubes 1.8 c. c. of the same salt solution is placed. To the first tube is
added 0.2 c. c. of spinal fluid to be tested and well mixed. 1 c. c, of
this 1 :10 solution is added to the second tube, mixed, and of this 1 c. c.
is added to the third tube, in consequence of which dilutions are ob-
tained running from 1:10, 1:20, 1:50. Now to each of these is
added 1 c. c. of the colloidal gold solution. Changes begin to take
place in five minutes, which usually reach their completion in about
a half hour. The spinal fluids must be free from blood and clear
from bacterial contamination.
Normal spinal fluids produce usually no reaction. The so-called
luetic curve is usually one in which the greatest precipitation occurs
in the tubes ending in 1 :40 to 1 :160. In suppurative lesions, etc.,
the strongest precipitation occurs in higher dilutions, say, from 320
to 280, which Lange has called "verschiebung nach oben." A so-
called paretic curve means complete precipitation in, say, from 110
to 160, with a gradual fading out toward the higher dilutions. A
complete precipitation means a colorless tube; partial precipitation
shades from pale blue to the complete red of the unaffected colloidal
gold. Fluids from early stages of syphilis without nervous system
involvement react usually like normal fluids.
That the reaction has unquestionably found a permanent useful-
ness cannot be doubted. It seems to be dependent entirely upon
the technique of making the colloidal gold solution.
CHAPTER XXI
COLLOIDS
BY STEWART W. YOUNG
Professor of Physical Chemistry, Stanford University, CaL
INTKODUCTOEY
IN attempting to give in the brief space of a single chapter any
adequate account of the present state of our knowledge in so vast a
field as that of colloid chemistry and physics one is confronted with
a rather difficult problem. In the present outline the attempt will
be made to get at some notion of the matter by a presentation first
of the more important generalizations which have been drawn, this
to be followed in each case by sufficient experimental evidence to
serve as illustration, together in some cases perhaps with certain
evidence which may seem to contradict in some degree such cur-
rent conceptions. The reason for this particular method of pres-
entation lies in the fact that new material is so rapidly accumu-
lating, much of which seems more or less at variance with
present accepted theories that it seems more than possible that
some of these fundamental generalizations may soon undergo ma-
terial modification if not reform. It would, therefore, seem ill-
advised, in presenting a brief resume to readers who are not
physicists or chemists, merely to present the present theories as
they are used to-day by workers in the field, and to sound a note
of warning that many, if not all of them, are not so securely
supported by broad evidence as to allow of very concrete prediction
being based upon them.
Definition. — .The fundamental distinction between the crystalloid
and colloid states of matter was first drawn by Thomas Graham 1 as a
result of his investigations into the phenomenon of dialysis. He
noted that in general those substances which when in solution did not
pass through the dialyzing membrane, or did so only very slowly, also
were characterized by the fact that when they separated from solu-
tion, either by precipitation or by evaporation, they did so in the non-
1 Graham. Phil. Trans., 1861, 183.
543
544 INFECTION AND RESISTANCE
crystalline or amorphous form. This class of bodies he named col-
loids, since glue (Greek KoXXa meaning glue) presented a typical case.
Colloid substances may appear in highly dispersed states, such as
dilute glue, arsenic sulphid suspensions, oil or rosin emulsions, milk
(casein in highly dispersed condition), and the like, in which case
they are spoken of as sols. They may also occur in the undispersed
or only slightly dispersed state, as the amorphous precipitated sul-
phids of the heavy metals, precipitated casein, or dry glue. In this
state they are spoken of as gels.
When a colloid substance has once been converted from the sol
or dispersed state into the gel or undispersed state, its properties
may differ greatly in different cases. Thus if a dispersed soap (soap
solution, or more correctly soap sol) be coagulated by the addition
of common salt, the coagulum or soap gel may be removed from the
salt solution, and if again placed in pure water it will redisperse
and again assume the sol condition. Such a colloid is spoken of as a
reversible colloid. If, however, an arsenious sulphid suspension be
put through precisely the same course of treatment it will, in the
last stage of the treatment, refuse to redisperse, and is therefore
spoken of as an irreversible colloid. Some authorities prefer to
speak of these two classes of colloids as emulsion and suspension col-
loids, respectively,2 since in general those colloids which are reversi-
ble tend to separate out in soft masses, and in general to gelatinize
rather than to flocculate, while the irreversible colloids rather tend
to truly flocculate and form very compact and frequently more or
less granular coagula. Since, however, we seem more likely at the
present time to suffer more from an excess of classification than from
a lack of it,. the attempt will be made to get along in this discussion
with the earlier nomenclature. It may, indeed, be added that
it is highly probable that the distinction between reversible and irre-
versible colloids is only one of degree. For example, many of the
metallic sulphids which are typically irreversible may be made
to some extent reversible by means of thorough washing and re-
treatment with hydrogen sulphid which had been originally used
in their preparation. It is probable that certain colloids are
apparently irreversible only because we do not truly reverse
the conditions.
Heretofore in this discussion the term "colloid" substance has
been used as if to imply that certain chemical individuals were
characteristically colloid, while others were not. It was much in this
sense that Graham used the term. Investigations since his time have
shown this to be a misconception, and it is now apparent that any
and all substances may be either colloid or crystalloid, the form they
assume depending upon treatment. Thus albumin may be crystal-
lized and common salt may be obtained in the state of a colloid solu-
2 V. Weimarn. Ztschr. Chem. Ind. Koll., 1908, 3, 26.
COLLOIDS • 545
tion or sol.3 Albumin, gelatin, and agar may be obtained crystalline
by proper regulation of temperature and the use of proper solvents,
as solutions of ammonium sulphate for albumin, and alcohol-water
mixtures of varying strengths for the two latter substances. Sodium
chlorid has been obtained in the colloidal condition by precipitating
it in a solution of sodium sulphocyanate by hydrochloric acid, each
of the reacting substances being dissolved in a mixture of amyl alco-
hol and ethyl acetate.
There is much evidence that leads to the belief that all colloid
systems are unstable. Van Bemmelen characterized them as systems
which never reached a state of rest, that is, were never in equili-
brium. The conditions which determine the appearance of a body
in the colloid or crystalline form lead to the suspicion that bodies
always separate from solution in the amorphous or colloidal condi-
tion and that all crystallization is a secondary phenomenon. The
conditions that are favorable for the transformation of a colloid into
a crystalline form are a considerable solubility and a considerable
rate of crystallization. Where either or both of these is at
a minimum the conditions are favorable for relative permanence
in the colloid condition. It is upon the basis of this prin-
ciple that von Weimarn succeeded in obtaining relatively stable
colloidal solutions of common salt and many other easily crystal-
lizable salts. Furthermore Doelter 4 has succeeded in converting
many wrell-known amorphous precipitates into crystalline bodies
by means of stirring, pressure, impact, and high temperature.
Among the substances thus transformed are aluminium, chromium,
and iron hydroxids, and the sulphids of arsenic, antimony,
and zinc.
With this much by way of introduction, we may now proceed to
a closer consideration of some of the better recognized properties of
colloid sols and gels. For convenience we shall first take up the
discussion of these systems from the more definitely physical point
of view, and later take up those properties which seem more defi-
nitely chemical.
Physical Properties of Colloids.— 1. FORM AND SIZE. — Current
opinion seems to be leading rapidly to the general acceptance of the
hypothesis that in liquid systems of two or more components we have
to do with a continuous series of conditions ranging from coarse sus-
pensions through suspensions of increasing fineness (increasing de-
grees of dispersion) to finally the molecular and ionic states of solu-
tion. The opinion is also growing that, although for certain practical
purposes the classification of all such systems in one way or another,
as in terms of the various degrees of dispersion, may be useful, the
8V. Weimarn. Ibid., 1910, 7, 92, and "Grundzii^e der dispersoid
Chemie," 107-108.
4Ztschr. Chem. Ind. KolL, 1910, 7, 86.
546 INFECTION AND RESISTANCE
excessive use of such classifications is likely to narrow rather than
broaden our conception of the whole subject matter of the field. It
would seem that the most stimulating point of view is to be reached
from the acceptance of the suggestion of Wolfgang Ostwald, that the
chief problem of colloid chemistry at the present time lies in deter-
mining the influences of the degree of dispersion upon the physical
and chemical properties of all liquid solutions, mixtures, suspensions,
or what-not. If this point of view be taken it follows that the form
and size of the particles in a disperse system are a matter of the first
importance.
If the degree of dispersion in a given system be not too great the
form of the particles may be readily observed under the microscope.
Such evidence shows that the spherical form predominates enor-
mously over all others, although under carefully controlled conditions
ovoid forms may appear, as in the case of gelatin and agar. These
ovoid forms are taken by von Weimarn (loc. cit.) and others, as evi-
dence of directive forces, and hence of incipient crystallization. If
the system be treated in such a way as to decrease the dispersion, as,
for example, if a reagent be added which tends to flocculate the col-
loid, but not in sufficient quantity to produce actual precipitation,
the decrease in dispersion may take place in two quite different
ways : first, the size of the particles may increase, as in the case of
oil emulsions; second, the particles do not coalesce but become at-
tached together in chains and groups which, in many cases, resemble
bunches of grapes. This sort of aggregation may go so far as to pro-
duce web-like structures. The jellying of gelatin has been shown to
be due to the development of such web structures. Glue shows much
less tendency in this direction, and if some acetic acid be added, as
in the preparation of commercial liquid glues, this web formation is
almost entirely absent, and the adhesive qualities are at the same
time greatly improved. It seems quite certain that both of the above
modes of aggregation are possible in one and the same system at
different stages in its condensation. Thus highly dispersed copper
sulphid becomes aggregated in its first stages of condensation by an
actual increase in the size of the spherical particles. After these
reach a certain fairly definite size further aggregation takes place
by the grouping together of these spheres. It is generally recognized
that all grouped and webbed structures are secondary.
A large number of very important investigations have been di-
rected toward the determination of the size of disperse particles
throughout the greatest variation in dimensions. The fact first
noted by Graham, that substances in colloidal solution show a very
small, and frequently almost negligible, rate of dialysis, points di-
rectly to the supposition that the particles in such solutions are in a
far less dispersed state than in solutions of crystalloid substances.
The rate of dialysis is directly determined by the rate of diffusion.
COLLOIDS 547
which, in turn, is inversely proportional to the square roots of the
masses of the diffusing bodies.
^Measurements of osmotic pressure in solutions also give an accu-
rate measure of the relative masses of dispersed systems where such
measurements can be successfully carried out, and a great deal of
work has been devoted to attempts to measure the osmotic pressure
of colloidal solutions. Great difficulties both of experimentation and
of interpretation are encountered in this field. As will soon be seen
a colloid particle stands in a very complex relationship to its sur-
rounding liquid, and furthermore it is a matter of extreme difficulty
to obtain a colloid solution free from electrolytes, which themselves
may create osmotic pressure or otherwise affect the measurements.
About the only conclusion which it is safe to draw at the present time
is that if colloid solutions show osmotic pressure at all the value of
it is very small compared to that shown by crystalloidal solutions of
substances of more or less like formula weights. This leads to the
conclusion that the particles in a colloid solution are in a state of
dispersion far less than that found in a typical crystalloidal solution.
For a most excellent resume of the present state of our knowledge in
this field the reader is referred to a recent book by Dr. L. Casuto,
of Pisa, entitled "Der Kolloide Zustand der Materie" (Steinkopf,
1913).
When the size of the disperse particles is sufficiently great they
may, of course, be measured under the microscope, and with the
advent of the ultramicroscope the limits of visibility of small bodies
has been very notably extended. The ultramicroscope is known in
several forms, the first having been devised by Siedentopf and Zsig-
mondy. All depend upon the production of powerful rays of light
in directions parallel to the surface of the microscope slide. In such
a field there will be no luminosity, provided the field is optically
empty, that is, contains no particles of sufficient size to produce a
dispersion of light. If, on the other hand, such particles are present,
the effect observed will be an illumination whose character will de-
pend upon the size of the particles. If the particles are of sufficient
size the illumination will show them individually as bright points
of dispersion, even though the particles are too small to be observed
of themselves, just as the stars are visible from the light which they
disperse, but cannot of themselves be seen.' If the particles are so
small that they are no longer able to disperse sufficient light to make
each particle appear as a bright point, there will, nevertheless, pro-
vided the particles are present in sufficient numbers, be produced a
diffuse luminosity throughout the field. These phenomena are
wholly analogous to those observed when a beam of light is passed
through a dark room in the atmosphere of which fine dust particles
are found. The path of the whole beam is made apparently uni-
formly luminous by the smaller particles, while occasionally there
548 INFECTION AND RESISTANCE
appear points of bright illumination, due to the presence of larger
particles. This is known as the Tyndall effect. The light which has
passed through such fields is found to have become polarized.
It is evident that, in a solution whose particles are sufficiently
large to become individually visible as points of light under the ultra-
microscope, it immediately becomes possible to determine the size of
the particles on the assumption that these are all of the same size.
The procedure consists in determining the following quantities: (1)
the total number of particles in a given volume by the usual blood-
count method; (2) the weight of the dispersed substance in a given
volume by a chemical or other analysis ; ( 3 ) the density of the dis-
persed substance, which is usually taken as equal to that in the undis-
persed state. This undoubtedly introduces an error in the computa-
tion, since, in all probability, the density increases in the dispersed
state, owing to increased compression by surface tension. This error
is probably small unless the degree of dispersion is very great. By
this method, particles in colloidal gold solutions have been observed
and counted whose diameters were as small as 10~6 mm. This rep-
resents about the limits of individual visibility under the ultramicro-
scope, that is, with particles much smaller than this the field appears
diffusely illuminated. This value is about one-hundredth that of the
wave-length of violet light, and about ten times that of the calculated
diameter of the ethyl alcohol molecule.
The rate of settlement under the influence of gravity has also
been used to determine the size of colloid particles. By means of
Stokes' law for the fall-rate of bodies through a viscous medium, a
comparatively simple equation permits of the calculation of the diam-
eter of the falling body when the fall-rate, the viscosity of the me-
dium, and the densities, respectively, of the dispersed substance and
of the medium are known. Perrin 5 used the same principle in the
preparation of suspensions in which the particles were all of more
or less the same size, using, however, regulated centrifugation in-
stead of simple settling under gravity.
There has also been developed, largely by Bechhod 6 another
method which throws some light on the relative sizes of particles,
and also offers a very interesting and valuable experimental weapon
for colloid investigation. This is the method of ultrafiltration. It
has been found possible to produce graded filters which allow of the
passage of particles below a certain size, and which restrain any
larger ones. These filters are made by impregnating ordinary filters
with gelatin and other colloidal solutions and drying with special
precautions. The permeability decreases with the concentration of
the gelatin or other substance used.
5 C. B., 146, 967, 1908.
«Ztschr. Chem. Ind. KolL, 2, 3, 1907; Die Kolloide in Biologic u. Medizin,
Dresden, 1912.
COLLOIDS 549
The investigations as to the size of particles all lead to two gen-
eral conclusions : first, suspensions and colloidal solutions in general
differ from one another mainly in degree of dispersion, at least up
to the limit of individual detectibility of the particles under the
ultramicroscope, beyond which point at present all is speculation,
although the presumption is strong and the belief is growing that
there is also no other fundamental distinction to be drawn between
colloidal and so-called true solutions ; second, it is always found that
unless special purification is resorted to a colloidal solution contains
particles of widely differing sizes side by side.
2. THE BEOWNIAN MOVEMENT. — About a century ago the Eng-
lish botanist, Brown, noticed that very small spores and other bodies
when suspended in water, and observed under the microscope, were
in a state of rapid oscillatory and rotary motion. This motion of
small masses of matter has come to be known as the Brownian move-
ment. It is noticed in colloidal solutions whose particles are not too
large, and at the same time are large enough to be individually de-
tectible under the ultramicroscope. As a result of the theoretical
considerations of Einstein,7 of Smoluchowski 8 and of Corbino,9
and of the experimental researches of Svedberg 10 and of Perrin,11
the Brownian movement has come to be considered as nothing more
nor less than a manifestation of that kinetic energy with which all
matter is endowed, and which forms the basis of the kinetic theory
of gases. A rapidly gyrating and oscillating colloid particle is there-
fore looked upon as a large scale picture of the state of the molecules
themselves. These investigations have probably done more than any-
thing save the development of the kinetic theory itself to place
molecular and atomic speculations on a firm basis of plausibility.
Svedberg' s investigations were instituted to determine the mean
velocity of colloid particles whose mass could be determined by the
ultramicroscopic method above referred to. Computing from these
factors the average kinetic energy of the particles, this, according to
Svedberg, gives the same value which would be computed for the
particle on the basis of the kinetic theory. Perrin attacked the prob-
lem from a somewhat different point of view. The number of gas
molecules in the atmosphere decreases from the surface of the earth
outward at a rate which is determinable by computations based on
the kinetic theory. Perrin set himself the task of determining the
rate of decrease in the concentration of the particles of a colloid solu-
tion, in which the particles were of uniform size, the concentrations
being determined at different levels in a cylinder in which the solu-
7 Ann. der Phys., 9, 417; 11, 170; 17, 549; 19, 371.
8 Ann. der Phys., 21, 756.
0 Nuovo Cimento, 20, 5.
10 Ztschr. f. Elektroc-hem., 12, 853, 1906.
11 C. E., 146, 967, 1908.
550 INFECTION AND RESISTANCE
tion had been allowed to stand until it had reached equilibrium with
the gravitational forces. The result was that the same law of dis-
tribution was found to hold in this case as in the case of the atmos-
phere. The kinetic theory is thus shown to apply quantitatively as
well as qualitatively to colloidal solutions.
3. ELECTRICAL PROPERTIES. — If a U-tube be filled with water,
electrodes placed in each arm, and these electrodes maintained at a
constant difference of potential either by a battery, dynamo, or other
source of direct current, it is noticed that there is a continual flow of
liquid in the tube, in one direction near the walls and in the opposite
direction in the interior of the tube. There is every reason to believe
that such currents will be set up in all cases whatever the nature of
the liquid or of the tube, although the current set up may in par-
ticular cases be very small and even very rarely approach or equal
zero. If we name the current along the walls simply the "current,"
and that through the interior the "countercurrent," then in the case
of glass and water the direction of the current is from anode to
cathode, and that of the countercurrent is from cathode to anode.
This phenomenon is explained by the hypothesis that at the surface
of contact between the glass and the water there is established a dif-
ference in electrical potential, the glass becoming negatively and the
water positively electrified. If this assumption is valid it follows
that if a particle of glass placed in water be subjected to the influence
of two electrodes placed in the water, it will, being negatively
charged, be attracted by the positively charged electrode (the anode)
and repelled by the negatively charged electrode (the cathode). The
result would be a wandering of the particle of glass through the solu-
tion toward the cathode. This result is confirmed by ample experi-
ments. Furthermore, the phenomenon is common to all particles in
all liquids, so far as is known, so that any colloidal solution placed
in a potential gradient will show wandering of its particles in one
direction or the other. Thus in water, ferric hydroxid, chromium
hydroxid (and most hydroxids in the colloidal state), methyl violet,
and some other dyes wander to the cathode. All colloidal metal solu-
tions, sulphur, the halogen salts of silver, chlorophyll, rosin, mastic,
.most dyes, and, in fact, the great majority of substances investigated
wander toward the anode. Albumin (and probably some other sub-
stances) wanders toward the cathode in acid solution, and toward
the anode in alkaline solution. As will be seen later, the hypothesis
of the existence of such electrical charges on colloid particles has
been of very great use in explaining many forms of conduct on the
part of dispersed systems.
4. SURFACE TENSION.— If a globule of mercury be divided into
two parts, these two parts will unite again if opportunity be given.
All the opportunity which is necessary, if the surfaces be clean, is to
bring the two parts into mechanical contact. The union of the sep-
COLLOIDS 551
arate parts may, however, occur in a variety of other ways, in fact,
in any way whatever whereby such union is physically possible.
Thus, if the separate portions be of different size, the smaller one
will have a higher vapor pressure than the larger, and evaporation
from the smaller to the larger will take place until the whole of the
smaller portion has transferred itself to the larger one, and the re-
union is therefore complete. There is every reason to believe that if
the two portions of unequal size were made electrodes in a galvanic
cell, and this cell were then short circuited, that the smaller portion
would go into solution and again deposit upon the larger one. In
case the two portions were of the same size, these forms of recom-
bination, with the exception of that of direct coalescence, would not
occur if all other conditions were kept constant, but a slight differ-
ence of conditions in respect to the two portions would start the act
of recombination, which would then in general proceed to comple-
tion. The same tendency is noticed with all substances. Thus in a
liquid small crystals disappear while larger ones grow at their ex-
pense, and it may be stated that, other influences for the moment
ignored, the most stable configuration which can be assumed by a
given mass of any substance is that in which all of the substance is
in one portion, and that portion is spherical in form. This is equiv-
alent to saying that all bodies so arrange themselves as to expose the
least possible surface. The force which tends to bring about this
condition is called surface tension. In so far as surface tension
alone is concerned it follows that any colloidal solution must be un-
stable, and tend to condense itself until all of the dispersed matter
has aggregated itself together into a single mass of spherical form.
But there are many other forces which may under certain con-
ditions act against surface tension. If the dispersed substance is
one that is crystalline, the directive forces of crystallization over-
come those of surface tension, and the form of stable configuration
will be that of the crystal instead of spherical, and equilibrium will
be established when all of the available substance has aggregated
itself together into one large crystal. We know, on the Other hand, a
great many colloidal solutions which seem to be quite stable even in
very high degrees of dispersion. To explain such cases we must look
for other forces working against the force of surface tension. If the
dispersed particles in a colloidal solution are all charged with the
same kind of electricity, they will then repel one another with a
force which will vary inversely as the squares of their distances from
one another. This repulsion will then tend to work against any
coalescence or other sort of union between the disperse particles. We
have already seen that colloidal particles are in general charged
either positively or negatively, and this may be taken to some extent
as explaining the stability of such systems. Equilibrium results
when the surface tension is just counterbalanced by the electrical
552 INFECTION AND RESISTANCE
repulsion. The extension of this idea has been of great value in
colloid investigation. The electrical repulsion will not, of course,
necessarily prevent the smaller particles from dissolving and deposit-
ing upon the larger ones, unless the solubility is affected by the
charge. Concerning this we know nothing. The fact that colloid
substances possess little or no solubility in the ordinary sense of the
word means such solution and deposition must, of necessity, be a
very slow process, and the colloid solution would thus appear to be
perfectly stable over very long periods of time. There are many who
believe that all such systems are only apparently stable, and that on
account of the absence of any sufficiently rapid means of transforma-
tion which would allow the stabilizing influences to operate rapidly
enough to be perceptible.
Chemical Properties of Colloids — 1. It is reasonable to suppose
that the chemical properties of colloid solutions are very much what
is to be expected from the chemical nature of the dispersed substance
as it is known under other conditions. The colloidal solutions of ar-
senic sulphid should therefore react very much as would be expected
of arsenic sulphid in general, except in so far as the substance is in a
finely divided state in the presence of a dispersing medium (water)
in which it is little soluble. Thus colloidal arsenic sulphid is soluble
in alkalies and alkaline sulphid just as is the massive form. If a rod
of zinc is suspended in a colloidal solution of arsenic sulphid there
takes place a slow reaction, lasting over weeks and even months,
whereby the sulphur of the sulphid unites with the zinc to form
colloidal zinc sulphid, while a black deposit, probably arsenic, is
found on the zinc. Chemical reactions with colloids are thus, as
a rule, very slow, as is to be expected, but otherwise not essentially
unusual.
2. The exact chemical composition of the disperse phase in a
colloidal solution is probably not definitely known in any case. In
the case of colloidal metal solutions, such as gold and silver, the sus-
pended particles seem to be practically pure metals, but in most cases
the composition is very problematical. The 'great variation in the
properties of such solutions with variations in the methods of prep-
aration are undoubtedly to a great extent due to small differences in
composition. Thus the properties of arsenic sulphid vary greatly
with* the extent to which free hydrogen sulphid is removed from the
solution, which is probably due to the differences in the amount of
hydrogen sulphid absorbed or otherwise held by the arsenic sulphid.
Linder and Picton believed that amorphous copper sulphid was a
definite compound of copper sulphid with hydrogen sulphid. It has
also been found that amorphous copper sulphid suspended in water
continually deposits free sulphur, the cupric sulphid being at the
same time largely converted to cuprous. It seems to be rarely or
never the case that the disperse phase may be looked upon as a sub-
COLLOIDS 553
stance of definite composition, being usually, if not always, a more
or less complex mixture of absorption products.
3. Although not usually pure substances, it is not at all im-
plausible to assume that the dispersed particles may, to some extent,
undergo ordinary' electrochemical ionization, in which case the par-
ticles would partake of the nature of enormously large ions. This
assumption is interesting as offering a purely electrochemical ex-
planation of the origin of the charge which is found on such par-
ticles, and it is to be said that frequently the effect of foreign sub-
stances on the electrical charges of suspended particles is explain-
able on this assumption. For further information on these matters
reference must be had to the papers of Duclaux,12 Jordi,13 and
P. P. von Weimarn.14
The Flocculation of Colloids by Electrolytes — 1. When neutral
salts are added to colloid solutions in gradually increasing amounts
there always follows sooner or later a precipitation of the dispersed
substance. If the salt solution be removed and pure water added
in its place, this decanted, and pure water again added, so as to
wash out the salt as thoroughly as possible, the final result may
be that the coagulum (or gel) becomes redispersed, or such redis-
persion may not occur. The outcome is determined both by the
nature of the colloid and by that of the neutral salt used.
2. If acids and alkalies are the electrolytes used, the relation-
ships are somewhat different. The addition of alkali to a negatively
charged colloidal solution renders it more stable, while a positively
charged colloid is flocculated. With acids the reverse of this condi-
tion holds ; that is, the positive colloid is stabilized and the negative
one is flocculated.
3. The concentration of the electrolyte required to flocculate a
given colloidal solution depends very greatly on the nature of the
electrolyte used. It has been generally considered that the cathion
of the electrolyte is the active agent in flocculating negative colloids,
while the anion is active in the case of positive ones. Thus acids
(hydrogen ions) are very effective in flocculating arsenic sulphid,
and alkalis (hydroxyl ions) are effective with ferric hydroxid, as
might be inferred from the preceding paragraph. Different cathions,
however, show very different degrees of precipitating or flocculating
power. Thus in the case of arsenic sulphid, if the flocculating power
of the potassium ion is taken as unity, that of calcium is about
twenty, while that of the aluminium ion is three hundred and fifty.
The flocculating power in this case increases very rapidly with the
12 "Thesis," Paris, 1904.
13 C. R., 136, 680, and 1,448; 137, 122; Bull. Soc. Chim., 31, 573.
14 Articles in Ztschr. Chem. Ind. Roll, 1906-1911. Also in "Grundziige
der dispersoid Chemie," Dresden, 1911.
554 INFECTION AND RESISTANCE
valence of the ion, a relationship which seems to be quite generally
true. It has been quite commonly considered that the effect of the
aiiion in the flocculation of negative colloids is negligible. That this
point of view is not tenable has recently been strikingly shown by
Sven Oden15 in his work on colloidal sulphur. He finds that the
effect of electrolytes on sulphur sols, which, like arsenic sulphid, are
negative colloids, is distinctly the resultant of two factors, one a floc-
culating effect on the part of the cathion, the other a dispersing effect
on the part of the aniou. It is not improbable that our views in
regard to the phenomena of electrolyte flocculation will undergo con-
siderable modification in the near future, as they are based on rather
scanty experimental evidence. Since the properties of sols of the
same materials differ very considerably with the most minute details
of their method of preparation, it is naturally difficult for the same
investigator even to obtain uniform results.
4. The actual concentration of a given electrolyte required to
flocculate a sol depends also very greatly on the nature of the sol
itself. Some sols are precipitated by very small concentrations of
electrolytes (three to four one-hundredths normal acid being usually
sufficient for arsenic sulphid), while gelatin, albumin, and protein
substances in general require far higher concentrations. So marked
is the difference in many cases that attempts have been made to clas-
sify colloids on the basis of their conduct in this respect. Thus col-
loids which are very sensitive to electrolytes are called suspension
colloids, while those that are not very sensitive are called emulsion
colloids. To the first class belong all the true, rather coarse-grained
suspensions, while the sols that yield soft gelatinous flocculates
usually fall into the second class. It is also very frequently true that
the latter type shows the phenomenon of reversible flocculation (see
ante). This classification is for many purposes quite useful, but
cannot be considered as very fundamental. For example, if the elec-
trolyte used be a salt of a heavy metal most of the so-called emulsion
colloids, such as albumin, are irreversibly flocculated.
5. The flocculation of sols by electrolytes is usually explained
as due to the phenomenon of absorption. That is, the flocculating
ion is absorbed from the solution by the dispersed particles. Since
Jn general the ion which is absorbed is the one whose electrical charge
is opposite to that of the dispersed particle the absorption results in
a reduction of the charge on the particle, and allows the aggregating
forces of surface tension to become operative. The evidence of the
validity of this assumption is considerable. Thus the flocculated
colloids always contain appreciable amounts of the ion which caused
the precipitation, which is prima facie evidence of the absorption.
Furthermore, the electrical charges on the particles may be meas-
ured by determining their rates of migration, and the effect of elec-
15"Inau£. Diss.," Upsala, 1913, pp. 118 et seq.
COLLOIDS 555
trolytes on this charge may 'also be observed. It is very commonly
true that the addition of a precipitating electrolyte to a sol reduces
materially the charge of the particles. This is not, however, always
true, and the relationships are more complicated than has generally
been assumed. Nor is it always true that the complete neutraliza-
tion of the electrical charge on dispersed particles results in floccula-
tion. For example, acetic acid may be added to arsenic sulphid sols
in such quantities as to completely neutralize the negative charges of
the particles and, further, so much acid may be added that the par-
ticles acquire a very considerable positive charge, all this without
the least signs of flocculation. Ultimately so much acid may be
added as to cause flocculation.
6. In the flocculation of sols by electrolytes there is frequently
observed a curious effect known as the "zone-phenomenon" It is
observed when increasing amounts of certain electrolytes are
added that at a certain concentration flocculation will be brought
about, while if the concentration be greater flocculation will not
occur, although still further increase of concentration will result in
another flocculation zone. The phenomenon is most common when
the electrolyte used is a salt that shows marked hydrolysis, such, for
example, as ferric chlorid. If a negative sol be treated with a solu-
tion of this salt it is obvious that there will be three precipitating
influences present, namely, the hydrogen ions and the colloidal ferric
hydroxid, both of which are formed by the hydrolysis of the ferric
chlorid and the unhydrolyzed ferric chlorid. Since the amounts of
these precipitating substances in a given solution vary with the con-
centration, and since each has its own concentration function in
precipitation, it will be seen that the zone phenomenon may be ac-
counted for in such cases. Since, however, the zone phenomenon
occurs in many cases where strongly hydrolyzed electrolytes are
not used, as, for example, in the agglutination of bacteria by
citric and some other acids, the explanation is not wholly sufficient.
There are also many curious phenomena concerned with the
action of electrolytes on sois which have as yet been very little
investigated, and which will probably throw considerable light on
the subject.
The Mutual Reactions of Colloids — The conduct of mixtures of
two different sols is of very great interest and variety, both in
the absence and in the presence of electrolytes. A number of
particular cases may be distinguished, and these will be taken up
seriatim.
1. When two positive or two negative sols are mixed together
nothing very much seems to happen. It is generally considered that
the addition of one sol to another of like electrical properties results
in no action. Whether this is wholly true or not is doubtful, but it
is at least true that in all cases investigated up to the present time
556 INFECTION AND RESISTANCE
neither individual nor mutual flocculation occurs. Whether the
presence side by side of two sorts of similarly charged disperse par-
ticles results in any change in the dispersion of either is not known.
Except to show that no mutual precipitation occurs, these cases
have been but little studied.
2. When two oppositely electrical colloids are mixed mutual
precipitation may or may not occur. The factor which in the main
determines the outcome is the relationship between the amounts of
the two colloids used. If neither is present in too great excess com-
plete mutual precipitation in general occurs. If either is present
in great excess precipitation does not in general occur.
3. The effect of a great excess of one colloid in preventing
mutual precipitation is very marked when the colloid in excess is
one of the emulsion-colloid type. Thus gelatin, a typical emulsion
colloid, when present in excess over another colloid very frequently
prevents all flocculation, even in fairly coarse-grained suspensions.
Advantage has been taken of this action in preventing scaling in
boilers. This scaling is due largely to the fact that lime salts held
in solution as bicarbonates are decomposed by heat, with the separa-
tion of calcium carbonate, at first in the highly dispersed state. This
gradually aggregates together and deposits on the interior of the
boiler as an amorphous flocculated colloid, which in time becomes
very compact, and in many cases crystalline. If, however, a small
amount of glue (impure gelatin) be added to the boiler water the
colloidal constituents of the water do not flocculate and compact,
but remain suspended, and may from time to time be blown off.
Another illustration is found in the preparations of photographic
emulsions. The silver halides flocculate very readily in pure water,
but in gelatin solution remain in a highly dispersed state, which is
necessary to the preparation of the plate. In this case, not only is
the suspension protected from flocculation., but also a degree of dis-
persion is reached which is far beyond anything attainable in pure
water. The same is true when lysalbinic acid is used to prevent
flocculation of colloidal silver. In pure water only very dilute sus-
pensions of metallic silver are obtainable, but in the presence of lysal-
binic acid suspensions containing as high as ninety per cent, of silver
are obtainable, the product being used medicinally under the name of
"argyrol." These are the phenomena known as "protective actions,"
and the gelatin, albumin, or other colloid which exerts the protective
action is spoken of as a "protective colloid/'
4. The protective action of certain colloids is not only exerted
against the tendency of the protected colloid to spontaneously floc-
culate, but also a certain and very great protection is offered against
flocculation by electrolytes. Thus Zsigmondy 16 was able to find a
definite measure of the protective action of certain colloids on the
16 Ztschr. f. analyt. Ch., 40, 697, 1902.
COLLOIDS 557
precipitation of gold suspensions by means of sodium chlorid. The
method was to find the amount of the protective colloid that was just
necessary to prevent the flocculation of a fixed amount of a given gold
sol by a fixed amount of the salt. In this way it was observed that
the protective action of different colloids is very different. Thus if
the amount of starch in solution which is necessary for protection
be taken as 2,500 the amounts of various other colloids required are
as shown in the following table :
Protective colloid .... Starch Dextrin Gum Arabic Albumin Gelatin Glue
Amount required 2,500 1,200 40 25 11
5. Very simple theories are devisable to explain the interactions
between different colloidal solutions. Thus two oppositely electrical
colloids may be considered to precipitate one another mutually be-
cause of the electrical attraction existing between all oppositely
charged particles. This results in bringing together the oppositely
charged particles with the formation of relatively neutral aggregates,
wiiich, as shown in the discussion of the flocculation by electrolytes,
is a condition favorable to precipitation. For very obvious reasons,
then, no flocculation would be expected when two like charged col-
loids are mixed. Many objections may be made to the unqualified
acceptance of this explanation. It offers, nevertheless, a valuable
leading idea when not accepted too dogmatically.
A number of factors probably contribute to the protective action
of many colloids. To some extent the effect may be purely mechani-
cal, since increased viscosity imparted by the presence of the protec-
tive emulsion colloid will shorten the mean free path of the floc-
culable particles, and thus materially lessen the probability of im-
pacts between them. Consequently flocculation wrill not occur as
readily. Further, the ultramicroscope gives considerable evidence
of the existence of another and very important factor. It seems
certain, at least in many cases, that the protective colloid arranges
itself in a film or coating around the flocculable particles, and in this
way prevents the aggregation of the particles. These factors are not,
however, sufficient to explain all cases of protective action. It must
be considered that in . some cases, at least, the protective colloid
exerts a truly dispersive effect, such as would result from a nullifica-
tion of surface tension forces. The ease with which a small amount
•of soap will emulsify a large amount of oil is difficult to explain on
any other hypothesis. When an oil is shaken with pure water little
or no emulsification results, while in the presence of the soap as a
protective colloid the same amount of work in shaking accomplishes
enormously greater results. In fact, the oil will spontaneously
emulsify by merely standing in contact with the soap solution. The
equilibrium condition of oil in contact with pure water is reached
558 INFECTION AND RESISTANCE
when the oil is very little, if any, dispersed, while in contact
with soap solution equilibrium is reached only when the oil
is highly dispersed. ' From the surface tension point of view this
would be expressed by saying that in contact with pure water
the surface tension is large and positive, while in contact with
soap solution the surface tension is large but negative. It is im-
possible to say to what extent these effects may be due to obscure
chemical action.
The Preparation of Colloidal Solutions — 1. This subject might
perhaps upon logical grounds have best been treated in an opening
paragraph. However, with the conclusions that have been reached
in the foregoing discussion the whole matter may be dismissed very
briefly. The conditions which must be fulfilled in order to obtain
colloidal solutions are in the main summarized in the following
paragraphs :
2. A medium must be chosen in which the given substance does
not reach to any great extent the maximum, molecular degree of dis-
persion— that is, in which what is usually called true solution does
not occur to any great extent. While, as pointed out at the begin-
ning of this discussion, there is no sharp line to be drawn between
colloidal and true solutions, the substances that are distinctly recog-
nizable at the present time as colloidal solutions carry particles which
are in the neighborhood of one thousand times as large as average
molecules. From media in which the dispersion is approximately
molecular the dispersed substance usually shows a strong tendency to
separate in the crystalline form, although it may frequently, if not
generally, first appear in the colloidal form, then more or less rapidly
becoming crystalline. The only distinction to be drawn is this:
from solutions in which the dispersion is very great, approximating
the molecular, separation in short time in the crystalline form is
favored ; from solutions in which the dispersion does not approximate
the molecular separation persistence in the amorphous state is fa-
vored.
3. A colloidal sol or gel, one or the other, may generally be pro-
duced from any substance in any medium in which the amount of
molecular dispersion is at a minimum by means of any reaction
whereby the new substance is produced from solution. Thus, for
example, arsenic sulphid does not disperse ' in water to molecular
extent in any considerable degree. Therefore, by mixing together
a solution containing an arsenic compound which is soluble, and a
solution of hydrogen sulphid, the resulting arsenic sulphid will
appear in the colloidal state. Whether it appears in the dispersed
state as a sol or in the flocculated state as a gel will depeud mainly
on the electrolyte content of the solutions which are used. Tf a solu-
tion of arsenic chlorid be used the resulting solution will contain
.considerable free hydrochloric acid, and the tendency wil* be for the
COLLOIDS 559
resultant arsenic svtlphid to appear in the flocculated or gel form.
If it is desired to prepare the substance in the sol form electrolytes
must be avoided. This may be done by using a solution of arsenic
trioxid instead of one of arsenic chlorid.
Any reaction which is brought about under the above conditions
will result in the formation of a colloidal product. The dialysis of
salts which form insoluble hydroxids simply allows the normal
reaction of hydrolysis to complete itself. The resulting hydroxids
appear in such a way as to fulfill the above conditions, and conse-
quently appear in the colloidal state. In this connection note the
preparation of colloidal sodium chlorid (see ante).
4. In an appropriate medium most if not all substances, crystal-
line or otherwise, may be brought directly, that is, without the inter-
vention of specific chemical reactions, into the colloidal state. In
some cases this may be accomplished by mechanical means. Thus
oils violently shaken with water disperse to some extent and form
emulsions of greater or less stability. By shaking glass, quartz, and
the like with various liquids in which they are virtually insoluble the
abrasion results in the formation of more or less dispersed systems,
usually not very stable.
Many metals may be brought into the dispersed state by causing
an electric arc to pass between points held under a liquid. This
electrical dispersion method has been very considerably used, but is
obviously confined to substances which are conductors of electricity.
On the other hand, many substances when merely brought into
contact with an appropriate medium spontaneously undergo disper-
sion. Thus gelatin, glue, tannin, and many other substances spon-
taneously disperse in water. Even crystalline substances frequently
do this. Thus soaps which have been crystallized from alcohol solu-
tions when brought into water disperse in the colloidal state. Crys-
tallized cuprous sulphid, the mineral known as chalcocyte, disperses
in the colloidal form in solutions of hydrogen sulphid.
Substances which go spontaneously into the dispersed colloidal
state are usually spoken of as "lyophyllic/' while those that tend to
spontaneously leave the dispersed state are called "lyopkobic." It is
evident that a substance may be lyophyllic with respect to one me-
dium and lyophobic with respect to another. Furthermore, a very
small change in the nature of the medium may cause the change
from a lyophyllic to a lyophobic colloid. Thus oils are lyophobic
with respect to water, but lyophyllic with respect to even very dilute
soap solutions.
Applications to Biology. — 1. When one considers the relatively
infrequent occurrence in biological systems of either crystalline sub-
stances, or of substances that may be readily made to crystallize
from water (the universal biological dispersing medium), it imme-
diately becomes evident that the chemistry and physics of such
560 INFECTION AND RESISTANCE
systems must be in the main colloidal. All biochemistry is thus in
the main colloid chemistry. Aside from mineral salts, urea, uric
acid, and a few other bodies, the reagents and products which are
active in life processes are all to be found in the living organism
in the colloidal state. While crystalline directive forces are in gen-
eral absent, we have nevertheless to deal in biological phenomena
with a great variety of directive forces of a wholly different charac-
ter. Colloidal substances in high degrees of dispersion such as
proteins and the like in the alimentary fluids are being continually
converted into active living tissues, a process manifestly involving
very definite directive forces, since the product (the tissue cells)
is an organized one, even though its organization is not similar to
that of a crystal. Thus the building of living tissue involves
among other things the conversion of sols of many sorts (alimen-
tary fluids, blood, etc.) into gels. In other words, the living
tissue is to a certain extent to be looked upon as a colloidal gel, dif-
fering, however, from laboratory gels in possessing a definitely or-
ganized cell structure. While it is true that in the spontaneous gel
formation with certain colloids, as, for example, gelatin, myelin, or
web structures are formed purely as a result of the physical and
chemical forces active, these cannot be said to bear any very strong
resemblance to living cells. It is thus apparent that life processes
differ very materially from those of the chemical laboratory. On the
other hand, it is true that many of the component reactions which go
to make up the life process may be very closely duplicated by labora-
tory means, and that already the study from the colloid chemistry
point of view of the reaction of many of the substances which go to
make up the living organism has given interesting and important
results. Furthermore, the whole field of colloid investigation has
been greatly stimulated by the hearty support and encouragement
which it has received from biologists. Some slight attempt will be
made here to illustrate by a few examples the far-reaching possi-
bility of explanation which colloid chemistry offers of certain phases
of biological science. The actual accomplishment in the field is
already so great that only a very limited discussion can be offered
here. .*.
2. The action of electrolytes on emulsions of bacteria is wholly
analogous to their action on colloidal suspensions. The bacterial
emulsions are very sensitive to flocculation by mineral acids, one thou-
sandth normal hydrochloric and sulphuric acids usually being suffi-
cient to cause complete clumping and settling of the bacteria. Neu-
tral salts, with the exception of those of silver, mercury, iron, and
aluminium, do not flocculate the bacteria. If, however, the bacteria
are first treated in the absence of electrolytes with an agglutinating
serum small concentrations of salt solutions will bring about floccu-
lation. It is also known that bacteria have the power of absorbing
COLLOIDS 561
agglutinins from sera, so that it is evident that what we have here
is a case of the production of a flocculable combination of bacteria
and agglutinin, neither component of which is alone flocculable.
Citric acid in concentrations ranging between one one-hundredth
and one eight-hundredth normal produces flocculation. In either
greater or less concentrations no flocculation is produced, which is an
illustration of the so-called zone phenomenon.
Like all other suspensions, bacteria are electrically charged, and
consequently wander in the electric field. Under all ordinary condi-
tions the charge which they carry is negative, from which in their
general conduct it might be expected that they would conduct them-
selves similarly to arsenic sulphid, which is also negatively charged.
This is found to be the case. Their rate of migration is reduced
by acids and evidently somewhat increased by alkalies, although
very little alkali is necessary to bring about disintegration of the
bacteria.
It has also been found that all colloids are more or less sensitive
to light in respect to their migration rates. Bacteria also show this
phenomenon, as they migrate notably slower in the light than in the
dark. It is quite possible that this reduction in their electrical
charge may to some extent be responsible for the bactericidal action
of light.
3. A very interesting application of the principles of colloidal
precipitation by electrolytes has recently been made by Loeb.17 He
finds that the eggs of the Fundulus, a small fish, are killed by being
immersed in a pure isotonic salt solution, in spite of the fact that
they normally develop in sea water. The factor which allows of their
development must therefore be sought in some other constituent of
the sea water which is absent in the pure salt solution. This Loeb
finds in the presence of small amounts of calcium salts. Further,
if a small amount of calcium chlorid be added to the pure salt solu-
tion the eggs will develop in it as well as in sea water. Loeb's ex-
planation is simple and very ingenious. He supposes that the sodium
chlorid is toxic, provided it can diffuse into the egg. In the ahsence
of calcium salts such diffusion is possible because the sodium chlorid
is not a sufficiently powerful colloid precipitant to make the mem-
brane about the egg impervious. It is, however, very well known
that calcium salts (and all bivalent cathions) are far more effective
colloid precipitants than sodium ions. Consequently the presence
of a relatively small amount of calcium chlorid in the salt solution
is sufficient to so condense the egg membrane as to make it impervious
to the sodium chlorid, and thus render the latter non-toxic.
4. Another interesting application of colloidal principles is
found in what is known as the Danysz phenomenon. It is found
that the neutralization of the toxicity of diphtheria toxin by the
17 Am. J. Physiol., 6, 411, 1902.
562 INFECTION AND RESISTANCE
antitoxin depends on the way in which the two are mixed. If a
quantity of toxin just sufficient to neutralize a fixed amount of anti-
toxin when it is added all at once be in another experiment added in
small installments, the resulting mixture will be found to be still
quite strongly toxic. This is quite analogous to what is found in the
interaction of many colloids. The amount of a given colloid re-
quired to neutralize and precipitate another depends greatly on the
way in which it is added.
5. Homer,18 Field and Teague,19 and Teague and Buxton 20 all
carried out interesting investigations directed toward determining
the migration directions (electrical charges of toxins and antitoxins).
Their conclusions were that all wandered toward the cathode, and
that all were therefore positively charged. If this is correct the
analogy between toxin and antitoxin reactions and those of simple
colloids is rather mutilated, since two positive colloids are not sup-
posed to react with one another. It is, however, more than possible
that the above experiments are misleading. In all cases agar dia-
phragms were used. Through these there would always occur a
streaming of water toward the cathode as a result of the electrical
potential between the agar and the water. This might well be so
great as to obscure, and even more than overcome any anodic wan-
dering that might occur. Furthermore, the conduct of proteins in
general in the electric field is a very complex one, and one that is
only just beginning to be understood. For these reasons it is at
present very dangerous to draw any very dogmatic conclusions.
6. In closing mention may be made of what seems to be an im-
munity phenomenon which seems rather clearly to be a case of pro-
tective colloid action. It is observed when an agglutinin is added to
a bacterial emulsion that if an excess of the agglutinin be added no
agglutination occurs. This is wholly analogous to the fact that,
while a small amount of gelatin will precipitate arsenic sulphid sus-
pension, a larger amount will not.
Conclusions — While great progress has been made in the field of
colloid investigation from the chemical and physical sides, and while
also many very striking analogies are to be found from the biological
side, it is nevertheless true that we are still very much in the dark
in regard to a great many matters. The one great difficulty which
lies in all such investigations is that it is a matter of very great diffi-
culty to duplicate results. The nature of any colloid sol or gel
depends so greatly upon its whole previous history, apparently
down to the least detail, that great discrepancies in experimental
results are found. Even the age of a sol is frequently a matter of
very great importance in determining its properties. For example
18 Berl. klin. Wochenschr., Vol. 41, p. 209, 1904.
19 Journ. Exp. Med., Vol. 9, pp. 86 and 223.
20 Ibid., Vol. 9, p. 254.
COLLOIDS 563
a freshly prepared gelatin solution will not precipitate arsenic sul-
phid, but it will do so after it has stood for some hours. What is
now greatly needed is more data on a greater variety of colloids that
have heretofore been investigated and work directed toward the
preparation of colloidal solutions of definite character. Until some-
thing has been accomplished in these directions all biological anal-
ogies and the like cannot be anything more than qualitative, and the
same holds true for many of the physical and chemical conclusions
which have been discussed in this chapter.
INDEX
Abderhalden, protective
fermenti of, 532,
533, 534
Abderhalden's reaction, 534
Abscesses, secondary,
caused by bac-
teria, 25
Absorption theory in tox-
in-antitoxin reac-
tion, 123
Acne, opsonic index in
vaccine treatment
of, 339
Adrenal cytotoxin, 92
Agglutination, absorption
experiments of
Castellani on, 232
acid, 246
value of, for differen-
tial purposes, 240,
247
action of salts in, 240
agglutinability of bac-
teria in, in agglu-
tinoids, 229
normal differences in,
between strains
of same species,
228, 229
agglutinins in, 223. 224
absorption of, 232
complete, impossi-
bility of. 232
heating of, 226
explanation of, 236
major, 229
normal, 233
explanation of. 234
qualitatively identi-
cal with "im-
mune" agglutin-
ins, 234
para- or minor. 229
agglutinogen in, 223, 224
effects of heating of,
226
localization of, in ec-
toplasmic layers
of cells, 224
agglutinoids in, 235
biological relationship
and, 230
Bordet's explanation of,
240
by means of cell body
proper, 225
by means of ectoplasmic
substances of bac-
teria, 224, 225
by means of flagella,
224. 225
cataphoresis of bacteria
in. 242, 243
description of, 218
effect of gelatin addition
in mastic solu-
tion on, 244
Ehrlich's interpretation
of process of, 234
Agglutination, Ehrlich's in-
terpretation o f
process of, dia-
grammatic repre-
sentation of, 235
Eisenberg and Yolk's in-
terpretation of,
235
Ficker's reaction in, 223
flocculation of colloids
and, 241
mechanism of, 241,
242
mutual, 242
group, 229
cause of, 230
diagnostic value of,
231
hemagglutination analo-
gous to, 236, 237
history of, 218
importance of electro-
lytes in, 240
in colloidal solutions, in-
hibition zones in,
245
in excess of agglutinin,
colloid phenom-
enon and, 562
in immune serum, 141
in motile and non-mo-
tile bacteria, 224,
225
in salt-free environ-
ment, by addition
of organic sub-
stances, 245
influence of immuniza-
tion with different
animal species
on, 232
with different species
of bacteria on,
232
influence of salts on
sensitized and un-
sensitized b a c -
teria in, 243, 244
experiments of Neisser
and Friedemann
in, 244
inhibition zones in, 162,
236
iso-agglutinins in. 237
grouping of, 237, 238
value of presence of,
239
methods of, 218 et seq.
Bordet's, 223
Ficker's reaction in,
223
Gruber-Widal, 220
macroscopic. 219
microscopic, 220
Proescher's, 220
thread reaction of
Pfaundler, 222,
223
nature of, 223
565
Agglutination, not associ-
ated with life of
bacteria, 222, 223
of "agglutinin" bacteria,
243
of bacteria in active im-
munization, 89
of capsulated bacteria,
243
phenomenon of, 218
power of, alterations in,
by cultivation in
immune serum,
228
effect of heating on,
226
spontaneous alteration
in, 227
pro-agglutinoid phenom-
enon in, explained
as protective col-
loid action, 236
pro-agglutinoid zone in
162
pro-agglutinoids in, 235
relation of flagellar
mechanism t o,
222
specificity of, 219, 229
diagnostic value of, in
group reaction,
231
limitations to, 229
thread reaction of
Pfaundler in, 222.
223
"two phase" theory of,
241
Agglutination reaction,
diagnostic use of,
220, 221, 222
flagellar mechanism in,
222
in diagnosis of dysen-
tery, 221
of glanders in horses,
222
of paratyphoid fever,
221
of pneumonia, 221
of streptococcus in-
fections, 222
of typhoid fever, 220,
221
nature of, 239 et seq.
precipitin reaction anal-
ogous to, 263 "
with dead bacteria.
223
Agglutinins. 223
absorption of, in agglu-
tination. 232
complete, impossibil-
ity of. 233
bacteriotropins and, 321
definition of, 89
group :
major, 229
para- or minor 229
566
INDEX
Agglutinins, heating of, ef-
fects of, 226
explanation of, 236
"immune," 234
in hemolytic serum, 93
nature of, 224
normal, 91, 233
explanation of, 234
qualitatively identical
with "immune"
agglutinins, 234
production of, 129
Agglutinogen, 223
effects of heating on,
226
localization of, in cell
body proper, 225
in ectoplasmic layer
of cells, 224
in flagellar substance,
224, 225
nature of, 224
Agglutinoids, 235
Aggressins, 326
action of, 21
obtaining of, 21
secretion of. by bacteria
in body, and vir-
ulence, 20-22
Albuminolysins, 193, 211,
401
Alexin, 137
a combination of soaps
and proteins. 175
absence of, from aqueous
humor of the eye,
170
analogy between fer-
ments or enzymes
and, 176
bactericidal powers of,
137 /
definition of, 80 ^
dependence of, on con-
centration, 176
extraction of, from leu-
kocytes and lym-
phatic organs,
304 et scq.
filtration of, 177
in hemolysis. 144
inactivation of, bv salt,
178
by salt-removal, 178
by shaking, 185
reversibility of. 184
Gramenitski's ex-
periments on, 184,
185
increase of, on clot, 172
influence of salts on
action of, 177,
178
leukocytic origin of, 168,
169 et seq.
multiplicity of, 154 et
sea.
Borders views on.
156
Ehrlich's views on.
155
nature of, 137. ir,4
chemical, 174. 175
physical, 177
presence of. in blood
plasma, 170-1TL'
in blood stream, 170-
172
in normal blood, Gen-
gou's view of.
303, 304
MetchnikofTs view
of, 302
Alexin, presence of. in nor-
mal blood, other
experimenters on,
303
production of. in liver,
173
in thyroid gland, 172
varieties of :
macrocytase, 169
microcytase, 169
Alexin fixation, 186
albuminolysin in, 193
identity of, with pre-
cipitins, 193
writer's opinion on.
193, 194
Bordet and Gengou's ex-
periment on, 186.
187
Bordet-Gengou phenome-
non in, 188 ct
stq.
Gay's experiments sup-
porting, 190
in diagnosis of infec-
t i o u s diseases.
188
in diagnosis of syph-
ilis. 188
Moreschi's e x p e r i -
ments supporting.
189
Bordetscher Antikorper
in, 194, 195
by immune animal sera
with their specific
antigens. 189
by protein and antipro-
toin sensitizers,
188
distinguished from com-
plement devia-
tion, 186
during hemolysis, nature
of, 176
Khrlich's (schematic)
conception of. 187
experiments of, on syph-
ilitic monkeys,
198. 199
forensic tests in. 211
in ana phyla xis. 392
in determination of na-
ture of unknown
protein. 211
delicacy of. 212. 213
technique of. 212
in diagnosis of glanders.
216
in diagnosis of gonor-
rheal infections.
216
in diagnosis of malig-
nant neoplasms.
213
von Dungern's method
of. 214
antigen production
for. 214
results of. 215
technique of, 214 et
seq.
in diagnosis of syphilis
in human beings,
199
in diagnosis' of tubercu-
losis, 217
in Wassermann reaction,
198
nature of, 192 et seq.
non-specific. 195
by heated normal
serum, 196
Alexin fixation, non-spe-
cific, by lipoidal
substances in tis-
sues, 195
by preserved normal
serum, 196, 197
by protein emulsion
and other ex-
tracts, 196
by unsensitized bac-
teria, 195
of specific precipitates,
190 et scq.
albuminolysin identi-
cal with. 193
writer's opinion on,
193, 194
Gay on, 190
Pfeiffer and Fried -
berger on, 190
Sachs on, 191
writer on. 193, 194
practical applications of,
198
precipitin reaction and,
190, 192
Dean on, 194
Gengou on. 192
Neufeld and Ilaendel
on, 194
writer on. 193, 194
specific antiprotein sen-
sitizers in. 192
with syphilitic serum in
antigens from
normal organs,
200
Alexin splitting. 178
Brand on. 179
by method of Ferrata,
178. 179
by method of Liefmann,
184
by method of Sachs and
Altmann, 184
effect of acid reaction on
fractious of, 181
end-piece in, 179 et
seq.
mid-piece in, 179 et
scq.
heat sensitiveness of
fractions of, 180
interchange of fractions
of. in different
animals, 182
is it the inactivation
of the mid-piece 'i
183
mid-piece only bound in
Wassermann reac-
tion. 181
physical occurrence of
fractions of, in
blood. 181
presensitizetl cell in,
180
properties of fractions
of, 179
quantitative relations
between fractions
of, 1S2, 183
Alexocytes. 168
Alkali-albuminate precipi-
tin, 260
"Alkalinity theory" of
immunity, 83, 84
Amanita phalloides, spe-
c i f i c antitoxin
from. 96
Amboceptor, 149
Bordet's definition o f,
159
INDEX
567
Amboceptor, complement-
ophile groups or
polyceptors o f
(Ehrlich), 149,
156
cytophile group of, 149,
152, 153
multiplicity of, 150, 151,
154
Ehrlich and Morgen-
roth on, 150, 151
quantitative determina-
tion of, in im-
mune serum, 160,
161
specificity of, 150
Amboceptor and comple-
ment, Ehrlich and
Sachs's views on
union of, 164. 165
Noguchi's measurement
of quantitative
relations of, 164
quantitative ratio be-
tween, 163. 164
Amebse, artificial, 294
Aiiaphylactic antibody, re-
lation of. to other
antibodies, 401
Anaphylactic intoxication,
peptone poisoning
and, 405
Anaphylactic shock, 363 et
seq. See also Ana-
phylaxis, clinical
manifestation of
effect of atropin and
other drugs on,
379
histamin as cause of, 406
Anaphylactin. 385
Anaphylatoxin, 22, 394,
415
action of alexin in. 425
with normal or in-
activated immune
serum, 425, 426
with salt solution, 426
inhibition of. by too
vigorous and pro-
longed reaction.
419
source of, 427
Anaphylaxis, 358
alexin fixation in. 392
analogy of immediate re-
action in serum
sickness to, 429
analogy of serum sick-
ness to, 430
analogy of tuberculin re-
action to, 444
anaphylactic poisoning,
nature of, 404 et
seq.
from precipitates. 394
proteid split products
of Vaughan and
Wheeler in, 405
symptoms of, similar
to peptone poi-
soning, 406
Anderson and Schultz's
work on. 365
anti-anaphylactic state
in. See Antiana-
phylaxis
antigen in, intervals be-
tween administra-
tions of, 376
identity of sensitizing
and toxic sub-
stances of, 387
Anaphylaxis, antigen i n,
identity of sensi-
tizing and toxic
substances of,
Doerr and liuss's
work on, 387
Wells' work on, 387
nature of, 370
path of introduction
of, 374
intracerebrally, 374
i n t r a-intestinally,
375
by feeding, 375
by rectum, 375
into large intes-
tine, 375
intravenously, 374
subcutaneously, 374
quantity of, adminis-
tered, 376 et seq.
specificity of. 371
degree of, 372
organ, 373
a u tosensitization
in, 373
species, 372
two separate sub-
xtances in, 386
Doerr and Russ's
experiments on,
387, 389
Gay and Adler's ex-
periments in. 3S6
Pick and Yamanou-
chi's experiments
in. 387
Arthus' work on, 361
asthma and. 436
Auer and Lewis's work
on. 364
autosensitization in, 373,
439
Besredka and Stein-
hardt's work on,
374 ct seq.
Besredka's theory of, 386
Besredka's work on, 375
ct seq.
Biedl and Kraus's work
on. 365. 368, 3(59
Bogomolez's work on,
371
Calvary's work on, 369
cell participation in, 390,
399 et seq.
clinical manifestations
of, 363
in dogs, 368
fall in blood pres-
sure in, 368
fall of temperature
in, 369
increase of lymph
flow in. 369
intestinal reaction
in, 370
lowered coagulabili-
ty of blood in. 36!)
in guinea pig, 363
alexin reduction in.
367
circulation symp-
toms in. 366
effect of atropin
and other drugs
on, 365
fall of temperature
in, 366
fever in, 366, 367
lowered coagulabil-
ity of blood in,
367
Anaphylaxis, clinical mani-
festations of, in
guinea pig, pulmo-
nary emphysema
in, 364
respiratory symp-
toms in, 364, 365
susceptibility o f
various breeds in,
368
temporary diminu-
tion of polynu-
clear leukocytes
in, 367
in rabbits, 368
clinical significance of,
428
dependence of, on pre-
ceding inocula-
tion, 360
diagnostic uses of, 446
diminution of alexin af-
ter anaphylactic
shock in, 392
significance of, 392
early work on, 359
Flexner's work on, 359
Friedberger's work on,
366
Friedemann's e x p e r i-
ments in, in vitro.
393
fundamental principles
of, 358
Gay and Southard's ar-
guments against
antigen - antibody
reaction theory
of. 385
Gay and Southard's the-
ory of, 386
Gay and Southard's work
on, 364
hay fever and, 436
in serum therapy. See
Serum sickness
in sudden attacks of ca-
tarrhal nasophar-
yngitis and con-
junctivitis, 437
in vaccine therapy, 434
incubation time in, 360,
376
Jobling and Petersen's
work on, 396
Lesne" and Dreyfus'
work on, 375, 376
Magendie on, 359
Manwaring's work on,
370
mechanism of anti-ana-
phylaxis in, 402
et seq.
desensitization in, 403
specificity in, 404
tolerance to anaphy-
lactic poison in,
403
Nicolle's theory of, 391
organ specificity in,
438
Otto's work on, 361
Pearce and Eisenbrey's
work on, 368, 369
Pfeiffer's work on, 366
phenomena related to,
406
toxic action of nor-
mal sera, 406
toxin hypersusceptibil-
ity, 409
Pick and Yamanouchi's
work on, 371
568
INDEX
Anaphylaxis. quantitative
methods applied
to study of, 389
Ranzi's work on, 367
relation of alexin to,
391
relation of antibodies
of, to other anti-
bodies, 401
Richet and Hericourt's
work on, 360
Richet and Portier's
work on, 360
Rosenau and Anderson's
work on, 362, 376
et seq.
sessile receptors, theory
of, 399
specificity of, 362, 363
sympathetic ophthalmia
and, 439
Theobald Smith phenom-
enon in, 361
Theobald Smith's work
on, 363
toxic action of normal
sera and, 406
toxin hypersusceptibility
and. 409
transference of, 362, 380
et seq.
true antigen-antibody re-
action, 388
tuberculin ophthalmo-re-
action and, 442
tuberculin reaction and,
440. See also Tu-
berculin reaction
tuberculin skin reaction
and, 442
Vaughan and Wheeler's
theory of mecha-
nism 'of, 390
Vaughan and Wheeler's
work on toxic
fraction of pro-
tein molecule in,
. 390
Vaughan's work on, 366,
367
Weichhardt and Schit-
• tenhelm's work
on. 369. 370
with bacterial extracts,
363
with normal serum, 362
with proteins, 362
Anaphylaxis. bacterial. 412
anaphylatoxin formation
in, 414. See also
under Anaphyla-
toxin
endotoxin theory of pro-
duction of. 414
Friedberger's e x p e r i -
ments in, 415,
416, 417
facts deduced from,
417
effect of excess of
bacteria adminis-
tered in, 418
effect of excess of
sensitization on
anaphylatoxin in,
418
effect of too pro-
longed exposure
in anaphylatoxin
in, 418
nature of bacterial in-
fections and, 420
et seq.
Anaphylaxis, bacterial,
N e u f e 1 d and
Dold's e x p e r i -
ments in, 419
serum anaphylaxis and,
413
Vaughan's theory of bac-
terial spitting as
cause of. 414, 415
Anaphylaxis, passive, 380
et seq.
Doerr and Russ's work
on, 381
duration of, 382
Friedemann's work on,
380. 381, 382
Gay and Southard's
work on, 380. 381
interval between injec-
tion of sensitized
serum and injec-
tion of antigen in,
382, 383
methods of production
of, 380, 381, 382
nature of reaction of. 382
Nicolle's work on, 380
Otto's work on, 380
Richet's work on. 380
Weill-Halie and Le-
maire's work on,
383
Anderson and Schultz's
work on anaphy-
laxis, 365
"Anergie," 509
Anthrax, relative suscepti-
bility of man and
animals to, 53
study of, in regard to
resistance and
immunity. 296
vaccination against, his-
tory of. 64
method of. 64
Anthrax bacilli, attenua-
tion of virulence
of, 18
virulence of. 15
Anti-alexin. 157
action of. 157
Anti-amboceptor, 152, 153
Anti-anaphylaxis, 362, 377
Besredka's work on, 375
et seq.
mechanism of. 402 et seq.
desensitization in, 403
tolerance to anaphy-
lactic poison in,
403
methods of producing,
377-378
Besredka and Stein-
hardt's methods,
377
Rosenau and Ander-
son's methods. 378
specificity of, 379. 404
"Anti-antibodies." 147
Antibodies. concentration
of. in lymphatic
organs in active
immunity. 101
in other organs in ac-
tive immuniza-
tion. 102
in active immunization,
85
locality of production
of, dependent on
locality of anti-
gen concentra-
tion, 102
Antibodies, normal, expla-
nation of, 234
origin of, 100
specificity of, 85
Antibody formation, body
cell in, 125
chemical nature of.
126
chemical action of anti-
gens in, 128
in active immunity, re-
moval of spleen
and, 101
mechanism of (side
chain theory), 124
by internal secretion
of body cell, 125
processes of metabol-
ism and, 125
overproduction of recep-
tors in, 128
principles of, 94. 96
Anticomplement, 157
action of, 157
Anticytophile interpreta-
tion of anti-sensi-
tization (Ehrlich
and Morgenroth),
153
controversy on, 153
"Antiformin." 70
Antigen-antibody reactions,
129. See also
Toxin - antitoxin
reaction
antibody production in
body cells in, 130
relationship between
susceptibility of
tissue and toxin-
binding properties
in, 131
side chain theory in,
129
specificity of, 98, 129
variety of antibodies in,
129
agglutinins, 129
antitoxins, 129
cytotoxins, 129
precipitins, 129
Antigenic properties of
cells, relation of,
to lipoid constit-
uents, 97
Antigens, action of, 95
active immunization
with, analogies to
drug tolerance,
99
characteristics of. vari-
ations in, 98, 99
classification of. 95
complex structure of.
Ehrlich and Mor-
genroth's concep-
tion of. 151
Definition of. 35. 94
"local" immunity in or-
gans directly in
contact with,
102
locality of production of
antibodies depen-
dent on locality
of concentration
of. 102
organ specificity of, 99
protein nature of, 97
specificity of, 98
Anti-isolysins. 147
, "Antiricin." 85
! Antisensibilisin, 386
INDEX
569
Antisensitization. a n t i -
cytophile interpre-
tation of (Ehrlich
and Morgenroth ),
153
controversy on. 153
Antisensitizers. 152, 153
non-specificity of, 154
Antitoxic serum, direct ef-
fect of, on toxin,
104
indirect protective ac-
tion of. against
toxin, 104
"normal" serum of Behr-
ing, 107
Antitoxin, chemical rela-
tions of. with
toxin, 114
definition of, 85, 86
diphtheria. See Diph-
theria antitoxin
production of. 129
by true toxins, 35
snake venom, 466
effect of heat on, 105
specific, substances in-
citing, 86
standardization of, 465
by means of toxin,
107
guinea pigs used in,
108
tetanus, production of,
4G5
use of. in passive im-
munization. 86
Antitoxin unit, diphtheria,
107
Antitoxinogen, 95
Antivenin, 466
Arrhenius and Madsen on
neutralization in
t o x i n-antitoxin
reaction, 120
Arthus, phenomenon of,
380
work of. on anaphylaxis,
361
Ascoli and Izar's work on
meiostagmin reac-
tion. 538, 539
Asiatic cholera, relative
susceptibility of
man and animals
to. 53
Asthma, anaphylaxis and,
436
Atrepsie. 56
Attenuation of bacteria by
chemicals, 66
by cultivation under
pressure, 66
by drying, 65
by heating. 65
by passage through ani-
mals. 65
by prolonged cultivation
above optimum
temperature. 65
by prolonged growth on
artificial media in
presence of own
metabolic prod-
ucts. 65
capsule formation in,
18
Auer and Lewis's work on
anaphylaxis, 364
Autocytotoxins, 93
Autogenous vaccines, 351
"Autohemolvsins," 146,
147
Auto-inoculation by mas-
sage in active
immunization, 340
Autolysins, 146
Autosensitization in ana-
.phylaxis, 373
Auxilysin, 167
Avian tuberculosis, relative
susceptibility of
animals to, 52
Bacillus botulinus, action
of, 4
Bacteria, adaptation of, in
body, 6, 7
agglutinability of, alter-
ations in, 226
by cultivation in im-
mune serum, 228
caused by heating,
226
normal differences
in, between
strains of same
species, 228. 229
spontaneous, 227
in agglutinoid, 229
agglutination in. See
under Agglutina-
tion
"agglutinin" bacteria,
agglutination of,
243
aggressin secretion of,
in body, and vir-
ulence of, 20-22
anti-opsonic properties
of, and antichem-
o t a c t i c sub-
stances, 325
» artificial cultivation of,
10
attenuation of, 18
methods of, 65. See
also under Atten-
uation
by laboratory ma-
nipulations. 17
capsulated. agglutination
of, 243
virulence of, 326
capsule formation in,
and virulence, 18
colloid phenomena and
action in, 560
destruction of. by cy-
tases in leuko-
cytes, 301. 302
by exudates. 300
by phagocytes, 300
different strains of. va-
riation in infec-
tion from. 15. 16
ectoplasmic hypertrophy
of. in relation to
virulence, 19, 20
effect of body tempera-
ture on invasive
powers of. 12
effect of cultural adap-
tation of. on vir-
ulence. 12
effect of path of intro-
duction of. on
infection, 12-14
on virulence of. 12-14
effect of quantity of, in-
troduced, on in-
fection. 14
entrance of. into body
tissues, 6
generalized action of, 24
; to phagocy-
of, 325
Bacteria, growth of, within
leukocytes, 298
in blood stream, 24
in localized infection, re-
action to, 26
through accidental
conditions, 25
in normal serum, resis-
tance
tosis
incubation of, 26
localized action
of, 23
measurement of relative
degrees of viru-
lence of, 15
negative charge of, in
suspension, 242
number of, introduced,
and relative viru-
lence, 15
occurrence of, 2
parasitic and saprophy-
tic, classification
of, 11
phagocytosis of. See
under Phagocyto-
sis
relative vi.-ulence of dif-
ferent strains of
same, 15, 16
resistance of, to phago-
cytosis, due to
nonabsorption of
opsonin, 326
resistance of living cell
to, 6
secondary abscesses
caused by, 25
selective action of, in
localized infec-
tion, 25
selective lodgment of, in
tissues, 40
sensitized, immunization
with, 68
sensitized and unsensi-
tized, influence of
salts on aggluti-
nation of, 243,
244
similar conditions pro-
duced by differ-
ent, 23
specificity of. and infec-
tion, 22
toxicity of, 423
use of, in active immu-
nization, 85, 87-
89
variation in virulence of,
when successively
passed through
animals. 16, 17
Bacterial anaphylaxis. See
Anaphylaxis, bac-
terial
Bacterial extracts, active
i m m u n i z ation
with. 69
extraction of bacteria
for. by mechan-
ical methods. 71
by permitting them
to remain in fluid
media. 70
Bacterial infections, con-
ceived as reac-
tion of body
against a foreign
antigen, 421
Bacterial precipitins, group
reactions in, 251
570
INDEX
Bacterial precipitins, group
reactions in, diag-
nostic value of,
252
partial or minor, 252
Bacterial products, active
i m m u n i z ation
with, 72
Bactericidal properties of
blood serum, 134
Bacterial proteins, 33
Bacterial toxemia, 423
Bacterial toxins, 28. See
also Toxins
action of, after distribu-
tion in body, 41
active immunization
with, 72
chemical structure of,
in relation to tox-
icity, 43
endotoxins. Sec Endo-
toxins
lesions produced by. in
course of excre-
tion, 45
local injury by, 45
nature of, 32
nature of union of, to
body cells. 44
nerves attacked by, 41
obtaining of, 32
from dead cultures, 32
from living cultures,
32
physical relationship
with body cells in
action of, 44
production of antitoxin
by, 35
selective action of, 40
principles of, 43
reasons for. 45
selective localization of,
39
specific distribution af-
ter introduction
of, 40
specific susceptibility of
tissues to. 46
chemicals inhibiting,
true. 33
analogy of, with en-
zymes, 36
bacteria producing. 34
characteristics of. 34
heat sensitiveness of,
36
incubation time of, 36
Bactericidal powers of
blood serum, 79,
134
alexin in, 137
nature of. 137
in vivo, 137
cholera experiments
in. 137
Bactericidal substances.
origin of, from
leukocytes, 169 et
seq.
Bacteriolysins, agglutinins
and, 321
in active immunization,
89
Bacteriolysis, 137
extracelluar theory of,
140
heat a factor in, 138
mechanism of. 138
immunity conferred by,
137, 138
Bacteriolysis, in immune
serum, 137, 138
Bordet's findings in,
140, 141
inactivation and re-
activation i n ,
140
Pfeiffer's phenomenon
in, technique of,
138 et seq.
specificity o f pro-
tection of. 137,
138
transference of power
of, 137, 138
leukocyte action i n,
168
leukocytes in, 140
Bacteriotropins, bacteri-
cidal sensitizers
and, 321 et seq.
normal opsonins and,
320 et seq.
presence of, in immune
sera without ly-
sins, 322
specificity of, 321
thermostability of, 320
Bail's aggressin 'theory, 21,
67
classification of para-
sites, 11
Bauer's modification of
Wassermann test,
209
Baumgarten's osmotic the-
ory of bactericidal
powers of blood,
135
Behring, Kitasato and
Wernicke. a n t i -
body theory in ac-
tive immunity of,
84
Besredka and Steinhardt's
work on anaphy-
1 a x i s, 374 et
seq.
on serum therapy of ty-
phoid fever, 478
Besredka's anti-endotoxic
serum in treat-
ment of typhoid
fever, 478
method of administra-
tion of antitoxin,
433
theory of anaphylaxls,
386
vaccines in prophylactic
i m m u n i z a tion
against plague,
489
work on antianaphylax-
is, 375 et seq.
Biedl and Kraus's work on
anaphylaxis, 365,
368, 369
Blood, non-putrefaction of,
79. 134, 256
phagocytic activities of,
in immunity. 79
Blood plasma, cell-free
inhibition of bac-
terial growth in,
in immunity, 79
presence of alexin in,
170-172
Blood serum, agglutina-
tion in, immune,
141
alexin in, 137
nature of, 137
Blood serum, anti-
bacterial powers
of, in immunity,
79, 134
anti-isolysins in, 147
autohemolysins in, 146.
147
bactericidal and agglu-
tinating powers
of, Wright's stud-
ies of, 328
bactericidal action of
immune, 81
alexin in, 137
early theories re-
garding, 134
in natural immuni-
ty, 79, 80
in viro, 137
cholera o x p e r i -
ments in, 137
mechanism of, 135
assimilation theory of,
136
by chemically un-
favorable environ-
ment. 135
osmotic theory of,
135
bacteriolysis in immune,
137
Bordet's findings in,
140, 141
cholera experiments
in, 137
summary of facts
in, 138
heat a factor in. 138
mechanism of, 13H
immunity conferred
by, 137, 138
inactivation and reac-
tivation in, 140
intracellular theory
of, 138
leukocytes in, 140
Pfeiffer's phenomenon
in. technique of,
138 et seq.
specificity of protec-
tion of, 137, 13S
bacteriolytic powers of
transferable, 137
cell-free, bacterial
growth in. 81
hemolysinogens in, 148
hemolysis in immune,
141
alexin or complement
in. 144
analogy of. to bacter-
iolysis, 142
Bordet's work on, 1.41
Ehrlich and Morgen-
roth on mechan-
ism of, 142
haptophore groups in,
142
relation of antigen,
amboceptor and
complement in,
143-145
work of Ehrlich and
Morgenroth o n,
143-144
work of Liefmann
and Cohn on. 145
isohemolysins in, 146
protective action of,
against bacteria,
50
Blood stream, presence of
alexin in, 170, 172
INDEX
571
Body fluids, bactericidal
powers of, in
natural immunity,
80
Bogomolez's work on ana-
phylaxis, 371
Borden's method of agglu-
tination, 223
Bordet, explanation of, on
agglutination. 240
findings of, on bacterio-
lytic power of
immune serum,
140, 141
o n neutralization i n
t o x i n-antitoxin
reaction, 122
views of, concerning re-
lation of antigen,
ainboceptor and
complement, 158,
159
concerning relation of
antigen, ambocep-
tor and comple-
ment, action of
complex in, 159
schematic representa-
tion, 159
Bordet-Danysz phenomenon
in neutralization
in toxin - a'n t i -
toxin reaction,
123
Bordet and Gengou's ex-
periment on alex-
in fixation, 186,
187
Bordet-Gengou phenomenon
in alexin fixation,
188 et seq.
Gay's experiments sup-
porting, 190
Moreschi's experiments
supporting. 189
Botujinus poisoning, ac-
tion of. 41
Botulinus toxin. 4
Bouillon Filtre (Denys),
357
Bovine colloid. 167
Bovine tuberculosis, rela-
tive susceptibility
of man and ani-
mals to. 52
Brownian movement in
colloids, 549
Buchner on bactericidal
power of blood in
natural immunity,
80
Calmette's investigations
in snake poisons,
46(5 ct seq.
ophthalmo-reaction. 442
Capsule formation in bac-
teria, by attenua-
tion. 18
virulence and, 18
Calvary's work on anaphy-
laxi's, 369
Carriers, bacillus, 2
Castellani, absorption ex-
periments of, in
agglutination. 232
Catarrhal nasopharyngitis
and conjunctivi-
tis, sudden at-
tacks of. anaphy-
lactic nature of,
437
Cell receptors, overproduc-
tion of, 152
structure of, 127
Cellular theory of immu-
nity, 136
Cercbrospinal meningitis,
epidemic, serum
therapy in, 471
early investigations in,
471
Flexner and Jobling's
work on, 472
Jochmann's investiga-
tions in, 471, 472
Kolle and Wassermann's
investigations in,
471
nature of action in, 473
results of. 472, 473
standardization of serum
in. 473
Chantemesse's early ex-
periments in se-
rum therapy of
typhoid fever, 477
Chemptaxis, 285
anaphylatoxiu and, 291
Eugelmaun's studies in,
287
influence of bacteria in.
288
influence of bacterial
extracts in, 288,
289
malic acid in, 286
of slime-molds or inyxo-
mycetes, 286
of spermatozoa of ferns,
286
Pfeffer's technique in,
286
physical explanations of,
292 ct seq.
selective, 291, 295
surface tension in. 293
"artificial aruebae"
and, 294
Chicken cholera, vaccina-
tion against, his-
tory of, 63
Cholera, active prophylac-
tic immunization
against, 487
Ferran's earlv investi-
gations in, 487
Haffkine's method in,
487, 488
Kolle's method in, 489
Strong's method i n,
489
Asiatic, relative suscep-
tibility of man
and animals to,
53
chicken, vaccination
against, history
of, 63
effect of path of intro-
duction of bac-
teria of, on infec-
tion, 14
experiments in, showing
bactericidal pow-
ers of blood se-
rum in, 137
hog, immunization with
bacterial products
in, 72
Cobra antitoxin, action of,
467
standardization of, 468
Cobra lecithid, 175
Cobra venom, action of, 467
Coctoprecipitin, 258
"Coefficient of extinction,"
332
Cole's work on serum
therapy of pneu-
monia, 477
Colics' law, 499
Colloid gold reaction, 539
Colloids, 543
application of phenom-
ena of, in elec-
trical field, 562
to action in animal
body, 560
to action in bacteria,
560, 561
to action with eggs of
Fundulus, 561
to biology, 559 et seq.
to Danysz toxin-anti-
toxin phenome-
non, 561
to nonagglutination in
excess of agglu-
tinin, 562
chemical properties of,
552, 553
chemical composition
in, 552
chemical reactions in,
552
electrochemical ioniza-
tion in, 553
classification of, 544
definition of, 543
emulsion, 544, 554
flocculation of, by elec-
trolytes, 553
acids and alkalies in,
553
concentration of elec-
trolyte in. 553
explanation of, 554
nature of sol in, 554
precipitin reaction an-
alogous to, 265
salts in, 553
zone phenomenon in,
555
gel, 544
Graham's work on, 543
irreversible. 544
lyophobic, 559
lyophylic, 559
mutual reactions of. 555
in two oppositely elec-
trical sols, 556
explanation of, 557
in two similarly elec-
trical sols, 555
protective action of
electrolyte in, 556
protective action of
great excess of
one colloid over
the other, 556
explanation of, 550
nature of, 544
physical properties of,
545 et seq.
Brownian movement
of particles in,
549
distribution of parti-
cles in, 549, r,r,o
electrical properties
in, 550
form of particles in,
546
kinetic energy in. 549
size of particles in,
measurement of,
546 et seq.
572
INDEX
Colloids, physical proper-
ties of, size of
particles in, mi-
croscopic, 547
osmotic pressure in,
547
rate of settlement
in, 548
ultrafiltration meth-
od in, 548
surface tension in, 550
et seq.
preparation of solutions
of, 558, 559
reaction in, analogous to
complement devi-
ation phenomenon
of Neisser-Wechs-
berg, 162
inhibition zones in,
162
reversible, 544
sol, 544
stability of, 545
suspension, 544, 554
Complement. See also
Alexin
amboceptpr and, Nogu-
chi's measurement
of quantitative
relations of,
164
quantitative ratio be-
tween, 163, 164
union of, Ehrlich and
Sachs's views on,
164, 165
definition of, 144
in hemolysis, 144
multiplicity of, 154 et
sea.
Borders views on,
156
Ehrlich's views on,
155
nature of, 154
chemical, 174
Complement deviation, 160
ct seq.
argument in favor of
Bordet's views,
162, 163
colloid reactions, analo-
gous to, 162
Gay's explanation of,
163
in hemolytic reactions,
163
Morgenroth and Sachs's
experiments sup-
porting. 163
pro-agglutinoid zone re-
action analogous
to, 162
Complement fixation, 186.
See also Alexin
fixation
in determination of na-
ture of unknown
protein, 211
delicacy of, 212,
213
technique of, 212
practical applications of,
198
test of, in diagnosis of
glanders, 216
in diagnosis of gonor-
rheal infections,
216
in diagnosis of malig-
nant neoplasms,
213
Complement fixation, test
of, in diagnosis of
malignant n e o -
plasms, von Dun-
gern's method of,
214
antigen produc-
tion for, 214
results of, 215
technique of, 214
et seq.
in diagnosis of syphi-
lis, 512
in diagnosis of tuber-
culosis. 217
Complement splitting, 178.
See also under
Alexin, splitting
of
Complementoid, 158
Complementophile group
of amboceptor,
149
Conglutinin, 167
Conjunctiva, susceptibility
of, to infection,
13
Corpus luteum cytotoxin,
92
"Cryptogenic tetanus," 5
Cultivation of bacteria, ar-
tificial, 10
Cytases, in phagocytes, de-
struction of bac-
t e r i a by, 301,
302
Cytolysins, 92
Cytolysis, 134
Cytolytic substances, ori-
fin of, from leu-
ocytes, 169 et
seq.
Cytophile group of ambo-
ceptor, 149, 152,
153
Cytotoxins, 92
specificity of, 92
Danysz effect in neutral-
ization in toxin-
antitoxin reac-
tion, 123
Danysz toxin-antitoxin
phenomenon, ap-
plication of col-
loid phenomena
to, 561
D a p h n i a. Metchnikoff's
study of, 274. 2JX5
phagocytosis in, 296
Dean's antiplague sera,
482
Denys and Leclef on im-
portance of phag-
ocytosis in im-
munity. 311, 312
Diphtheria, active immu-
nization in, with
toxin - antitoxin,
460
relative susceptibility of
man and animals
to, 53
Diphtheria antitoxic ser-
um, normal, 107
Diphtheria antitoxin, 448
antitoxin production in,
457
concentration of, 459
presence of, in blood of
normal individ-
uals, 450
Diphtheria antitoxin, pre-
servation of, 108
speed in administration
of, 449
speed in absorption of,
on intravenous injec-
tion, 451, 452
on subcutaneous in-
jection, 451, 452
speed of diagnosis for,
necessity of, 4,>.>
stability of, 108
standardization of, 457
by means of toxin,
107
early attempts in, 107
statistics showing reduc-
tion of mortality
with, 448
toxin production for, 454
choice of culture in,
454
cultivation of strain
in, 454. 455
culture medium in,
455
"maturing" of toxin
in, 455
testing of toxin in,
456
Theobald Smith's
method of, 456
unit of, 107
Diphtheria bacillus, action
of, 4, 5
Diphtheria toxin, action
of, 41
construction of, 118
determination of diph-
theria immunity
with. 464
normal, 107
stability of, 108
unit of, 107
Diphtheria toxin-antitoxin,
neutral mixtures
of, 460
Behring's method of im-
munization with,
4(50
advantages of, 461
chief value of, 461
danger of anaphylaxis
in, 461
determination of pres-
ence of free tox-
in or antitoxin in
convalescents fol-
lowing treatment
with. 464
human susceptibility to,
461
production of homolo-
gous antitoxin in
h u in a n beings
with, for passive
i m m u n i zation,
462
results of treatment
with, 462
standardization of anti-
toxin, 462, 463
limes-necrosis of toxin
in, 463
Komer's method of,
463
toxic action of, 460
Doerr and Kuss's experi-
ments on two sep-
arate substances
in anaphylactic
antigen, 389
INDEX
573
Doerr and Russ's work on
passive anaphy-
laxis, 381
Dochez and Gillespie's
work on serum
therapy in pneu-
monia, 477
Drug tolerance, analogy
between, and ac-
tive immunization
with antigens, 99
Dunbar's work on hay
fever, 430
von Dungern's method of
alexin fixation in
diagnosis of ma-
lignant tumors,
214
antigen production in,
214
results of, 215
technique of, 214 et
seq.
view of precipitin re-
actions, 268, 271
"Dust cells" of the lungs
in phagocytosis,
279
Dysentery, agglutination
reaction in diag-
nosis of, 221
Ehrlich, conception of
alexin fixation
(schematic) o f,
187
of relation of antigen,
amboceptor and
complement, 149,
150
interpretation of agglu-
tination by, 234
diagrammatic repre-
sentation of, 235
on multiplicity of alexin
or complement,
155
side chain theory of, in
toxin - antitoxin
reaction, 124
Ehrlich's. analysis, 104
Ehrlich's "antiricin," 85
Ehrlich and Morgenroth on
multiplicity o f
amboceptor, e x-
ample of, 150,
151
Ehrlich-Sachs phenomenon
in sensitization,
165
Bordet and Gay's inter-
pretation of, 166,
167
Eisenberg on residue an-
tigen and anti-
body in precipitin,
reaction, 268
Eisenberg and Volk's in-
terpretation o f
agglutination. 235
Endocarditis, malignant, 24
Endolysins, 305
Endotoxins, 33. 34. 39
characteristics of, 37
recent work on, 423
toxic cleavage products
of. 38
Engclmann's studies in
chomotaxis, 287
Enzymes, analogy of, with
true bacterial
toxins, 36
Enzymes, in phagocytosis,
endocellular and
extracellular, 305
leukocytic, 524
serum, 523
Epiphanin reaction, 537
Epithelioid cells, action of,
in phagocytosis,
284
Epitoxoids, definition of,
112
Epitoxonoids, 124
Erythrocyte laking, 91
"Exhaustion theory" of
immunity, 83
Exotoxins, 33. See also
Toxins, true
bacteria producing, 34
characteristics of, 34
chemically indefinable
nature of, 35
Fermentation, infectious
disease and, 1
micro-organisms caus-
ing, 1
Ferments, protective, in
animal body, 523
Abderhalden's e x p e r i-
ments with, 532,
533, 534
Ferran's investigations in
active prophylac-
tic immunization
against cholera,
487
Ferrata, experiments of,
i n complement
splitting, with
salt solution,
179
Ficker's reaction in agglu-
tination, 223
Flexner's observations on
anaphylaxis, 359
on serum therapy in
cerebros pinal
meningitis, 472
Flexner and A moss' work
on poliomyelitis,
497, 498
Flexner and Lewis' work
on poliomyelitis,
495
Forensic alexin fixation
tests, 211
Forensic determination of
unknown pro-
teins, 211
delicacy of, 213
technique of, 212
Fornet and Miiller, ring
test of, for pre-
cipitin blood
tests, 257
Friedberger, experiments
of, in bacterial
anaphylaxis, 415
on the nature of bac-
terial infections,
421 et seq.
work of, on anaphylaxis,
366
Friedemann, experiments
of, on anaphylaxis
in vitro. 393
work of, on passive ana-
phylaxis, 380, 382
in rabbits, 398
Fundulus, Loeb's experi-
ments with eggs
of, 561
Garbat and Meyer's work
OQ serum therapy
of typhoid fever,
479
Gastric juice, action of, on
stomach itself, 6
Gastrotoxin, 92
Gay and Adler's experi-
ments on two
separate sub-
. stances in ana-
phylactic antigen,
386
Gay and Southard's objec-
tions to antigen-
antibody theory
o f anaphylaxis,
385
theory of anaphylaxis,
385
work on anaphylaxis,
364
on passive anaphy-
laxis, 380
Gay's sensitized killed vac-
cines in prophy-
lactic typhoid
fever immuniza-
tion, 486
Gay's work on injection of
non-specific sub-
stances in infec-
tious diseases, 521
Gels, 544
Giant cells in phagocy-
tosis, 280
foreign body, 280
tuberculous, 280
Glanders, alexin fixation
test in diagnosis
of, 216
in horses, agglutination
reaction in diag-
nosis of, 222
relative susceptibility of
man and animals
to, 53
Gold reaction, colloidal,
539
Gonoccoccus, relative sus-
ceptibility of man
and animals to,
53
Gonorrheal infections,
alexin fixation
test in diagnosis
of, 216
Gottstein-Mathes' work on
serum therapy of
typhoid fever,
479
Graham's work on colloids,
543
Gramenitski's experiment
in reversal of
alexin inactiva-
tion by heating.
184, 185
Grohman on inhibition of
bacterial growth
by cell-free blood
plasma in immu-
nity. 79
Gruber-Widal reaction in
diagnosis, 220
Haemotoxins, 21
Haffkine's early work on
prophylactic im-
m u n i z a t i o n
against plague,
489
574
INDEX
Haffkine's method in active
prophylactic im-
munization
against cholera,
488
Half parasites, 11
Haptines, 129
of the second order, 235
of the third order, 150
varieties of, 129
Haptophore group in toxin,
110
action of, 110
Haptophore groups in
hemolysis, 142
Hay fever, anaphylaxis
and. 436
Dunbar's study of, 43G
reaction in, 436
anaphylactic nature
of, 437
toxic nature of. 437
treatment of, 437
Heat alkali-precipitin. 259
Heat-precipitins. 259
Hemagglutination, agglu-
tination of bac-
teria by serum
analogous to, 23(5,
237
Hemoglobinuria, paroxys-
mal, 147
autohemolysins in, 147
hemolysis in, 147
Hemolysinogens, human.
148
nature of, 148
Ilemolysins, anti-isolysins
in, 147
autolysins in, 146, 147
isohemolysins in, 146
specific, definition of. 92
specific inciting of, in
animal. 91
Hemolysis, alexin or com-
plement in, 144
amboceptor in. action
of. 149 et xeq.
anti-amboceptor in, 152,
153
anti-cytophile interpre-
tation of anti-
sensitization in
(of Ehrlich and
Morgenroth). 153
controversy on. 153
antisensitizer in, 152,
153
anti-isolysins in. 147
experiments of Liefmann
and Colin on. 145
in immune serum. 141
analogy of. to bacteri-
olysis. 142
Bordet's work on. 141
Ehrlich and Morgen-
roth on mechan-
ism of, 142
haptophore groups in,
142
relation of antigen,
amboceptor and
complement i n,
143-145
multiplicity of ambocep-
tor in, 150. 151,
154
recapitulation of views
of Ehrlich and
Morgenroth on,
152
Hemolytic properties of
normal serum, 91
Hemolytic reactions, com-
plement deviation
in, 163
Hemolytic serum, agglu-
tinins in, 93
Ehrlich and Morgen-
roth's conception
of neutralization
of, by antilysin
or anti-ambocep-
tor reacting with
cytophile group,
153
precipitins in, 94
Hemolytic substances, ori-
gin of, from leu-
kocytes, 169 et
seq.
Hepatotoxin, 92
Hiss, investigations of. on
therapeutic use of
leukocyt e ex-
tracts. 308, 309
Hiss and Zinsser's work on
injection of non-
specific substances
in infectious dis-
eases. 521
Hog cholera, immunization
*v i t h bacterial
products in, 72
Ilogyes method of treat-
ment in rabies,
494
Holobut's work on bac-
terial anaphy-
laxis, 413
Hopkins' method of stand-
ardization of vac-
cines, 353
"Horror autotoxicus," 147
Human isolysins. See Iso-
lysins, human
Humoral theory of immu-
nity, 136
Hydrophobia, active .pro-
phylactic immu-
nization against,
492. See also un-
der Rabies
Hyperleukocytosis, specific.
309
Hypersusceptibility. See
Anaphylaxis
toxin, and anaphylaxis,
409
"Ideal fall." in syphilis,
503
Immune bodies, structure
of. 127
Immune serum. See also
Serum, immune
agglutination in. 141
bacteriolytic power of
transferable. 137
bacteriolytic properties
of, Bordet's find-
ings in. 140, 141
direct neutralization of
toxin - antitoxin
reaction, the pro-
tective power of,
124
hemolysis in. 141
alexin or complement
in. 144
analogy of. to bac-
teriolysis, 142
Bordet's work on. 141
Ehrlich and Morgen-
roth on mechan-
ism of, 142
Immune scrum, hemolysis
i n . haptophore
groups in, 142
relation of antigen,
amboceptor and
complement i n,
143-145
work of Ehrlich and
Morgenroth o u,
143-144
work of Liefmann and
Cohn on, 145
phagocytosis in, 90
specific, agglutination of
bacteria in. 89
precipitin formation
in, 90
Immune isolysins, 148
Immunitats Einheit, 107
Immunity, 49
acquired, 60
artificially, 63
definition of, 63
history of, 61
increased phagocytosis
and. 299
active, relation of pha-
gocytosis to. 32!)
cellular theory of, 130
definition of, 3
diphtheria, determina-
tion of, wit h
diphtheria toxin,
464
"high tide" of, 340
humoral theory of, 136
lasting, diseases in
which one attack
conveys, 60
diseases in which one
attack does not
convey, 61
local, in organs directly
in contact with
antigens. 102
in skin infections, 103
natural, 49
cellular theory of. 80
definition of. 50, 62
humoral theory of,
80
inflammation in, 78
mechanism of, 78
theories concerning,
78-82
bacterial destruc-
tion by phago-
cytic cells, 78
bacterial growth
in cell - free
blood serum, 81
bactericidal pow-
er of blood in
natural immu-
nity, 80
bactericidal pow-
er of normal
blood in nat-
ural immunity.
79
bactericidal prop-
erties of extra-
vascular plas-
ma or serum,
81
inhibition of bac-
terial growth
by c e 1 1 - f r e e
blood plasma,
79
intracellular de-
struction o f
bacteria, 78
INDEX
575
Immunity, natural, mecha-
nism of, theories
concerning, phag-
ocytic activities
of blood, 79
phagocytic activities
of blood in, 79
principles of, 50
body temperature
in, 51
cultural conditions
for bacteria in
body a factor in,
56
increased invasive
powers of bac-
teria in, 56
individual differ-
ences in, 58
inheritance in, 56
racial differences in,
55
relative resistance
of animals in, 51
species resistance
in, 51
Pfeiffer phenomenon in,
138
phagocytosis in, 90
Immunization, 60
against snake venoms,
466
history of.
ImmunizatioK. active.
also Vaccine
apy
against anthrax, 64
against chicken cholera,
63
against small-pox, 62
agglutination of bacteria
in. 89
"alkalinity theory" of,
83. 84
antibacterial. 85, 87
antibodies in, 85
bodies in. fundamen-
tal principles of
theory of, 84
origin of. 100
antitoxic, 85
as a therapeutic meas-
ure, action of. in
generalized sys-
temic infections,
347
in local infections,
346
in successive local
infections, 347
value of. 346
in acute diseases,
350
in s u b a c u t e or
chronic cases. 349
as prophylactic measure.
value of. 345, 346
autogenous vaccines in,
351
auto-inoculation by mas-
sage in. 340
bacteria used in, 85, 87-
89
bacteriolysins in. 89
by means of living but
attenuated cul-
tures. 65
concentration of anti-
bodies in lym-
phatic organs in,
101
in other organs in.
102
Immunization, active, defi-
nition of, 63
"exhaustion theory" of,
«3
"high tide" of immunity
in, 340
in diphtheria, with tox-
in-antitoxin mix-
tures, 460. See
also Diphtheria
toxin-antitoxin
invasion of bacteria in,
mechanism of re-
action in tissue
cells against, 103
locality of production of
antibodies depen-
dent on locality
of antigen con-
centration in,
102
negative phase in, 338
second injection in,
338
successive inoculations
in, 338, 339
summation of, 338
339
non-bacterial anti-toxin-
stimulating sub-
stances in, 86, 87
"osmotic theory" of, 84
phagocytosis in, 90
phenomena following. 82
precipitin formation in,
90
of tissue cells
to invasion in,
103
removal of spleen in.
and antibody-for-
mation. 101
"retention theory" of. 83
second positive phase
in, 339
specificity of antibodies
in. 85
summation of positive
phase in. 339
tuberculins in, 355
vaccines in. 351
production of, 351
sensitized. 355
with dead bacteria,
351
with living bacteria,
351
standardization o f,
353
Hopkins' method of,
353
Wright's method of,
352, 353
with antigens, analogy
between drug tol-
erance and. 99
with bacterial extracts,
69
extraction of bacteria
for. by mechan-
ical methods. 71
by permitting them
to remain in fluid
media. 70
with bacterial products,
72
with dead bacteria. 68
methods used in kill-
ing bacteria for,
68
with fully virulent cul-
tures in sublethal
amounts. 66-68
Immunization, active, with
.sensitized b a c -
teria, 68
Immunization, active pro-
phylactic, in man,
483
against cholera, 487. See
also under Cholera
against paratyphoid
fever, 484
against plague, 487. See
also under Plague
against poliomyelitis,
495
against rabies, 492. See
also under Rabies
against small-pox, 490.
See also under
Small-pox
against syphilis, 498
against typhoid fever,
484. See also
under Typhoid fe-
ver
Immunization, passive, 74
antitoxins in, 86
definition of. 64
history of. 74
in diphtheria. See Diph-
theria antitoxin
in diseases caused by
bacteria which do
not form soluble
toxins, 468. See
also under Serum
therapy
therapeutic application
of. 75
toxin-antitoxin reaction
in. 104
underlying principles of,
75
Immunized animals, bac-
teriolysis in, 137
summary of facts in,
138
Incubation of bacteria. 26
Infection, acquired resis-
tance to, 60
adaptation of bacteria
in tissues in. 6, 7
aggressin secretion of
bacteria in body
and, 20-22
body temperature and. 51
capsule formation of
bacteria and. 18
chronic, adaptation of
bacteria in, 8
clinical manifestations
of, 28
conjunctiva susceptible
to, 13
criteria governing. 3
cultural conditions for
bacteria in body
and. 56
defense of intestinal
tract in, 12
defense of mucous mem-
branes in. 12. 13
defense of skin in, 12
definition of. 5, 6
different, produced by
same bacteria 23
effect of body tempera-
ture on invasive
powers of bac-
teria in. 2
effect of cultural adap-
tability of bac-
teria on, virulence
of, 12
576
INDEX
Infection, effect of path of
introduction o f
bacteria on, 12-14
effect of quantity of
bacteria intro-
duced on, 14
entrance of bacteria i-i
body tissues in, 6
focus in, 7-9
from bacteria in blood
stream, 24
generalized, 24
increased invasive pow-
ers of bacteria a
factor in, 56
incubation of bacteria
in. 26
individual differences
and. 56
inheritance and resis-
tance to, 56
localized. 23
reaction in, 26
selective action of
bacteria in. 25
through accidental
conditions. 25
natural resistance
against, 49
of various diseases, rel-
ative susceptibil-
ity of man and
animals to. 52
protective action of
blood serum
against, 50
protective action of leu-
kocytes against,
50
protective action of tis-
sues against, 50
ptomains and. 28
ptomains a s indirect
cause of. 31
racial differences and.
55
resistance of living cell
•to. 6
secondary abscesses in,
25
secondary modifying fac-
tors in, 2
selective lodgment of
bacteria in body
and. 40
similar, produced by dif-
ferent bacteria,
23
species resistance to. 51
specificity of bacteria
and. 22
susceptibility to, racial
differences in, 55
relative, 51
variation in, of differ-
ent strains of
same bacteria, 15,
16
variation in degree of,
in bacteria suc-
cessively passed
through animals,
16, 17
without infectious dis-
ease, 6, 7, 517
Infectious disease, defini-
tion of, 6. 8
influence of injections of
non-specific sub-
stances upon, 520
Inflammation, process of,
and phagocytosis,
280 et seq.
Inflammation with p y o -
genie staphylo-
cocci. 281
with tubercle bacilli,
283
Influenza, relative suscep-
tibility of man
and animals to,
53
Inheritance, a factor in re-
sistance to infec-
tion, 56
iso-agglutinins in blood
serum influenced
by, 58
Inhibition zones in colloid
reactions, 162
in precipitation and ag-
glutination, 162
of sera in agglutination,
236
Intestinal tract, defence
of, in infection,
12
Iso-agglutinins, 237
grouping of. 237. 238
in blood serum, 58
value of presence of. 239
Isohemolysins, 146
Isolysins, human, 148
grouping of, 148
testing of, for trans-
fusion. 149
iso-agglutinins analogous
to. 237
Isoprecipitins, 255
Jacobsthal's ultramicro-
scopic method of
finding precipi-
tates in syphilitic
sera, 204
Jenner, Edward, experi-
mentation of, for
i m m u n i z a t ion
against small-pox,
62
Jobling's work on serum
therapy in cere-
brospinal menin-
gitis, 472
Jobling and Petersen's
work on anaphy-
laxis, 396
Jochmann's investigations
in serum therapy
of epidemic cere-
brospinal menin-
gitis, 471, 472
Kolle's method of prophy-
lactic vaccination
in cholera, 489
Kolle and Otto's investi-
gations in pro-
phylactic immuni-
zation against
plague, 489
Kolle and Wasserrnann's
investigations in
serum therapy of
epidemic cerebro-
spinal meningitis,
471
Kraus, Rudolf, discovery
of specific precip-
itins by. 248
Kraus and Doerr's study
of bacterial ana-
phylaxis, 412, 413
Kraus and Stenitzer's se-
rum in treatment
of typhoid fever,
479
L+, definition of, 109
method of determination
of. 109
L0, definition of, 109
constancy, 110
method of determination
of. 110
Laking, erythrocyte, 91
''Landsteiner phenomenon"
of autohemolysis
in hemoglobiu-
uria. 147
Leishmann's technique for
determination of
opsonic index, 32i>
"Leistungskern," definition
of, 126
Leprosy, relative suscepti-
bility of man and
animals to, 54
Lesne and Dreyfus' work
on anaphylaxis,
375. 376
Leukine. 305
^Leukocidins, 21
Leukocyte extracts, thera-
peutic use of, 308 I
et seq.
Leukocytes, alexin extrac-
tion from, 304 et
seq.
growth of bacteria in,
298
in bacteriolysis. 140
action of, 168
in leukocytosis. action
of. 275. 276
in phagocytosis. 324
origin of bactericidal
and hemolytic
substances from,
168, 169 et seq.
oxidasc in, 527
phagocytic powers of, 50
proteolytic enzymes
from, in phagocy-
tosis. 306, 307,
525. 526 -,
Leukocytosis. 200
bacteria decreasing, 290
bacteria increasing. 290
sources of leukocytes in,
290
Leukoproteasos, 306, 307
Leukotoxin, 92
Limes-necrosis (L-n). 463
L i p o i d constituents of
cells, relation of,
to antigenic prop-
erties, 97
Lister on phagocytic activ-
ities of blood in
natural immunity,
79
Liver, production of alexin
in, 173
Loeb's experiments with
eggs of Fundulus,
561
Lubarsch on bactericidal
properties, of ex-
travascular plas
ma or serum in
immunity, 81
Liidke's work on serum
therapy in ty-
phoid fever, 479
INDEX
577
Lustig's antiplague se-
rum, 482
Lysins, production of, 130
Macrocytase, 109, 301
Magendie on anapbylaxis,
339
Malta fever, relative sus-
ceptibility of man
and animals to, 53
Manwaring's work on ana-
phylaxis, 370
Markl's s e r u m in treat-
iiu'nt of plague,
481
Marmorek's work on serum
therapy of strep-
tococcus infec-
tions, 474, 475
Measles, relative suscepti-
bility of man and
animals to, 54
Meat poisoning, 4, 31
Meiostagmiu reaction, 538
Ascoli and Izar's experi-
ments in, 538, 539
value of, in diagnosis,
538
Meningitis, epidemic cere-
brospinal, serum
therapy of, 471.
Kee also under
C e r e b rospinal
meningitis, . epi-
demic
Metabolism, processes of,
compared with
those of antibody
formation, 125
Metchnikoff and Besredka's
living sensitized
vaccines for pro-
phylactic typhoid
immunization, 486
Metchnikoff on bacterial
growth in cell-
free blood serum,
81
theory of, on bacterial
destruction by
phagocytic cells
in natural immu-
nity, 78
Metchnikoff's soured milk
therapy, 31
Microcytase, 109, 301
Minimum lethal dose, defi-
nition of, 108, 109
method of determination
of, 109
M L, D, definition of, 108.
109
method of determination
of. 109
Morgenrotlfs toxin-HCl
modification In
t o x i n-antitoxin
reaction, 100
Mucous membranes, de-
fence of, in infec-
tion, 12. 13
Mushroom, specific anti-
toxin from, 97
Narcotics, reduction of
Ehagocytosis by.
99
Natural immunity. &ec
Immunity, nat-
ural
i Necroparasites, 11
"Negative" phase in active
immunization, 33o
second injection in,
338
successive inoculations
in, 338, 339
"summation" of, 338,
339
Neisser and Friedemann,
experiments of,
on influence of
salts on sensitized
bacteria in agglu-
tination, 244
Neisrror and Wechsberg,
phenomenon • of,
100 et fsgy.
analogous to colloid re-
actions, 102
argument in favor of
Bordet's views,
102
Gay's explanation of,
103
Morgenroth and Sachs's
experiments sup-
porting, 103
pro-agglutinoid zone re-
action analogous
to, 102
Neisser's studies in syph-
ilis, 500, 509
Neoplasms, malignant,
alexin fixation in
diagnosis of, 21 ;i
von Dungern's method
of, 214
antigen production
for, 214
results of, 215
technique of, 214 et
seq.
Nernst on views of Arrhen-
ius and Madsen
on neutralization
in toxin-antitoxin
reaction, 122
Neufeld and Dold's experi-
m e n t s in bac-
terial anaphylax-
is, 419
Neufeld and Ilaendel's
work on serum
therapy of pneu-
monia, 470
Neurotoxin, 92
in snake venom, 407
Nicolle's theory of ana-
phylaxis, 391
work on passive anaphy-
laxis. 380
Noguchi's modification of
the Wassermann
test, 208, 209
schematic presentation
of, 209
"Normal" diphtheria anti-
toxic serum, 107
"Normal" diphtheria tox-
in, 107
"Normal" serum. &ec un-
der Serum
agglutinins in, 91
hemolytic properties of.
i/l
opsonins in, 91
toxic action of, and an-
aphylaxis. 400
Nuttall on bactericidal
power of normal
blood in natural
immunity, 79
Nuttall's experiments on
determining zoo-
logical classifica-
tions by means of
precipitin reac-
tion, 254, 255
Ophthalmia, sympathetic,
Elschnig's expla-
nation of, as ana-
phylactic r e a c-
tion, 439
Opium, reduction of pha-
gocytosis by, 299
Opsonic action, phagocy-
tosis due to, 313
Opsonic index, determina-
tion of, Leisn-
inaun's technique
for, 329
Simon, Lamar and
Bispham's tech-
nique of, 332, 3;^.>
Wright's technique
for, 330 et sea.
difficulties in, 332,
333
value of, 333
fluctuation of, in un-
treated patients
under influence of
exercise of dis-
eased parts, 340
in autoinoculations by
massage, 340
in sera of normal and
infected individ-
ual, comparison
of, 334
in serum therapy, com-
parison between
that in exudate of
infected foci and
blood serum, 340
in staphylococcus infec-
tions, 334
during vaccine treat-
ment with dead
s t a p hylococcus
cultures, 335
in treatment of gonor-
r h e a 1 arthritis
with autoinocula-
tion by massage,
340
in vaccine therapy, im-
provement and,
341
of acne, 339
of staphylococcus fu-
runculosis, 335
of sycosis, 330, 337
of tuberculosis, 341-
343
value of, in control-
ling therapeutic
vaccinations. 344
in showing degree
and conditions in
which vaccination
is successful, 344,
345
vaccine therapy and, 328
et seq.
value of, in therapeusis,
338
Opsonic powers of normal
serum, 314
reduction of, by heat,
314
Opsonins. 311. See also
Phagocytosis
578
INDEX
Opsonins, definition of, 313
immune, bactericidal
sensitizers and,
321 et seq.
increase of, 315
heated, increase of
power of, by ad-
dition of fresh
normal serum, 318
reactivation of, by
addition of alex-
in, 318
normal and, 320 et
scq.
resistance of, to heat,
315
specificity of, 321
thermostability of, 320
normal, 91
cooperation of heat-
stable and heat-
sensitive body in,
319
instability of, 314, 315
nature of, 316
similarity of, to alex-
in or complement,
316, 317
specificity of, 318
production of, in thyroid
gland, 173
qualitative difference be-
tween normal and
immune, 315,
316
specific thermostable, in
normal serum.
317
Organ specificity of anti-
gens, 99
"Osmotic theory" of im-
munity, 84
Otto's work on anaphy-
laxis. 361
in passive anaphylaxis,
380
Oxidase, in proteolytic en-
zymes, 527
Pancreas cytotoxin, 92
Panum's theory of intra-
cellular destruc-
tion of bacteria
in natural immu-
nity, 78
Parasites, biological transi-
tion of sapro-
phytes to. 5
Bail's classification of,
11
Parasitic bacteria, 4
Paratyphoid fever, agglu-
tination reaction
in diagnosis of,
221
immunization against,
484
vaccination against. 484
Paroxysmal hemoglobin-
uria. 147
hemolysis in, 147
Partial absorption method
of E h r 1 i c h in
measurement of
toxin -, antitoxin
combination, 115
Pasteur, "exhaustion the-
ory of," 83
experimentation of, on
i m m u n ization
against chicken
cholera, 63
Pasteur, work of, on immu-
nization against
anthrax, 64
on prophylactive im-
munization in ra-
bies, 492
Pathogenic bacteria, adap-
tation of, in tis-
sues, 6, 7
entrance of, in body tis-
sues, 6
saprophytic nature of
certain, 4
Pathogenic micro-organ-
isms, definition
of, 3
occurrence of, 2
resistance of living cell
to, 6
Pearce and Eisenbrey's
work on anaphy-
laxis, 368, 369
Persensitized cells, 180
Petterson's investigations
on therapeutic use
of leukocyte ex-
tracts, 308
Pfaundler's thread reac-
tion in agglutina-
tion, 222, 223
Pfeiffer on causes of bac-
terial anaphylax-
is, 414
work of. on anaphylaxis.
366
"Pfeiffer phenomenon" in
active immuniza-
tion, 89
in bacteriolysis. tech-
nique of, 138 et
seq.
Metchnikoff's view of
phagocytosis in
peritoneal exu-
date and, 302
Phagocytes, 276
fixed, 276
macrophages, 277
microphages, 277
motile, 276
Phagocytic index. 331
Phagocytosis, 272
acquired immunity and.
298
a 1 e x i n extraction in.
from leukocytes
and lymphatic or-
g a n s , 304 et
seq.
chemotaxis in. 285
influence of bacteria
in. 288. 289
influence of bacterial
extracts in. 288
malic acid in, 286
of slime-molds or myx-
omycetes, 286
of spermatozoa of
ferns, 286
Pfeffer's technique in.
286
destruction of bacteria
in. 297, 300
by alexin (or cytasr)
in leukocytes,
301. 302
action of, 302
Metchnikoff's inter-
pretation of, 302
by exudates. 300
by phagocytes, 300
destruction of red blood
cells by, 276
Phagocytosis, differences in
degree of. due to
bacteria, 325
differences in phagocytic
energy in, due to
leukocytes in, 324
differences in virulence
of bacteria, de-
pendent on their
resistance to leu-
kocytes in, 325
digestion among proto-
zoa and, 274
"dust cells" in, 279
early investigations in.
272
endothelial cells in, 278.
279
enzj'mes in, endocellular
and extracellular.
305
eosinophile cells in, 27<s
factors determining, 311
tixateur or sensitizer in.
action of, in im-
munized animals.
301
giant cells in, 280
foreign body. 280
tuberculous, 280
in daphnia, 296
in higher animals, 296
in immune serum, 315
bacteriolysins in, bac-
tericidal sensitiz-
ers and, 321 ft
neq.
heated, opsonic action
in, increase of, by
addition of fresh
normal serum, 31. S
increase of. 311
attributed to "stim-
ulins," 311
with addition of
leukocytes, 312
opsonin contents a
factor in, 313
opsonins in, increase
of, 315
normal opsonius
and. 320 et seq.
specificity of, 321
thermostability of,
320
in immunity, 90
in normal serum, op-
sonins in, cooper-
ation of h e a t-
sensitive body in,
319
nature of, 316
similarity of, to
alexin, 316, 317
specific thermosta-
ble. 317
specificity of, 318
in process of inflamma-
tion, 280 et seq.
with pyogenic staphy-
lococci. 281
with tubercle bacilli,
283
increase of. by injection
of leukocyte ex-
tracts, .308 et seq.
in increased resis-
tance, 329
with acquisition of
immunity, 299
intracellular digestion
and, 274
in vertebrates, 275
INDEX
579
Phagocytosis. leukocytes
in, 324
action of, 275. 276
polynuclear. 278
leukocytosis in, 290
bacteria decreasing,
290
bacteria increasing,
290
lymphocytes in, large.
measure of degree of. in
active immuniza-
tion. 329
Leishmann's technique
for, 329
Simon, La mar and
Bispham's tech-
nique for, 332.
333
value of, in therapeu-
sis, 338
Wright's technique
for, 330 e t
seq.
difficulties in, 332,
333
value of, 333
mechanism of process of,
280 et seq.
Metchnikoffs early in-
vestigations o n,
273. 274
normal and immune op-
sonic action in.
quantitative dif-
ferences between,
324
normal degenerative and
retrogressive proc-
esses and. 27(5
observation of, in vitro,
313
of micro-organisms, with
or without cul-
ture media. 297
opsonins in, 311. See
also Opsonins
phagocytes engaged in,
varieties of, 276
fixed, 276
macrophages, 277
microphages, 277
motile. 276
process of inflam-
mation in, 280 et
seq.
proteolytic enzymes from
leukocytes in. 306,
307
qualitative difference be-
tween normal and
immune opsonic
substances in. 315,
316
reduction of phagocytic
activity in. 298
by growth of bacteria
within leukocytes.
298
by protection of bac-
teria from pha-
gocytes, 299
by use of narcotics,
299
relation of, to active im-
munity, 329
relation of virulence to,
312
removal of extravasa-
tions of blood
and. 275
Phagocytosis, resistance of
bacteria to, due to
non-absorption of
opsonin, 326
resistance of infected
subject and, 296,
297
resistance of virulent
bacteria to, in
normal serum, 325
spontaneous. 313
tissue cells in, 278
varieties of body cells
engaged in, 278
dependent on nature
of invading sub-
stance, 279
Pick and Yamanouchi's ex-
periments on two
s e p a ra t e sub-
stances in ana-
phylactic antigen,
387
work on anaphylaxis,
371
von Pirquet and Schick's
studies of serum
sickness, 428 et
seq.
von Pirquet's tuberculin
skin reaction. 442
Placentar cytotoxin, 92
Plague, active prophylactic
i in in u n iz ation
against, 489
Besredka's vaccines in,
490
Haffkine's early work
on, 489
Kolle and Otto's in-
vestigations i n,
489
Rowland's vaccine in,
490
Strong's investigations
in, 490
relative susceptibility of
man and animals
to, 53
serum therapy of, 480
Dean's serum in, 482
Lustig's serum in. 482
Markl's serum in, 481
Rowland's serum in,
482
value of, 482, 483
Yersin. Calmette and
Borrel's investi-
gations in, 480
Yersin's serum in, 480.
482
value of, 481
"Plasmines," 72
Pneumococcus infection,
relative suscepti-
bility of man and
animals to, 54
Pneumococci, mutation of,
474-
Pneumonia, agglutination
reaction in diag-
nosis of, 221
serum therapy in. Cole's
work on, 477
Dochcz and (iilles-
pie's work on, 477
nature of action in.
476, 477
Neufeld and Ilaendel's
work on. 476
Poison-ivy, specific anti-
toxin from, 97
Pollantin. 437
Poliomyelitis, carriers in,
496
cause of, 495
epidemiology of, 495
Flexner and Amoss'
work on, 497, 498
Flexner and Lewis'
work on, 495
infection and immunity
in, 495
Landsteiner's and Pop-
per's work on,
495
relative susceptibility oi'
man and animals
to, 55
Polyceptors (Ehrlicli)./l7M"'
Precipitation, 248
inhibition zones in, 162
Precipitin reaction, 248
against heated proteins.
258 et seq.
coctoprecipitin in, 25,s
experiments on, 260-
262
h e a t-alkali-precipitin
in, 259, 260
native precipitin in,
259
70° precipitin in, 259
Schmidt's experiments
on, 260, 261
agglutination reaction
analogous to,
263
analogy of, to colloidal
flocculation, 265
autocytotoxins in, 263
bacterial precipitins in.
partial or minor,
252
specificity of, 251, 252
von Dungern's views of,
268, 271
Ehrlich's conception of,
264
electrolytes in, effect of.
265
forensic blood test in,
257
ring test of Fornet and
Miiller in, 257
group reactions of bac-
terial precipitins
in, 251, 252
diagnostic value of.
252
heat precipitins in, 259
heated precipitating ser-
um, effect of
mixed sera in,
266-267
protective action of,
266
inhibition zones in, 265,
266
isoprecipitins in, 255
medico-legal value of,
254
non-specific partial reac-
tions in, elimina-
tion of, 254
Nuttall's experiments on
determining zoo-
logical classifica-
tions by, 254, 255
organ specificity in, 262.
263
precipitinogen in. 249
chemical nature of,
249
effect of heating on,
258
580
INDEX
Precipitin reaction, precipi-
tinogen in, non-
protein, 249, 250
obtaining of, 249
precipitinoids in. for-
mation of, 265
precipitins in, delicacy
of, 253
determination of po-
tency of, 253
inactivation of, by
heat, 264
effect of, in bac-
terial filtrates, 264
production of, against
unformed p r o -
teins. 252, 253
methods of, 251
of specific, 249
by pepton, 251
effect of heating
on, 249
in animal sera by
foreign protein,
248
structure of (Ehrlich),
264
zymophore group In,
264
effect of heat on,
265
quantitative proportions
in, effect of, 265,
266
relative concentration of
reacting bodies
a factor in, 265.
266
residue antigen and an-
tibody in, 267 et
seq.
explanations of, 268
et seq.
experiment on, 269
salts in. effect of, 2(55
species determination by
means of, 253.
254
species specificity in. 262
specificity of, 248, 253.
254
vegetable proteins deter-
mined by. 255
zoological classifications
by means of, 254,
255
Precipitin tests, methods
of performing, 255
et seq.
forensic blood test in,
. 257
ring test of Fornet
and Miiller in, 257
Precipitinogen. 249
chemical nature of. 249
effect of heating on,
258
non-protein, 249, 250
nature of, 250
obtaining of, 249
Precipitinoids. 265
Precipitins. 248
against heated proteins,
258 et seq.
coctoprecipitin. 258
experiments on, 260-
262
heat - alkali-precipitin,
259. 260
native precipitin. 259
70° precipitin. 259
Schmidt's experiments
on, 260, 261
Precipitins, bacterial, group
reactions in, 251
partial or minor, 252
specificity of, 251, 252
definition of, 90
delicacy of, 253
determination of potency
of, 253
heat, 259
in hemolytic serum. 94
inactivation of, by heat,
264
effect of, in bacterial
filtrates, 264
isoprecipitins, 255
organ specificity of, 262,
263
production of, 129, 249
against unformed pro-
teins, 252. 253
methods of. 251
specific, by pepton.
251
effect of heating on,
249
in animal sera by
foreign protein,
248
"species," specificity of,
262, 263
specific, 248
discovery of, by Ru-
dolf Kraus. 248
structure of (Ehrlich),
264
zymophore group in, 264
e ff e c t of heat on,
265
Pregnancy, diagnostic-
value of Abder-
halden's protec-
tive ferments in,
534
Pro-agglutinoid phenome-
non in agglutina-
tion explained as
protective colloid
action, 236
Pro-agglutinoid zone In
agglutination, 1(52
complement deviation re-
action analogous
to. 162
Pro-agglutinoids. 235
Prof eta's law, 499
Prophylactic immuniza-
tion, active, in
man, 483
against cholera, 487.
See also under
Cholera
against paratyphoid fe-
ver. 484
against plague. 489. See
also under Plague
against rabies. 492. *SVe
also under Rabies
against small-pox, 490.
See also under
Small-pox
against typhoid fever,
484. See also un-
der Typhoid fever
"Protection." 195
"Protein fever," 367
Proteins unknown, alexin
fixation test in
determination o f
nature of. 211
delicacy of. 213
technique of. 212
Proteolysis. 524
Prototoxoids, 115
Protozoa, digestion among,
and its relation
to phagocytosis.
274
Ptomains, as indirect cause
of infection, 30
chemistry of, 29
classification of, 29
definition of, 31
relation of, to infection,
28
Putrefaction, chemistry of,
29
micro-organisms causing,
Pyemia, 25
Rabies, active prophylactic
i in m u n i zation
against, 492
Hogyes method of treat-
ment in, 494
Pasteur's work on, 492
preparation and attenu-
ation of virus for.
493
treatment of patients in,
493
Ranzi's work on anaphy-
laxis, 367
Rattlesnake poison, anti-
toxin for, 468
Reaction, colloidal gold.
539
epiphanin, 537
meiostagmin. 538
toxin-antitoxin, 104
Receptors, cell, overpro-
duction of. 152
complementophile, 149
cytophile, 149
definition of, 126
of third order, 150
sessile, in anaphylaxis,
399
structure of, 127
Resistance. tiee also Im-
munity
acquired. 60
bactericidal properties
in serum and, 21)7
body temperature and,
cellular theory of, 80
cultural conditions for
bacteria in body
and, 56
degree of phagocytosis
and. 296, 297
humoral theory of, 80
increased invasive pow-
ers of bacteria
and. 56
individual differences
and. 58
inheritance a factor in,
56
local, in skin infections,
103
natural, against infec-
tion, 49
racial differences in, 55
species, to infection, 51
vital, 524
"Retention theory of im-
munity," 83
Rhus toxicodendron, spe-
cific antitoxin
from, 97
Richet and H6ricourt'9
work on anaphy
laxis. 360
INDEX
581
Richct and Portier's work
0 n anaphylaxis,
360
Richet's work on passive
anaphylaxis, 380
Riein, "protein-free," 97
Romer's method for diph-
theria antitoxin
s« t a ndardization,
401
"Root-tubercle" bacilli, 7
Rosenau and Anderson, re-
searches of, in
bacterial anaphy-
laxis, 412
work on anaphylaxis,
362, 374 et seq.
Rosenow on variations in
streptococci, 474
Roux and Yersin, experi-
mental immuniza-
tion in hog chol-
era by, 72
Rowland's antiplague se-
rum, 482
vaccine in prophylactic
1 m m u n i zation
against plague,
489
Russell's vaccines for pro-
phylactic immuni-
zation against ty-
phoid fever, 485
Salt-inactivation of alexin,
178
Salts, effect of, in agglu-
tination, 243
in precipitation, 265
Saprophytes, biological
transition of, to
parasites, 5
occurrence of, 2
pathogenic powers of
certain, 4
pure, 11
Saprophytic micro-organ-
isms, definition of,
4
Scarlet fever, relative sus-
ceptibility of man
and animals to,
54
Schmidt, precipitation ex-
periments of, on
heat precipitins,
260, 261
Sensibilisin, 387
Stnsibilisinogen. 387
Sensitization. 134, 359
Bordet's views on, 162,
163
complement deviation in,
160
Ehrlich and Sachs's phe-
nomenon in, 165
Bordet and Cray's in-
terpretation of,
166, 167
Ehrlich and Sachs's
views on, 164,
165
Neisser-Wechsberg phe-
nomenon in. 160
Sensitized bacteria, immu-
nization with. 68
Sensitized tuberculin. 357
Sensitized vaccines, 355
Sensitizer. See Ambocep-
tor
Bordet's definition of,
159
Sensitizer, quantitative de-
termination of. in
immune serum,
160, 161
Septicemia, chronic, adap-
tation of bacteria
in, 7
secondary foci in, 7
Serum; See also Blood
serum
activities of, 523 et sc<i.
antitoxic, direct effect
of, on toxin, 104
indirect protective ac-
tion of. against
toxin, 104
bactericidal powers of,
79, 134
bactericidal properties
of, in immunity,
81
resistance and, 297
cell-free, bacterial
growth in, 81
immune, bacteriotropins
in. without lysins,
322
heated, reactivation
of. by addition of
alexin. 318
opsonins in, bacteri-
cidal sensitizers
and, 321 et seq.
heated, increase of
power of. by ad-
dition of fresh
normal serum,
318
increase of, 315
normal opsonins
and. 320 et seq.
specificity of. 321
phagocytosis in, 91,
increase of, 311.
See also Phagocy-
tosis
normal, agglutinins in,
91
antic omplementary
properties of, 196
hemolytic properties
of, 91
o p s o n i c powers of,
314
reduction of, by
heat, 314
opsonins in, 91
cooperation of heat-
stable and heat
sensitive body in,
318, 319
nature of, 316
resistance of. to
heat, 315
similarity of, to
alexin. 316, 317
specificity of, 318
thormostability of.
320
specific thermostable
opsonins in, 317
normal antitoxic, diph-
theria, 107
opsonins in normal and
immune, qualita-
tive differences in,
315, 316
reactions, physical fac-
tors in. 535
Serum sickness, 428
analogy of anaphylaxis
with, 430
Serum sickness, antibody
formation in, 431
incubation time in,
431
methods of administra-
tion of antitoxin
to avoid, 432
Besredka's method of,
433
by alteration of serum,
432
Friedberger and Mita's
method of, 434
in animal experimen-
tation, 433
with concentrated an-
titoxin, 432
von Pirquet and Schick's
studies of, 429 et
seq.
symptoms of, 428
accelerated reaction of
von Pirquet and
Schick in, 429
after first injection,
428
after second injection,
429
immediate reaction in,
429
analogy of, to ana-
phylaxis, 429
Serum therapy, anaphylax-
is in. See Serum
sickness
in diphtheria, 448. See
also Diphtheria
antitoxin
in diseases caused by
bacteria which do
not form soluble
toxins, 468
action of serum upon
extensive infec-
tion in. 469
antibacterial action in,
468
in epidemic cerebrospinal
meningitis, 471
See also under
Cerebros pinal
meningitis, e p i -
demic
in plague, 480. See also
Plague, serum
therapy of
in pneumonia, 476. See
also pneumonia,
serum therapy
in
in streptococcus infec-
tions, 473. See
also under Strep-
tococcus i n f e c -
tions
in typhoid fever, 477.
See also Typhoid
fever, serum ther-
apy of
Side chain theory, 124
anti-body production in
body cells in, 130
body cell in, 125
chemical nature of.
126
"Leistungskern" in,
126
side chains or recep-
tors in, 126
chemical action of anti-
gens in. 128
definition of side chains,
in, 126
582
INDEX
Side chain theory, diagram
showing cell-re-
ceptors and im-
raune bodies
(Ehrlich) in. 127
in toxin-antitoxin reac-
tion, 124
overproduction of recep-
tors in, 129
flow of, into blood a
cause of immu-
nity, 128
physical mechanism of,
124
recapitulation of. 130
relationship between
susceptibility of
tissue and toxin-
binding properties
in, 131
Skin, defence of, in infec-
tion, 12
Skin infections, local im-
munity in, 103
Small-pox, active prophy-
lactic immuniza-
t i o n against.
490
Jenner's discovery of,
490
production and prep-
aration of vac-
cine for, 491
history of experimenta-
tion in immuniza-
tion against, by
Jenner, 62
relative susceptibility of
man and animals
to, 54
vaccination for, history
of, 62
principles of. 62
Smith, Theobald, investiga-
tions of, in ana-
phylaxis. 363
phenomenon of, in ana-
phylaxis. 361
Snake venoms. 36
action of, 467
Activation of. by endp-
complement in
blood cells, 174
by sera, 174
antitoxins for. 466
e ff e c t of heat on,
ion
immunization against,
466
K y e s ' experiments in.
174, 175
neurotoxins in. 467
peculiarities of. 466
toxin-antitoxin combina-
tion in, stability
of. 105
toxin-antitoxin reaction
with. 105. 467
filtration experiments
in. 105
neutralization theory
of, 105
time element in,
105
toxin-HCl modification
of, effect of heat
on, 106
Sols. 544
Species resistance, 51
Specific hyperleukocytosia
309
Specificity, definition of, 76
Spermatotoxins, 92
Spleen, , removal of, and
antibody - forma-
tion in active im-
munity, 101
and susceptibility to
infection. 101
Standardization of anti-
toxin, guinea pigs
used in. 108
minimum lethal dose in,
109
Standardization of diph-
theria antitoxin,
by means of tox-
in, 107
early attempts at. 107
unit in, 107
Standardization of sera,
Pfeiffer's method
of, 139
Standardization of tetanus
antitoxin by
means of toxin,
107
Standardization of vac-
cines. 352
Hopkins' method of, 353
Wright's method o f,
352. 353
Staphylococcus furunculo-
sis, opsonic index
in vaccine treat-
ment of, 335
Staphylococcus infections,
opsonic index in.
334
during vaccine treat-
ment with dead
s t a p h y lococcus
cultures, 335
relative susceptibility of
man and animals
to, 54
Stern's modification of
Wassermann test.
209
Stimulins. 311
Strauss test, 25
Streptococci variations in,
474
Streptococcus infections,
agglutination re-
action in diag-
nosis of, 222
relative, of man and
animals, 54
serum therapy in 473
difficulties in, owing
to variations of
streptococci, 473,
474
early investigations in,
474
Marmorek's work on,
474. 475
nature of action in,
475
standardization of se-
rum in. 476
value of, 475
Strong's investigations in
prophylactic im
munization
against plague.
490
method of prophylactic
vaccination i n
cholera, 489
Sub-infection, 24, 234
"Summation of negative
phase" in active
i m m u n i z ation
338. 339
'Summation of positive
phase" in active
immunization, 339
Surface tension in chcmo-
taxis, 293
Susceptibility, body tem-
perature and, 51
cultural conditions for
bacteria in body
and, 50
increased invasive pow-
ers of bacteria
and, 56
individual differences
. and. 58
inheritance a factor in,
56
racial differences in,
55
relative, of animals to
infection, 51
species resistance to in-
fection and, 51
Sycosis, opsonic index in
vaccine treatment
of, 336, 337
Syntoxoids, 116
Syphilis, autoinoculation
in. 500. 502
Colles' law, 499
colloidal gold reaction
in, 539
diagnosis of, Bordet-Gen-
gou phenomenon
in, 188
by alexin fixation.
198, 199, 512
by direct precipita-
tion of syphilitic
serum by emul-
sions of lecithin
and of sodium
glycocholate, 204
"ideal fall" in, 503
immunity in, 498, 511.
512
Profeta's law. 499
reinfection in, 502
reinoculation in, 500,
502
relative susceptibility of
man and animals
to, 52, 505, 508
resistance of animals to,
508
superinfection in, 499
transmission of, to ani-
mals, 505
ultramicroscopic method
of finding precipi-
tates in sera of.
204
vaccination against,
511
Wassermann reaction in,
diagnostic value
of. 210, 211, 512.
513
technique of, 207
Bauer's modification
of. 209
Noguchi's modifica-
tion of. 208. 209
schematic presen-
tation of. 209
Stern's modification
of, 209
T.. definition of. 109
Temperature, body, and re-
sistance to infec-
tion, 51
INDEX
583
Tetanus, "cryptogenic," 5
relative susceptibility of
man and animals
to, 53
toxin, fixation in, by
brain tissues, lol
lipoidal substances
a factor in, 133
proteolytic enzymes
a factor in, 133
temperature a fac-
tor in, 132
Tetanus antitoxin, produc-
tion of, 405
standardization ot. 4(>>
by means of toxin, 10 <
Tetanus bacillus, action ot,
4, 5
Tetanus toxin, action ot,
41
Thread reaction of Pfaund-
ler in agglutina-
tion, 222, 223
Thymotoxin, 92
Thyroid gland, production
of alexin in, 172
production of opsonins
in. 173
Toxemia, 10
bacterial, 423
Toxicity, definition of, 11
Toxic unit, definition of,
109
Toxin-antitoxin, diphthe-
ria. See Diph-
theria toxin-anti-
toxin
Toxin-antitoxin combina-
tion, chemical re-
lations of, 114
effect of heat on, 106
in snake venom, stabil-
ity of. 105
toxin-HCl modification
of, 106
effect of heat on
106
measurement of, by par
t i a 1 absorption
method of Ehr
lich, 115
stability of, 105
valency of componen
parts of, 114
Toxin-antitoxin reaction,
104
analogy between chemi-
cal reactions and.
118, 119
antibody production in
body cells in, 130
body cells in, 125
chemical nature of.
126
chemical action of anti-
gens in, 128
concentration of rea-
gents in, 107
degrees of toxicity in.
123
direct neutralization the
protective power
of, 124
effect of heat on, 105
effect of temperature on.
104
mechanism of, 104
neutralization in. ab-
sorption theory
of, 123
Arrhenius and Mad-
sen on, 120
Bordet on, 122
Toxin-antitoxin reaction,
neutralization in,
Bordet' -D a n y s z
phenomenon in,
123
von Dungern's views
on, 124
Danysz effect in. 123
phenomena of, 119 ct
seq.
overproduction of recep-
tors in, 128
flow of, into blood a
cause of immu-
nity, 128
physical mechanism of,
124
quantitative relations in,
106, 123
relationship between sus-
ceptibility of tis-
sue and toxin-
binding properties
in, 131
side chain theory in,
124
specificity of, 124 129
speed of action of, 10 <
time element in, 105
with snake venom, lOo
filtration experiments
in, 105
neutralization theory
of, 105
time element in, 105
Toxin, bacterial. See Bac-
terial toxins
chemical relations of,
with antitoxin,
114
deterioration theory of
110, 111
definition of, 32
differences in combining
avidity of. Ill
diphtheria, construction
of, 118
normal, 107
direct effect of antitoxic
serum' on, 104
epitoxoid form of, 112
indirect effect of anti
toxic serum on
104
structure of, 110
toxoid, 110
prototoxoids in. 115
syntoxoids in, 116
toxon. 113
true, 33
analogy of. with en
zymes, 36
bacteria producing. 3
characteristics -of. 34
chemically indefinab
nature of. 35
diseases for whic
some investigatoi
claim. 471
heat sensitiveness o
36
incubation time of. £
production of ant
toxin by. 35
Toxin hypersusceptibilit
anaphylaxis an
409
Toxin spectra, constru
tion of, 116
definition of, 116
measurement of, 11
117
principles of, 116
oxin unit, definition of,
109
diphtheria, 107
oxoids, definition of, 110
"oxons, action of, 113
definition of, 113
structure of, 113
oxophore group of toxin,
110
action of, 110
ropins, 311
ubercle bacilli, effect of
body temperature
on virulence of, 12
uberculosis, avian type,
relative suscepti-
bility of animals
to, 52
bovine type, relative sus-
ceptibility of man
and animals to, 52
complement fixation test
in diagnosis . of,
217
human type, relative
susceptibility o f
man and animals
to, 52
meiostagmin reaction in
diagnosis of, 538
of cold-blooded animals,'
immunity of
warm-blooded ani-
mals to, 52
opsonic index in, 341-
342, 343
Tuberculin ophthalmoreac-
tion, anaphylactic
nature of, 442
Tuberculin reaction, anal-
ogy of, to ana-
phylaxis. 444
anaphvlactic nature of.
'440
Bail's experiments with
passive sensitiza-
tion in, 445
diagnostic value ot, 444
Koch's experiments in,
441
nature of, 440
Babes' interpretation
of, 441
Koch's interpretation
of, 441
Wassermann and
Brucks interpre-
tation ot. 441
specific antibody forma-
tion in, 444
Tuberculin skin reaction,
anapnylactic na-
ture of, 442
von Pirquet's interpreta-
tion of, 443
Tuberculins. 355
Bouillon Filtre (Denys).
357
New Tuberculin (TR
and TO), 356
New Tuberculin Bacilla-
ry Emulsion. 35 <
Old Tuberculin (Koch).
355
sensitized Tuberculin,
357
Tumors, malignant, alexin
fixation in diag-
nosis of, 213
von Dungern's method
of. 214
antigen production
for. 214
584
INDEX
Tumors, malignant, alexin
fixation in diag-
nosis of, results
of, 215
technique of, 214 et
seq.
organ-specific qualities
in, 373
Typhoid bacilli, attenua-
tion of virulence
of, 18
toxicity of, 423
Typhoid carriers, 3
Typhoid fever, adaptation
of bacteria in, 8
agglutination reaction
for diagnosis of,
219
macroscopic method,
219
microscopic method,
220
effect of path of intro-
duction of bac-
teria of, on infec-
tion. 14
meiostagmin reaction in
diagnosis of, 53*
prophylactic immuniza-
tion against, ac-
tive, 484
early experimentation
in, 484
living sensitized vac-
cines used in. 48(5
results of* in United
States Army. 4S5
Russell's vaccines in,
485
sensitized killed vac-
cines in. 486
relative susceptibility of
man and animals
to, 54
serum therapy in. 477
Besredka's anti-endo-
toxic serum in,
478
Chantemesse's early
experiments i n,
477
Garbat and Meyer's
work on. 479
Kraus and Stenitzer's
serum in, 479
Gottstein - M a t h e s '
work on. 479
Liidke s work on. 479
nature of reaction in,
478
vaccine treatment in,
486
Typhus fever, relative sus-
ceptibility of man
and animals to, 54
Ultramicroscope, 547
Vaccination, prophylactic,
in man. 483
in anthrax, 64
in cholera. 487. See also
under Cholera
in paratyphoid fever,
484
in plague, 489. See also
under Plague
in rabies. 492. See also
under Rabies
in small-pox. 491. See
also under Small-
pox
Vaccination, prophylactic,
in smallpox, his-
tory and general
principles of, 62
in typhoid fever, 484,
486. See also un-
der Typhoid fever
Vaccine therapy. See also
Immunization, ac-
tive
anaphylaxis in, 434
as a therapeutic meas-
ure, action of, in
local infections,
346
in generalized sys-
temic infections,
347
in successive local
infections. 347
value of. in acute dis-
eases. 350
in subacute or
chronic cases, 349
autoinoculations by mas-
sage or exercise
in, 340
"high tide" of immunity
in, 340
"negative" phase in, 338
second injection in,
338 -
successive inocula-
tions in, 338, 339
summation of, 338,
839
opsonic index in, 328 et
seq.
comparison between
that in exudate of
infected foci and
blood serum. 340
improvement and,
341
in tuberculosis, 341-
343
Leishmann's technique
for determination
of,. 329
Simon, Lamar and
Bispham's tech-
nique for deter-
mination of. 332,
333
value of. 338
in controlling thera-
p e u t i c vaccina-
tions. 344
in showing degree
and conditions in
which vaccination
is successful, 344,
345
Wright's technique for
determination of,
330 et seq.
difficulties in, 332,
333
value of. 333
relation of phagocytosis
to. 329
second positive phase in,
339
"Summation of positive
phase" in. 339
tuberculins in, 355
value of, as prophylactic
measure, 345,
346
as therapeutic meas-
ure, 346
Vaccines, autogenous, 351
production of, 351
Vaccines, production of,
with dead bac-
teria, 351
with living bacteria,
351
sensitized, 355
standardization of, 352
Hopkins' method of,
353
Wright's method of,
352, 353
Vaughan's work on ana-
phylaxis, 366. 367
on bacterial anaphylaxis,
Vaughan and Wheeler, pro-
teid split prod-
ucts of, in ana-
phylactic poison,
405
theory of, on mechanism
of anaphylaxis,
390
work of. on toxic frac-
tion of protein
molecule in ana-
phylaxis. 390
Virulence, aggressin secre-
tion of bacteria
in body and. 20-22
capsule formation of
bacteria and, 18
definition of, 11
dependent on resistance
of bacteria to leu-
kocytes in pha-
gocytosis. 325
differentiated from toxi-
city, 11
ectoplasmic hypertrophy
of bacteria in re-
lation to, 19. 20
effect of body tempera-
ture on, 12
effect of cultural adap-
tation of bacteria
on, 12
effect of path of intro-
duction of bac-
teria on, 12-14
effect of quantity of bac-
teria introduced
on, 14
increase of, by attenua-
tion of bacteria,
17
measurement of relative
degrees of. 15
of capsulated bacteria,
326
relation of, to phagocy-
tosis, 312
relative to number of
bacteria i n t r o -
duced, 15
specificity of bacteria
and, 22
variation in. of bacteria
successively
passed through
animals, 16. 17
of different strains of
snme bacteria, 15,
16
Virulins. 22. 326
"Virus fixe" in treatment
of rabies. 492
Vital resistance, 524
Wassermann reaction, 198
alexin fixation principle
in, 198
INDEX
585
Wassermann reaction, alex-
in fixation prin-
ciple in, not by
union of specific
syphilitic antigen
with spirochaeta
pallida antibodies,
204
alexin titration in, 200
antigen preparation for,
200
by addition of choles-
terin, 201
by method of Brown-
ing and Cruik-
Khank, 201
by method of Noguchi,
200
by methods of Forges
and Meier. 200
by methods of Weil
and Braun. 200
titration in, 202
diagnostic value of, 210,
211
in diagnosis of syphilis,
198
in diseases other than
syphilis, 210
in normal organs, 200
Klausner theory in,
204
Wassermann reaction, pre-
cipitation in, by
addition to syphi-
litic serum to leci-
t h i n emulsions,
204
produced with syphilitic
serum in antigens
from normal or-
gans, 200
specific antigen from
spirochaita pallida
cultures unsuit-
able in, 203
spinal fluid used in per-
f o r m a n c e of,
210
technique of perform-
ance of, 207
Bauer's modification
of, 209
Noguchi's modification
of, 208, 209
schematic presenta-
tion of, 209
refrigerator method in,
208
schematic presentation
of, 207
Stern's modification of,
209
theories of, 204, 205
Wassermann reaction, titra-
tion of hemolytic
amboceptor o r
sensitizer in, 205
ultramicroscopic method
of finding precip-
itates in syphili-
tic sera in, 204
Weichhardt's epiphanin re-
action, 537
Weigert's law of overcom-
pensation. 128
Widal's reaction, 220
Wright's method of stand-
ardization of vac-
cines, 352, 353
Wright's technique for de-
termination of op-
sonic index, 330
et 8€Q.
difficulties in, 332, 333
value of, 333
Wright's studies of bac-
tericidal and ag-
glutinating pow-
ers of blood se-
rum, 328
Yellow fever, susceptibilitv
to, 55
Yersin anti-plague serum,
480-482
YD 23432
BIOLOGY
LIBRARY
UNIVERSITY OF CALIFORNIA IJfeRARY