MR
V
PROTEIN SPLIT PRODUCTS
IN
BY
VICTOR C. VAUGHAN, M.D., LL.D.
DEAN OF THE DEPARTMENT OF MEDICINE AND SURGERY OF THE UNIVERSITY OF
MICHIGAN
VICTOR C. VAUGHAN, JR., M.D, A.B.
IN CHARGE OF THE TUBERCULOSIS WORK OF THE DETROIT BOARD OF HEALTH
JUNIOR ATTENDING PHYSICIAN TO HARPER HOSPITAL, DETROIT
J. WALTER VAUGHAN, M.D., A.B.
JUNIOR ATTENDING SURGEON TO HARPER HOSPITAL, DETROIT
ILLUSTRATED
LEA & FEBIGER
PHILADELPHIA AND NEW YORK
1913
Entered according to the Act of Congress, in the year 1913, by
LEA & FEBIGER,
in the Office of the Librarian of Congress. All rights reserved.
PREFACE
THE investigations recorded in this volume, begun
nearly fifteen years ago, were inaugurated in consequence
of certain fundamental ideas or theories held by the writer,
and these have directed and dominated all our labors along
this line. The purpose of this work has been to solve
scientific problems, rather than to discover practical
applications. The latter, so far as they have in any way
influenced our studies or even received our attention, have
been only incidental. Quite naturally our theories have
been more or less modified, and have developed as the work
has progressed, but the essentials and fundamentals have
not been materially altered. From time to time these
theories have been given in more or less detail, notably
in an address at the opening of the Medical Department of
the University of Toronto in 1905 (Canadian Jour, of
Med. and Surg., xviii, 283) and in the Shattuck lecture for
1906 (Boston Med. and Surg. Jour., civ, 215). However,
it may be well to restate briefly the original conceptions
which impelled us to begin and continue these studies.
The only essential and constant difference between
living and non-living matter is that within the molecules
of the former there is constant metabolism, while in the
latter no such process operates. We are to conceive of
the living molecules as made up of numerous atoms, and
each atom surrounded by its electrons; atoms and elec-
IV PREFACE
irons in ceaseless motion, and groups of atoms being con-
stantly cast out of the molecule and replaced by new groups
split off from outside matter. As soon as a molecule becomes
the seat of assimilation and excretion, it is no longer dead;
it lives. As a result of assimilation it acquires the property
of building up its own structure; then polymerization
follows and reproduction in its simplest form begins. The
one phenomenon always manifested by living matter, and
never exhibited by non-living matter, is metabolism.
When matter becomes endowed with life it does not
cease to be matter; it does not lose its inherent properties;
it is not released from the laws that govern its structure,
its attractions, and its motions. In studying living things
it should be borne in mind that they are material in com-
position and subject to the fundamental laws that govern
matter, and possessed of those properties essential to
matter.
Matter is alive when it feeds and excretes. The living
molecule not only absorbs; it assimilates. It chemically
alters what it absorbs, and within limits, it may be altered
by what it absorbs. Atomic groups taken into living
molecules enter into new combinations. The living mole-
cule is not stabile, but is highly labile. Its composition is
never constant, and it is never in a condition of equilibrium.
It is in constant chemical reaction with outside matter.
Apart from other matter it could not exist. There is a
constant interchange of atoms between it and external
matter. A condition, best designated as latent life, may
exist without interchange of atoms. This is exemplified
in spores, seeds, and ova. Matter existing in these forms
may be awakened into activity by proper stimuli; active
life begins with the interchange of atoms.
PREFACE V
Why is there this constant change of atomic groups
between the living molecule and outside matter? It is
for the purpose of supplying the living molecule with
energy. It is probable that in the absorption of energy
by the living molecule, oxygen is released from its combina-
tion with carbon or hydrogen, and is attached to nitrogen,
while in the liberation of energy the reverse takes place.
Nitrogen seems to be the master element within the living
molecule. It is by virtue of its chemism that groups are
torn from non-living matter, taken into the living molecule,
and assimilated by atomic rearrangement; and furthermore,
it is on account of the lability of the compound thus formed
that potential energy is converted into kinetic and work
is accomplished. A nitrogen side-chain serves as a receptor
and transmitter of oxygen, and thus the traffic in energy
within the living molecule goes on rhythmically. It is not
to be supposed that the nitrogen side-chain, which serves
as the receptor and transmitter of oxygen, consists of so
simple a body as nitrogen or nitrogen oxide, but it is probably
a highly complex nitrogenous body in which the location
of the nitrogen is central, as suggested by Allen. Nor is it
probable that only oxygen is broken off from the pabulum,
but substances containing this element. This is the way
in which the living molecule keeps up its constant, rhythmic
traffic in energy, absorbing heat by assimilation, and giving
it off by dissociation. Each living molecule has not only
one, but many of these nitrogenous groups that act as
receptors. Moreover, metabolism within the molecule is
not confined to the absorption of oxygen, and the casting
out of non-nitrogenous products of combustion. The
whole molecule is labile, and there is probably in every
living molecule a nitrogenous, as well as a non-nitrogenous,
vi PREFACE
metabolism. Nitrogen absorbed with the oxygen is, in
part at least, utilized in replacing the waste in this element,
and the carbon brought into the molecule at the same
time is in part detached by the free valences in the carbo-
hydrate groups, and used to repair loss in this part of the
molecular structure. In the living molecule it is probable
that nitrogenous metabolism proceeds much more slowly
than the carbon and hydrogen metabolism, but both move
rhythmically, and the tempo depends upon the swing of
the atomic groups that constitute the molecule, and this
rate can be changed, hastened or retarded, by alterations,
either physical or chemical, in the medium in which life
resides. When the molecule is in active life its food is
prepared for it by ferments, and it is quite certain that
these ferments have their origin in the nitrogenous metab-
olism of the living molecule.
The keystone or archon of the protein molecule is our
poison. It is common to all protein molecules. It is the
primary group. One protein differs from another in the
secondary and tertiary groups. Ordinary proteins are not
poisonous, because in them the chemism of the primary
group is satisfied by combination with secondary groups.
Strip off the secondary groups and the primary becomes
poisonous on account of the avidity writh which it combines
with the secondary groups of other molecules.
The specificity of proteins resides in the secondary groups
of their molecules, and all specific protein reactions are
due to these groups. This is true of agglutination, precipi-
tin, and lytic reactions. Biological relationship between
proteins is dependent upon the chemical structure of their
molecules. The poisonous part of a protein is its primary
group; the sensitizing part is found among the secondary
PREFACE vi i
groups. The former ^ physiologically the same in all
proteins. There probably are chemical differences in the
primary groups of varied proteins, and it is possible that
fine physiological differences may be detected by more
careful study, but the primary group is the ring about
which all proteins are built, or at least, it contains this
ring; and just as innumerable compounds may be built
with the benzol ring as a nucleus, so all proteins are con-
structed about a common centre. The secondary groups
are not identical in any two kinds of proteins. There may
be one or more common to the two, but in some respects
there are differences.
The cell is not the unit of life; life is molecular. The cell
is not only made up of protein molecules, but its form and
function are determined by the chemical structure of its
constituent molecules. The lines along which the spore,
seed, or ovum develops are determined by the chemical
structure of its ^proteins. Growth in other directions is
impossible, and this accounts for stability in reproduction.
However, gradual changes in the chemical structure of
living proteins occur, and in these lies the basis of organic
evolution.
The basic points of our theory, as stated above, will be
in evidence throughout this volume. The experimental
wyork devoted to the development of this theory could not
have been done without the aid of able assistants who
have devoted much time to it, and all without adequate
reward. Besides those associated with me in the prepa-
ration of this volume, special mention is due Drs. Sybil
May Wheeler and Mary Leach. The former gave eight
years and the latter two years of most devoted and skilful
service to the elaboration of the problems discussed here.
vill PREFACE
I regard the studies recorded in this volume as the mere
beginnings of work which should be developed. I dare say
that our record contains many imperfections and possibly
some errors. Future studies will perfect the former and elimi-
nate the latter. Attempts to solve the problems stated in
this volume have occupied many years and filled them with
the interest and pleasure that always come to those who
seek to widen the fields of the known.
THE SENIOR AUTHOR.
ANN ARBOR, 1913.
CONTENTS
CHAPTER I
INTRODUCTION
Bacteria are participate proteins; all true proteins contain a poi-
sonous group; the chemical nucleus contains the poison; when proteins
are disrupted the poisonous group may be set free; the patho-
genicity of a bacterium is not determined by its capability of forming
a poison, but is dsterminsd by its ability to grow and multiply
in the animal body; any foreign protein which can grow and mul-
tiply in the body of a given animal is pathogenic to that animal;
the infectious diseases result from the parenteral digestion of proteins;
natural bacterial immunity, that which follows an infectious disease,
and that induced by vaccination, result from inability of the organ-
ism to grow and multiply in the animal body; protein sensitization
and bacterial immunity, apparently antipodal, are in reality identical;
protein sensitization consists in the development of a new function
in certain body cells — that of elaborating a specific, proteolytic
ferment; a foreign protein introduced into the blood is distributed
through the tissues; vaccines are protein sensitizers; toxin and
bacterial immunities are different; the protein poison is not a toxin;
it is not specific; it elaborates no antibody; it develops a specific*
ferment; different proteins tend to accumulate in predilection places;
the symptoms of the infectious diseases are largely determined by the .
organ or tissues in which the foreign protein accumulates; the poison
elaborated in all the infectious diseases is the same; when a cell in
the animal body is permeated by a foreign protein, the former strives
to elaborate a ferment by which the latter is destroyed; this we
believe to be a biological law 18
CHAPTER II
THE GROWTH OF MASSIVE CULTURES OF BACTERIA
The large tanks and the preparation of bacterial cellular substances 29
x CONTENTS
CHAPTER III
PRELIMINARY EXPERIMENT OF BACTERIAL CELLULAR SUBSTANCES
They consist essentially of complex proteins, each of which contains
a poisonous group 37
CHAPTER IV
THE CHEMISTRY OF BACTERIAL CELLULAR SUBSTANCES
Their protein, nuclein, carbohydrate, fatty and amino constituents 52
CHAPTER V
THE CLEAVAGE OF PROTEINS WITH DILUTE ALKALI IN SOLUTION
IN ABSOLUTE ALCOHOL
Bacterial, vegetable, and animal proteins can be split into poison-
ous and non-poisonous parts; the former are non-specific, the latter
are specific; the exact nature of neither of these portions is yet known 95
CHAPTER VI
ACTION OF ANIMALS
The action of the living bacillus, of the dead bacillus, and of the
poisonous split product 119
CHAPTER VII
THE PRODUCTION OF ACTIVE IMMUNITY WITH THE SPLIT
PRODUCTS OF THE COLON BACILLUS
The establishment of a certain degree of tolerance with the poi-
sonous product; this is non-specific; the production of a mild degree
of immunity with the non-poisonous product; this is specific . . 137
CHAPTER VIII
THE SPLIT PRODUCTS OF THE TUBERCLE BACILLUS AND THEIR
EFFECTS ON ANIMALS
The cellular substance ; the cell poison ; the cell residue ; the precipi-
tate poison; the precipitate residue; the final filtrate; the action of
these on animals; the effects of the tuberculo-poison; sensitization
with tuberculo-protein; the relation of tuberculo-sensitization to
immunity 164
CONTENTS Xl
CHAPTER IX
THE ANTHRAX PROTEIN
Literature, investigations; the anthrax cellular substance, like
other proteins, contains a poison; sensitization with anthrax protein 189
CHAPTER X
THE CELLULAR SUBSTANCE OF THE PNEUMOCOCCUS
Difference in virulence in strains; properties and effects on animals
of the cellular substance; the action of the poisonous portion; auto-
lysis of the pneumococcus; sensitization with pneumococcus protein 205
CHAPTER XI
PROTEIN SENSITIZATION
Introduction; definition; the sensitizer; all true proteins sensitize;
volatile sensitizers; the sensitizing group in the protein molecule;
the effects on different animals; period of incubation; the anaphyl-
actic state; the reinjection; symptoms; the mechanism of anaphyl-
axis; passive anaphylaxis; anti-anaphylaxis; the Arthus phenomenon;
anaphylaxis and toxic sera; the toxogens; anaphylaxis in vitro; the
poison; j8-iminazolylethylamin; the kyrins; anaphylatoxin ; physio-
logical action of the protein poison; general physiological action of
proteins; sensitization is cellular; theories; theory of Friedberger;
1 theory of Vaughan and Wheeler; theory of Nolf 214
CHAPTER XII
PARENTERAL DIGESTION
The disposition of peptones; the fate of proteins introduced
directly into the circulation; the' poisonous action of proteins; egg-
white injected into the stomach of a rabbit may be in part absorbed
unchanged; egg-white injected into the rectum of a rabbit may be,
in part at least, absorbed unchanged; egg-white injected into the
peritoneal cavity of a rabbit may be absorbed unchanged; egg-
white injected intravenously in a rabbit quickly disappears from the
circulating blood; egg-white injected intravenously in a rabbit may
be detected in the peritoneal cavity, in the bile, and in certain
organs after it has disappeared from the circulating blood; the 'njec-
tion of a large amount of egg-white intravenously in a rabbit may
prove fatal; the blood is a digestive fluid; proteolytic digestion in
the blood is regulated by the accumulation of digestive products . 342
xii CONTENTS
CHAPTER XIII
PROTEIN FEVER
The production of acute, intermittent, remittent, and continued
fevers by the injection of foreign proteins; fever results from the
parenteral digestion of proteins; the sources of fever in the paren-
teral digestion of proteins; fever per se is a beneficent process . . 373
CHAPTER XIV
SPECIFIC FERMENTS OF THE CANCER CELL
Extra- and intracellular ferments; a ferment developed in animals
by injections of cancer protein; the nature and action of this ferment 416
CHAPTER XV
THE PHENOMENA OF INFECTION
How bacteria grow; how bacteria cause disease; the phenomena
of the period of incubation; the phenomena of the active period of
the infectious diseases; the germicidal properties of the blood; the
phenomena of tubercular infection; the tuberculin test; vaccines and
sensitization; sensitization and idiosyncrasies to food and medicine . 436
PROTEIN POISONS
CHAPTER I
INTRODUCTION
MANY years ago the senior contributor to this volume
began a research on the chemistry of bacterial cellular
substance. This work has grown and the progress made
has been reported from time to time in current scientific
and medical literature. Able assistants have rendered
valuable service, and as the research has developed it has
been correlated with that done along similar lines in other
laboratories. We feel that the time has come when the
more important facts ascertained along this and related
lines, by all investigators, should be classified and proper
deductions drawn from them. We are the more inclined
to do this because we believe that the proper interpretation
of the results obtained opens up a view of the etiology and
development of both immunity and disease, which has
hitherto not been appreciated. We have thought it best
to state briefly in this introduction some of the most impor-
tant points dwelt upon in the volume. We have done this
somewhat dogmatically, hoping that they will impress the
reader and hold his attention while they are more fully
detailed in subsequent chapters.
1. Bacteria are essentially particulate, specific proteins.
Bacteria are usually classified as microscopic plants, but we
have sought diligently for the presence of cellulose in their
structure, with uniformly negative results. We have
shown that some bacteria, at least, contain two carbohy-
2
18 PROTEIN POISONS
drates, but neither of these gives the reactions characteristic
of cellulose. One of these is certainly a constituent of the
nucleic acid group, having the same relation, or at least a
similar relation, to the other members of this group as
exists in the nucleins and nucleoproteins found in the
vegetable and animal world. Our studies together with
those of other investigators render it quite certain that
bacterial cellular substance yields the nuclein bases on
hydrolysis. The position of the second carbohydrate in
the molecular structure has not been determined with
certainty. It is thought possible that it is simply stored
in the cell as a reserve food supply; but that this is not
true is indicated by the fact that it cannot be removed by
simple solvents, and that its separation is secured only
after disruption of the molecular structure. We are inclined
to the opinion, subject to change as the result of more
exact knoAvledge, that the second carbohydrate group found
in at least some bacteria is an essential constituent of the
protein structure. Our work on the amino-acids, both
mono- and di-amino, makes it certain that the greater part
of the bacterial cell is made up of true proteins. We have
not only isolated and identified many of the amino-acids,
but we have shown that they exist in widely different
proportions in different species of bacteria, just as they do
in different proteins obtained from plants and animals.
While fats and waxes are found in relatively large amount
in certain bacteria, notably in bacillus tuberculosis, we
see no reason for concluding that they are essential con-
stituents of the living molecule. That they are specific
products of the life activities of certain bacteria we are
convinced, but we have seen no reason for believing that
they are essential constituents of the bacterial molecules.
We conclude that, chemically, bacteria, at least those
with which we have worked, are nucleoproteins or glyco-
nucleoproteins. While bacteria are morphologically simple
in structure and without differentiation in parts, chemically
they are complicated in structure, quite as much so as
many of the tissues of the higher plants and animals. The
INTRODUCTION 19
demonstration that bacteria are not only proteins, but
relatively complex proteins, is a matter of marked impor-
tance. It shows that in many of their life processes they
must bear a close resemblance to the cells of the higher
animals; that they require the same kind of food, which
they select, assimilate, and excrete in much the same way;
that the conditions of life are much the same; what is favor-
able to one bearing a like relation to the other, and what
proves injurious to one having a like effect upon the other.
2. All true proteins contain a poisonous group. At first
we found that the cellular substance of certain pathogenic
bacteria could be split up with the liberation of a poisonous
substance, then we tested non-pathogenic bacteria, then
animal and vegetable proteins, and all with the same result.
Not only do all these contain a poison, but so far as its
gross effects on the higher animals have been studied, the
same poison. We have held that when we know more
about these poisonous oodies obtained from all proteins,
it will be found that chemically they are not identical, but
physiologically they are so closely similar that up to the
present time we have not been able to distinguish one
from the other by the symptoms induced. The poison
obtained from the typhoid bacillus, that from egg-white,
and that from edestin of hemp-seed kill animals in the same
doses, with the same symptoms and with the same lesions.
This is striking evidence of the similarity in the structure
of the protein molecule, whether it be of bacterial, animal,
or vegetable origin. One cannot resist the temptation to
formulate a theory to fit these facts. Indeed, the theory
Unfolds itself and may be briefly expressed as follows: All
proteins are constructed on the same model and contain a
chemical nucleus, archon, or key-stone. This is the poison-
ous group and is practically the same in all proteins. One
protein differs from all others in its secondary and possibly
its tertiary groups. In these lies the specificity of proteins.
Living proteins function through their secondary and
tertiary groups. When the primary group is detached
from its own subsidiary and specific groups it manifests
20 PROTEIN POISONS
its poisonous action through the avidity which it has for
the secondary groups of other proteins. These are thus
detached from their normal positions and consequently
the living protein is deprived of its capability of functioning
normally. This is only a theory, but it is one which naturally
suggests itself.
3. The chemical nucleus does not become a poison until
stripped in part at least of its secondary groups, and the
intensity of its poisonous action is determined by the thorough-
ness with which the secondary groups have been removed.
The protein molecule may be regarded as a highly complex
neutral salt, made up of many basic and acid groups.
One of these components, it may be either a basic or an
acid group (or it may have within itself both a basic and
an acid group), is the chemical nucleus of the molecule. In
its natural condition its chemism is satisfied by nicely
adjusted combination. When this combination is disrupted,
which may be accomplished either by chemical agents or
by enzymes, the chemical nucleus is set free, more or less
completely, and to the extent that it is released from com-
bination, it becomes, in the presence of living proteins, a
poison because it disrupts the same. We have shown by
direct experiment that the protein poison may be at least
partly neutralized by being kept for some days in the
presence of an alkaline carbonate at 37° C.
4. When proteins are submitted to the action of disrupting
agents there is the possibility of the chemical nucleus being
set free more or less completely, and to the extent that it is
detached it becomes a poison. We have found that this
occurs when proteins are carefully disrupted by either
dilute acid or dilute alkali. So far as our work has gone
the best agent with which to disrupt the protein molecule
and obtain the largest yield of poison is a 2 per cent, solu-
tion of caustic soda in absolute alcohol. This is a crude
procedure and much of the poison is destroyed in the
process. The disruption easily extends beyond the point
where the poison is set free and much of the product sought
is destroyed. In peptic digestion the poison becomes active
INTRODUCTION 21
at about the stage of the formation of peptone, and it has
long been known that peptones are quite highly poisonous
when administered parenterally. This also is' a crude
method of obtaining the poison, and with all the work that
has been done along this line, we do not know whether the
peptone is itself poisonous or whether its poisonous action
is due to admixture with some other digestive product.
We do know that as alimentary digestion proceeds the
protein poison itself is destroyed. Indeed, we had no
conception of the small amount of protein necessary to
furnish a lethal dose of the poison, until we submitted
proteins to the blood sera and organ extracts of sensitized
animals. Then we found that 1 mg. of protein may supply
enough poison to kill a guinea-pig when injected intra-
venously. But to produce the poison in this way necessi-
tates a delicate adjustment between substrate and ferment
which is imperfectly understood, and consequently inade-
quately controlled, and we can know that we have produced
the poison in any given experiment only by its effect on
an animal. Thus it happens that after years of study we
are still quite ignorant of the true nature and chemical
composition and structure of the protein poison.
5. The pathogenicity of a bacterium is not determined by
its capability of forming a poison. Non-pathogenic bacteria
yield just as much of the protein poison as the most highly
pathogenic, and the proteins of our food contain the same
poison that is found in pathogenic bacteria.
6. The pathogenicity of a bacterium is dependent upon its
ability to grow and multiply in the animal body. Any micro-
organism which can grow and multiply in an animal body
is pathogenic to that animal. Growing and multiplying
in the animal body means that the invader converts the
proteins of the animal into its own proteins, transforms
native into foreign proteins, and the accumulation of foreign
proteins can result only from the destruction of the native.
There are two conditions which determine whether or not a
foreign protein can grow and multiply in the animal body:
One is the capability of the invader of digesting and utilizing
22 PROTEIN POISONS
the proteins of the body. All living cells grow by means
of their own digestive ferments, and these must act upon
the pabulum within their reach. If the ferment of the
bacterial cell cannot digest and prepare food for the bac-
terium from the body proteins, then the invading bacterial
cell dies. The second factor in determining whether a
given bacterial cell will grow in the animal body is the
effect of the ferments of the body cells on the invader. If
these are rapidly and thoroughly destructive there is no
bacterial development, and the organism is innocuous.
The prodigiosus is not pathogenic, but the cellular sub-
stance of this bacillus obtained by growth on artificial culture
media is highly poisonous to animals. This is true, with
modification as to degree, of the cellular substance of all
non-pathogenic bacteria. It is not the lack of poison in
the substance when placed under conditions favorable to
its growth, but it is its inability to grow under unfavorable
conditions. The smallpox virus is pathogenic to the unvac-
cinated but non-pathogenic to the vaccinated, because by
vaccination there has been developed in the body a ferment
which destroys the smallpox virus before it can develop.
7. Any foreign protein which can grow and multiply in the
body of a given animal may prove pathogenic to that animal.
Our idea of the development of an infectious disease may
be stated as follows: An infective agent is any protein
which possesses the capability of growth in the animal
body. What these essentials are have been stated under 6.
We will take as illustration typhoid fever. The infective
agent is the typhoid bacillus, a specific, particulate protein.
It is infective because by means of its digestive ferment it
can feed upon the proteins of man's body. This means
that it can convert man's proteins into typhoid proteins
and thus multiply its kind. Moreover, it is not, imme-
diately on its entrance in man's body, destroyed by the
ferments of the body cells. Having found admission to the
body it proceeds to grow and multiply. This continues
through the period of incubation, which in this disease is
somewhere about ten days. During this period of incuba-
INTRODUCTION 23
tion there is no effective resistance on the part of the body
cells to the growth and multiplication of the foreign pro-
tein. During this time the man is not sick, and we conclude
therefore that it is not the growth of the foreign protein
which per se gives rise to the symptoms of typhoid fever.
However, during this time the body cells are being pre-
pared for their combat with the foreign protein. This
preparation consists in the development in certain of the
body cells of a new function, that of elaborating a new and
specific ferment which will digest and destroy the foreign
protein. When this new ferment begins its action the
first symptoms of the disease appear. The active stage of
the disease, with its symptoms and the lesions induced,
marks the period over which the parenteral digestion of
the foreign protein extends. Death may come from the
too rapid Jbreaking up of the foreign protein and the conse-
quent liberation of a fatal dose of the protein poison, which
is always formed on the disruption of the protein molecule,
or it may result from some lesion induced by the products
of this disruption, such as perforation and hemorrhage, or
it may follow from chronic intoxication and consequent
exhaustion. In case of recovery the individual is for a
time at least immune to the typhoid bacillus because his
body cells are now able to elaborate and make immediately
effective the specific ferment which destroys the typhoid
protein.
8. The infectious diseases result from parenteral protein
digestion. Parenteral digestion, like all fermentative pro-
cesses, is influenced in its rate of progress by many condi-
tions, among which may be mentioned the relation between
amount of ferment and substrate, the physical condition
of the substrate, and the presence of the fermentation
products. These influences upon parenteral digestion are
not easily ascertained, and consequently not as yet measur-
able or controllable. The liberation of heat as measured by
body temperature has recently received attention, and we
can say in a general way that fever is one of the most easily
recognizable effects of the process. While natural infec-
24 PROTEIN POISONS
tion is due to living proteins, we have recently learned
that experimental fever can be induced by repeated injec-
tions of foreign proteins and by changes in size of dose and
in intervals between doses, fever of any desired type can
be induced.
9. Natural immunity to any infection is due to inability
of the infecting agent to grow in the animal body. This, of
course, does not include toxin immunity, which is due to
the presence in the body of an antitoxin or of something
which destroys or neutralizes the toxin. The inability of
an infection to multiply in the animal body has been
explained under 6.
10. The immunity which is due to recovery from an infec-
tion is the result of the development in the body during the
course of the infection of a specific ferment which imme-
diately destroys the infection on renewed exposuret As has
been stated, the cells of the body acquire the function
of developing the specific ferment, and this function is
awakened and made immediately effective on subsequent
exposure. This new function developed in the body cells
by disease may continue throughout life, or it may be lost
after a period which is variable in different diseases. The
immunity induced by one attack of yellow fever is believed
to continue through life; that from smallpox generally
holds through life, and but few have typhoid fever more
than once. Some infectious diseases, such as pneumonia,
are apparently not followed by immunity. In most instances
it seems that the immunity induced by one attack of an
infectious disease is not absolute, but only relative, and
may be overcome by severe or prolonged exposure to a
virulent form of the infection.
11. Immunity established by vaccination is similar to
that induced by an attack of the disease. The vaccine is the
same protein that causes the disease. It must be so modified
that it will not induce the disease, but yet so little altered
that it will stimulate the body cells to form a specific ferment
which will promptly and quickly destroy the infecting
agent on exposure. The smallpox virus is modified by
INTRODUCTION 25
passage through the cow. The anthrax bacillus is con-
verted into a vaccine by growth at high temperature.
The typhoid bacillus is killed by heat. In all these instances
the protein is so little changed in being converted from an
active, infecting agent into a vaccine that it still sensitizes
the animal to itself in the unmodified state. It seems
reasonable to conclude that the protein retains its capa-
bility of sensitizing so long as there is no radical alteration
in its chemical structure. The results secured by vaccina-
tion with killed typhoid bacilli prove that in this instance
at least a vaccine is not necessarily a living organism. The
possibility of obtaining vaccines from the split products of
pathogenic proteins has led to some of the investigations
detailed in this volume, and while we do not claim success
in this particular, we think that continued efforts in this
direction are justified.
12. Protein sensitization and bacterial immunity, appar-
ently antipodal, are in reality identical. This statement first
made by us in 1907 has since met with wide acceptance,
and will be discussed in detail in later chapters.
13. Protein sensitization consists in developing in the
animal body a specific proteolytic ferment which digests the
same protein on reinjection. Protein sensitizers may be
living or dead, particulate or in solution. Soluble proteins
sensitize more readily and more fully than those not in
solution. The development of the specific proteolytic
ferment in sensitization is due to the action of the foreign
protein upon the body cells. There is developed in certain
body cells a new function, that of elaborating this new
ferment. In order for a given body cell to be thus influenced
by a foreign protein, the latter must come in contact with
the former. Cell permeation by the foreign protein is
probably essential to the perfect elaboration of this process.
14. When a foreign protein is introduced into the blood of
an animal it soon leaves the circulating fluid and is distributed
throughout the tissues. The truth of this has been demon-
strated by the researches of independent investigators, and
will be detailed later. This is true of both particulate and
26 PROTEIN POISONS
soluble proteins, but the distribution is more prompt and
effective with soluble than with particulate proteins. This
explains why the former are more efficient sensitizers than
the latter. It will be understood that a protein relatively
insoluble in vitro may become more readily soluble in vivo.
15. Vaccines are protein sensitizers. One of the most
important problems in scientific medicine now awaiting
solution is that of the preparation and employment of
vaccines. The term vaccine — from vacca, a cow — was
first used when Jenner employed the infection of cowpox
to induce immunity to smallpox. Since that time the use
of the word "vaccine" has been extended to include every'
form of preventive inoculation. Through the researches
of Wright vaccine therapy has been developed, and now
vaccines are employed not only in the prevention but in
the treatment of disease. Under 11 we have spoken of
the employment of vaccines in inducing immunity, and
now we wish to speak briefly of vaccine therapy. As we
understand it, these two uses of vaccines depend upon the
same principle. The action of the vaccine is the same in
both instances. The protein of the organism responsible
for the diseased condition and that of the vaccine must be
identical or closely related bodies. Both must be protein
sensitizers. In most, if not in all, of the systemic infectious
diseases the infecting organism sensitizes the body either
throughout or over large areas. It seems to us to treat
such diseases with vaccines is irrational, and we believe
that much harm has been done by such attempts. There
are, however, local infections in which the area of sensitiza-
tion is limited and circumscribed. Such diseases may be
treated rationally with vaccines, provided such agents
can be obtained in such forms that they will act both
effectively and harmlessly. The future of vaccine therapy,
in our opinion, depends upon our ability to secure such
vaccines. That we have not yet fully established our
ability to obtain vaccines that are both harmless and
effective we are ready to admit. This does not mean,
however, that all efforts to accomplish this should be dis-
INTRODUCTION 27
continued. In our own attempts in this direction we have
met with enough encouragement to lead us to be hopeful
of ultimate success, while admitting present failure In
our opinion, it is not only unwise, but unjustifiable to treat
advanced cases of tuberculosis with tuberculin or other
tuberculo-sensitizer. So long as the disease is strictly localized
and the body in general is not sensitized, such treatment
can find reasonable justification. Our theory is that in
strictly localized infections the proper use of a specific
sensitizer may cause the more general and abundant forma-
tion of a specific proteolytic ferment which may aid in the
destruction of the infecting organism. In our opinion
sensitization consists in the development of a new function
in the body cells. In strictly local infections this new
function has been developed only in the infected area, and
to establish a like function in more distant cells may be
beneficial. The present tendency on the part of the pro-
fession to employ all kinds of bacterial proteins as vaccines
is, in our opinion, not only unscientific, but wholly without
justification. It should be clearly understood that with
every protein injected into the body a most potent poison
is introduced, and caution in the use of vaccines is not
out of place.
16. Toxin immunity and bacterial immunity are radically
different. This is a point upon which we shall frequently
touch, and we hold that attempts to describe one of these
forms of immunity in terms of the other have not only
been unwarranted by the facts, but have led to unnecessary
confusion.
17. The protein poison is not a toxin. It elaborates no
antibody, and while its repeated use in non-fatal doses may
establish a certain tolerance, it gives no immunity com-
parable in either nature or degree with that obtained by
like employment of toxins.
18. The protein poison is not specific.
19. The tolerance which may be secured by the protein
poison is not specific.
28 PROTEIN POISONS
20. The sensitization developed by a protein is specific,
but is not due to the poisonous group in the protein. As we
have stated, the specificity of a protein is not due to its
poisonous group, which is much the same in all proteins,
but to its secondary groups, for it is in these that one protein
differs from all others.
21. Different proteins find in the body certain predilection
places in which they are most prone to accumulate. The
pneumococcus accumulates in the lungs, the smallpox
virus in the skin, the typhoid bacillus in the spleen, and
mesenteric glands; the tubercle bacillus finds its most
frequent location in the lungs, but it has been a parasite
so long that it may grow on any human tissue.
22. The symptoms of a given disease are largely determined
by the location of the foreign protein. The most skilful
physician may not be able to tell what organism is respon-
sible for a meningitis. The symptoms of acute miliary
tuberculosis and those of typhoid fever are much alike.
It is the location of the infection rather than the exact
nature of the infecting agent which gives rise to the more
or less characteristic symptoms of the several infectious
diseases.
23. The poison elaborated in all the infectious diseases is
the same. It is the protein poison, and it is physiologically
the same whatever its source, whether it comes from coccus,
bacterium, spirillum, or protozoan. The specificity which
characterizes the infectious diseases is not due to the poison
formed, but to the protein cause and the specific ferment
produced.
24. When a cell in the animal body is permeated by a
foreign protein, the former strives to elaborate a ferment by
which the latter is destroyed. We believe this to be a biological
law, and we think that it lies at the foundation of a correct
understanding of many of the problems of immunity and
disease.
CHAPTER II
THE GROWTH OF MASSIVE CULTURES OF
BACTERIA
HAVING decided to study the chemistry of the bacterial
cell, it soon became evident that we must devise some way
of obtaining this substance in large quantity and fairly free
from admixture with foreign material. Bacteria had been
grown only in test-tubes, Petri dishes, and Roux flasks,
and none of these methods of growth gave the amount of
material necessary to promise any satisfactory investiga-
tion. As a medium, agar suits the purpose admirably,
because the bacterial growth can be detached and washed
from the surface of this medium quite free from admixture,
but it remained to devise some means of obtaining a large
surface so protected as to be guarded against contamina-
tion. At first we tried the Roux flasks, and by inoculating
one hundred of them with the colon bacillus, allowing the
cultures to grow for from two to three weeks at room tem-
perature, or for a shorter time in the incubating room, and
then washing off the growth with alcohol, we secured a
somewhat bulky and promising volume of bacterial, cellular
substance; but when this had been thoroughly washed
with sterile salt solution, extracted with alcohol and ether,
dried and weighed, we found the total yield, under the
most favorable circumstances; was not more than three
grams. This enabled us to make some preliminary experi-
ments and to demonstrate that the dead bacterial cells,
thus prepared, gave all the general color reactions for pro-
teins and were highly poisonous to animals, but the possi-
bility of making any satisfactory chemical study was not
promising. Moreover, the labor and care necessary to
30
PROTEIN POISONS
FIG. 1
Tank with raised lids.
FIG. 2
Tank with lids lowered.
GROWTH OF MASSIVE CULTURES OF BACTERIA 31
remove with anything like completeness the bacterial
growth from one hundred Roux flasks were greater than
we could afford to exert more than a few times; therefore,
attempts to secure the desired quantity of bacterial cellular
substance by this method were abandoned. We then tried
FIG. 3
The incubating room, lids lowered.
growing the colon bacillus in ordinary moist chambers, used
as Petri dishes, but with the greatest care many of these
cultures became contaminated. Finally, we devised the
large copper double tanks which have proved wholly satis-
factory and have supplied abundant growths, easily obtain-
32
PROTEIN POISONS
able and free from contamination. This tank, photographs
of which are here given, was first put into operation in 1900,
and was described in the following year1 (Figs. 1, 2, 3, 4).
A copper tank ten feet long, two feet wide, and four
FIG. 4
The incubating room, lids raised.
inches deep, with a trough around the edge one inch deep,
has a cover which, when lowered into place, rests in the
trough. This tank is supported by an iron frame of gas
piping, the legs of which rest on rollers, so that the whole
]Trans. Assoc. Amer. Phys., 1901, xvi, 217.
GROWTH OF MASSIVE CULTURES OF BACTERIA 33
may be easily moved about the room. An inner tank, two
inches shorter and two inches narrower, also provided with
a trough that runs around the edge, sits in the large one,
and is supported two inches from the bottom of the larger
one by iron cross-bars. The bottom of the outer tank and
the seal trough on its edges are filled with water. The seal
trough of the inner tank is filled with glycerin. Both lids
are raised and lowered by wire ropes passed through pulleys
fixed in the ceiling. The iron frame supporting the tanks
may be of any desired height. In our incubating room we
have a nest of six tanks, three of which are on frames four
feet high and three on frames two feet high. This economizes
space, as the lower ones can be rolled under the higher ones.
Both lids are supplied with vent tubes which are plugged
with cotton in sterilization. Twenty liters of 3 per cent,
agar is placed in the inner tank; both lids are lowered into
their respective troughs, and with large gas burners at full
blast underneath the apparatus is a sterilizer. After three
sterilizations on successive days the medium is inoculated
by pouring a liquid culture through the vent tubes in the
lid of the inner tank. Then with upper lid lowered into
the water trough and gentle heat, which may be controlled
by a thermoregulator, it becomes an incubator. With a
number of tanks in a small room it is better to heat the room
to the desired temperature, thus regulating the heat, than
it is to heat each tank separately.
When the growth has reached its maximum, the time
necessary for this varying with the organism grown and the
temperature maintained, both lids are raised, the growth
is detached from the subjacent agar with sterilized bent
glass rods, sterile salt solution added if necessary, and the
bacterial mass is drawn by means of a water pump into a
sterilized receiver.
The tanks are inoculated from special glass bulbs in which
the organism has been grown for some days. With the
colon bacillus we have usually employed Uschinsky's
solution, or some modification of it, in the inoculating bulbs,
in order that there may be no trace of foreign protein in
3
34 PROTEIN POISONS
the bacterial growth. While it was highly desirable that
this should be done at least once in order to demonstrate
that the protein reaction given by the bacterial substance
was not due to some constituent of the culture medium,
ordinarily beef-tea cultures may be employed. As will be
seen later, we did grow the colon bacillus once in liquid
Uschinsky medium for the purpose of fully satisfying
ourselves that the protein material obtained did not come
from the culture medium.
After removal from the tanks the bacterial cellular sub-
stance may be washed with various fluids. As a rule, we
have washed once or twice with sterile salt solution by
decantation and then repeatedly with alcohol, beginning
with 50 per cent, and increasing to 95 per cent. The sub-
stance is then placed in large soxhlets and extracted first
for one or two days with absolute alcohol, and then for
three or four days with ether. These extractions with
alcohol and ether should be thorough in order to remove
all traces of fats and waxes.
After extraction, the cellular substance is ground, first
in porcelain, then in agate mortars, and passed through
the finest meshed sieves. If there be bits of agar in the
bacterial cellular substance, which is seldom the case, it is
separated by the sieve and discarded. The one who grinds
the cellular substance should wear a mask in order to pro-
tect himself; notwithstanding this precaution, several
workers have been acutely poisoned, especially with the
typhoid bacillus. Of course, there is no danger of infection,
as the material, after the treatment already described,
contains no living bacilli. The finely ground cellular sub-
stance in the form of an impalpable powder may be kept
in wide-mouthed bottles in a dark place, and if so kept
it retains its toxicity for years, but when long exposed
to the light, even if kept perfectly dry, it becomes less
poisonous.
The yield from the tanks varies with the organism, but
generally amounts to from 60 to 80 grams of the purified
cellular substance for each tank, and with six tanks in
GROWTH OF MASSIVE CULTURES OF BACTERIA 35
operation, and with a crop every three weeks, one may
obtain several kilograms within a few months.
Three successive crops of the colon bacillus have been
grown on the same agar, resterilizing and reinoculating
after each harvest, but the third crop is not abundant. It
has been found to be well to follow the example of the
scientific agriculturalist and rotate the crops. Colon
grows well after typhoid, but typhoid does not grow well
after colon. Five crops have been obtained in the following
order: (1) pneumococcus, (2) typhoid, and (3) three suc-
cessive colon growths, or better, non-pathogenic bacteria
following one colon growth.
Many non-pathogenic organisms, the colon, typhoid,
pneumococcus, and diphtheria organisms, have been grown
on the tanks, and so simple is this method of obtaining
bacterial cellular substance in large amount that any intel-
ligent person, after some experience, may repeatedly go
through the whole manipulation, producing growth after
growth, without contamination. In the laboratory it is
best to have one man make a specialty of producing these
growths.
The anthrax and tuberculosis cellular substances with
which we have worked have not been produced in the
tanks. The former has been grown in Roux flasks and the
latter in glycerin beef-tea cultures. After these growths
have been obtained, their further preparation has been the
same as that already outlined.
Prepared, as described, the bacterial cellular substances
form fine, white, or yellowish-white powders. This is true
even of the chromogenic bacteria, such as b. violaceus and
b. prodigiosus, the pigment being removed from the cells
by its solubility in alcohol. One of our former students,
Detweiler,1 studied some of these pigments, but these do
not concern us at present, because they constitute no part
of the cellular protein. The same is true of the other bodies
soluble in alcohol and ether. The extracts made with these
i Trans. Assoc. Amer. Phys., 1902, xvii, 246.
36 PROTEIN POISONS
solvents contain fats, waxes, pigments, and possibly other
substances, but in this review we are interested solely in
the cellular proteins.
Microscopic examination of the powdered bacterial
substances show the bacilli, mostly intact, though many
are more or less broken by the attrition to which they have
been subjected. The cells still take the ordinary stains,
but the tubercle bacillus and others of this group are no
longer acid-fast, showing that the property of retaining
the stain when washed with mineral acid is due to some
constituent removed by the alcohol and ether.
CHAPTER III
BACILLUS coli communis was selected in the earlier
experiments for the following reasons: (1) It is easily
obtained at any time from the normal feces of man. (2) It
is quite stable in artificial cultures, varying but little, if
transplanted from day to day, in its effects upon experi-
mental animals. (3) It elaborates no extracellular poison, at
least under ordinary conditions and in beef-tea cultures.
Our early findings were reported in 1901, * and these,
confirmed and enlarged by subsequent work, will be briefly
reported as follows:
1. The poison is contained within the bacterial cell
from which it does not, at least under ordinary conditions,
diffuse into the culture medium.
This was demonstrated by the following experiment,
which was repeatedly made, and always with the same result :
Beef-tea cultures of the colon bacillus, grown for three
weeks or longer, in the incubator, were filtered through
porcelain. From 8 to 10 c.c. of the clear, sterile filtrate
was injected intra-abdominally in guinea-pigs. The animals
thus treated were restless and evidently in pain for some
minutes after the injection, probably due to the volume of
the fluid and its slightly irritating character, but gave no
other evidence of any effect of the injection.
As controls to the above, other guinea-pigs received
intra-abdominally 0.25 c.c. of the same culture unfiltered,
and all died within twelve hours. It will be understood
1 Trans. Assoc. Amer. Phys., xvi, 201.
38 PROTEIN POISONS
that these animals died from infection, and the object in
inoculating them was to show that the culture contained
living, virulent bacteria, while its filtrate was without
effect on animals. However, it might be claimed that the
poison, although in solution in the beef-tea, will not pass
through porcelain. This suggestion is reasonable, and
calls for further experimentation ; consequently an un weighed
portion of the dead cellular substance of the colon bacillus
was suspended in water, heated in the autoclave at 154°
under 2 kilos of pressure, and filtered through porcelain.
Four cubic centimeters of this clear, sterile filtrate injected
intra-abdominally into a guinea-pig caused death within
thirty-six hours, and section showed the same lesions that
are found after death from either the living bacillus or the
dead cellular substance. This demonstrates that when
the bacterial cells have been disrupted by superheated
steam, their poisonous constituent becomes to some extent
soluble in water, and may be passed through porcelain.
Furthermore, experiment showed that colon cultures when
boiled in open dishes and filtered through porcelain supplied
inert filtrates. This indicates that the disrupting effect
of a high temperature is necessary to the extraction of the
poison from the cell.
Filtrates from living cultures of the diphtheria bacillus
contain a toxin which is a secretion of the living micro-
organism. The colon bacillus produces no such active
toxin. Old, dead cultures of the colon or typhoid bacillus
may contain soluble poisons, but these are not secretions
of the living cells. They come from the autolysis of the
dead cells, and, as we shall see later, they are not properly
toxins, capable of producing antibodies, but are chemical
poisons. Moreover, the toxin of the diphtheria bacillus is
specific, while the cellular poison is not.
2. The poison is not extracted from the bacterial cell
by dilute saline solution, alcohol, or ether, either at ordi-
nary temperature or at the boiling-point of these fluids.
Extracts of the cellular substance of the colon bacillus with
these agents were repeatedly made, filtered, evaporated
BACTERIAL CELLULAR SUBSTANCE 39
in vacua, taken up in a small volume of water, and injected
into animals without effect. That the extraction of the
cellular substance of the colon bacillus with alcohol and
ether has no destructive action on the intracellular poison
is shown by the fact that this material after extraction with
these agents does not lose any of its toxicity. Furthermore,
it may be stated that prolonged boiling of the cellular sub-
stance of the colon bacillus with alcohol or ether, or with
the two successively, neither sets free nor destroys the
intracellular poison. These agents dissolve the fats, waxes,
and coloring matter from bacterial cells, but do not remove
or destroy the intracellular poison. As will be seen later,
this is also true of the intracellular poison of those bacteria
that produce a soluble toxin in their cultures. Moreover,
it will later appear that the poisonous group not only in
bacterial but in vegetable and animal proteins is soluble
in absolute alcohol after it has been well detached from
the other groups in the protein molecule, but it is not
removed from its place in the complex molecule by either
alcohol or ether. Indeed, the poisonous group in the protein
molecule is not removed from its attachment to other
groups by any purely physical solvents, and the molecule
must be disrupted by high temperature, chemical agents,
or enzymes before its poisonous constituent can be extracted
by physical solvents.
3. The cellular substance of the colon bacillus may be
heated with water without destruction of its poisonous group.
Two hundred milligrams of the cellular substance of the
colon bacillus was suspended in 10 c.c. of water in a tube,
which was then sealed and heated at 184° for thirty minutes.
On opening the tube, the milky content was found on
microscopic examination to contain granular debris with a
few unbroken cells. Portions of this heated substance
injected into guinea-pigs caused death, and autopsy revealed
the same lesions that are seen after death from either the
living bacillus or the unbroken cellular substance.
Another portion of the content of this heated tube was
placed in a centrifuge and separated into a deposit and a
40 PROTEIN POISONS
supernatant fluid, the latter being somewhat opalescent.
Guinea-pigs were treated with both portions, and all died.
This demonstrates that superheated steam disrupts the
bacterial cells, but does not destroy the intracellular poison.
Heating the cellular substance of the colon bacillus in
physiological salt solution to 140° in the autoclave does
not destroy or even weaken the poison. The cell substance
used in this experiment had been prepared six years pre-
viously. One gram was thoroughly mixed with 100 c.c. of
salt solution. Two cubic centimeters of this mixture, con-
taining 20 mg. of the cell substance, killed guinea-pigs of
300 grams' weight when injected intra-abdominally, and
0.5 c.c. or 5 mg. made the animals very sick. The emulsion
was then heated in the autoclave to 140° and held at this
temperature for ten minutes. One cubic centimeter of
this heated emulsion, containing 10 mg. of the cell sub-
stance, killed guinea-pigs of 300 grams' weight, and 0.1 c.c.
or 1 mg. made the animals sick. Animals treated with the
heated emulsion died more promptly and from smaller
doses than those treated with the unheated preparation.
Evidently the heating prepared the cell substance so that
it was more promptly split up by the ferments of the body.
When heated to this temperature .a part of the poison
passes into solution. This was shown by filtering the
heated emulsion through hard paper. The filtrate was
clear, slightly acid to litmus, gave the biuret, Millon, and
«-naphthol tests, and killed guinea-pigs of 300 grams'
weight. The animals that had the unheated suspension
showed marked peritoneal inflammation, with bloody exu-
date. In those having the larger doses (40 mg. of the cell
substance) the inflammatory condition extended to the
muscular walls of the abdomen. In those that had the
heated suspension the inflammatory condition was much
less marked, there being only a slight serous exudate, less
and less stained with blood as the amount of the cell sub-
stance injected decreased. In those killed with the filtrate
there was no evidence of peritoneal inflammation. This
furnishes a beautiful illustration of the nature of inflam-
BACTERIAL CELLULAR SUBSTANCE 41
mation as caused by bacterial cells. When such cells are
disrupted wholly by the body cells, and in a restricted
locality, there is marked destruction of the local body
cells, and this is the condition which we designate as " local
inflammation." On the other hand, when the bacterial
cells are disrupted and the cellular poison made soluble
before being introduced into the body of the animal, there
is no local reaction, or no special local reaction, and conse-
quently no recognizable inflammatory conditions.
While heating suspensions of the cellular substance in
water and salt solution does not lessen its toxicity, the
poison passes into solution only partially. The heated
suspension is more than twice as poisonous as the filtrate
from the same. Some of the poison actually goes into
solution and the filtrate from hardened paper may be
perfectly clear, provided the first portion be returned to
the filter, but the greater part of the poison is removed by
filtration through paper. It is worthy of note that heated
suspensions of the colon cellular substance, filtered or
unfiltered, easily become contaminated, and apparently
furnish acceptable culture media.
4. Dilute (0.5 per cent.) solutions of the caustic alkalies
disrupt the cellular substance of the colon bacillus slowly
and imperfectly.
This is shown by the following: 100 mg. of the cellular
substance was boiled for five minutes in an open dish with a
0.5 per cent, solution of sodium hydroxide. The fluid
was centrifuged and the deposit found to be still poisonous,
while the supernatant fluid was without effect. However,
stronger solutions (2 per cent.) of alkali completely disrupt
the bacterial cell and dissolve the poison after prolonged
heating. As will appear later, the method finally selected
for splitting off the poisonous group consists in heating
the cellular substance with a 2 per cent, solution of sodium
hydroxide in absolute alcohol.
5. Boiling with a 0.2 per cent, dilution of hydrochloric
acid has but little effect upon the bacterial cell or its
contained poison.
42 PROTEIN POISONS
One hundred milligrams of the cellular substance was
boiled in an open test-tube with 10 c.c. of a 0.2 per cent,
dilution of hydrochloric acid. This induced no visible
alteration in the bacterial cells, as seen under the micro-
scope, and the injection of this material into animals caused
death in the usual time and with the usual findings.
6. Heating the cellular substance for an hour in an
open dish on the water-bath (about 80°), with from 1 to 5
per cent, solutions of hydrochloric acid, breaks up the cells
but does not wholly destroy the toxicity of the cell content;
however, prolonged boiling with 1 per cent, or stronger
dilutions of hydrochloric acid does destroy the poison.
Five hundred milligrams of the colon cellular substance
was heated on the water-bath for one hour with 500 c.c.
of a 5 per cent, solution of hydrochloric acid and then
decanted through a hard filter. The filtrate was clear
and colorless and gave no appreciable precipitate when
dropped into absolute alcohol, but that the acid had dis-
solved some part of the cellular substance was shown by
the response of the filtrate to the biuret test.
The undissolved material was suspended in a dilute
solution of sodium bicarbonate, sufficient to neutralize the
acid, and injected into guinea-pigs which died in the charac-
teristic way, and showed the usual lesions.
7. The bacterial cellular proteins are, so far as their
toxicity is concerned, quite resistant to the action of pepsin
and trypsin.
A given sample of the cellular substance of the colon
bacillus was tested upon a large number of guinea-pigs, in
order to determine the minimum lethal dose, which was
found to be for half-grown animals 0.5 mg. given intra-
abdominally, and 1 mg. subcutaneously. This material
was then subjected for three days to an artificial gastric
juice, the efficiency of which was demonstrated simul-
taneously by its action on coagulated egg-white. The
soluble and insoluble parts were separated and their toxicity
tested. One-half milligram of the undigested part given
intra-abdominally did not kill but 1 mg. did; while 1 mg.
BACTERIAL CELLULAR SUBSTANCE 43
given subcutaneously no longer killed but 2 mg. did. Of
the part that was dissolved in the acid pepsin solution,
doses up to 100 mg. had no effect. A like result was obtained
with the cellular substance of the typhoid bacillus. The
amount of cellular substance left undigested after three
days' exposure to the acid-pepsin was about 10 per cent, of
that originally taken. The conclusion is that the gastric
juice slowly digests the bacterial cellular proteins, and in
so doing destroys the poison.
With trypsin the effect is somewhat different. The
cellular protein goes into solution more rapidly, and at
least a part of the poison goes into solution without complete
loss of its properties. The parts of both the colon and
typhoid cellular substance that passed into solution after
three exposures to trypsin killed in doses of from 35 to 40
mg. given intra-abdominally, while the undigested portion
killed in doses of from 4 to 7.5 mg.
One gram of the cellular substance given to a rabbit
through a stomach-tube had no recognizable effect on the
animal.
At one time early in these investigations we had an idea
that the poison in the colon cell resisted peptic digestion,
and we therefore quite naturally suspected that it might
be a nuclein. This belief was founded upon the following:
The growth on fifty Roux flasks was removed, extracted
with 96 per cent, alcohol so long as the alcohol took up
coloring matter, then dried, placed in a beaker, and stirred
with 1 liter of 0.2 per cent, hydrochloric acid dilution in
which 0.5 grams of active pepsin had been dissolved. The
beaker with content was kept in the incubator for two
days, with occasional stirring. The undigested portion
on microscopic examination was found to be amorphous,
but still easily stained with methylene blue. It was col-
lected on a filter, washed thoroughly with 96 per cent,
alcohol, dried at 100°, and pulverized. One hundred milli-
grams of this powder was shaken with 50 c.c. of water,
forming an acid, colloidal mixture. On adding sodium
bicarbonate to a faintly alkaline reaction, the substance
44 PROTEIN POISONS
dissolved to an opalescent fluid. This was heated in order
to insure sterilization, and injected into guinea-pigs, which
it killed within from six to twenty-four hours. One milli-
gram and even less of this undigested portion killed the
animals thus treated, but subsequent investigation showed
that the poisonous portion, when fully detached from the
other constituents of the protein molecule, kills in a few
minutes, and we concluded that the undigested part con-
sisted of several groups still attached, or, in other words, of
a larger and more complex group of which the poison is
only a part. The action of the proteolytic enzymes on
bacterial cells deserves a more thorough study than we
have given it.
These preliminary studies quite convinced us so long ago
as 1901 that a typical colon bacillus, obtained from normal
human feces, does not elaborate in its cultures a soluble
poison, but that its cells do contain a highly active body.
Moreover, these studies indicate that the poison of the colon
bacillus exists in the essential proteins of the bacterial
cell, and that it cannot be isolated until these proteins are
broken up into their constituent parts. In other words,
the poison consists of one or more groups in the protein
molecule. Since the colon bacillus may grow in a medium
consisting solely of inorganic matter and a small amount
of some organic compound, as asparagin, its protein must
be formed synthetically, and its poison, as a constituent
of its protein, must be developed in the same way. One
is forced to the conclusion that the poison of this bacillus,
at least, does not result from the cleavage action of the
bacterial cell or its soluble ferments on the constituents
of the medium in which it grows, but that it is built up
synthetically, and is set free only when the cellular protein
is disrupted. In other words, the harmful action of bacillus
coli communis upon animals is not due directly to the
growth and multiplication of the organism in the animal
body, but to the breaking up of the bacterial protein and
the consequent liberation of its poisonous group.
Subsequent and more extended research has shown that
BACTERIAL CELLULAR SUBSTANCE 45
the stability of the protein molecule varies within wide
limits, and that the facts learned in the study of the special
strain of the colon bacillus do not hold strictly true in
every particular with all proteins. This was expected, and
we hoped to obtain from these preliminary studies nothing
more than certain standards by which our findings in more
extended studies might be measured. With some strains
of the colon, and in many more of the typhoid bacillus,
we have obtained evidence of the presence of soluble poisons
in old cultures. For the most part at least these come from
autolysis of the bacterial cell. This is a subject to which
we shall return in recording the development of these
researches.
The findings in our early studies quite naturally developed
several inquiries, some of which may be formulated as
follows: If the protein of the colon bacillus contains a
poisonous group, may not the proteins of other pathogenic
bacteria contain similar groups, and if the proteins of patho-
genic bacteria contain poisonous groups, why should not the
proteins of non-pathogenic bacteria possess like constituents.
If bacterial proteins contain poisonous groups, why should
not other proteins, such as those of vegetable and animal
origin, contain like groups, and if all proteins possess in
their structure, poisonous bodies, how is it that the animal
world, including man, lives so largely on proteins ? Attempts
to solve these questions have taken our time and energy,
and given us much pleasure. The results of these labors
constitute the principal record of this volume.
Marshall and Gelston1 made an exhaustive study of the
toxicity of the cellular substance of the colon bacillus. At
first they employed material coarsely ground in a porcelain
mortar. This was suspended in water, boiled to insure
complete sterilization, and then injected intra-abdominally
in guinea-pigs. Up to 1 part of poison to 40,000 parts of
body weight all animals treated in this way died. When
the proportion was reduced to 1 to 50,000 and less, none of
1 Trans. Assoc. Amer. Phys., 1902, xvii, 298.
46 PROTEIN POISONS
the animals died. The same powder more finely ground in
an agate mortar killed 15 out of 16 animals up to 1 to
75,000; out of 28 pigs it killed 9 at 1 to 100,000; out of 8
it killed 5 at 1 to 200,000; out of 34 it killed 4 at 1 to
2,000,000. Heating the cellular protein for fifteen minutes
under 2 kilos of pressure at 134° did not appreciably lessen
its toxicity. The finely ground powdered substance killed
rabbits when injected intra-abdominally : 2 out of 4 at 1
to 75,000; 4 out of 7 at 1 to 100,000; 4 out of 12 at 1 to
200,000; and 2 out of 18 at 1 to 2,000,000.
One gram of the cellular protein was incinerated and the
whole ash injected intraperitoneally in a guinea-pig without
effect. The toxic effect when subcutaneously injected is
practically the same as when employed intraperitoneally,
but death is longer delayed.
The intracellular poison of the diphtheria bacillus was
studied by Gelston,1 using different strains, among which
was the well-known extracellular toxin producer desig-
nated as Park No. 8. These were grown in Roux flasks,
as it was found that this organism does not grow well in
the tanks, supposedly on account of limited aeration.
The growth scraped from the surface of the agar was placed
in physiological salt solution and heated for two and one-
half hours at 50° to secure sterilization. This suspension
was then poured onto hard filters and placed in an ice-box
until filtration was complete. The mass on the filter was
washed with sterile physiological salt solution, dried on
porous plates over sulphuric acid in vacua, and reduced
to a fine powder in agate mortars. It is worthy of note
that the physiological salt solution contained small amounts
of the extracellular toxin, and that still larger quantities
could be obtained by macerating the agar, on which the
bacillus had grown, with salt solution. It will be observed
that this cellular material was not extracted with alcohol
and ether. It killed guinea-pigs when injected subcu-
taneously or intraperitoneally up to 1 to 33,000. On post-
i Trans. Assoc. Amer. Phys., 1902, xvii, 308.
BACTERIAL CELLULAR SUBSTANCE
47
mortem examination all animals dying from intraperitoneal
injection showed marked congestion of the mesentery,
omentum, peritoneum, and adrenals; also numerous ecchy-
moses and hemorrhagic effusions in the serous coat of the
stomach, intestine, and peritoneum. The spleen, liver,
and kidneys were slightly enlarged and dark. In the
kidneys a dark line sharply divided the cortex from the
medulla. The heart was in diastole. When mixtures of
diphtheria antitoxin and suspensions of the cell substance
in salt solution' were injected into guinea-pigs, death followed
as promptly as when the dead germ only was given. One
thousand immunity units failed to protect against 10
minimum fatal doses of the cell substance as shown by the
following :
TABLE I
Amount
Immunity
No.
Weight.
injected.
Antitoxin.1
Antitoxin.2
units.
Result.
1
155
47 . 1 mg.
0
.0005
c.c.
0
.00031
c.c.
0
.118
+
2
135
40.9 mg.
0
.0050
c.c.
0
, 00290
c.c.
1
175
+
3
200
60.6 mg.
0
.0242
c.c.
0
.01940
c.c.
5
,700
+
4
150
46.0 mg.
0
.0500
c.c.
0
.03000
c.c.
11
750
+
5
135
40.9 mg.
0
,2000
c.c.
0
10900
c.c.
47
000
+
6
175
53.0 mg.
0
.5000
c.c.
0
35000
c.c.
117
500
+
7
120
37.0 mg.
1
.0000
c.c.
0
48000
c.c.
235
000
+
8
120
37.0 mg.
2
0000
c.c.
0
96000
c.c.
470,
000
+
In these experiments the minimum lethal dose of the
cellular substance was 1 to 33,000, and 1 c.c. of antitoxin
was equivalent to 235 immunity units.
With another sample of cell substance and of antitoxin,
the following figures were obtained:
No. Weight.
1 257
2 280
3 245
Amount
injected.
55.39 mg.
0.05 c.c.
52.80 mg.
TABLE II
Antitoxin.
2.5000 c.c.
0.0025 c.c.
control
Toxin Anti- Immunity
cor. toxin, cor. units. Result.
2.5680 1000 +
0.056 0.0028 1 —
In this series No. 2 received 10 minimum lethal doses of
a filtered bouillon culture of the same bacillus from which
1 Antitoxin for 250 grams body weight.
2 Antitoxin for actual body weight.
48 • PROTEIN POISONS
the cellular substance had been obtained, and 1 immunity
unit of antitoxin. These experiments show that while
diphtheria antitoxin protects against the extracellular
toxin it fails to protect against the intracellular poison.
Suspensions of the diphtheria cellular substance when
exposed to a temperature of 50° or higher, for fifteen minutes
or longer, gradually decrease in toxicity, but this is not
wholly lost after exposure to 122° in the autoclave for
thirty minutes. The minimum lethal dose of our prepara-
tion having been found to be 1 to 33,000, it "was heated to
122° for 20 minutes, and then proved to be 1 to 6400. It
is worthy of note that the several strains of the diphtheria
bacillus employed yielded cellular substances that varied
widely in toxicity. As a rule, one that supplied a potent
extracellular poison yielded relatively an indifferent intra-
cellular poison. Possibly this is due to the greater lability
of the molecular structure of the former, which leads to
the partial breaking down of the protein molecule in the
heating resorted to in order to sterilize the growth.
The cellular substance of the anthrax bacillus was pre-
pared and studied by J. Walter Vaughan.1 This kills
guinea-pigs in only relatively large doses, and this fact
indicates that the intensity of the infectious properties
of a microorganism is not, always at least, measured by
the potency of its intracellular poison. The bacillus pro-
digiosus is non-pathogenic to the higher animals, not from
its inability to furnish a poison, but because it cannot
grow and multiply in the animal body; while, on the other
hand, the anthrax bacillus is highly infectious to some of
the higher animals, not from the intensity of the poison
which it elaborates, but rather from the fact that in these
animals this bacillus finds conditions favorable to its growth
and multiplication.
Detweiler2 prepared and demonstrated the poisonous
action of the cellular substances of b. prodigiosus, b.
violaceus, sarcina aurantiaca, and s. lutea.
1 Trans. Assoc. Amer. Phys., 1902, xvii, 313.
2 Ibid., 257.
BACTERIAL CELLULAR SUBSTANCE 49
The relative toxicity of the finely powdered cellular sub-
stances of certain bacteria in proportion to body weight of
animal, is shown in the following figures:
Bacillus anthracis 1 to 1,700
Sarcina lutea 1 to 2,050
Micrococcus pneumonias 1 to 10,000
Sarcina aurantiaca 1 to 25,500
Bacillus violaceus 1 to 26,500
Bacillus diphtherias 1 to 33,000
Bacillus typhosus 1 to 40.000
Bacillus pyocyaneus 1 to 50,000
Bacillus coli 1 to 75,000
Bacillus prodigiosus 1 to 90,000
Numerous and varied attempts to immunize animals
with the bacterial substances were made. Guinea-pigs,
rabbits, and goats were used in these experiments. The
following quotation is taken from the report Marshall and
Gelston1 made in 1902:
"1. Although guinea-pigs and rabbits acquire immunity
to the germ substance but slowly, yet if sufficient time be
given, and the animal be allowed to recover before a second
dose is administered, a fair degree of immunity may be
obtained after many months of treatment. Unless the
animal be allowed to recover completely before a second
injection is given, death generally results. In these ex-
periments for the purpose of inducing immunity, we
employed the finely divided germ substance which had
been used in determining the toxicity. Our previous
experiments had shown that one part of the finely divided
powder to 75,000 parts of body weight in a guinea-pig
was surely fatal. Guinea-pig No. 93 received in the fifteenth
injection a dose of 28 mg. (1 to 17,321), and recovered from
the same. Guinea-pig No. 109 received on the ninth injec-
tion 25.5 mg. (1 to 22,745), and recovered. Rabbit No. 17
received on the fifteenth injection 372.3 mg. (1 to 4780),
and recovered. Rabbit No. 29 received on the thirteenth
injection 293.3 mg. (1 to 6409), and recovered.
1 Loc. cit.
50 PROTEIN POISONS
"2. Study of the blood serum from animals partially
immunized to the germ substance led to the following
results :
" (a) The blood serum of rabbit No. 2, which had received
on the twelfth injection 562.3 mg. (1 to 3503), obtained
twenty days after the last injection, had no bacteriolytic
action on the organisms contained in a beef-tea culture
of the colon bacillus of the same strain as that which had
been used in immunizing the animals.
" (6) Tubes of immune serum and others of normal
serum were inoculated with virulent cultures of the colon
bacillus, and placed in the incubator at 37.5°. Both tubes
showed good and equal growth within seven hours.
" (c) Tubes of immune and normal serum were inoculated
writh virulent cultures of colon and typhoid bacillus, and,
from these gelatin plates were made at the end of one
minute, thirty minutes, one hour, and forty-eight hours.
The number of colonies on the colon plates did not diminish
while those on the typhoid plates did with both sera, thus
showing that the immune serum had no specific action.
" (d) Immune serum, when mixed with filtered cultures
of the colon bacillus, gives a slight precipitate, which,
however, is also given by normal serum.
" (e) The immune serum gives very positive agglutinating
reactions with suspensions of the dead colon germ used in
immunizing the animals. Suspensions of from 1 to 50 mg.
of the germ substance in 1 c.c. of physiological salt solution
are completely precipitated in from three to five minutes
on the addition of an equal quantity of immune serum.
Control tubes to which normal serum had been added gave
negative results, inasmuch as complete subsidence did not
occur in these writhin fourteen days.
" (/) The immune serum exerts, when mixed with suspen-
sions of the germ substance and injected into the abdominal
cavity of rabbits, a slight protective action, as is shown by
the following:
Rabbit No. 1, 850 grams, had 17.0 mg. (1 to 50,000), 2 c.c., I. S., — .
Rabbit No. 2, 700 grams, had 14.0 mg. (1 to 50,000), 2 c.c., N. S., +.
Rabbit No. 3, 840 grams, had 16.8 gm. (1 to 50,000), 2 c.c., beef-tea, +.
BACTERIAL CELLULAR SUBSTANCE 51
" (g) The immune serum of the rabbit exerts no protec-
tive action for the guinea-pig against the germ substance.
"3. The immunity obtained by treating animals with
the germ substance is apparently of short duration, and
if the interval between the administration of doses be
prolonged, death is likely to follow even when there is no
increase in the dose. It will thus be seen that the attempt
to immunize animals to the germ substance of the colon
bacillus is beset with difficulties. If the intervals be too
short, and the animal has not fully recovered, death is
likely to result, and the same will also probably happen
when the interval is unduly long. Even after a marked
degree of immunity has been obtained, this is apparently
lost within a few weeks, and a repetition of a dose of the
same size causes death. The following table, showing next
to the last, and the last injections given to certain animals,
will illustrate our meaning:
January 19, 1902, rabbit No. 17, 1780 grams, had 372.3 mg. (1 to 4780),
and recovered.
March 29, 1902, rabbit No. 17, 1880 grams, had 273.5 mg. (1 to 5048), .
and died March 31.
January 18, 1902, rabbit No. 29, 2000 grams, had 293.3 mg. (1 to 6819),
and recovered.
March 29, 1902, rabbit No. 29, 1880 grams, had 293.3 mg. (1 to 6409),
and died March 31.
January 6, 1902, pig No. 93, 485 grams, had 28 mg. (1 to 17,329), and
recovered.
March 29, 1902, pig No. 93, 645 grams, had 28 mg. (1 to 19,464), and
died March 31. "
Later these experiments were repeated and extended to
goats by V. C. Vaughan, Jr., and Gumming, with prac-
tically the same results. We concluded that the capability of
the animal to bear increased doses of the cellular substance
was not sufficiently marked to be designated by the term
immunity, and it was decided to recognize it as increased
tolerance. This opinion we still hold.1
1 It will be seen from this work as here recorded that we met with the
phenomena of protein sensitization or so-called anaphylaxis as early as
1902 in our studies on the bacterial cellular substances, but that we failed
to follow it up and indeed did not attach much importance to it.
CHAPTER IV
CHEMICAL STUDIES OF BACTERIAL CELLULAR
SUBSTANCE
Proteins. — Xencki and Schaffer1 obtained the cellular
substance from a mixed culture of putrefying bacteria,
dried it to a constant weight, first on the water-bath and
then at 110°, pulverized, and extracted with alcohol and
ether. The residue thus obtained was extracted on the
water-bath with 0.5 per cent, potassium hydroxide. From
the alkaline extract a protein, designated as mykroprotein,
was precipitated by neutralization and saturation with
sodium chloride. Mykroprotein when freshly precipitated
was found to consist of amorphous flakes soluble in water,
but losing in solubility when dried at 110°. It contains
52.32 per cent, of C, 7.55 per cent, of H, 14.75 per cent,
of N, and neither sulphur nor phosphorus. In aqueous
solution it gives an acid reaction, and is not precipitated
by alcohol, but is precipitated by picric acid and other
alkaloidal reagents. It gives the biuret and Millon reac-
tions, but not the xanthoproteic. On being fused with
potash it furnishes ammonia, amylamin, phenol, valerianic
acid, leucin, and traces of indol and skatol.2 Later, Xencki3
attempted to prepare mykroprotein from anthrax. He
obtained from anthrax spores a substance which he desig-
nated as anthrax protein, closely related to plant casein
and animal mucin, soluble in alkalies, but insoluble in
water, acetic, and dilute mineral acids. Like mykroprotein,
it contains no sulphur. Dyrmont4 made an analysis of
1 Jour. f. prakt, Chem., 1879, xx, 443.
2 Ibid., 1881, xxiii, 302. 3 Berichte, 1884, xxvii, 2605.
4 Arch. f. exper. Path. u. Pharm., 1886, xxi, 309.
BACTERIAL CELLULAR SUBSTANCE 53
anthrax protein, with the following results: C, 52.1 per
cent.; H, 6.82 per cent.; N, 16.2 per cent., and a trace of
ash. We now know that both mykroprotein and anthrax
protein are cleavage products obtained from bacterial
cellular substances through the action of alkali.
Brieger1 found in Friedlander's pneumococcus a protein,
partially soluble in water and precipitated on boiling,
containing less nitrogen than is found in mykroprotein.
Lewith,2 reporting work done by Hellmich on a hay bacillus
grown on synthetic medium, stated that a globulin was
extracted from the cellular substance by neutral salts at
ordinary temperature. This probably was in fact no part
of the cellular protein. Subsequent extraction with dilute
alkali gave a protein body described as an albuminate and
said to resemble casein.
Buchner3 demonstrated that certain bacterial cells con-
tain pyogenetic bodies. These are extracted from the
cellular substances tjiith dilute alkalies from which they
are precipitated with dilute acids. The amount of protein
obtained in this way varied greatly with the species of
bacteria. Bacillus pyocyaneus gave the most abundant
yield, supplying 19.3 per cent, of the dried cells. By
heating on the sand-bath under a reflux condenser, or
in an autoclave at 120°, filtering through sand, and precipi-
tating with absolute alcohol, he obtained a more soluble
protein.
Brieger and Frankel,4 Proskauer and Wassermann,5 and
Dzierzgowski and Rekowski6 prepared so-called toxal-
bumins from diphtheria cultures, and made ultimate
analyses of the same, but, as we now know, these were
mixtures and gave us no information concerning the com-
position of the bacterial cell.
1 Zeitsch. f. physiol. Chem., 1885, ix, 1.
2 Arch. f. exper. Path. u. Pharm., 1890, xxvi, 341.
3Berl. klin. Woch., 1890, xxvii, 673; ibid, 1084; Munch, med. Woch.,
1891, xxxviii, 841.
< Berl. klin. Woch., 1890, xxvii, 241, 268.
6 Deutsch. med. Woch., 1891, xvii, 585.
6 Arch. d. Sci. biol., 1892, i, 167.
54 PROTEIN POISONS
Hammerschlag1 extracted tubercle bacilli with dilute
alkali and precipitated a protein with ammonium sulphate.
Buchner2 extracted tubercle bacilli with from 40 to 50 per
cent, glycerin, and obtained what he believed to be the
active principle of tuberculin, but which in reality consisted
of a mixture of the autolytic products of this bacillus.
Hoffman3 reported the isolation of six proteins from the
tubercle bacillus, but these were mixtures. Weyl4 believed
that he had succeeded in separating the membrane from
the protoplasmic content of the bacterial cell, and from the
latter he obtained a body which he designated as a toxo-
mucin, but there is no proof that the bacterial cell has
any such structure as he supposed. Ruppel5 prepared
from the tubercle bacillus a body that he named tuber-
culosamin.
Vandervelde6 reported the presence of nuclein in bacillus
subtillis, and Dreyfuss,7 basing his opinion on the behavior
of bacteria toward the basic aniline dyes, concluded that
nuclein is a constituent of all bacteria. Gottstein,8 finding
that various bacteria decompose hydrogen peroxide, both
during life and after death, concludes from this, from the
presence of phosphorus, and from the affinity of bacteria
for basic aniline dyes, that they contain nuclein. Nishi-
mura9 reported the finding of nuclein in a water bacillus
grown on potato. The bacterial cells were removed from
the potatoes, extracted with alcohol and ether, heated under
a reflux condenser with 0.15 per cent, sulphuric acid, and
then heated in an autoclave at 105°. From this acid extract
he obtained 0.17 per cent, xanthin, 0.08 per cent, adenin,
and 0.14 per cent, guanin. Lustig and Galeotti10 prepared
1 Monats. f. Chem., 1899, x, 9; Centralbl. f. klin. Med., 1891, xii, 9.
2 Munch, med. Woch., 1891, xxxviii, 45.
3 Wien. klin. Woch., 1894, 712.
4 Deutsch. med. Woch., 1891, xvii, 256.
s Zeitsch. f. Physiol. Chem., 1898, xxvi, 218.
6 Ibid., 1884, viii, 367.
7 Ibid., 1893, xviii, 358.
8 Virchow's Archiv, 1893, cxxxiii, 302.
9 Arch. f. Hygiene, 1893, xviii, 318.
10 Deutsch. med. Woch., 1897, xxiii, 228.
BACTERIAL CELLULAR SUBSTANCE 55
a nucleoprotein from the pest bacillus. From the effect
of methylene blue on this bacillus and its microscopic
appearance after extraction, they conclude that the alkali
removes the nuclein without destruction of the cell mem-
brane. Our studies do not indicate the existence of a mem-
brane in any bacterial cells. Galeotti1 extracted an organ-
ism similar to bacillus ranicidus with 1 per cent, potassium
hydroxide, and obtained a protein body containing from
11.99 to 12.21 per cent, of nitrogen and from 0.94 to 1.16
per cent, of phosphorus, and which he believed to be a
nucleoprotein. The percentage of phosphorus increased
after several reprecipitations.
Aronson2 extracted the diphtheria bacillus with from
one-tenth to one-fifth normal alkali in the cold, at 100°
and at 130°, precipitated these extracts, first with acetic
acid, then with acidulated alcohol, sometimes with alcohol
to which ether and a little acetic acid had been added.
The precipitate with acid furnished a white powder, giving
the biuret, xanthoproteic, and Adamkiewicz reactions.
Xanthin bases, a pentose, and an albumin were found
among its decomposition products, thus proving the presence
of a nucleoprotein. On the addition of alcohol to this
acid filtrate, a new precipitate was formed, and this on
purification yielded nucleic acid, from which xanthin
bases, a pentose, and a phosphate were obtained. Blandin3
obtained a nuclein and a nucleo-albumin from typhoid
cultures, but these may have come from constituents of
the medium or from the bacterial cells. Klebs4 concluded
that nuclein makes up a large part of the tubercle bacillus.
He extracted the bacilli with ether and benzol, digested
with hydrochloric acid and pepsin and dissolved the residue
in alkali. By precipitating the alkaline extract with alcohol,
he obtained a nuclein containing from 8 to 9 per cent, of
phosphorus. Hahn5 rubbed moist tubercle bacilli with
1 Zeitsch. f. physiol. Chem., 1898, xxv, 48.
2 Arch. f. Kinderheilkunde, 1900, xxx, 23.
3 La Riforma Medica, April 17, 1901.
4 Centralbl. f. Bakteriol., 1896, xx, 488.
6 Munch, med. Woch., 1897, xliv, 1344.
56 PROTEIN POISONS
sand, mixed with water, 20 per cent, glycerin, or physio-
logical salt solution, to a dough consistency, and subjected
the mass to a gradually increased pressure of from 400 to
500 atmospheres. The clear, expressed fluid contained a
large quantity of coagulable protein, which decomposed
hydrogen peroxide, but lost this property on being heated.
It gave the protein tests and behaved like a nucleoprotein.
Ruppel1 obtained a nuclein containing 9.42 per cent, of
phosphorus from the residue left after the preparation of
his tuberculosamin. This he called tuberculinic acid, and
he believed that it exists in the cell partly combined with
the tuberculosamin and partly free. Levene2 prepared
three proteins from tubercle bacilli grown on protein-free
medium. The dried bacilli were ground for two or three
days in a porcelain mill, then extracted repeatedly for two
days with an 8 per cent, solution of ammonium chloride.
These proteins coagulated at from 50° to 64°, 72° to 75°,
and 94° to 95° respectively. Ammonium sulphate pre-
cipitated all of them; sodium chloride only the first; 50
per cent, magnesium sulphate the first; magnesium sul-
phate to saturation the second, but not the third. It
required less acid to precipitate the first, but 0.2 per cent,
hydrochloric acid precipitated all three. The third was
richer in phosphorus than the others, and Levene con-
cluded that the tubercle bacillus consists principally of
nucleoproteins, one of which differs from the others in
that it is not precipitated by magnesium sulphate and does
not give the biuret reaction. He called attention to the
coincidence between the coagulation temperature of the
first protein and that necessary for the sterilization of the
bacillus. He believed that tuberculin is a specific substance
having the constitution of a nucleoprotein. He also made
a study of tuberculinic acid, finding but little of this free
in mannite synthetic cultures, but considerable in beef-
broth cultures. Samples differ in composition, and experi-
ments suggest that tuberculinic acid is less stabile than
1 Loc. cit.
2 Jour. Med. Research, July, 1901, 135; Medical Record, 1898, liv, 873.
BACTERIAL CELLULAR SUBSTANCE 57
any other known nucleic acid. De Schweinitz1 thought
that a nucleo-albumin is the fever-producing agent in the
tubercle bacillus. Maragliano2 made an aqueous extract
of the tubercle bacillus by digesting it on the water-bath
and obtained a poisonous substance. This comes from the
autolytic cleavage of the bacillus.
Carbohydrates. — One of the earliest studies of the chemistry
of bacteria was made by Scheibler3 upon leuconostoc mesen-
teroides. The viscous growth of this germ in beet juice,
after extraction with alcohol was boiled with milk of lime.
The filtrate furnished a gum which was regarded as an
anhydride of dextrose, since by slow hydrolysis it is con-
verted into the latter. This substance, to which the name
of dextran was given, is a white, amorphous powder, soluble
in water and dextrorotatory, having three times the rotary
power of cane sugar. Scheibler stated that this germ con-
tains ash, fat, water, dextra, and a substance containing
nitrogen, believed to be protagon or some closely related
body. Kramer4 separated from the slime of bacillus viscosus
sacchari two modifications of a carbohydrate of the formula
C6Hi0O5. Both were optically active, and on being boiled
with acid reduced Fehling's solution. Ward and Green5
found that a species of bacterium from Madagascar sugar-
cane secretes invertose. In sugar solutions it produces
a viscous growth that gives an opalescent solution in water,
which, when treated with alcohol, yields a bulky flocculent
precipitate, found to contain two carbohydrates, one of
which gives an osazone, and is optically active, while the
other is inactive. They regard these bodies as related to,
but not identical with, Scheibler's dextran, and are not
certain whether they are products of the vital processes
of the organism or are cleavage products. Vincenzi6 could
1 Bulletin No. 7, Bureau of Animal Industry; Jour. Amer. Chem. Society,
1897, xix, 782.
2 Berl. klin. Woch., 1899, xxxvi, 385.
3 Zeitsch. f. Rubenzuckerindustrie, 1874, xxiv, 309.
4 Monats. f. Chem., 1889, x, 467.
6 Proc. Roy. Soc., 1899, Ixv, 65.
6 Zeitsch. f. physiol. Chem., 1887, ix, 181.
58 PROTEIN POISONS
find no evidence of cellulose in bacillus subtillis, but Drey-
fuss1 heated masses of this organism to 180° with concen-
trated alkali and from this extract obtained a substance
that reduced Fehling's solution and from which crystals
of glucosazone were prepared. From this Dreyfuss con-
cluded that cellulose is present. However, this conclusion
is hardly justifiable. Like results were obtained from
pyogenic bacilli. Hammerschlag2 concluded that the
tubercle bacillus contains cellulose. The cell substance,
previously extracted with alcohol, ether, and 1 per cent,
potassium hydroxide, was dissolved in concentrated sul-
phuric acid, diluted, and boiled, after which it reduced
Fehling's solution. A second portion was treated with
potassium chlorate and nitric acid, but most of the substance
remained undissolved. A third portion was partly dissolved
in ammoniacal copper solution. Hammerschlag stated
that if one assumes that the nitrogenous material in the
tubercle bacillus is all protein and that the protein contains
16 per cent, of nitrogen, this bacillus contains 36.9 per
cent, of protein, 28.1 per cent, of cellulose, 27 per cent, of
substance soluble in alcohol, and 8 per cent, of ash. Nishi-
mura3 thought that he found hemicellulose in a water
bacillus, in prodigiosus, and in staphylococcus pyogenes.
De Schweinitz and Dorset4 extracted dried tubercle bacilli
with alcohol, digested the residue with 1.25 per cent, sodium
hydroxide for from forty to sixty minutes, washed the
residue, then digested with 1.25 per cent, sulphuric acid,
washed, dried, and ignited. The loss by ignition they
calculated should give the cellulose. Accordingly, they
reported 6.95 per cent, cellulose in the tubercle bacillus.
However, this conclusion is hardly accepted by these
authors themselves, since in the same paper they state that
cellulose is probably present in small amount in the tubercle
bacillus, and not present in the bacillus of glanders. Brown5
boiled "the membrane" of bacterium xylinum twenty
1 Loc. cit. 2 Loc. cit. 3 Loc. cit.
4 Jour. Amer. Chem. Soc., 1895, xvii, 605; ibid., 1896, xviii, 449; ibid.,
1897, xix, 782; ibid., 1898, xx, 618.
6 Jour. Chem. Soc., 1886, xlix, 432; ibid., 1887, li, 643.
BACTERIAL CELLULAR SUBSTANCE 59
minutes with 10 per cent, potassium hydroxide and found
that this digested the bacteria but left "the film" appar-
ently unchanged. The residue was washed with dilute
hydrochloric acid, then with water, and treated with
bromine according to Miiller's method for obtaining cellu-
lose. The product seemed to be identical with that from
cotton, dissolving in ammoniacal copper solution and in
strong sulphuric acid. It gave a reducing sugar, dextro-
rotatory, even when grown on media containing only levoro-
tatory substances. Analysis showed a close agreement
with cellulose, and Brown regarded this as a cellulose proper,
differing from the metacellulose usually found in yeast and
the fungi. No trace of dextran was found. Brown regarded
the formation of cellulose as a process of assimilation and
not of fermentation. Bendix1 extracted dried bacilli with
5 per cent, hydrochloric acid over the free flame, cooled,
made alkaline, then acidulated with acetic acid in order to
precipitate the protein. The filtrate gave with phenyl-
hydrazin an osazone which when purified melted at 153°
to 155°, thus showing it to be pentosazone. It gave the
orcin and optical tests for pentose. He obtained pentose
from diphtheria and tubercle bacilli, also from mixed fecal
bacteria, but not from the typhoid bacillus. The pentose
exists in the nucleoprotein. Aronson2 found a nucleo-
protein containing pentose in alkaline extracts of diphtheria
bacilli. This author stated that the residue after complete
extraction with alkali contains carbohydrate which is
dextrorotatory and yields an osazone; it is neither cellulose
nor chitin.
Meyer3 came to the conclusion that in some species of
bacteria fat is stored up, while in others complex carbo-
hydrates take their place. One species grown on barley
gave a substance, colored blue by iodine, and easily soluble
in malt diastase and in saliva. Bacillus subtilis gave
a body that is colored red by iodine and is dissolved by
saliva, and on boiling with dilute sulphuric acid. Meyer
1 Deutsch. med. Woch., 1901, xxvii, 18.
2 Loc. cit. 3 Flora, 1889, 432.
60 PROTEIN POISONS
thought' that this might be either glycogen or amylodextrin.
He also obtained a substance that he considered a mixture
of much amylodextrin and a little /3-amylose. It is easily
extracted from the cell with water. It may be remarked
here that our researches have shown that substances
removable from the cell by physical solvents constitute no
part of the actual cell protein. The carbohydrate in the
essential part of the cell is a constituent of the protein
molecule. All carbohydrates, fats, waxes, and inorganic
salts that may be washed out of the cell substance are
either no part of the cell proper or result from autolytic
changes in the cell molecules.
Levene1 obtained a glycogen-like body from the tubercle
bacillus. The cell substance was extracted with salts, or
better, with alkali, the albumins removed from the extract
with picric and acetic acids, the nuclein and carbohydrate
carried down together with alcohol, and then separated by
means of copper chloride. The glycogen thus obtained is
soluble in water, gives the iodine color test, and reduces
Fehling's solution after being boiled with dilute mineral
acid. Emmerling2 prepared chitin or a closely related body
from the zooglea of bacterium xylinum. From 110 grams
of moist, impure material he secured 0.2 gram of crystal-
line glucosamine hydrochloride. Helbing3 concluded that
chitin makes up a large part of the tubercle bacillus, and
to this constituent he attributed the peculiar staining
properties of this organism. He was clearly wrong in this
inference. All the early work on the carbohydrate con-
stituent of the bacterial cell, when the material was grown
on media containing carbohydrate, must be regarded as not
possessed of practical value.
Fat, Wax, etc. — In the earlier studies of the chemistry
of bacterial cells it was assumed that the alcoholic and
ethereal extracts consisted of fats exclusively. Kramer4
1 Loc. cit. 2 Berichte, 1899, xxxii, 541.
3 Deutsch. med. Woch., xxvi, Vereinsbeilage, 1900, 133.
4 Arch. f. Hygiene, 1891, xiii, 71; ibid., 1893, xvi, 151; ibid., 1895, xxii,
167; ibid., 1897, xxviii, 1.
BACTERIAL CELLULAR SUBSTANCE 61
noted that such an extract had the appearance of fat and
melted not much over 40°. Hammerschlag1 obtained from
the tubercle bacillus free fatty acids that melt at 63° and
concluded that the fat of this organism consists mainly of
tripalmatin and tristearin, and that it contains little or
no triolein. Nishimura2 obtained from the alcoholic and
ethereal extracts of his water bacillus a putty-like mass
with the properties of lecithin. Meyer3 found that the
fat in bacillus tumescens gradually increases until spore
formation occurs, when it disappears, the spores also being
free from fat. Klebs4 found in the tubercle bacillus 20.5
per cent, of a red fat, melting at 42°, and 1.14 per cent, of a
white fat melting above 50°; the latter being insoluble in
ether, but soluble in benzol. De Schweinitz and Dorset"'
saponified fats from the tubercle bacillus and from the
melting-points of the acids concluded that the fat of this
organism contains palmitic and arachidic acids, while
that of the glanders bacillus contains oleic and palmitic.
They also found a crystalline acid, for which they suggested
the name tuberculinic acid, though this is quite different
from Ruppel's nucleic acid. This new fatty acid was
obtained mainly from the culture medium, only in small
amounts from the bacilli. The crystals are prismatic or
needles, melting at 161° to 164°, readily soluble in water,
alcohol, and ether, and not responsive to the biuret test.
Analysis showed close correspondence to the formula,
C7H10O4. The authors called attention to the similarity in
composition and properties of this body to teraconic acid,
and suggested that this may be the substance which is
responsible for the coagulation necrosis and that it is the
temperature reducing substance. In a later paper they
described a crude fat extracted from the tubercle bacillus
and from which they obtained an acid melting at 62°,
unchanged by recrystallization. In concluding they decided
that the fat of the tubercle bacillus consists principally of
a glyceride of palmitic acid with a minute amount of the
1 Loc. cit. 2 Loc. cit. 3 Loc. cit.
4 Loc. cit. 5 Loc. cit.
62 PROTEIN POISONS
glyceride of a volatile fatty acid to which cultures of this
bacillus owe their characteristic odor, also a very small
amount of an acid (probably lauric) melting at 42° to 43°,
and an unusually high melting acid, one apparently with a
larger carbon content than any before noted in plants.
Ruppel1 obtained three extracts from the tubercle bacillus
by using successively cold alcohol, hot alcohol, and ether.
The first contains free fatty acids and a fat melting between
60° and 65°, easily saponified and decomposed into a free
acid and a higher alcohol. The second contained a waxy
mass, saponified with difficulty, and which seemed to be
the ester of a fatty acid and a high alcohol. The third
melted at 65° to 67°, and had an odor resembling that of
beeswax. Aronson2 obtained from tubercle bacilli, by
means of a mixture of five parts of ether and one of abso-
lute alcohol, a yellowish-brown tenacious mass, constituting
from 20 to 25 per cent, of the dried bacilhis. From the
growth of several hundred liters of culture 70 grams were
secured. This contained 17 per cent, of free fatty acids.
The remainder was wax, not acid and glycerin, but esters
of acid and alcohol insoluble in water. Most of this wax is
not in the cells, but lies around and between them. Levene3
found almost 30 per cent, of fat or wax in tubercle bacilli.
Kresslig4 extracted tubercle bacilli successively with ether,
chloroform, benzol, and alcohol, and obtained 38.95 per
cent, of fatty and waxy substances. Repeated extraction
with chloroform gave a dark brown mass of the consistency
and color of beeswax and melting at 46°. He found 14.38
per cent, of free fatty acid, 77.25 per cent, of neutral fat
and esters of fatty acids, and some volatile fatty acid,
probably butyric. He decided that the fat of the tubercle
bacillus is quite different from that obtained from any
other source.
Reducing Action of Bacteria. — Although the general
subject of the reducing action of bacteria scarcely falls
1 Loc. cit. 2 Berl. klin. Woch., 1896, xxxv, 484.
3 Loc. cit. 4 Centralbl. f. Bakteriol., 1901, xxx, 897.
BACTERIAL CELLULAR SUBSTANCE 63
within the domain of this work, it may be well to mention
the results of a few investigations. W. Smith1 found that
many bacilli, including the colon, decolorize methylene
blue, sodium indigo sulphate, litmus, etc. He concluded
that this reducing action is common to all bacteria, both
aerobic and anaerobic; that the velocity of reduction
depends upon the number of bacteria and the temperature;
that it is a function of the bacterial plasma, and that the
reducing substance does not diffuse into the culture medium,
but that the cell retains this property for a time after
death. Klett,2 testing this reducing action of bacteria on
sodium silicate, tellurate, and some other salts, also con-
cluded that the reducing agent exists in the cell, and is not
found among the cleavage products. Jegunow3 showed
that hydrogen sulphide is formed by the reducing action
of bacteria on sulphates and on organic bodies contain-
ing sulphur. Sulphur bacteria oxidize hydrogen sulphide
and store sulphur in the form of oily spheres, which may
constitute as much as 90 per cent, of the cell substance.
This sulphur is oxidized to sulphuric acid, thus serving as
a source of energy in the vital processes of the bacterium.
The sulphuric acid is neutralized by carbonates and sepa-
rated as a sulphate; then by bacterial activity the sulphate
is reduced, thus forming a complete cycle. If the bacteria
can obtain no sulphur they use that stored up in their
cells and die in from one to two days.
The above is a resume of the work done on the chemistry
of bacterial cells up to the time when our work was begun.
It should be clearly understood that we are not now con-
cerned with the cleavage products of bacteria produced in
the media in which they groAv. This subject is discussed in
Cellular Toxins by Vaughan and Novy (fourth edition, 1902).
Moisture, Ash, and Nitrogen. — Leach,4 in studying the
chemistry of the cellular substance of the colon bacillus
1 Centralbl. f. Bakteriol., 1896, xix, 181.
2 Zeitsch. f. Hygiene, 1900, xxxiii, 137.
3 Annuaire g6ologique et mineralogique de la Russie, 1900, ii, 157.
* Jour. Biol. Chem., 1906, i, 463.
64 PROTEIN POISONS
prepared in our laboratory, makes substantially the follow-
ing statement. The cell substance, prepared by the method
already described, takes up moisture readily and holds it
tenaciously, but may be dried to constant weight by heating
small amounts in a steam-drying oven for many days at from
85° to 95°. If the temperature falls to 60°, it may absorb
moisture even in the oven. One sample was heated to 105°
during working hours for three days and kept in a desiccator
during the intervals; it increased in weight. Drying in
vacua over sulphuric acid is, on the whole, the most satis-
factory method, although it may require days and even
weeks.
The dried cellular substance burns with a flame, forming
volatile and liquid products, giving off odors characteristic
of nitrogen compounds, and finally leaving a greenish ash.
Two determinations gave the following results:
0.346 gram gave 0.0296 gram of ash, or 8.55 per cent.
0.496 gram gave 0.0431 gram of ash, or 8.68 per cent.
Values reported for other bacteria vary from 3 per cent,
in putrefactive organisms to 13 per cent, in prodigiosus, or,
by using special media, to nearly 30 per cent, in the cholera
bacillus; while in the tubercle bacillus the ash has been
found to vary from 1.77 to 5.92 per cent, according to
conditions. The ash from the colon bacillus, as we have
prepared it, contains sodium, potassium, small amounts of
calcium, aluminum, copper, and phosphates. A slight
residue insoluble in acid is probably silica. Sulphate is
present in so small an amount that it may escape detection,
and chloride has not been found. In comparison with the
data obtained with other bacteria, these findings are note-
worthy only in the absence of magnesium, and in the
presence of copper and aluminum. Presumably the former
comes from the tanks and the latter from the agar.
Phosphorus was the only constituent of the ash quanti-
tatively determined. The ash was dissolved in nitric acid,
the phosphate precipitated with ammonium molybdate,
dissolved in ammonia, precipitated with magnesia mixture,
BACTERIAL CELLULAR SUBSTANCE 65
and weighed as pyrophosphate. The following results
were obtained:
Weight of Weight of
sample. phosphate. Weight of P. Per cent, of P.
0 . 496 gram 0 . 0475 gram 0 . 01323 gram 2 . 68
0 . 346 gram 0 . 0380 gram 0 . 0 1 059 gram 3 . 06
The mean of these determinations, 2.87 per cent., agrees
quite closely with Levene's finding, 2.67 per cent, of phos-
phorus in the tubercle bacillus from mannite cultures.
Most observers report smaller results, but it should be
noted that our samples are free from fat and wax, and
therefore the percentage is higher than if calculated for
the cells not previously extracted with alcohol and ether.
In view of the fact that the cellular substance, notwith-
standing the washings to which it has been subjected,
cannot be regarded as chemically pure, we have not
wasted time in making elementary analyses. Wheeler
has collected the nitrogen and ash determinations made
in bacterial cellular substance in this laboratory and has
arranged them in the following table:
Substance. Per cent, of nitrogen. Per cent, of ash.
Typhoid 11.55 5.70
Colon 10.65 8.615
Colon 8.38
7.20 (air-dried)
Tuberculosis 10.55 11.47
9.27 (air-dried) 9.98 (air-dried)
Anthrax 10.285 7.76
Subtilis 5.964 10.83
Proteus vulgaris 6.791 10.88
Ruber of Kiel 10 . 655 6.71
Megaterium 8.349 10.18
Pyocyaneus 10.843 9.04
Violaceus 11.765 6.90
Sarcina aurantiaca .... 11.460 6.40
As will be seen, the nitrogen varies from 5.964 per cent,
in subtilis to 11.765 per cent, in violaceus, and the ash
from 5.7 per cent, in the typhoid bacillus to 11.47 per cent,
in the bacillus tuberculosis.
5
66 PROTEIN POISONS
Nicolle and Alilaire1 give the following table, showing the
percentage of water, nitrogen, substance soluble in acetone,
and phosphorus in the bacterial cells named:
Organism. *g ^ *3 g co_; E
.+3 X +3 0) -*J CO O
fl do " fe "-1 « ^ *-•
S g« g.nP s.a.2 go
B
ll
PH
._
PH
n 55o
•
PH
t, <a o
£
&
(- 3
&
PH
B. glanders .
76.49
10.47
11.69
8.59
3.10
2.530
8.0
B. chicken cholera
79.35
10.79
7.54
6.30
1.24
2.370
7.5
B. cholera
73.38
9.79
8.70
6.77
1.93
2 370
7.5
B. of Shiga . . .
78.21
8.89
12.80
10.57
2.23
1.570
5.0
B. proteus .
79.99
10.73
10.87
7.10
3.77
1.580
5.0
B. typhoid .
78.93
8.28
15.44
10.64
4.80
1.160
3.5
B. anthrax .
81.74
9.22
6.31
1.48
4.83
0.948.
3.0
B. pseudotubercu-
losis ....
78.83
10.36
15.63
10.31
5.32
0.793
2.5
B. pneumonia .
85.55
8.33
15.45
7.36
8.06
0.790
2.5
B. coli ....
73.35
10.32
15.25
11.77
3.48
0.790
2.5
B. prodigiosus .
78.00
10.55
9.00
6.60
2.40
0.474
1.5
B. psittacosis .
78.05
9.55
11.08
• 7.03
4.05
0.474
1.5
B. diphtheria .
84.50
7.04
5.23
1.81
0.158
0.5
B. pyocyaneus .
74.99
9.79
15.77
10.67
5.10
0.157
0.5
B. lymphangitis . .
77.90
9.17
6.83
2.53
4.30
0.157
0.5
Froberg's yeast
69.25
10.00
4.22
2.92
1.30
0.000
0.0
Chlorella vulgaris .
63.06
3.96
21.10
12.81
8.29
0.000
0.0
Carbohydrates. — In no case have we been able to detect
cellulose in the bacterial cell substance. Wheeler made
special search for it in sarcina lutea. Twenty grams of
substance was autoclaved with 25 parts (500 c.c.) of 10
per cent, potassium hydroxide at 120°, first for thirty
minutes and then for an hour. There remained a consider-
able residue which no longer gave the protein reactions,
but did respond to the carbohydrate test with a-naphthol,
although it did not reduce Fehling's solution even after
prolonged boiling with dilute hydrochloric acid. Cellulose
could not be detected by any of the tests employed.
Schweitzer's reagent failed to dissolve it, and it gave no
1 Annales de 1'Institut Pasteur, 1909, xxiii, 547.
BACTERIAL CELLULAR SUBSTANCE 67
color with iodine even after treatment with sulphuric acid.
A portion was dried, and heated with soda lime, when it
evolved a gas which turned red litmus paper blue, thus
indicating nitrogen which had been reduced to ammonia.
The odor of burning feathers also indicated the presence
of nitrogen. From these results it was concluded that the
residue left after extraction of the cellular substance with
10 per cent, potassium hydroxide at 120° contains a car-
bohydrate, but there is nothing to indicate that it is cellu-
lose. Leach made a search for cellulose in the cells of the
colon bacillus, with like negative results.
There are two carbohydrate bodies in bacterial cellular
substances. One of these furnishes a reducing sugar after
being boiled with dilute mineral acid, while the other does
not. The former may be extracted from the cells with
either alkali or acid, better with the former. In Wheeler's
studies of sarcina lutea the portion soluble in 10 per cent,
potassium hydroxide was filtered through paper, acidified
with hydrochloric acid, and treated with three volumes of
95 per cent, alcohol, which produced an abundant white,
curdy, sticky precipitate. Precipitation by means of
alcohol with acetic and picric acids was also tried, but did
not prove satisfactory. Purification was attempted by
repeated solution and precipitation with acidified alcohol,
but the quantity was diminished each time on account of
its relative solubility in dilute alcohol. After filtering,
washing, and drying in an atmosphere of carbon dioxide,
the powder obtained weighed only 0.7872 grams. It con-
tained phosphorus, responded to the carbohydrate test,
and reduced Fehling's solution after prolonged boiling with
dilute mineral acid. Its phosphorus content was deter-
mined and found to be 0.861 per cent., which is too low to
indicate the presence of nucleic acid or nuclein in anything
like a pure condition. The same investigator at one time
took 300 grams of the colon germ substance and heated
on the water-bath with six liters of 2 per cent, potassium
hydroxide. The extract was filtered through paper and
acidified with acetic acid. The precipitate produced,
68 PROTEIN POISONS
presumably protein, was filtered out after standing twenty-
four hours, and was so small that it was lost on the filter
paper.
The filtrate was then poured into three volumes of 95
per cent, alcohol acidified to the extent of 0.5 per cent, with
hydrochloric acid. This formed a heavy, white, curdy,
fibrous precipitate, which was filtered, washed acid-free
with alcohol and then with ether. It was purified by
repeated solution in 0.5 per cent, alkali and precipitation
with alcohol. Finally there was obtained a fine white
powder amounting to something less than 10 per cent, of
the original cellular substance, but much had been lost by
its partial solubility in dilute alcohol. This powder consists
almost wholly of a carbohydrate which is converted into
a reducing sugar after prolonged boiling with dilute mineral
acid. However, it contained 5.9 per cent, of ash and 0.194
per cent, of phosphorus. Solutions of this powder give
none of the protein reactions, with the- exception of the
xanthoproteic, to which they responded imperfectly.
Leach prepared the same body, but with a higher phos-
phorus content, from the colon bacillus. The cellular
substance, after repeated extraction with dilute (1 to 5
per cent.) sulphuric acid, was extracted upon the water-
bath or over a free flame with from 2 to 4 per cent, of
sodium hydroxide. *The alkaline extract, after filtration,
was neutralized with hydrochloric acid and poured into 95
per cent, alcohol. A light colored, flocculent precipitate
was obtained. This turned dark on the exposure to air
incident to filtration. It was twice dissolved in 0.5
per cent, potassium hydroxide and reprecipitated with
acidified alcohol. Each time the fresh precipitate was
white or nearly so, but the utmost care in filtering, even
in an atmosphere of carbon dioxide, did not prevent its
turning dark. The solution in alkali gave the xantho-
proteic and furfurol tests, but neither the biuret nor Millon
test. Copper chloride gave a precipitate, but picric acid
and platinum chloride did not. The solution was accord-
ingly acidified with picric and acetic acids, copper chloride
BACTERIAL CELLULAR SUBSTANCE 69
added, and the mixture poured into three volumes of
alcohol. An aqueous solution of the precipitate thus
obtained did not reduce Fehling's solution, but after boiling
with hydrochloric acid it reduced both Fehling's and
Nylander's solutions, and also gave the furfurol test, thus
showing the presence of a carbohydrate. The original
powder burned readily, puffing up and glowing as does
nucleic acid, then fusing and leaving a dark ash. Two
determinations of phosphorus gave the following:
Weight of Weight of
sample. pyrophosphate. Weight of P. Per cent, of P.
0 . 4723 gram 0 . 0524 gram 0 . 01450 gram 3 . 09
0.6469 gram 0.0685 gram 0.01895 gram 2.93
In our attempts to extract the poisonous groups from
bacterial proteins this carbohydrate gave us great trouble.
It is readily soluble in water, whether acid or alkaline, and
more or less soluble in alcohol, the degree of solubility
depending upon the strength of the alcohol. In absolute
alcohol it is insoluble, but it cannot be precipitated com-
pletely from aqueous solution by the addition of alcohol.
Concentrated solutions and residues obtained by evapora-
tion in vacuo are sticky and unsatisfactory in all attempts
at purification. As we ascertained after much experimenta-
tion, the poisonous group in the protein molecule is freely
soluble in absolute alcohol, and finally when we disrupted
the protein molecule with a dilute solution of alkali in
absolute alcohol we secured a complete separation of the
poisonous group from both carbohydrates. Therefore the
best material in which the bacterial or other protein carbo-
hydrates can be studied is the non-poisonous portion after
complete removal of the poisonous group by heating the
protein repeatedly with a 2 per cent, solution of sodium
hydroxide in absolute alcohol. This method will be dis-
cussed in detail later, but it needs to be stated here that
when this is done the poisonous group, free from any trace
of either carbohydrate, goes into solution in the alkaline
alcohol, while the non-poisonous part, or the haptophor, as
70 PROTEIN POISONS
we have designated it, containing all the carbohydrates,
remains insoluble in this menstruum.
Leach1 has studied the carbohydrate in the haptophor
portion of the cellular substance of the colon bacillus.
Gram samples of the haptophor portion were dissolved in
water containing a little alkali, neutralized with hydro-
chloric acid to definite strength and heated on a water-bath
in a flask with a reflux condenser. The hydrolyzed solution
was neutralized and titrated with Fehling's solution. Al-
though there is undoubtedly some pentose present, there
is no proof that the reducing substance is all carbohydrate.
However, for purposes of comparison the reducing sub-
stance was calculated as xylose. In order to find conditions
giving the maximum yield, amount and strength of acid
as well as time of boiling were varied as shown in the
following table:
REDUCING POWER OF COLON HAPTOPHOR
No. of Amount of Per cent, of Hours Per cent, calculated
sample. HC1. HC1. boiled. as xylose.
1 26.0 c.c. 1.0 1 7.05
2 38.8 c.c. 2.5 1 16.45
3 38.5 c.c. 2.5 2 21.56
4 . 38.5 c.c. 2.5 4 23.12
5 72.0 c.c. 2.5 3 23.93
6 72.0 c.c. 2.5 9 23.53
As shown by these figures the maximum amount of
reducing substance was obtained by using 2.5 per cent,
acid, and boiling for three hours. Longer heating changes
the result very little.
Attempts were made to separate this carbohydrate from
the other constituents of the haptophor of the colon bacillus.
A 5 per cent, aqueous solution of the haptophor was poured
into four volumes of absolute alcohol, containing 10 c.c. of
hydrochloric acid and 100 c.c. of ether per liter. After
settling, the supernatant liquid was siphoned off and the
precipitate (known as G) collected with suction, washed
i Jour. Biol. Chem., 1907, iii, 443.
BACTERIAL CELLULAR SUBSTANCE 71
with alcohol containing ether, then with ether, dried, and
pulverized. The yield from 50 grams of haptophor was 19
grams. This was twice dissolved in water made faintly
alkaline with sodium acid carbonate, and reprecipitated by
alcohol containing hydrochloric acid and ether. The final
precipitate G weighed 16 grams. With water, G forms an
emulsion, acid in reaction and cleared by the addition of
alkali. The biuret test is negative, Millon doubtful, xantho-
proteic, Adamkiewicz, a-naphthol, and orcin tests are all
positive, the carbohydrate tests being very marked. After
boiling with acid there is copious reduction of Fehling's
solution. G was tested for glycogen, with negative results.
One gram of G boiled two and one-half hours with 72 c.c.
of 2.5 per cent, hydrochloric acid, gave 38.63 per cent, of
reducing substance calculated as xylose. A second gram
boiled for five hours yielded 43.77 per cent.
The second carbohydrate, or, more properly, the second
substance giving the a-naphthol test in bacterial proteins,
is not converted into a reducing sugar on being boiled with
dilute mineral acid. In Wheeler's work with sarcina lutea
it remained as a residue after extracting the cell material
with 10 per cent, potassium hydroxide at 120°. This
residue responded to the a-naphthol test, did not give the
protein reactions, and did contain nitrogen. The same
investigator also found this body in alkaline extracts of
the residue left after extraction with dilute acid. In her
work with the cell substance of the colon bacillus, Leach
makes the following statement touching this body: "The
residue, after repeated extraction with dilute sulphuric
acid (from 1 to 5 per cent.), was treated with 2 to 4 per
cent, sodium hydrate either upon the water-bath or over
a free flame. In every case the substance went into solution
readily, leaving only a slight coating on the filter. The
slight residue gave no protein test, contained no nitrogen,
but gave test for carbohydrate. In one case it was removed
from the filter, and the organic matter approximately
determined. The total residue was about 0.4 gram, equiva-
lent to 0.8 per cent, of the cell substance used. The organic
72 PROTEIN POISONS
matter .was only 0.15 gram, equivalent to 0.3 per cent, of
the original." This body is also found in the dilute acid
extracts of cellular proteins, as is shown by the following
additional quotation from Leach: "Some earlier investi-
gations in this laboratory upon the toxicity of the colon
germ showed the desirability of studying the action of
dilute acid upon the cell substance. Accordingly, samples
were treated with 1 per cent, sulphuric acid under varying
conditions. On filtering, a light brown or straw-colored
fluid was obtained. This readily reduced nitric acid and
gave the typical xanthoproteic color on the addition of
ammonia. In no case was there more than a slight biuret
test, and there was too much sulphate present for a satis-
factory Millon test. The a-naphthol test for furfurol was
positive. Alcohol gave a voluminous precipitate, A, which
will be described more fully under another heading. The
alcoholic filtrate, B, was neutralized with sodium hydroxide,
the sodium sulphate filtered out, together with some organic
matter mechanically carried down, and the liquid distilled
under diminished pressure at 30° to 38°. The liquid residue,
C, left after distillation, turned yellow on heating with
potassium hydroxide, but gave neither the biuret nor
Millon test. Again, the xanthoproteic and a-naphthol
tests were positive, but it failed to reduce either Fehling's
or Nylander's solution (after boiling with dilute mineral
acid). It yielded precipitates with ammonium molybdate,
phosphomolybdic acid, ammoniacal silver nitrate, and
picric acid. A guinea-pig was injected with 5 c.c. of C
with no apparent effect."
We are inclined to attribute the sticky, mucilaginous
properties of both acid and alkaline extracts of bacterial
cellular substances to these bodies giving the furfurol
reaction, and here regarded as carbohydrates. Further-
more, we are of the opinion, though this cannot be con-
sidered as conclusive, that the one yielding a reducing sugar
after boiling with dilute mineral acid exists in the cellular
molecule as a constituent of the nucleic acid group, while
the other is a part of the protein component.
BACTERIAL CELLULAR SUBSTANCE 73
The reducing carbohydrate is present in the bacterial
cellular substance examined in this laboratory in a minimum
of something over 10 per cent.; in the colon haptophor in
about 24 per cent. ; and in precipitate G from the haptophor
in about 44 per cent. So far, we have not obtained it free
from phosphorus. The percentage of the other furfurol
giving body we have no means of determining even approxi-
mately, although we can safely say that the amount is
much smaller. The one yielding a reducing sugar probably
exists in the nucleic acid group as a pentose.
Nuclein Bodies. — In her work on sarcina lutea, Wheeler1
makes the following statement: "So far as the xanthin
bases are concerned, Nishimura2 found 0.17 per cent, of
xanthin, 0.08 per cent, of adenin, and 0.14 per cent, of
guanin in his water bacillus. It has been suggested that in
Nishimura's experiments these bases might have come
from the potato upon which his organism was grown, but
inasmuch as the potato contains only a very small per-
centage of protein, this is not likely. Lustig and Galleotti3
report guanin obtained from the pest bacillus, and Galleotti4
says that a nucleoprotein separated from the bacillus
ranicidus yielded xanthin bases, although the percentage
of nitrogen was low.
"I have gone through the process of testing for xanthin
bases four times. Three times the acid extracts were
carefully precipitated with powdered silver nitrate crystals
until a drop of the solution gave a yellow instead of a white
precipitate with barium hydrate. The precipitate was
filtered out, washed, dried, and then worked up for xanthin
bases. The fourth time the process was almost the same,
the difference being that 33^ per cent, acid extract had
been made. This was first almost neutralized with barium
hydrate, the barium sulphate filtered out, carefully washed
out and boiled with water, and then the slightly acid
extract was precipitated with silver sulphate instead of
silver nitrate. The first silver nitrate precipitate was
1 Trans. Assoc. Amer. Phys., 1902, xxvii, 265.
2 Loc. cit. 3 Loc. cit. 4 Loc. cit.
74 PROTEIN POISONS
investigated according to the method given by Kruger and
Solomon;1 as no satisfactory separation was thereby ob-
tained, the last three precipitates were separated according
to the method of Kossel and his pupils, as outlined by
Hammersten,2 for the separation of the four bases, xanthin,
hypoxanthin, guanin, and adenin. The precipitate was
dissolved as completely as possible in boiling nitric acid
(specific gravity, 1.1), a little urea having been added
to prevent nitrification, filtered hot, concentrated some-
what, and allowed to cool. On cooling, only a slight
residue of the guanin-hypoxanthin-adenin portion separ-
ated out. On decomposing this precipitate, treating with
ammonia and evaporating, the amount obtained was so
small that it was not possible to make separation of the
bases, but ammoniacal solution produced a comparatively
heavy flocculent organic precipitate. The nitric acid filtrate
containing the xanthin portion was precipitated with
ammonia. A heavy, reddish-brown, mucilaginous precipi-
tate came down, but was not sufficient in quantity or
sufficiently free from impurities to justify an ultimate
analysis."
Leach3 obtained from 1 per cent, sulphuric acid extracts
of the colon cellular substance a body containing 7.33 per
cent, of phosphorus. It gave none of the protein color
reactions except the ubiquitous xanthoproteic. It could
hardly be anything else than a nucleic acid. However, the
percentage of nitrogen was only 8.98, and no known nucleic
acid contains so small an amount of nitrogen. In the same
extracts she obtained indications of the presence of two
xanthin bases, xanthin and guanin. The evidence of the
existence of nuclein bodies in the haptophor of the colon
cellular substance will be given later.
Diamino-acids. — Wheeler failed to obtain any evidence
of the presence of arginin or histidin in sarcina lutea, but
in each of five attempts she secured quite convincing proof
1 Zeitsch. f. Physiol. Chem., 1898, xxvi, 373.
2 Physiol. Chem., p. 120, as translated by Mendel.
3 Jour. Biol. Chem., 1906, i, 463.
BACTERIAL CELLULAR SUBSTANCE 75
of the existence of lysin. She says: "As to lysin, I have
obtained in every instance a yellow, thick, oily body, where
lysin picrate should be formed, which, however, could not
be crystallized. This oily body was shaken up with ether
to remove excess of picric acid, but when I attempted to
purify it by redissolving it in alcohol, it was no longer
completely soluble, inasmuch as a part of it hardened into
a solid which seemed somewhat crystalline, and the remainder
was precipitated by alcohol. However, it was found to be
readily soluble in water, especially in hot water, and although
no crystals were obtained, on concentration of the aqueous
solution, the same heavy oil separated. In order to obtain
the hydrochloride, if possible, the oily substance was treated
with hydrochloric acid in a little water, and the picric
acid shaken out with ether. From the solution on concen-
tration an imperfectly crystalline mass was obtained. It
may be that I have not lysin, but there is undoubtedly
present some organic body which, in its chemical behavior
at least, is very similar to lysin."
In searching for the hexon bases in the cellular substance,
Leach proceeded as follows : " The cell substance was stirred
with nine times its weight of 33.33 per cent, sulphuric
acid, allowed to stand overnight, then heated in an evapor-
ating dish on the water-bath. When all danger of frothing
was over, the mixture was transferred to a flask fitted with a
reflux condenser, and boiled on a sand-bath for eight hours
one day and six the next. After cooling and filtering, some
water was added to the filtrate, and it was neutralized by
the addition of the calculated amount of barium hydrate.
When the barium sulphate had completely settled, the
supernatant liquid was siphoned off, the precipitate stirred
up with boiling water, heated to boiling, settled overnight,
and again siphoned. This was repeated until the wash
water was nearly colorless. The extract and wash water
were united, acidified with acetic acid,1 concentrated on
1 If there is a large excess of barium present, it is well to remove it by
carbon dioxide.
76 PROTEIN POISONS
the water-bath, cooled, and filtered to remove any tyrosin
and leucin that may crystallize out. The filtrate was
diluted to about one and one-half liters for each 100 grams
of cell substance, made decidedly acid with nitric acid,
and 20 per cent, silver nitrate added as long as it gave a
precipitate. It was left overnight to settle, and the silver
precipitate of xanthin bases filtered out. To this filtrate
excess of silver nitrate and barium hydrate were added to
remove arginin and histidin. After their removal, silver
and barium were precipitated by hydrochloric and sulphuric
acids, these inorganic precipitates boiled out with water
several times, the filtrate and wash water united, and con-
centrated. The solution, which should contain some 5
per cent, of acid, was treated with a 50 per cent, solution
of phosphotungstic acid as long as it gave an immediate
precipitate. The precipitate was rubbed up with 5 per
cent, sulphuric acid, carefully washed with the same
solution, and filtered with suction. The heavy white
precipitate was again rubbed up with water, hot saturated
solutions of barium hydrate added, until the mixture was
no longer acid, settled overnight, and the supernatant
liquid siphoned off. The precipitate, consisting of barium
phosphate, tungstate, etc., was washed several times with
hot barium hydrate solution, decanted, and finally filtered
by suction. The filtrate and wash water were united, and
barium was removed as carefully as possible, first by running
in carbon dioxide, and then by adding ammonium carbonate
to the solution. This precipitate, like all the other inorganic
ones, was boiled out several times with water, and the
washings added to the original filtrate. The resulting
liquid was concentrated nearly to dryness on the water-
bath, the residue taken up with water, filtered to remove
barium carbonate, and again concentrated to a thick
syrup.
"The alkaline syrup was vigorously stirred with alcohol,
and then with an alcoholic solution of picric acid. Some-
times a crystalline precipitate came down at once, sometimes
there was a viscous mass like molasses candy, which became
BACTERIAL CELLULAR SUBSTANCE 77
granular or crystalline after long kneading and stirring, as
Fischer and Weigert suggest. When picric acid would no
longer give a precipitate even on standing, the crystals
were filtered out by suction, washed with alcohol, and
dried on a porous plate. On concentration the alcoholic
mother liquid became gummy and viscous, but no more
crystals were obtained. The crude picrate was recrystallized
from hot water several times. On dissolving there was
much sediment which mainly filtered out, but on concen-
tration more appeared upon the sides of the vessel. The
loss of substance by the first crystallization was very large;
as it became pure, however, it crystallized like an inorganic
salt. All mother liquors were treated with hydrochloric
acid to remove picric acid, reprecipitated with phospho-
tungstic, the precipitate worked up as before, and a further
crop of crystals obtained. The crystals are slender, yellow,
silky, felted needles or prisms. On heating in a melting-
point tube the substance begins to change color at 216°,
and is very dark at 230°. Heated side by side with lysin
picrate from fibrin and from gelatin, they agree within a
degree. The authorities all agree that lysin picrate turns
black at 230° to 232°, while Kutscher and Lohmann also
say that it begins to change color at 215°.
"To change the picrate into the chloride, 2 grams were
dissolved in 33 c.c. of hot water, 5 c.c. of concentrated
hydrochloric acid added, cooled, the picric acid filtered out
and washed with water containing hydrochloric acid. The
filtrates were shaken out with ether as long as there was
any yellow color. The solution should be colorless or
nearly so; if it is not, it can be decolorized by treatment
with animal charcoal. The solution was evaporated nearly
to dryness, first on the water-bath, and finally in a desic-
cator. When down to a thick syrup, stirring gave crystals.
These were recrystallized out of hot water containing
hydrochloric acid, giving long prisms, which melt at 192°,
again agreeing with the corresponding salt from gelatin
and fibrin. Henze says that lysin chloride becomes soft
at 193° and melts at 195°; Lawrow says that it has no sharp
78 PROTEIN POISONS
melting-point, but begins to melt at 194° to 195°. Hender-
son collected samples melting from 190° to 200°, prepared
by different individuals from widely different sources.
By careful purification he obtained from each sample a
product melting at 192° or 193°. Thus it would appear
that the apparent discrepancy in the melting-point is due
to impurities. Reactions, crystalline form, properties, and
melting (or decomposing) show that the picrate and chloride
from the germ are identical with lysin picrate and chloride
from gelatin and from fibrin. Thus the presence of one of
the hexon bases in the bacterial cell has been demonstrated,
and another point of resemblance between bacterial and
other proteins has been established."
Mono-amino-acids. — The phosphotungstic acid filtrate ob-
tained by Leach in her work on the hexon bases was turned
over to Wheeler, who has made the following report on it:
"From the solution phosphotungstic acid was removed
with barium hydrate and carbon dioxide used to remove
excess of barium. By concentration and crystallization
bodies were obtained resembling tyrosin and leucin under
the microscope. These were purified by repeated recrystal-
lization from water or from ammoniacal water, the tyrosin
being so much less soluble than the leucin that they could
be separated by difference of solubility. It was necessary
to boil the leucin fraction with animal charcoal to remove
the coloring matter. The tyrosin formed the characteristic,
colorless, silky needles, many grouped in the characteristic
sheaves. As it became more pure the needles were longer
and longer, and grouped in the sheaves less positively.
After many purifications the crystals melted at a constant
temperature, though this was difficult to determine, since
tyrosin melts with decomposition. The melting-point
maintained after each of two or three recrystallizations
was 288°, unconnected. The correction was 8.13°, which
made the corrected point 296.13°. A Kahlbaum prepara-
tion of tyrosin in the laboratory melted within a degree of
the same point, and agreed in the chemical tests, to be
mentioned presently. Richter gives the melting-point of
BACTERIAL CELLULAR SUBSTANCE 79
tyrosin as 235°; Cohn as 295°; while Fischer says that with
rapid heating the corrected point is 314° to 318°.
"The tyrosin obtained gives a Hofmann test with
Millon's reagent. It gives Scherer's test with nitric acid
and sodium hydrate on platinum foil, and also a beautiful
Piria test with sulphuric acid, then barium carbonate and
ferric chloride, which test is characteristic for tyrosin.
"The leucin crystallized in the characteristic knobs or
balls. As it became purer it crystallized more and more in
shining, white, very thin plates, sometimes in radial groups,
sometimes not. The crystals were finally obtained with a
practically constant melting-point, 262° to 263° or corrected,
268.5° to 269.6°. The pure laboratory leucin (Kahlbaum)
melted at the same point. Schwanert, Hammersten, and
others give the melting-point for active leucin as 170°;
that for the inactive form is given as 270°. Fischer says
the melting-point is 293° to 295° (corrected) if heated
quickly in a closed tube. Cohn gives 275° to 276°. The
leucin obtained melted with darkening and decomposition.
With careful heating in an open tube it sublimed with
the characteristic white, woolly deposit. It also responded
to Scherer's test on platinum foil with nitric acid and
sodium hydrate, which test Hammersten says is charac-
teristic for leucin."
Agnew1 has made the following contribution to the
mono-amino-acid content of the cellular proteins of the
colon and tubercle bacilli.
The material used in this research consisted of the cellular
substance of the bacteria. Growths from massive cultures
were placed in large Soxhlets and extracted for three days
with absolute alcohol and for the same time with ether.
The protein bacterial substance thus freed from everything
soluble in alcohol and ether was ground into a powder and
passed through a fine-meshed sieve. For the preparation
of the amino-acids Fischer's method slightly modified was
employed. The cellular protein was boiled with three and
1 Unpublished research.
80 PROTEIN POISONS
one-half times its weight of strong hydrochloric acid under a
reflux condenser until it failed to respond to the biuret
test. The humus substance was filtered out, repeatedly
extracted with hydrochloric acid, and the concentrated
extracts added to the filtrate from the humus. From this
filtrate glutamic acid was isolated as a hydrochloride by
saturation with hydrochloric acid gas and standing for some
days in the ice-box. The deposited glutamic acid was
collected on a filter with the aid of a pump. The filtrate
was concentrated in vacua to a syrup, diluted with an equal
volume of alcohol, and esterified by saturating with hydro-
chloric acid gas, the solution being warmed to complete
the esterification. The alcohol was distilled off at a tem-
perature under 40°, more alcohol added, and removed by
distillation, this being repeated three times. At this point
glycocoll, as the hydrochloride of the ethyl ester, crystal-
lized out on account of difficult solubility in alcohol. For
convenience, the thick syrup containing the esters of the
hydrochlorides of the amino-acids was divided into portions
and each treated separately but in the same way. The
syrup was diluted with half its volume of water, cooled in a
freezing mixture, and made slightly alkaline with sodium
hydrate. This alkaline solution was extracted with ether,
the liquid being kept alkaline by the addition of a few drops
of a saturated solution of sodium hydrate. The liquid was
then converted into a paste by the addition of solid potas-
sium carbonate and repeatedly shaken with ether. The
combined ethereal extract was dried by shaking with solid
potassium carbonate for a short time, and by being left for
twelve hours over fused sodium sulphate. The ether was then
distilled off and the free esters subjected to fractional dis-
tillation in vacua. The highest vacuum that I was able to
obtain varied from 20 to 30 mm. The pasty mass left after
the distillation of the esters, according to Abderhalden's
recommendation, was made acid with hydrochloric acid,
concentrated, and the inorganic salts allowed to crystallize.
The filtrate freed from these crystals was again esterified.
furnishing a further but smaller amount of esters.
BACTERIAL CELLULAR SUBSTANCE 81
Hydrolysis of the Cellular Protein of Bacillus Coli Com-
munis. — Three hundred and fifty grams of this cellular
protein was thoroughly mixed with 1100 c.c. of hydrochloric
acid, of specific gravity 1.19, and allowed to stand over-
night. The next morning it was found to be a frothy mix-
ture of deep purple color (Liebermann's reaction). It was
placed in a 2-liter flask connected with a long reflux con-
denser, and boiled on the sand-bath for six hours, when it
no longer gave the biuret reaction. It was then filtered
through heavy paper, giving a clear brown filtrate and
leaving a heavy deposit of humus on the paper. The latter
was repeatedly extracted with dilute hydrochloric acid,
the extracts concentrated, and added to the filtrate The
humus, air-dried, weighed 77 grams, making 22 per cent,
of the cellular protein.
The filtrate was concentrated in vacua to half its volume,
placed in a freezing mixture, saturated with hydrochloric
acid gas and left in the ice-box for two days. By this time
the hydrochloride of glutamic acid had crystallized out.
An equal volume of ice-cold alcohol, previously saturated
with hydrochloric acid gas, was added, and the glutamic
acid collected, dried over solid sodium hydrate, and
weighed. It yielded 10.5 grams of 3 per cent, of the cellular
protein.
The filtrate from the glutamic acid was esterified with
hydrochloric acid gas, followed by gentle heat. The alcohol
was distilled off in vacuo, an equal volume added, again
saturated with the gaseous acid, heated, and removed by
distillation, this being repeated several times. The fluid
was left in the ice-box for forty-eight hours, when crystals
were deposited. These were purified by reprecipitating
them from alcoholic solution with hydrochloric acid gas.
The melting-point was 144° and the crystals were identified
as the ethyl ester of glycocoll. The yield of glycocoll was
1 gram.
The filtrate from the glycocoll containing the hydro-
chlorides of the esters was treated as already stated and
fractioned with the following yields:
6
82 PROTEIN POISONS
Bath temperature. Yield.
Fraction I 40° to 60° 16.5 grams
Fraction II 60° to 80° 8.5 grams
Fraction III 80° to 100° 22.5 grams
Fraction IV 100° to 130° 20.0 grams
Fraction V 130° to 160° 16.5 grams
The dark red residue left in the distillation flask was
dissolved in hot alcohol, decolorized with animal charcoal,
filtered, and from the filtrate 0.5 gram of leucinimid was
obtained.
Fraction I. — This was saponified by boiling with five
times its weight of water for five hours with a reflux con-
denser. On concentration two crops of crystals were
obtained. The first was dissolved in the smallest possible
amount of hot water, then treated with an equal volume
of alcohol, and left in the ice-box for twenty-four hours.
The white crystalline mass, tasting sweet and closely
resembling alanin, weighed 1.75 grams.
The second crop was obtained by evaporation to dryness,
and weighed 1 gram. It was dissolved in 3 c.c. of hot
water; this was poured into an equal volume of hot absolute
alcohol, and set in the ice-box for twenty-four hours. White
needle-like crystals, sweet to the taste and arranged in
bundles, formed. These were recrystallized, washed with
absolute alcohol, dried, and weighed. The yield was 0.5
gram. These crystals sublimed on heating and were prob-
ably valin, although it will be seen that neither of these
bodies has been positively identified.
Fraction II. — This on being saponified by boiling with
five times its weight of water for five hours under a reflux
condenser and concentrated, yielded a crystalline mass
that contained no prolin, since no part of it was soluble in
hot absolute alcohol. The crystalline mass. was dissolved
in 700 c.c. of water and boiled for an hour with an excess of
freshly prepared copper oxide. On concentration of the
blue filtrate, a difficultly soluble salt was obtained. This
was crystallized and the percentage of copper in it found
to be 20.35, thus identifying it as the double salt of copper
with leucin and valin.
BACTERIAL CELLULAR SUBSTANCE 83
Calculated per cent, of Cu in Cu-leucin . . . . 19 . 66
Calculated per cent, of Cu in Cu-valin . . . .25.52
Calculated per cent, of Cu in Cu-leucin-valin . . 20.48
The more soluble copper salt was freed from its copper
with hydrogen sulphide. The filtrate furnished, on treating
it with 80 per cent, alcohol and allowing it to stand in the
ice-box, 0.5 gram of alanin.
Fraction III. — After saponification by boiling for five
hours under reflux condenser with five times its weight of
water, this fraction yielded 7.3 grams of solid. The absence
of prolin was demonstrated by the fact that hot absolute
alcohol dissolved nothing. The solution was boiled with
copper oxide which took up all the substance. The per-
centage of copper was found to be 20.38, thus showing the
compound to be the double salt of copper with leucin and
valin.
Fractions IV and V. — When these were mixed with
water a brownish oil separated. This was filtered off and
saponified by heating for several hours on the water-bath
with an excess of baryta. When the barium had been
removed there was obtained 0.5 gram of phenylalanin. It
was found difficult to purify this, but finally we did so, and
found the melting-point to be 262.5°. The filtrate from the
oily ester of phenylalanin was also saponified by heating
on the water-bath with an excess of baryta. The absence
of aspartic acid was shown by failure to obtain an insoluble
barium asparaginate. When the barium was removed a
small amount of glutamic acid was found in the filtrate.
The pasty mass left after the extraction of the esters
was neutralized with hydrochloric acid and prepared for a
second esterification, but owing to an accident this was
not completed. No ultimate analyses were made in this
or the subsequent hydrolyses, and we have relied for the
recognition of the amino-acids on (1) the boiling-point of
the esters, (2) crystalline form, (3) melting-point, and (4) the
percentage of copper in the copper compounds.
In this hydrolysis we have accounted for only a little
more than 10 per cent, of the nitrogen, distributed as
84 PROTEIN POISONS
follows: glutamic acid, 3 per cent.; glycocoll, 0.33 per cent.;
alanin, 1 per cent.; valin, 1.6 per cent.; leucin, 2 per cent.;
and phenylalanin, 0.2 per cent.
The cellular protein of the bacillus coli communis with
which we did this work contained 13.74 per cent, of moisture
and 7.2 per cent, of ash, or 8.38 per cent, of ash in the
moisture-free substance.
Hydrolysis of the Cellular Protein of the Bacillus Tuber-
culosis.— Five hundred grams of this substance was hydro-
lyzed after the manner already given. The air-dried humus
from this substance weighed 120 grams or 24 per cent, of
the cellular substance, or 27 per cent, of the moisture-free
substance. The humus was found to contain 14.34 per cent,
of moisture, 0.15 per cent, of ash, and 1.69 per cent, of
nitrogen.
Before hydrolyzing this substance samples were taken,
and the following determinations made.
Percentage of moisture 12.07
Percentage of ash . 7 . 08
Percentage of ash in moisture-free substance 8.05
Nitrogen was found distributed as follows:
Percentage of nitrogen in cell substance 9 . 27
Percentage of nitrogen in moisture-free cell substance . . . . 10.55
Percentage of nitrogen in ash-free cell substance 9.98
Percentage of nitrogen in ash- and moisture-free cell substance . 11 .47
Total amount of nitrogen in 500 grams of cell substance . 46 . 3500 grams
Total amount of nitrogen in the hydrolyzed fluid . . . 36.3075 grams
Total amount of nitrogen in the extract of humus . . 4 . 0530 grams
Total amount of nitrogen in the extracted humus ... 2 . 0388 grams
Total amount of nitrogen in 500 grams of cell substance . 46.3500 grams
Total amount of nitrogen in the hydrolytic product . . 42 . 3993 grams
Total amount of nitrogen lost in hydrolysis .... 3 . 9507 grams
Amount of nitrogen in hydrolyzed fluid 36 . 3075 grams
Amount of nitrogen in extract of humus 4 . 0530 grams
Amount of nitrogen in fluid to be esterified .... 40.3605 grams
It is thus seen that there is in the fluid to be esterified,
40.3605 grams of nitrogen, and it is supposed that this
exists in the form of mono- and diamino-acids.
BACTERIAL CELLULAR SUBSTANCE 85
Glutamic acid was separated as the hydrochloride.
Chlorides of ammonium and sodium were present in large
amount, but were easily separated on account of their
greater solubility in water. We obtained 1 gram of glutamic
acid, equivalent to 0.2 per cent, of the cellular substance
or 0.23 per cent, of the moisture-free substance.
The filtrate from the glutamic acid on being esterified, the
esters extracted, dried, and distilled, yielded the following:
Temperature Temperature Weight of
of bath.
of vapor.
distillate.
Fraction I
50° to 85°
35° to 60°
14 grams
Fraction II .
85° to 100°
60° to 85°
33 grams
Fraction III
100° to 140°
85° to 105°
34 grams
Fraction IV .
140° to 180°
105° to 130°
12 grams
Fraction V .
180° to 210°
130° to 160°
12 grams
The pasty mass left after extraction of the esters was
acidified with hydrochloric acid, evaporated, the salt
filtered out from time to time, and when brought to a thick
syrup it was diluted with an equal volume of absolute
alcohol and the esterification repeated. However, the yield
from this esterification was exceedingly small, not more
than a few drops for each fraction.
From the residue left after distillation, we obtained 2
grams of leucinimide, equivalent to 0.4 per cent, of the
cellular protein or 0.45 per cent, of the moisture-free sub-
stance.
Fraction I. — This was saponified by being boiled with
five times its weight of water for five hours under a reflux
condenser. It was evaporated to dryness and the white
crystalline mass was dissolved in 25 c.c. of hot water,
treated with an equal volume of hot absolute alcohol and
left in the ice-box for two days when a white crystalline
mass separated. This was purified and yielded 7 grams of
alanin, equivalent to 1.4 per cent, of the cellular substance
or 1.57 per cent, of the moisture-free substance.
Fraction II. — From this there was obtained by the method
already described a copper-leucin-valin compound repre-
86 PROTEIN POISONS
senting 3.7 grams of leucin and 12.4 grams of valin. Prolin
was not present.
Fraction III. — From this there was secured by the
formation of the copper salt 5.4 grams of leucin and 10.6
grams of valin.
Fractions II and III gave a combined yield of 9.1 grams of
leucin and 23 grams of valin, equivalent to 1.82 per cent,
of leucin and 4.6 per cent, of valin in the cellular substance
or 2.04 per cent, and 5.17 per cent, respectively in the
moisture-free substance.
Fractions IV and V.— Each was shaken with three times
its weight of cold water and filtered through a damp paper,
leaving a brown oil. The oil, after being washed twice
with cold water, was saponified with excess of baryta.
From this there was obtained 2.5 grams of crude phenyl-
alanin. This was purified and the melting-point found to
be 263°. The yield of phenylalanin amounted to 0.5 per
cent, of the cellular substance or 0.56 per cent, of the
moisture-free substance. From the filtrate from the oil a
few crystals which were probably glutamic acid, were
obtained, but the amount was too small for identification.
From 500 grams of the cellular protein of the bacillus
tuberculosis the following substances in the amounts and
percentages given were obtained.
Per cent, of Per cent, of dry
Substance. Amount. cell substance, cell substance.
Humus . . . . 120.0 grams 24.00 27.00
Glutamic acid . . 1.0 gram 0.20 0.225
Alanin . . . . 7.0 grams 1 . 40 1 . 57
Leucin . . . . 9.1 grams 1.82 2.04
Valin .... 23.0 grams 4.60 5.17
Phenylalanin . . 2.5 grams 0.50 0.56
Leucinimide . . 2.0 grams 0.40 0.45
It will be seen that of the total of 40.3605 grams of nitrogen
in the fluid esterified there has been recovered in the form
of mono-ammo-acids only 5.2 grams of nitrogen or 12.88
per cent. However, under the best conditions one cannot
hope to obtain more than a part of the mono-amino-acids
present and the diamino-acids probably take up a consider-
able part of the nitrogen.
BACTERIAL CELLULAR SUBSTANCE 87
Hydrolysis of the Non-poisonous Portion of the Cellular
Protein of the Bacillus Tuberculosis. — Five hundred grams
of this non-poisonous bacterial split product, known in this
laboratory as "residue" or haptophor, was hydrolyzed.
The humus was found to constitute 29 per cent, of the air-
dried, or 40 per cent, of the moisture-free substance, the
percentage of moisture being 27.5.
The distribution of nitrogen was studied with the fol-
lowing results:
Percentage of nitrogen in residue 4 . 59
Percentage of nitrogen in moisture-free residue 6 . 34
Percentage of nitrogen in ash-free residue 5 . 54
Percentage of nitrogen in ash- and moisture-free residue . . . . 8 . 29
Amount of nitrogen in the hydrolyzed fluid .... 18 . 980 grams
Amount of nitrogen in the extract of humus .... 1 . 888 grams
Amount of nitrogen in the extracted humus .... 2 . 030 grams
Total 22.898 grams
Total amount of nitrogen in the 500 grams of residue . 22.990 grams
Total amount of nitrogen in the products of hydrolysis . 22 . 898 grams
Loss during hydrolysis 0 . 092 gram
Amount of nitrogen in the hydrolyzed fluid .... 18 . 980 grams
Amount of nitrogen in the extract of humus .... 1 . 888 grams
Amount of nitrogen in the fluid to be esterified ... 20 . 868 grams
No glutamic acid could be obtained from the residue.
The result of the fractional distillation is shown as follows :
Temperature of bath. Yield.
Fraction I 40° to 60° 21. 5 grams
Fraction II 60° to 80° 5.0 grams
Fraction III 80° to 100° 8.0 grams
Fraction IV 100° to 130° 9.0 grams
Fraction V 130° to 170° 10.0 grams
By the hydrolysis of 500 grams of tubercle residue the
following substances were obtained in the amounts and
percentages given:
Per cent, of Per cent, of
Substance. Amount. residue. dry residue.
Humus .... 145. 00 grams 29.00 40.00
Alanin .... 4.00 grams 0.80 1.10
Leucin . . . . 6 . 40 grams 1 . 28 1 . 76
Valin .... 2. 30 grams 0.46 0.63
Phenylalanin . . 1.70 grams 0.34 0.46
88 PROTEIN POISONS
It has been possible in this hydrolysis to recover only
8.19 per cent, of the nitrogen in the fluid esterified. Doubt-
less we should have obtained a greater percentage had
we been able to secure a higher vacuum. Fischer does his
work with a vacuum of 1 to 2 mm. of mercury, while the
best wre could obtain with the facilities at our command
varied from 20 to 30 mm.
While the results of this work are not so satisfactory as
one might wish it does indicate that the proteins of the
two bacilli studied are different in their chemical compo-
sition. This is shown by the distribution of the amino-
acids as is indicated by the following figures:
Colon. Tuberculosis.
Glutamic acid . . . 3.00 per cent. 0.20 per cent.
Glycocoll 0.33 per cent. 0.00 per cent.
Alanin 1.00 per cent. 1.40 per cent.
Valin 1 . 60 per cent. 4 . 60 per cent.
Leucin 2 . 00 per cent. 1 . 82 per cent.
Phenylalanin . . . . 0.20 per cent. 0.50 per cent.
Wheeler has reported as follows upon the mono-amino
acids of the toxophor group:
The poisons selected for this work were those from
tuberculosis, typhoid, and colon germ substances, an,d for
comparison that from egg albumin. For the tuberculosis,
the poison from 900 grams germ substance was used — 296
grams; for the typhoid, 100 grams of the poison; for the
colon, that from 300 grams of germ substance, estimated
as 61.5 grams; while for the albumin, that from 200 grams
of the protein was employed, yielding, it was estimated,
93 grams of crude poison. These poisons were hydrolyzed
by boiling under a reflux condenser for fourteen hours
with concentrated hydrochloric acid. The nitrogen of
each extract was determined by the Kjeldahl method,
using an aliquot part of each as a sample, and giving the
following results :
NITROGEN OF ACID EXTRACT OF POISONS
Source of poison Per cent, of N.
Tuberculosis 9.895
Typhoid 10.380
Colon . . . 10.185
Egg albumen 11.477
BACTERIAL CELLULAR SUBSTANCE
89
From this point Fischer's ester method for obtaining the
individual mono-amino-acids was carried out.
This method is so well known that it is not necessary to
outline it here other than to say that after the amino-acids
have been produced by cleavage of the protein with concen-
trated acid, the hydrochloride of their ethyl esters is formed,
and later the free esters are separated by distillation in the
highest possible vacuum. These are then saponified and
the amino-acids crystallized and purified.
The efficiency of this method depending in large measure
upon the vacuum secured, the yields here presented for
cleavage of the poisons might have been materially increased
with a better vacuum, the highest one possible with the
apparatus at hand varying from 20 to 30 mm. The following
table shows the results of the distillations of the free esters,
both the bath and vapor temperatures being given, the
amount of the distillates, and the yield of crude crystals
after saponification.
DISTILLATION OF THE ESTERS OF THE MONO-AMINO-ACIDS FROM PROTEIN
POISONS
Tuberculosis Poison
Temperature
Temperature
Amount of Weight of
Fraction.
of oil bath.
of vapor.
distillate. crude crystals.
1
40° to 60°
20° to 40°
5 . 0 c c 2.0 grams
2
60° to 80°
40° to 60°
5.0 c c
2.0 grams
3
80° to 100°
60° to 80°
25 . 0 c c
16.0 grams
4
100° to 130°
80° to 100°
25 . 0 c c
16.0 grams
5
130° to 160°
25 . 0 c c
7.0 grams
Typhoid Poison
Fraction.
Temperature
of bath.
Temperature
of vapor.
Amount of
distillate.
Weight of
crude crystals.
1
25° to 60°
10° to 20°
12.0 c.c.
0.3158 gram
2
60° to 80°
20°
8.0 c.c.
1 . 2558 grams
3
80° to 110°
20°
2.0 c.c.
4
110° to 130°
95° to 108°
7.0 c.c.
6 . 0000 grams
5
130° to 145°
108° to 110°
4.0 c.c.
6
145° to 200°
138° to 185°
4.0 c.c.
No distillate passed over between 20'
between 110° and 138°.
and 95°, inside temperature, or
Temperature
Temperature
Amount of
Weight of
of bath.
of vapor.
distillate.
crude crystals.
40° to 60°
28° to 41°
1.5 c.c.
0.07 gram
60° to 80°
41° to 56°
1.0 c.c.
0.06 gram
80° to 104°
56° to 84°
4.0 c.c.
2.00 grams
104° to 120°
84° to 88°
2.0 c.c.
0.63 gram
120° to 160°
88° to 139
5.5 c.c.
90 PROTEIN POISONS
Colon Poison
Temperature 1
Fraction.
1
2
3
4
5
The yield from the colon poison was exceedingly small, due to the fact
that at one stage of the process part of the solution was lost.
Albumin Poison.
Temperature Amount of Weight of
Fraction. of bath. distillate. crude crystals.
1 40° to 60° 5 c.c. 0.2638 gram
2 60° to 80° 3 c.c. 0.3338 gram
3 80° to 100° 8 c.c. 4. 0000 grams
4 100° to 130° 10 c.c. 6.0000 grams
5 130° to 160° 7 c.c. 7. 0000 grams
After repeated recrystallizations these crude products
were obtained in a state of chemical purity. From the
tuberculosis poison, fractions 1 and 2 yielded needle-shaped
crystals, soluble in water and alcohol, sweet to the taste,
and containing 15.773 per cent, of nitrogen, the average
of four determinations by the Kjeldahl method. Alanin,
C3H7NO2, has all these properties and contains 15.73 per
cent, of nitrogen, thus identifying the crystals as alanin.
Fisher, Fraenkel, and others do not give a melting point for
d-alanin, saying that it is not sharp, due to the presence of
a mixture of the optically active and racemic forms. The
melting-point of the crystals from the tubercle poison
varied from 268° to 280° corrected, showing no constant
temperature. In fractions 3 and 4 the crystals were beau-
tiful, shiny, satiny plates, sweet to the taste, soluble in
water, but almost insoluble in alcohol. These sublimed
readily, melted with decomposition, and contained 11.976
per cent, of nitrogen (average of eight determinations).
These properties and the nitrogen correspond with valin,
a-aminoisovalerianic acid, C5HiiNO2, which contains 11.965
per cent, nitrogen. Frankel gives the melting-point of
BACTERIAL CELLULAR SUBSTANCE 91
valin as 298°, corrected, when heated in a closed tube,
decomposition taking place at the same time. The valin
from the tubercle poison melted as high as 296.28°, cor-
rected, but after continued recrystallizations the melting-
point dropped as low as 285°, and was never reliable.
Whether this was due to a partial racemization on repeated
heating is not known. Heated in a closed tube the melting-
point of the final product was 275.8° to 278.2°. As is well
known, valin closely resembles leucin in its properties, so
that it is very difficult to demonstrate the existence of one in
the presence of the other. On page 78, by another method,
the presence of leucin in the poison has been shown, but
by the Fischer method of ester distillation valin seems to be
the one obtained. The presence of leucin was further
demonstrated by the fact that from the final residue left
after the esters had been distilled, crystals of its decomposi-
tion product, leucinimide were obtained. This crystallized
from dilute alcohol in the form of needles and melted at
295.4°. Cohn gives the melting-point of leucinimide as 295°,
Frankel as 262°. From fraction 5 of the tubercle poison a
qualitative test only was obtained for phenylalanin, the
quantity obtained being too small for complete purification.
After evaporation of the ethereal solution of the thick,
oily ester, according to the method, the ester is saponified
by twice evaporating with hydrochloric acid. It is then
evaporated with ammonia, dissolved in a small amount of
water, and poured into a large volume of absolute alcohol,
which precipitates the phenylalanin. From this precipitate
the qualitative test was obtained, according to Frankel,
by dissolving in dilute sulphuric acid and adding an excess
of potassium dichromate, producing the characteristic
odor of phenylacetaldehyde and showing thus the presence
of phenylalanin. From fraction 5, after removal of the
phenylalanin, upon saponification with barium hydrate,
there was obtained, after the barium had been removed,
rhombic hemihedral crystals which had a distinctly sour
taste. These, after purification, showed 9.54 per cent, of
nitrogen, the average of two Kjeldahl determinations,
92 PROTEIN POISONS
identifying them as glutamic acid, C5H9NO4, which has 9.52
per cent, of nitrogen. Frankel gives the melting-point of
glutamic acid as 202° to 202.5°, or quickly heated, 213°, with
decomposition. The product above obtained melted in an
open tube at 242° to 245°, in a closed tube at 236° to 238°.
From fraction 1 of the typhoid poison was obtained alanin,
with characteristic properties, as described above. These
crystals showed 15.633 per cent, of nitrogen and melted
at 267° to 271°. Fractions 2, 3, 4, and 5 contained only
valin, which showed 11.932 per cent, of nitrogen, the average
of four determinations. This melted at 287° to 290.6° in
an open tube, 278° to 280° in a closed one. From fraction 6
the qualitative test for phenylalanin was obtained as from
the tubercle poison.
Owing to the small yield of esters and crystals, fractions
1 and 2 from the colon poison could only be determined
qualitatively. Both fractions, however, showed needle-
shaped crystals and a sweet taste, which in conjunction writh
the temperature at which their esters distilled indicated
alanin. Fractions 3 and 4 gave characteristic valin crystals
containing 11.942 per cent, of nitrogen and melting at
283.4° to 285° in an open tube, or 274° to 277° in a closed
tube. Phenylalanin was obtained qualitatively from this
as from the two preceding poisons. After its extraction
from fraction 5, and after saponification with barium
hydrate and its removal, crystals in the form of rhombic
plates and prisms, insoluble in alcohol, were obtained.
These corresponded with those of aspartic acid, and as the
quantity was not sufficient for purification by recrystalli-
zation, the copper salt was formed with copper acetate.
This was obtained in the form of needles, very difficultly
soluble in cold water, difficultly in hot, which again
corresponded with the properties of aspartic acid.
When the crystals from the fractions obtained from
albumin poison were examined the result was not different.
Fractions 1 and 2 produced characteristic alanin crystals
with 11.75 per cent, of nitrogen, the average of four deter-
minations. The melting-point was 277° to 279.6°. Valin,
BACTERIAL CELLULAR SUBSTANCE 93
with form and properties as already given, was obtained
from both fraction 3 and 4, containing 11.935 per cent, of
nitrogen, and showing a melting-point of 282° to 286.4° in
an open tube, and of 279° to 283 in a closed tube. Like-
wise from fraction 5 the heavy oil of phenylalanin ethyl
ester was obtained, and from this as in the other cases the
qualitative test for phenylalanin by the production of
phenylacetaldehyde. The remaining portion of fraction 5
yielded the same rhombic plates and prisms as described
under the colon fractions, and which are like those of aspartic
acid properly obtained at this point if present. The copper
salt was again formed, the same needles, very difficultly
soluble in cold water, difficultly in hot, being obtained.
The amount of crystals was too small for further identi-
fication.
From this it will be seen that mono-amino-acids are
obtained from the protein poisons after hydrolysis with
strong acid. It is not claimed that these are the only
mono-amino-acids present, or that all of these have been
sufficiently identified, but in consideration of the fact that
those discussed were found in the proper fraction according
to Fischer's separation and according to the boiling-points
of their esters, that the crystalline form and qualitative
properties corresponded, and that, when it could be deter-
mined, the percentage of nitrogen was close to the theo-
retical, it seems fair to conclude that the following tabulation
is not far from correct:
MONO-AMINO-ACIDS OF THE PROTEIN POISONS
Tuberculosis
poison.
Alanin
Valin
Phenylalanin
. Typhoid
poison,
alanin
valin
phenylalanin
Colon
poison,
alanin
valin
phenylalanin
Albumin
poison,
alanin
valin
phenylalanin
Leucinimide
Aspartic acid aspartic acid
This is sufficient to establish the point for the proof of
which the method was employed, that is, the protein nature
94 PROTEIN POISONS
of the poisonous group of the protein molecule. Attention
is called also to the comparative simplicity of the group and
to the great similarity of acids obtained from the different
poisons. This accords well with the great similarity and
non-specificity of their physiological action.
It is interesting that the final residue left after distilla-
tion of the esters gives still a very intense Millon reaction,
which cannot be ascribed to the presence of tyrosin.
It will be clearly understood that this work does not
show that the active agent or agents in the "crude soluble
poison" is or are protein in nature.
CHAPTER V
THE CLEAVAGE OF PROTEINS WITH DILUTE
ALKALI IN SOLUTION IN ABSOLUTE
ALCOHOL
THE researches detailed in the preceding pages seem to
establish the following propositions:
1. The cellular substances of bacteria consist largely of
proteins that yield split products identical with those
obtained by the hydrolysis of vegetable and animal proteins.
It has been shown that the bacterial cellular substances,
when broken up with mineral acids or alkalies, furnish
ammonia, mono-amino and diamino nitrogen, one or more
carbohydrate groups, and humin substances. It seemed
therefore logical to conclude that the bacterial cell consists
largely of proteins.
2. The proteins of the bacterial cell contain at least
one group which when injected intra-abdominally, subcu-
taneously, or intravenously in anmials, has a markedly
poisonous effect.
3. This poisonous group may be detached from the cell
protein by hydrolysis with either dilute acids or alkalies.
4. The dilute alkali furnishes the better means of ex-
tracting the poisonous group.
5. When the bacterial protein is broken up with alkali
in dilute aqueous solution, at least two groups are split
off and pass into solution. These are the carbohydrate
and the poisonous groups. Both are soluble in water and
in dilute alcohol, and their separation, when the cell protein
is disrupted by alkali in aqueous solution, is difficult and
unsatisfactory.
6. Since the carbohydrate group is insoluble in absolute
alcohol, while the poisonous group is more readily soluble
96 PROTEIN POISONS
in this menstruum than in water, it was decided to attempt
to disrupt the cell protein with a solution of alkali in absolute
alcohol. Another idea also acted as a determining factor
in attempting this method of hydrolysis, and in fact it
was at that time the dominating factor. The effect of the
poisonous group on animals so closely resembles that of
neurin that it was thought that the two might be identical,
or at least that the poisonous body might contain neurin.
Knowing that neurin can be heated without decomposition
in alkaline alcohol was, therefore, a reason for trying this
method.
7. Previous experiments had demonstrated the advantage
of extracting the cell substance thoroughly with alcohol and
ether before submitting it to hydrolysis. This frees the
material from fat, wax, and other substances soluble in
alcohol or ether, and since it had been shown that these
are no part of the cell protein it is beneficial to get rid of
them in toto before hydrolysis is attempted.
The following preliminary trials were made by Vaughan
and Wheeler (in the fall of 1903) in order to compare
hydrolysis with aqueous and alcoholic solutions of alkali.
Two samples, of 10 grams each, of the cellular substance
of the colon bacillus were taken. This material had pre-
viously been thoroughly extracted with alcohol and ether.
One sample was mixed with 250 c.c. of a 1 per cent, aqueous
solution of sodium hydroxide and the other with the same
volume of an absolute alcohol solution of the same substance
in the same strength. These mixtures were heated in
flasks, fitted with reflux condensers, for one hour on the
water-bath. Ten cubic centimeters of the clear filtrate
from each was evaporated, the aqueous solution to 5 c.c.
and the alcoholic to dryness, and then taken up in 5 c.c.
of water. Each was carefully neutralized with dilute
hydrochloric acid and injected into the abdominal cavity
of a guinea-pig. Both animals developed in a characteristic
manner the first and second stages of poisoning with the
split product, but neither died. This experiment showed
that the poison was present in both extracts, and, so far
THE CLEAVAGE OF PROTEINS WITH ALKALI 97
as we could judge by the development and intensity of
the symptoms, in similar amounts. That the poison could
be extracted by alkaline alcohol was proved. ' However,
the yield was not satisfactory, and a second test was made,
and in this the strength of the alkali was doubled. These
were treated as before, and the pigs that received the
injections developed the characteristic symptoms and
died. The one that had the alcoholic extract died within
six, and the other within eight minutes. This confirmed
the hope that the alcoholic alkali was quite as efficient as
the aqueous in the extraction of the poisonous group.
While the aqueous extract contained a large amount of
the carbohydrate group, it was found that the alcoholic
extract, after evaporation to dryness and solution in water,
gave the biuret, Millon, and xanthoproteic tests, but failed
wholly to give the Molisch reaction. The carbohydrate
group had been split off in both samples, but being insoluble
in absolute alcohol, it remained with the insoluble portion
of the cellular substance.
The above and many other experiments have demon-
strated that the best method, so far devised, for extracting
the poisonous group from the cell protein, or, as subsequent
work has shown, from any protein, is by means of a 2 per
cent, solution of sodium' hydroxide in absolute alcohol.
If satisfactory results are obtained, the alcohol used in the
extraction must be absolute. If it is not, more or less of
the carbohydrate will be mixed with the poison; a sticky
mass will be obtained, and the patience of the experimenter
will be taxed severely. Previous thorough extraction of
the protein with alcohol and ether for the removal of fats,
waxes, and other substances soluble in these agents, is
also essential to satisfactory work.
The method for preparing the bacterial cellular substance
has been given, but it may be well to give here some details
for the preparation of egg-white before splitting it up into
poisonous and non-poisonous proteins.
Fresh eggs (we have usually taken twenty dozen at a
time) are broken and the whites dropped into a beaker or
7
98 PROTEIN POISONS
precipitating jar, then poured with constant stirring into
four volumes of 95 per cent, alcohol. This stands with
frequent stirring for two days, then the alcohol is decanted,
and replaced with the same volume of absolute alcohol.
This is allowed to stand for from one to two days, when
the coagulated albumin is collected on a filter, allowed to
drain, then placed in large Soxhlets and extracted with
ether for from one to two days. It is then ground in porce-
lain mortars and passed through fine meshed sieves. This
gives a beautifully white powder which may be kept in
bottles in stock from which portions are taken for the
purpose of hydrolyzing it.
Twenty dozen eggs yield about 735 grams of this powder,
a little more than 3 grams per egg.
A weighed portion of the protein, prepared as above, is
placed in a flask, covered with from fifteen to twenty-five
times its weight of absolute alcohol in which 2 per cent, of
sodium hydroxide has been dissolved. The flask, fitted
with a reflux condenser, is heated on the water-bath for
one hour, when it is allowed to cool and the insoluble portion
collected on a filter. After thorough draining the insoluble
part is returned to the flask and the extraction repeated.
It has been found that three extractions are necessary in
order to split off all the poisonous group. The temperature
of these extractions is 78°, the temperature of boiling
absolute alcohol. By this method the protein is split into
two portions, one of which is soluble in absolute alcohol
and is poisonous, while the other is insoluble in absolute
alcohol and is not poisonous.
A large number of protein bodies, bacterial, vegetable,
and animal, have been split up in this way and no true
protein has failed to yield a poisonous portion. Among
the proteins with which we have worked the following may
be mentioned: egg-white, casein, serum albumin, edestin,
zein, Witte's peptone, Macquaire's peptone, de Chapoteaut's
peptone, the tissue of cancers, and the cellular substance
of bacillus coli communis, b. typhosus, b. anthracis, b.
tuberculosis, b. Moelleri (timothy), sarcina lutea, b. ruber
THE CLEAVAGE OF PROTEINS WITH ALKALI 99
of Kiel, b. proteus, b. subtilis, b. megaterium, b. pyo-
cyaneus, b. pneumonise, and b. diphtheriae. Gelatin con-
tains no poison, but gelatin is an albuminoid and gives the
Millon test imperfectly, if at all. Nicolle and Abt1 found
that Defresne's peptone does not yield a poison when
treated by our method, and we have confirmed this finding.
It would be interesting to know whether this peptone is
made from gelatin or from a true protein. The probabilities
are that in peptic digestion a point is reached when the
poisonous group in proteins is disrupted. In fact, as has
been stated (page 42), we have shown that the poison in
the cellular substance of the colon bacillus is slowly digested
and destroyed by digestion with pepsin-hydrochloric acid.
Therefore, it is not strange that certain peptones fail to
yield a poisonous body when disrupted with dilute alkali
in absolute alcohol. ' Witte's peptone, so-called, as is well
known, is not a peptone, but an albumose.
This poison, like the whole protein of which it is a part,
is formed synthetically by the living cell. In case of the
colon poison we demonstrated this by growing the bacillus
in Fraenkel's modification of Uschinsky's medium, which
has the following composition:
Water 10,000 parts
Sodium chloride 50 parts
Asparagin 34 parts
Ammonium lactate 63 parts
Di-sodium hydrogen phosphate 20 parts
After a week's development the contents of these flasks
were poured into from two to three volumes of 95 per cent,
alcohol. The precipitate was filtered out and put into
absolute alcohol; next it was extracted in Soxhlets with
ether, dried, and powdered. This powdered cellular sub-
stance, when split up with 2 per cent, sodium hydroxide
in absolute alcohol, furnished the poison, the action of
which was demonstrated on guinea-pigs. Moreover, the
poison obtained in this way gave all the protein reactions
1 Annales de 1'Institut Pasteur, February, 1908.
100 PROTEIN POISONS
hereafter described as being obtained from the poison from
agar-grown cultures. This demonstrates that the poison is.
an integral part of the cellular substance, and it is evident
that the bacterial cell must synthetically produce this
protein body during its growth from the chemical con-
stituents of the medium.
When the protein is split up by dilute alkali in absolute
alcohol according to the method described, the poison is
in solution in the alkaline alcohol. The preparation is
filtered and the filtrate neutralized with hydrochloric acid,
avoiding an excess of acid. This throws down the greater
part of both base and acid as sodium chloride, which is
removed by filtration. In this way a solution of the poison
in absolute alcohol is obtained. This is evaported in
vacuo at 40°, redissolved in absolute alcohol to remove
traces of sodium chloride, and again evaporated in vacuo
at 40° or less. Evaporation may be done in an open dish,
but the toxicity of the substances is somewhat decreased
when this is done. The poisonous part of the protein
molecule when obtained in this way and powdered, when
there is no water present, forms a dark brown scale which
pulverizes into a lighter brown powder.
It should be clearly understood that we regard this
method of extracting the poisonous group from the protein
molecule as by no means ideal. We know that it is crude
and that much of the poison is destroyed in the process.
In disrupting a protein by our method with dilute alkali
in absolute alcohol, ammonia is given off and the odor of
this gas is apparent even at the end of the third extraction.
An effort was made to discover how much nitrogen was
converted into ammonia in the process. A device was
arranged for conducting the ammonia into standard acid,
and four 10-gram samples of Witte's peptone were extracted
with 2 per cent, sodium hydrate in absolute alcohol, one
for three hours in a current of air, the others in a current
of hydrogen for two and one-half, eight and one-half, and
nineteen and one-half hours respectively. At the end of
each operation the excess of acid was titrated with deci-
THE CLEAVAGE OF PROTEINS WITH ALKALI 101
normal sodium hydrate, and the percentage of nitrogen
calculated. The relative toxicity of the split products was
determined. In every case ammonia was still being pro-
duced when the process was interrupted. Again, a 10-gram
sample of the poison from egg ablumen was boiled for
fifty-four and one-half hours with 2 per cent, alcoholic
alkali to ascertain if ammonia could be split from the
poison itself. The results of this work are shown in the
following table:
AMMONIA PRODUCED BY CLEAVAGE OF PKOTEIN WITH DILUTE ALKALI IN
ABSOLUTE ALCOHOL.
Per cent, of
Time in Atmos- N given Rate per
Sample. hours.
Witte peptone 3.0
Witte peptone 2,5
Witte peptone
Witte peptone
Poison
8.5
19.5
54.5
phere.
air
H
H
H
II
off.
hour.
Toxicity.
0.4305
0.1435
Diminished.
0.3956
0.1582
Greater than
that in air.
0.7383
0.0868
Diminished.
1.0517
0.0539
Diminished.
1.4800
0.0270
Diminished
by half.
The albumin poison as ordinarily obtained contains
13.74 per cent, of nitrogen. By the fifty-four and one-half
hours' heating with alcoholic alkali, 10.77 per cent, of its
nitrogen was converted into ammonia. After this treatment
the poison still gave a good Millon test, but no longer the
biuret.
It is probable that by continued heating in the same
manner quite all of the nitrogen could be separated, though
it is noticeable that the rate was greatly diminished as the
time lengthened. The decrease in toxicity with the evolu-
tion of ammonia suggests that this group is essential to the
toxicity of the poison. This seems to be highly probable.
Properties of the Crude Soluble Poison. The poison split
off from the protein molecule by the method above given
is designated as "the crude soluble poison;" "crude"
because it is undoubtedly a mixture of chemical bodies,
and "soluble" in contradistinction to the bacterial cellular
substance, from which it was first prepared, and which is
poisonous, but not soluble.
102 PROTEIN POISONS
The brownish toxic powder, varying in shade of color
somewhat with the protein from which it has been obtained,
has a peculiar odor. It is highly hydroscopic, and the
poisonous .portion is freely soluble in water. The solubility
of the whole powder, however, varies with the protein
from which it is obtained, and possibly with the length
of time that it has been exposed to the alkali in the alcohol.
Any portion insoluble in water should be removed by
filtration, and in some instances we have found filtration
through porcelain necessary. Generally the powder dis-
solves in water with a slight opalescence easily removed by
filtration through paper. In all cases we have found the
portion insoluble in water free from toxic effect. Aqueous
solutions of the poison are decidedly acid to litmus, the
acidity being due to some organic body and probably
not to the poison itself. On neutralization with sodium
bicarbonate a brownish, non-toxic precipitate is formed.
Prolonged contact with alkali, as we shall see later, lessens
the activity of the poison, and even neutralization has
some effect, which is more marked the longer the prepara-
tion stands. We are inclined to attribute this to the forma-
tion of a salt with the acid poison and the alkali. The
poison is freely soluble in alcohol, more readily than in
water. Alcoholic solutions on long standing deposit small
brownish sediments which we have always found to be
inert. When an alcoholic solution is evaporated, there is
a part of the residue that is insoluble in absolute alcohol.
These portions also are devoid of toxic effect. Alcoholic
solutions have been kept for five years without recog-
nizable loss in toxicity, and even aqueous solutions decom-
pose very slowly. The poison is soluble in methyl as well
as in ethyl alcohol. It is insoluble in ether, chloroform,
and petroleum ether. Each of these removes a small
amount of fatty substance, which is non-toxic, but they
do not dissolve an appreciable quantity of the poison.
From its alcoholic solution the poison is precipitated by
ether, but contact with ether decreases its toxicity to such
an extent that this method is not applicable in attempts
at purification.
THE CLEAVAGE OF PROTEINS WITH ALKALI 103
The "crude soluble poison" is soluble in strong mineral
acids, and such solutions remain clear on being boiled and
on dilution with water. However, a few drops of mineral
acid added to an aqueous solution cause a precipitate,
which seems to indicate that the acidity of the aqueous
solution is caused by the presence of some organic acid.
The poison diffuses slowly through collodion sacs both
within the animal body and when suspended in distilled
water. The following experiments bear on this point:
Two hundred milligrams of the crude soluble poison from
the cellular substance of the typhoid bacillus dissolved in
20 c.c. of water was placed in each of two collodion sacs
which were then suspended in distilled water. At the
end of twenty-four hours, the Millon reaction was given
by the dialysate. This was replaced every twenty-four
hours by fresh distilled water, and the dialysis continued
for ninety-six hours. At the end of this time the combined
dialysates were concentrated to dryness, the residues dis-
solved in absolute alcohol, filtered, and again evaporated.
The brown, sticky residue, thus obtained, dissolved in
water, was acid in reaction, had the characteristic odor,
and when injected into a guinea-pig, killed in twenty minutes
with typical symptoms, thus showing that the poison does
diffuse through a collodion sac. So slowly, however, does
it diffuse that at the end of ninety-six hours it was not
wholly removed from the sac. In another experiment
one gram of the same poison in 8 c.c. of water was put into
a collodion sac which was introduced into the abdominal
cavity of a medium-sized rabbit. After twelve days, the
animal not being visibly affected, the sac was removed
and found to contain 6 c.c. of a clear fluid which looked
more like blood serum than anything else. Five cubic
centimeters of this injected into the abdomen of a guinea-
pig had no effect. We conclude from this that the poison
had diffused from the sac, but so slowly that it was disposed
of by the animal's body without recognizable discomfort.
Notwithstanding the ready solubility of the crude soluble
poison in absolute alcohol, we must regard it as either
104 PROTEIN POISONS
being a protein itself or as being mixed with one or more
proteins. Its aqueous solutions give all the protein color
reactions with the important exception of that of Molisch.
It is worthy of note that the part that separates from
alcoholic solution on long standing is inert and does not
give the protein reactions, while the solution does not
decrease in toxicity. This indicates that the protein is
permanently soluble in absolute alcohol. The Millon
reaction shows most perfectly and persistently whenever
the poison is found. It is generally believed by physio-
logical chemists that this reaction is given by all benzene
derivatives in which one hydrogen atom has been replaced
by a hydroxyl group, and it is also generally supposed that
tyrosin is the only oxyphenyl compound in the protein
molecule, therefore this reaction is presumed to show the
presence of tyrosin. This is interesting in view of the
fact already stated that gelatin, which contains no tyrosin,
or but little, yields no poison. The fact that the poison
contains no carbohydrate, as shown by its failure to respond
to the Molisch test, an exceedingly delicate test, is, in our
opinion, strong evidence that the cleavage in the protein
molecule induced by dilute alkali in absolute alcohol at
the temperature of 78° follows along structural lines. If
the change were one of simple degradation without chemical
cleavage it wrould be difficult to explain the absolute failure
of the carbohydrate test in the crude soluble poison. It
seems quite evident from our work that in the process the
complex protein molecule is split into several groups, one
of which is the poison and another is a carbohydrate, the
former being freely soluble in absolute alcohol, while the
latter is insoluble in this reagent. It should be stated that
the crude, soluble poison not only fails to respond to the
Molisch test, but it also fails to reduce Fehling's solution
after prolonged boiling with dilute mineral acid.
The crude soluble poison gives the biuret test beauti-
fully, therefore we must say that the poison either is itself
a biuret body or is mixed with such a body. As is well
known, the biuret test is regarded as the landmark between
THE CLEAVAGE OF PROTEINS WITH ALKALI 105
proteins and their simpler non-protein disruption products,
and, so long as a disrupted protein continues to give the
biuret test it must still be classed among the proteins. It
will certainly be understood that the pure poison may
not be a protein, but until it is purified sufficiently to fail
to give the biuret test it must be regarded as a protein.
The poison responds nicely to the Adamkiewicz or gly-
oxylic acid test. Hopkins and Cole have shown quite
convincingly that this color test depends upon the presence
of tryptophan or indol-amino-propionic acid; therefore, while
we have made no direct search for tryptophan in our poison,
we assume its presence on account of the unequivocal
response to this test.
When the poison is boiled with concentrated hydrochloric
acid to which a drop of concentrated sulphuric acid has
been added, the powder passes into solution and a violet
color results, thus giving Liebermann's test. At one time
Hofmeister believed this to be a carbohydrate reaction in
which furfurol and the aromatic oxyphenyl radicals take
part, but Cole has shown that this, like the Adamkiewicz
test, also once regarded as a carbohydrate test, is due to
the tryptophan group. We are quite convinced that our
soluble poison contains no carbohydrate, and we regard
the fact that it does respond to the Liebermann test as a
strong confirmation of the error of Hofmeister's explanation
of this test, and in favor of the explanation given by Cole.
When heated with strong nitric acid the powdered poison
goes into solution, more or less yellow according to the
amount used, and this becomes orange on the addition of
ammonia, thus giving the xanthroproteic test and indicating
the existence of aromatic radicals.
The ordinary test for sulphur in proteins, that of heating
with excess of sodium hydrate in the presence of a small
amount of acetate of lead, is not given by the portion of
the protein split off "by alkali in absolute alcohol. If,
however, a portion of the substance in a test-tube is fused
with metallic sodium and the cooled mass treated with
water, a few drops of a freshly prepared solution of sodium
106 PROTEIN POISONS
nitroprussiate added to a part of the clear filtrate, a beauti-
ful violet color is produced, indicating the presence of sulphur.
Also, if the other part of the clear filtrate be treated with a
lead acetate solution, lead sulphide is precipitated. If the
solution be acidified before lead acetate is added a faint
but unmistakable odor of hydrogen sulphide is detected.
It is known that sulphur may exist in the protein molecule
in at least two forms, one part being readily split off with
dilute alkali as a sulphide, the other being obtained only
when the disruption of the protein molecule is carried much
farther. It is still a question whether or not both of these
sulphur groups come from cystin. Since the nitroprussiate
reaction is very delicate, no conclusion as to the amount
of sulphur can be drawn from this test, and although a
good precipitate of lead sulphide is formed, the amount of
sulphur in the poison is probably not large, since Leach
failed entirely to find sulphur in the ash of the colon bacillus,
though both the cellular substance and the non-poisonous
portion, as well as the poison, respond to the nitroprussiate
test for sulphur and also give the lead sulphide precipitate
in the clear acidified filtrate from the fused mass.
A solution of this toxic substance is not coagulated by
heat in acid, neutral, or alkaline solution, though, as already
stated, a few drops of a mineral acid added to an aqueous
solution causes the appearance of a considerable precipitate,
which is not soluble on heating or on the further addition
of acid. This precipitate is produced regardless of the
previous removal of the opalescence from the aqueous
solution.
Among the metallic salts, copper sulphate produces no
precipitate and ferric chloride only on heating. Silver
nitrate naturally precipitates any trace of chlorides present,
but after the addition of an excess of ammonia there still
remains a small precipitate. Potassium ferrocyanide gives
a precipitate, also potassium bismuth iodide in acid solution.
Lead acetate, mercuric chloride, and platinum chloride
all produce heavy precipitates. With lead acetate and
mercuric chloride, however, after removal of lead and
THE CLEAVAGE OF PROTEINS WITH ALKALI 107
mercury with hydrogen sulphide from their respective
precipitates and filtrates, the protein reactions are given
by the filtrates, and here also is found the poison in each
case. From 10 to 15 per cent, of the crude poison can be
precipitated by the use of platinum chloride in either water
or alcoholic solution. All attempts to crystallize this
precipitate failed, as only a small part of it is dissolved by
hot water, and the insoluble part is unaffected by any of
the ordinary solvents. The protein reactions are given
by the platinum precipitate, by both soluble and insoluble
parts, but not by the filtrate. The poison is found in the
insoluble part of the precipitate after removal of the plati-
num by hydrogen sulphide, its toxicity being markedly
increased. The other parts, after removal of the platinum,
are inert.
The most active products have been obtained by precipi-
tation from solution in absolute alcohol with alcoholic solu-
tions of the chlorides of platinum, mercury, and copper and
removal of the base from the precipitate with hydrogen
sulphide. By this method we have obtained a body which
kills guinea-pigs of from 200 to 300 grams' weight in doses
of 0.5 mg. given intravenously.
From a water solution of the poison, bodies giving protein
reactions may be salted out by the addition of ammonium
sulphate or sodium chloride to saturation, but in neither
case is the separation complete, the filtrates still responding
to the protein color tests after removal of the neutral
salts. In case of salting out with ammonium sulphate, the
solubility of both parts is thereby lessened and the toxicity
diminished, possibly on account of decreased solubility,
though both parts exhibit some poisonous action, and like-
wise both show the protein color tests.
Phosphotungstic, phosphomolybdic, and picric acids all
give abundant precipitates. Since these reagents are also
used in the precipitation of alkaloidal bodies, the precipitates
.with phosphomolybdic and phosphotungstic acids were
further examined, the possibility suggesting itself that the
toxic body might be alkaloidal in nature, and that the
108 PROTEIN POISONS
protein part might be entirely separate from the poison.
A sample was precipitated with phosphomolybdic acid
in acid solution, the precipitate removed, washed, and
dissolved in ammoniacal water. This solution was then
shaken with amyl alcohol, but the alcohol was not colored
and the residue obtained on concentration was so slight as
to be practically nothing. Another sample was precipitated
with phosphotungstic acid, the solution being acid in
reaction. The precipitate was allowed to settle, removed
by filtration, washed with acidulated water, decomposed
with a saturated solution of barium hydrate, and the
remaining insoluble part filtered out. So far as possible,
the barium was removed from the filtrate with carbon
dioxide, alternating with concentration, and further addi-
tion of carbon dioxide. The solution was then allowed to
concentrate to dryness, when the residue was dissolved in
absolute alcohol, leaving barium salts behind. On con-
centrating the slightly opalescent solution, more barium
salts came down during the process and were filtered out.
The dry residue was taken up in water and ammonium
carbonate used to precipitate the barium that still remained.
After removing the barium carbonate by evaporating on
the water-bath, both carbon dioxide and ammonia were
expelled, the solution again becoming acid. Dryness being
reached, absolute alcohol was once more used, leaving
undissolved a small amount of inorganic material. In this
way the final residue after evaporation of the alcohol was
practically freed from inorganic impurities. Sulphuric
acid no longer gave a barium precipitate in water solution.
The amount obtained by this method was very small and
an exceedingly small part of the original toxic powder.
Since the substance obtained in this way still gave good
Millon's, biuret and xanthoproteic reactions, it is fair to
say that it was not alkaloidal. The very small amount
obtained by this method given to a guinea-pig intra-
abdominally made the animal sick, but did not kill.
Either phosphotungstic acid does not precipitate the toxic
body or else the amount obtained was less than a fatal
dose.
Should the poison consist of an alkaloidal body existing
as a salt in the acid solution, the possibility of extracting
the base with ether or chloroform, after the solution had
been made alkaline with ammonia, is apparent. This
was tried with negative results. To a water solution of
colon poison, acid in reaction, ammonia was added, drop
by drop, to a slightly alkaline reaction, the mixture shaken
with ether, the ether separated and evaporated. The residue
remaining was non-toxic. The ammoniacal water solution
was next shaken with chloroform, the slightly colored
chloroform drawn off and evaporated at low temperature,
leaving a small amount of a dark, thick, semiliquid, which
was not poisonous either as it was or after faintly acidify-
ing with hydrochloric acid. The water solution remaining
being still poisonous, it is evident that the toxic part is not
an alkaloidal body capable of being extracted directly.
Potassium bismuth iodide in acid solution of the crude
soluble poison produces an abundant precipitate, apparently
more or less soluble in excess, and soluble in ammoniacal
water.
Kowalewsky has shown that uranyl acetate will com-
pletely remove from various albuminous fluids every trace
of protein giving a biuret reaction, while Jacoby and others
have used this reagent for the removal of proteins from
faintly alkaline solutions. Abel and Ford used it to remove
protein from an extract of poisonous fungi. In a slightly
alkaline solution of albumin poison, uranium acetate gave
an abundant precipitate, but not a complete separation,
as both precipitate and filtrate still gave the Millon and
biuret tests, and the filtrate, after removal of excess of
uranium with a solution of di-sodium hydrogen phosphate,
filtration, evaporation, solution in alcohol, and reevapora-
tion, was still poisonous. In acid solution, the precipitation
was complete, the filtrate no longer giving the protein
reactions.
Freshly prepared metaphosphoric acid also produced
an abundant precipitate, but not a complete separation,
the filtrate showing both Millon and biuret reactions.
110 PROTEIN POISONS
Likewise a heavy precipitate is produced by the use of a
saturated solution of picric acid, but the poison is not in the
precipitate, which gives only a very poor Millon test after
removal of the picric acid, and no biuret. Hofmeister has
given a method for introducing iodine into the molecule
of egg albumen. This was tried with the poison split from
egg albumen. The iodized compound no longer gave either
the Millon or biuret reactions, and while it affected animals
more or less, they did not die, and the symptoms were not
those induced by toxin poisoning. The iodine seemed to
have entered into chemical combination in the poison
molecule, and to have thus changed its characteristics.
The iodized body was freely soluble in absolute alcohol,
and in alkaline water, not in water alone, and was precipi-
tated by acid water from alcoholic solution, also on acidi-
fying an alkaline water solution. Though it no longer
responded to the Millon and biuret reactions, a good test
for nitrogen was obtained after fusing with metallic sodium.
An attempt was made to benzoylate the poison by the
Schotten-Baumann method, using albumin poison. Prac-
tically no precipitate was obtained. From the filtrate in a
part soluble in hot alcohol there were obtained shiny, glis-
tening plates or flat needles which matted together under
suction, and had much the appearance of some of the fatty
acids. These were insoluble in water or very difficultly
so, if at all, difficultly soluble in cold alcohol, readily in hot.
They gave no Millon test, no biuret, no Molisch, and con-
tained no nitrogen. After recrystallization from alcohol
they melted constantly at 62°. Palmitic acid melts at 62°
and boils at 339° to 350° (Mulliken). A Merck preparation
of palmitic acid melted at 62° and boiled at about 345° to
350°. Our crystals had not yet boiled at 360°, though
above 300° there was some decomposition. From the
remainder of the filtrate there was obtained from the part
soluble in cold alcohol a non-crystallizable body, giving
both Millon and biuret tests and containing 9.335 per cent,
of nitrogen, and from the part soluble only in water, likewise
a non-crystalline compound, with 9.66 per cent, nitrogen,
THE CLEAVAGE OF PROTEINS WITH ALKALI 111
and showing both Millon and biuret tests, but not seriously
affecting animals in usual doses.
The nitrogen in a number of the crude poisons has been
determined by Gidley in this laboratory as follows:
PERCENTAGE OF NITROGEN IN PROTEIN POISONS.
Per cent, of N.
Source of poison. in crude poison.
Colon bacillus 13.49
Typhoid bacillus 11.52
Tubercle bacillus 11.00
Pyocyaneus 10.50
Ruber of Kiel 10.495
Subtilis 8.12
Megaterium 8.595
Proteus vulgaris 10.17
Yellow sarcine 6.145
Egg albumen (Leach) 13 . 74
Serum albumin 10.48
Edestin 12.78
Zein 10.69
Witte peptone 11.14
De Chapoteaut peptone ........ 12.735
To study the distribution of the nitrogen, determinations
were made in both the colon and albumin poisons, of the
ammonia nitrogen, the mono-amino, and diamino nitrogen,
by the method already described under cleavage with
dilute mineral acids. The following are the results:
DISTRIBUTION OF NITROGEN IN PROTEIN POISONS.
Total N Mono-
Source of Total of acid Ammonia amino Diamino
poison. poison. extract. N. N. N.
Colon bacillus . 13.49% 10.185% 1.525% 6.472% 1.753%
Egg albumen . 13.74% 11.477% 0.745% 7.999% 1.400%
It will be seen that the greater part of the nitrogen is to
be found in mono-amino combination. From the phospho-
tungstic filtrates, from both the albumin and colon poisons,
containing the mono-amino acids, crystalline bodies were
obtained. Judged by the strong Millon test, tyrosin was
112 PROTEIN POISONS
undoubtedly present, but the crystalline masses were
largely leucin, and no tyrosin was obtained in purified
form. From the crude crystals, after many and repeated
crystallizations, what was thought to be leucin was obtained
pure, melting at 264° to 265° uncorrected, or 269.42° to
270.46° corrected. The crystals were thin plates charac-
teristically grouped, and sublimed readily. From another
5 per cent, sulphuric acid extract of albumin poison was
obtained a large mass of crystals in characteristic tyrosin-
like sheaves, and giving a deep Millon reaction. These
were undoubtedly tyrosin, though at the time no melting-
point was taken.
Properties of the Haptophor or Non-poisonous Group. — Leach1
has investigated this split product with the following
general results: After cleavage of the protein with alkaline
alcohol, the haptophor remains undissolved. It is collected
on a filter, then transferred to Soxhlets, and for some days
extracted with 95 per cent, alcohol. This is for the purpose
of removing as thoroughly as possible the alkali which it
has absorbed from the alkaline alcohol. This cannot,
however, be wholly washed out by this method, and it is
possible that in part it is held chemically. After this
extraction the substance is easily reduced to a fine brownish
powder. On burning it puffs up, gives off the characteristic
odor of nitrogenous compounds, and leaves a copious ash
containing phosphate. The solubility of the haptophors
from different proteins differs widely; that from egg-white
is wholly soluble in water, while that from the cellular
substance of the tubercle bacillus is only sparingly soluble.
However, it is only the part soluble in water from any of
these haptophors that is of special interest. The studies of
Leach, referred to, were made with the non-poisonous
portion of the colon bacillus. This is mainly soluble in
water, giving an opalescent solution from which a light-
colored sediment is deposited on standing, leaving a clear,
golden brown solution. The sediment is not soluble in
1 Jour. Biolog. Chem., 1907, iii, 443.
THE CLEAVAGE OF PROTEINS WITH ALKALI 113
either dilute alkali or acid in the cold, but is soluble in
alkali on boiling. The clear, aqueous solution of the hapto-
phor is alkaline from sodium hydrate held either mechanic-
ally or chemically; it is precipitated by mineral acids and
by alcohol. It responds to the biuret, xanthoproteic,
Millon, and Adamkiewicz tests. Millon's test is not very
satisfactory, and in some samples has failed altogether,
even after care has been exercised in neutralizing the alkali.
It is quite evident that the substance or substances in the
protein molecule to which the Millon test is due are for
the most part found in the toxophor group. However,
the readiness of response to this test varies greatly in the
different haptophors. The haptophor substance does not
reduce Fehling's solution directly, but does so readily and
abundantly after prolonged boiling with dilute hydro-
chloric acid. The presence of carbohydrate in the hapto-
phor has already been discussed (page 70). Tests with
a-naphthol, phloroglucin, and orcin give positive results.
Ammonium molybdate gives an organic precipitate, but
no evidence of free phosphoric acid. The preliminary tests
show the presence of protein, nucleic, and carbohydrate
groups. Comparing these results with those obtained in
the study of the toxophor, the following statements may
be formulated: (1) The toxophor is freely soluble in abso-
lute alcohol, the haptophor is insoluble in this menstruum.
(2) The toxophor contains no carbohydrate, all of which is
found in the haptophor. (3) The toxophor freely responds
to the Millon test, while the haptophor does so slightly and
in some instances not at all. (4) The toxophor contains no
phosphorus, or but little of this element, while the hapto-
phor is rich in phosphorus. (5) The toxophor from different
proteins seems to be the same, possibly with unrecognizable
differences in chemical structure, while the haptophor of
each protein differs from that from all other proteins.
Leach1 gives the following table showing the percentages
of ash, nitrogen, and phosphorus in the haptophor of the
colon bacillus:
1 Loc. cit.
ash.
ash.
N.
P.
free.
free.
N:P.
8.61
10.65
2.87
26.08
5.56
2.34
7.52
3.99
2.38
20.36
15.66
6.76
3.61
8.02
4.28
1.87
30.74
4.87
1.50
7.03
2.16
3.25
15.05
8.48
4.95
2.25
5.41
2.46
2.20
1.66
4.65
1.74
4.73
1.77
2.67
5.50
1.36
3.43
1.35
3.48
1.37
2.53
14.00
27.67
5.98
2.68
8.27
3.71
2.23
2.08
5.50
1.79
5.62
1.83
3.07
11.71
3.74
3.16
2.47
3.28
2.70
1.28
3.47
5.35
1.58
5.55
1.64
3.39
114 PROTEIN POISONS
Fixed Inorganic N ash P ash Ratio
Ash.
Cell substance .
Haptophor . . 33.25
Prep. A. . . . 26.76
Prep. B. . . . 35.34
Prep. D. . . . 15.38
Prep. G. . . . 6.99
Prep. G. pur. . 5.50
Prep. H. . . . 35.91
Prep. K2 . . . 7.57
Prep. M. . . 11.71
Prep. Ma . . 8.30
Explanation of the Table. — Ash, residue heating at low
redness. Fixed ash, residue after heating to full heat of
powerful burner. Inorganic ash, ash less calculated amount
of PO4. N, nitrogen by Kjeldahl-Groening method. P,
phosphorus by the Neumann method. N and P ash free,
reckoned free from "inorganic ash." N:P, quotient of
column 4 divided by column 5. A, portion of haptophor
dissolved by acid alcohol. B, portion of haptophor not
dissolved by acid alcohol. D, substance precipitated by
acid alcohol from solution of B in aqueous alkali. G,
substance precipitated by acid alcohol from aqueous solu-
tion of haptophor. H, obtained by concentration of the
alcoholic filtrate from G. K, substance precipitated by
dilute acetic acid from aqueous solution of haptophor. K2,
same as K, except that strong acid was used. M, precipi-
tated by alcohol from filtrate from K. M2, precipitated
from filtrate from K2.
Leach states: "As these preparations are all mixtures,
the absolute values are worth nothing taken singly, but
the comparative values, especially the ratio of N to P, as
given in the last column, are of interest. The determina-
tions were made for the sake of tracing the nucleo com-
pounds. There are many indications of nucleic acid, but
the amount of both nitrogen and phosphorus is much too
small. The ratio between them is, however, quite within
the range for nucleic acids from other sources, as may be
seen by comparison in the following table. Moreover, the
THE CLEAVAGE OF PROTEINS WITH ALKALI 115
nucleic acid and the nucleates are the only nucleo com-
pounds in which the ratios are at all comparable with
those given in the preceding table. Nuclein contains a
little less phosphorus than any of these preparations from
the germ, while other nucleo compounds are much richer
in nitrogen and poorer in phosphorus. It is perhaps worthy
of mention that contact with mineral acid apparently
breaks up the nucleic acid, the phosphoric acid going into
solution; thus, preparation A gives evidence of phosphorus
in inorganic combination, while G does not."
Substance.
Source.
Observer.
N.
P.
N:P.
Nucleic acid
Salmon sperm
Miescher
15.24
9.62
1.58
Nucleic acid
Sea urchin sperm
Mathews
15.34
9.59
1.60
Nucleic acid
Yeast
Miescher
16.03
9.04
1.77
Nucleic acid
Pancreas
Bang
18.20
7.67
2.37
Nucleic acid
Thymus
Kostytschew
15.55
9.25
1.69
Nucleic acid
Thymus
Kostytschew
15.26
9.30
1.65
Nucleic acid
Wheat embryo
Osborne and
15.88
8.70
1.83
Harris
Inosinic acid
Muscle
Haiser
16.00
8.60
1.86
Clupein nucleate
Mathews
21.06
6.07
3.48
Nucleohiston
Thymus
Huiskamp
18.37
3.70
4.97
Nucleoprotein
Thymus
Huiskamp
16.42
0.95
17.30
Nucleoprotein
Pancreas
Umber
17.82
1.67
10.65
Nuclein
Pancreas
Umber
17.39
4.48
3.88
Ba a-nucleate
Thymus
Kostytschew
12.83
7.63
1.68
Ba /8-nucleate
Thymus
Kostytschew
10.16
8.48
1.38
In a later paper, Leach1 has made a study of the hapto-
phor of egg-white. The percentages of ash, nitrogen, phos-
phorus, and sulphur in egg-white and its split products
are given and compared with the cellular substance of the
colon bacillus in the following table:
Egg-white .
Toxophor
Haptophor .
Cell substance
Toxophor
Haptophor .
Ash.
2.48
1.14
13.57
8.61
2.33
33.25
Inorganic
ash.
2.066
12.80
26.08
N.
14.48
13.74
12.67
P.
0,135
Trace
0.253
10.65 2.870
11.15 Trace
5.56 2.340
N ash P ash S ash
S. free. free. free.
2.66 14.70 0.138 2.73
2.19 13.90 .. 2.22
2.79 14.53 0.290 3.20
7.52 3.990
1 Jour. Biol. Chem., 1908, v, 253.
116 PROTEIN POISONS
Leach split up edestin, casein, egg-white, and colon
cellular substance with alkaline alcohol. The insoluble
part of each gave the various protein color tests, Millon's
reaction less satisfactorily than the others. On stirring
with water, the edestin preparation was entirely soluble,
there was a slight flocculence with the casein preparation,
the others were mainly but not wholly soluble. Addition of
a little sodium hydroxide increases the solubility. Mineral
acids give precipitates with the casein and egg preparations.
The most marked difference was found on testing for
carbohydrates. As edestin contains no carbohydrate, its
preparation showed no evidence of such a group. Although
casein is said to contain no carbohydrate, it has been found
to respond to the Molisch test, and so does its haptophor.
As was to be expected, the egg preparation gives evidence
of hexose and not pentose. The lead sulphide reaction
shows the presence of loosely combined sulphur in the
preparations from egg and edestin, not in the ones from
casein and the colon bacillus.
Samples of the haptophor of egg-white were stirred with
water, filtered, and attempts made to separate protein and
carbohydrate in the filtrate by means of uranium acetate.
The acetate was added both with and without sufficient
alkali to keep the solution alkaline. A copious precipitate
resulted in both cases and this was filtered out with some
difficulty. The slight excess of uranium was removed
from the filtrate by the addition of sodium phosphate.
The filtrate gave evidence of carbohydrate, but the separa-
tion was not sufficiently sharp, and that method was
abandoned. Acidifying until there was a slight permanent
precipitate, the addition of either ethyl or methyl alcohol
cleared the solution. Phosphotungstic acid precipitated
both protein and carbohydrate. In short, no method was
found that would remove the protein from the solution and
leave the carbohydrate. It is perhaps a legitimate infer-
ence that the combination of the two is a chemical one.
Samples were subjected to hydrolysis and titrated with
Fehling's solution. The proteins and possibly other bodies
THE CLEAVAGE OF PROTEINS WITH ALKALI 117
present interfered with the reaction, but by adding the
solution all or nearly all at once it was possible to obtain
comparative results. Experiments with the haptophor of
the colon bacillus had shown that the maximum reduction
was obtained by boiling for two and one-half hours with
2.5 per cent, hydrochloric acid (see p. 70).
Three grams of the haptophor of egg-white was mixed
with 200 c.c. of water, and 20 c.c. of 25 per cent, hydro-
chloric acid. A second sample was prepared in the same
way except that it was filtered before adding the acid.
Both were boiled with reflux condenser. After boiling half
an hour and then at intervals of three hours, aliquot parts
were removed, neutralized, titrated with Fehling's solution,
and the amount of reducing substance calculated. Other
samples were hydrolyzed with sulphuric acid, with less
satisfactory results. These preliminary experiments indi-
cated that the reducing substance is all present in the
portion soluble in water, and that the maximum yield,
which if calculated as dextrose, is about 9 per cent., is
obtained by boiling from ten to twelve hours, and until the
mixture no longer gives the biuret test.
Accordingly, 25 grams of the egg-white haptophor was
shaken for two hours on a shaker with ten times its weight
of water, filtered, 200 c.c. more of water added, the solution
neutralized with hydrochloric acid, then 50 c.c. of 25 per
cent, hydrochloric acid added, thus making approximately
a 5 per cent, solution of material in 2.5 per cent. acid. This
was boiled with a reflux condenser for ten or twelve hours,
until the solution no longer gave the biuret test. It was
then filtered, leaving very little on the filter. The clear,
red-brown filtrate was cooled, neutralized with sodium
hydroxide, and benzolated by the Schotten-Baumann
method. The mixture became very warm, but was cooled
by surrounding the flasks with pounded ice and salt. When
the reaction ceased, the compound settled nicely, and was
filtered by suction after standing two or three hours. The
precipitate was washed with water containing a little
ammonia, and treated with boiling water, in which a large
118 PROTEIN POISONS
portion was freely soluble. On cooling and concentrating
the alcoholic solution, a fine yield of crystals was obtained.
The crystals from several samples were united and recrys-
tallized from hot absolute alcohol until the solution was
clear and colorless. Macroscopic bundles of needles were
thus obtained, showing very characteristic grouping. They
were washed in alcohol and in ether, dried upon porous
plates, the operations being repeated until samples from
two recrystallizations melted side by side within 1° or 1.5°.
The crystals are pure white, readily soluble in benzol,
chloroform, and in glacial acetic acid as well as in alcohol,
and melt at 203°. When boiled with sodium hydroxide,
ammonia is given off; after removing benzoic acid by boil-
ing with hydrochloric acid, the resulting product reduces
Fehling's solution.
0.4150 gram gave 0.00891 gram N, corresponding to 2.14 per cent. N.
0.4220 gram gave 0.00962 gram N, corresponding to 2.279 per cent. N.
Average is 2.213 per cent. N.
These characteristics suffice to identify the compound
as glucosamin benzoate which Pumm reports as melting
at 203°. Kueny prepared different benzoates of glucosamin
by varying the conditions of the experiment. The one most
readily formed was the tetrabenzoate, melting at 199°
when recrystallized from alcohol, and at 207° when re-
crystallized from glacial acetic acid. He tried by various
methods to prepare a pentabenzoate, but without success.
Langstein prepared glucosamin benzoate from egg-white,
which, after once recrystallizing from hot alcohol, melted
at 201° to 202°, and gave 1.95 per cent, of nitrogen. The
theoretical amount of nitrogen in the tetrabenzoate is
2.35 per cent. Thus, the benzoate prepared from the
haptophor of egg-white agrees with glucosamin benzoate
prepared from glucosamin and from egg-white, at least as
well as those preparations do with each other. Numerous
observers have found glucosamin in egg-white, and this
work shows that it remains in the haptophor when egg-white
is disrupted by alkaline alcohol.
CHAPTER VI
ACTION ON ANIMALS1
IT will be interesting and instructive to compare the
effects of the living bacillus, the dead cellular substance,
and the soluble poison on animals. •
The Action of the Living Bacillus. — When a guinea-pig is
inoculated with a fatal dose of the living colon germ, prac-
tically no symptoms whatever are noticeable for a period
varying from five to twelve hours, according to the size
of the dose given. This may be considered as the period
of incubation and is roughly proportional to the amount of
living germ injected. We have always worked with a
bacillus 1 c.c. of a twelve-hour or older, bouillon culture of
which has invariably proved fatal to guinea-pigs within
twenty-four hours. If 1 c.c. of such a culture is given, no
effects will be seen for a period of from ten to twelve hours.
If, on the other hand, 2 c.c. of the same culture be injected,
the animal will begin to manifest symptoms of illness in
from eight to ten hours, and if larger doses are given the
symptoms will become apparent in a shorter time. This
period of incubation undoubtedly represents the time
taken for the bacillus to multiply and to be destroyed to
such an extent that sufficient poison may be liberated
through its disintegration to produce noticeable toxic
effects in the animal. This period of incubation is, therefore,
in reality the crisis of the disease and the outcome depends
solely on whether all bacteria have been destroyed before
a lethal dose of the poison has been set free or not. It is
1 This chapter is a reproduction, without material change, of an article
by Victor C. Vaughan, Jr., published in the Jour. Amer. Med. Assoc. in
1905.
120
PROTEIN POISONS
during this period that individual resistance and acquired
immunity are important factors acting by causing increased
bacteriolysis and the destruction of all bacilli before a
fatal dose of poison has been set free. During this time
the temperature of the animal may rise to a greater or less
extent or may remain stationary; the animal remains
active, eats; its coat is not roughened and it appears in
FIG. 5
102
101
iooc
\
Temperature curve of guinea-pig after inoculation with 1 c.c., sixteen-
hour bouillon culture of the colon bacillus. Death occurred twenty hours
after inoculation.
all respects as well as a normal animal. At the end of this
period, however, the appearance changes. The animal
becomes less active. It remains in one corner of its cage;
its coat becomes roughened; it hangs its head and apparently
enters into a state of stupor. At the same time the rectal
temperature begins to fall abruptly, as can be seen from
a study of Fig. 5.
ACTION ON ANIMALS 121
Indeed, this fall of body temperature is often the first
marked symptom and, when occurring to a marked degree,
it is invariably a bad omen. The body temperature will
often fall from 101° to 94° F. or even lower within from two
to four hours, and this fall is progressive and continuous
until the animal's death, immediately preceding which a
temperature as low as 87° or 86° F. is not uncommon. At
the same time the animal shows signs of the most marked
peritoneal inflammation, as is evidenced by rigidity and
spasm of the abdominal muscles on pressure. At autopsy,
the only gross lesion present is a marked hemorrhagic
peritonitis with a large amount of bloody fluid containing
intact red corpuscles and leukocytes in the peritoneal
cavity. The parietal and visceral peritoneum are studded
with minute punctiform hemorrhages. Hemorrhage is an
especially prominent feature in the great omentum and is
present to a less marked degree in the mesentery.
The Action of the Cellular Substance. — The dead bacterial
substance used in the following work was obtained by
growing a large amount of the colon germ on tanks filled
with agar for a period of two weeks at room temperature.
At the end of this time the growth was removed from the
tanks and extracted with absolute alcohol and ether. The
crude bacterial substance thus obtained was reduced to a
fine powder by pulverization in an agate mortar, and was
then ready for use.
It is interesting to note that the person who did the
pulverizing was often quite seriously poisoned during the
process unless he took the precaution of wearing a mask
which hindered the inhalation of the powder. The symp-
toms of such poisoning were exceedingly interesting. The
first thing noticed was a marked irritation of the nasal
mucous membrane and a huskiness of the voice, due no
doubt to the mechanical irritation of the inhaled powder.
This was followed by a feeling of depression and malaise,
and chilly sensations. Occasionally a decided chill would
be experienced. It is unfortunate that no accurate obser-
vations of temperature were taken in these cases. Nausea
122 PROTEIN POISONS
and even vomiting were occasionally noted. After a period
of discomfort varying from six to ten hours, during which
the patient often complained of dull pain in the various
joints, recovery would rapidly and completely take place.
On examining the powder obtained in this manner
microscopically we found that it consisted of colon bacilli
which still retained their morphological characteristics and
could still be stained by aniline dyes. On the other hand
cultures made from this powder have, of course, never
given a growth. In other words, the bacillus has not been
broken up by this treatment, but simply has been deprived
of life and of the power of reproduction. It is worthy of
note that neither by the action of alcohol, ether, physiological
salt solution, distilled water, nor any simple solvent have
we been able to extract a poison from the colon bacillus.
Nor, again, can a poison be split off by the action of heat
even when the germ substance is heated to 184° C. in a
sealed tube for thirty minutes. It is only when we make
use of agents which will chemically break up the colon
bacillus that we are enabled to obtain a poison apart from
the rest of the cellular substance. The powdered bacterial
substance is not soluble, but can be held in suspension in
normal salt solution and, since it can be boiled without
appreciably affecting its toxicity, suspensions were always
heated to 100° for fifteen minutes before injection in order
to insure sterility.
This coarsely powdered cellular substance killed guinea-
pigs when injected intraperitoneally in doses of 1 to 40,000,
body weight, and invariably proved fatal within twelve
hours, usually causing death at the end of from six to eight
hours. On the injection of a fatal dose of the cellular
substance intraperitoneally, we noticed that the most
marked change was in the length of the period of incu-
bation. Thus, whereas in the case of the living germ from
eight to twelve hours passed before noticeable symptoms
appeared, in that of the dead germ substance the animal
almost invariably showed symptoms of illness at the end
of four hours. In regard to the character of these symptoms
ACTION ON ANIMALS 123
it may be stated that they are similar in all respects to
those induced by the living bacillus. The temperature
remains the same or may rise slightly during the first two
hours. At the end of the four hours it has begun to fall,
and there is a decided drop from then on until the time of
death, provided the dose given is a fatal one. If a non-fatal
dose has been injected intraperitoneally the temperature,
as will be seen from Fig. 6, has reached a minimum at the
end of from six to eight hours and has returned to normal
again in from twelve to twenty hours.
Moreover, as a general rule, it may be stated that the
fall in non-fatal cases seems to be directly proportional to
the amount of bacterial substance injected. That this
.should be the case seems to be olily natural when we con-
sider the fact that in this instance we have largely done
away with that factor which is known as the individual
resistance of the animal. As has been previously mentioned
in the case of the living bacillus, the individual resistance
plays an important part in determining the amount of
poison which will ultimately be set free in the body. For
example, whereas 1 c.c. of a twelve-hour culture of our
colon bacillus invariably proved fatal, 0.25 c.c. never did.
The explanation of this is to be found in the fact that with
the smaller dose all animals were able to cause disintegra-
tion of all bacilli injected before a fatal dose of poison was
set free. If now 0.5 c.c. be given some would recover, while
others would die. In this case we would speak of the former
as possessing a greater individual resistance than the latter.
This simply means that, in the first instance the animal
has possessed a sufficient quantity of bactericidal substance
directly available to cause disintegration of all bacilli before
the latter have multiplied to a sufficient extent to furnish
enough poison to kill the animal on its liberation. On the
other hand, those animals which succumbed did not possess
quite enough of the bactericidal substance, or at least did
not possess it in a form available for immediate use. When,
however, the dead bacterial substance is given the dose
of poison which the animal receives is a certain definite
amount and is not capable of subsequent increase.
124
PROTEIN POISONS
Accompanying the fall in temperature there is apparent
lassitude, stupor, and roughening of the coat. In cases in
which many times the fatal dose has been given, the animals
occasionally die within from four to six hours with convul-
sions, a feature which can now and then be observed after
the injection of large quantities of the living bacillus. At
autopsy we find a picture similar in all respects to that
following inoculation with the living colon bacillus. There
is a marked hemorrhagic peritonitis, the peritoneal cavity
FIG. 6
101
100°
99
98
97°
96
95
94
Temperature curve of guinea-pig after intraperitoneal injection of
non-fatal dose of crude bacterial cell substance.
containing bloody fluid, together with unabsorbed bacterial
cell substance, and the omentum and mesentery showing
numerous punctiform hemorrhages. It is needless to
state that in all cases cultures were made from the peritoneal
cavity and heart's blood immediately after death, and
these proved to be sterile. From this we see that practically
the sole difference between the effects following inoculation
with the living bacillus and the injection of the dead bac-
terial substance is a shortening of the period of incubation
ACTION ON ANIMALS 125
due, no doubt, to the fact that the intracellular poison is
liberated much more rapidly and in greater concentration
in the second case. As will be seen later, it is not so much
the absolute quantity of the poison which is injected that
determines the result, as the amount which is active at a
given time.
The Action of the Soluble Poison. — When doses of this
powder are given intraperitoneally in amounts varying
from 8 to 60 milligrams, according as to whether we have
been careful to remove most of the common salt or not, a
fatal result follows in guinea-pigs in from thirty to sixty
minutes. Within fifteen minutes after injection the temper-
ature begins to fall and sometimes within half an hour has
reached 94° F. or even lower. At first, after an interval of
from five to ten minutes immediately following the injec-
tion, the animal appears restless, runs about the cage, and
shows a great tendency to scratch itself, this undoubtedly
being due to itching sensations in the skin caused by irrita-
tion of the peripheral nerves. The animal then begins to
show evidence of lack of coordination, which is rapidly
followed by partial paralysis, which is especially marked
in the hind extremities. This stage lasts for from five to
ten minutes, during the later part of which the animal
usually lies quietly on one side. From this state the animal
passes into what one might term the convulsive stage.
These convulsions are usually clonic in nature and, as a
rule, at first involve only the neck muscles, the head being
momentarily drawn backward on the back. At first these
convulsions are but slight in degree and are separated by
considerable intervals of time. Soon, however, they become
much more frequent and of much greater severity. Gradu-
ually they become more and more general in their extent,
until all the muscles of the body become involved in violent
clonic convulsions. This stage when present presages a
fatal outcome; rarely an animal recovers after reaching
the convulsive stage. During a convulsion, or occasionally
in the interval of calm, respiration ceases. The heart,
however, continues to beat, at first with perfect regularity
126 PROTEIN POISONS
and no acceleration; indeed, the rate seems to be somewhat
slower than normal. Gradually the beat becomes more
and more feeble, the rate and regularity being preserved
to the end. It is usually only after an interval of from
three to four minutes after the cessation of respiration that
the heart ceases to beat. As has been previously stated, a
fatal issue, if it occurs at all, always results within one
hour after injection and usually within from thirty to forty
minutes. This is to a large extent independent of whether
the dose is the minimum lethal one or two or three times
that amount. It is certainly entirely independent of the
size of the pig. Death, of course, results at slightly different
times with different batches of the poison, but even in this
case the interval of time between injection and a fatal
issue does not vary to any great extent. A dose wrhich has
proved to be the minimum fatal dose for one pig will almost
surely prove to be the same for another. In other words,
we have done away practically entirely with the period of
incubation, and the poison acts so rapidly that individual
resistance plays no part; hence, the animal acts almost
with the exactitude of a chemical compound into which
for all practical purposes it has been converted. The period
of incubation has ceased to exist since the poison is no longer
contained within either the dead or the living bacillus, but
is present in a free and uncombined form, capable of uniting
immediately with those body cells for which it may possess
a special affinity.
At autopsy no special gross lesions can be made out.
The peritoneum is smooth and shiny throughout, and there
is not the slightest evidence of either hemorrhage or even
marked congestion in the omentum or mesentery. This is
very important and in marked contrast to the hemorrhagic
peritonitis found after injection of either the living or the
dead colon bacillus. We are inclined to believe that it is
the distinguishing feature between the injection of the
poison in a comparatively free and in a combined state.
At one time we attempted to obtain the poison by a simpler
method, omitting the extraction of the crude substance
ACTION ON ANIMALS 127
with ether. The result was that on evaporation of the
alcoholic filtrate we obtained a sticky residue which it
was utterly impossible to pulverize or to weigh. We were
compelled, therefore, to content ourselves with evaporating
it to a sticky mass, which was then immediately dissolved
in water. The solution of the substance thus prepared
was . very poisonous, but, as a rule, took from one to two
hours or even longer to bring about a fatal result. The
animals showed the roughening of the coat and the stupor
characteristic of the living and dead bacillus, but not as a
rule seen in the case 'of the soluble poison. ..Furthermore,
the majority of the animals showed during life unmistakable
signs of peritoneal inflammation. They died in convulsions.
At autopsy an intense hemorrhagic peritonitis was present,
which was particularly prominent in the omentum and
mesentery, and hemorrhage was often present in the cap-
sules of the liver and the spleen. From the fact that death
was slower in these cases and that the symptoms were
more like those seen after inoculation with the living bacillus,
we are inclined to believe that in this instance the poison,
although split off from the bacillus itself, still exists in
combination with some other cell group, and that it is
essential that this combination be broken up before the
poison can be set free and can act on the body cell.
Another interesting fact in this connection is furnished
by the action of the poison in solutions which have been
rendered strongly alkaline by the addition of sodium bicar-
bonate. As has been previously stated, the aqueous solu-
tions of the poison are slightly acid in reaction, and in order
to avoid the irritative effects which might follow their
injection into the peritoneal cavity, they were neutralized
or rendered slightly alkaline by the addition of sodium
bicarbonate.1 At first no attempt was made to secure
perfect neutralization, with the result that sometimes
we were making use of neutral, while again slightly or
1 The precaution of neutralizing the soluble poison, when properly
prepared, is unnecessary as it has no appreciable irritative action.
128
PROTEIN POISONS
decidedly alkaline solutions were employed. It was soon
noticed, however, that the results obtained in the three
cases were very different. Thus, whereas in the neutral
or faintly alkaline solution the injection of 60 mg. of the
powder invariably killed, in the case of a stronger alkaline
solution the same amount did not cause a fatal result,
although the animals were very ill. From this fact it
became evident that some change had taken place in the
poison on standing in alkaline solution. In order to study
this change more in detail, experiments were conducted
with solutions of different degrees of alkalinity, with the
results found in the following tables:
TABLE III. — RESULTS WITH SOLUTION OF POISON BARELY NEUTRALIZED
WITH SODIUM BICARBONATE AND PLACED IN INCUBATOR
Solution
Result of Time of
No. of
Dose of
kept in
Weight
injec- death after
animal.
poison.
incubator.
of pig.
tion. injection.
1
60 mg.
Fresh
325 gm.
+ 30 minutes
2
60 mg.
2 hours
330 gm.
+ 20 minutes
3
60 mg.
20 hours
370 gm.
+ 20 minutes
4
60 mg.
2 days
320 gm.
+ 15 minutes
5
60 mg.
4 days
350 gm.
+ 20 minutes
6
60 mg.
6 days
350 gm.
+ 45 minutes
7
60 mg.
8 days
320 gm.
+ 20 minutes
TABLE IV. — RESULTS WITH SOLUTION OF POISON RENDERED DECIDEDLY
ALKALINE WITH SODIUM BICARBONATE AND PLACED IN INCUBATOR
Solution
No of
Dose of
kept in
Weight
animal.
poison.
incubator.
of pig.
1
60 mg.
Fresh.
350 gm.
(barely
neut.)
2
60 mg.
Fresh.
370 gm.
decided-
ly alk.
3
60 mg.
2 hours
325 gm.
4
80 mg.
4 hours
310 gm.
5
120 mg.
24 hours
280 gm.
160 mg. 3 days 350 gm.
Result of
injection.
Very sick
for 2 hours
Sick
+
+
Time of
death after
injection.
30 minutes
Recovered
Recovered
Recovered
More than
5 hours.
7 hours.
ACTION ON ANIMALS
129
TABLE V. — RESULTS WITH SOLUTION OF POISON RENDERED DECIDEDLY
ALKALINE wiTri SODIUM BICARBONATE AND KEPT AT ROOM
TEMPERATURE
Time at
No. of
Dose of
room tem-
Weight
animal.
poison.
perature.
of pig.
1
60 mg.
Fresh
350 gm.
2
60 mg.
12 hours
265 gm.
3
90 mg.
2 days
280 gm.
4
120 mg.
2 days
460 gm.
5
120 mg.
7 days
405 gm.
6 160 mg. 7 days
440 gm.
Result of
injection.
Sick
Sick
Sick for 5
hours
Sick for
several hours
Time of
death after
injection.
35 minutes
Recovered
Recovered
20 minutes
Recovered
Recovered
From the above tables it will be seen that the degree of
alkalinity of the solution, and especially the length of time
that the poison has stood in alkaline solution are very
important factors in determining its toxicity. Thus in
Table III, in which the solution was barely neutralized, the
poison seems to have retained its full potency after eight
days in the incubator, whereas, in the case of the strongly
alkaline solution, the potency has decreased markedly
within from twenty-four to forty-eight hours. Again,
there are great differences to be seen depending on whether
the strongly alkaline solution has been kept at room tem-
perature or at that of the incubator, the decrease in toxicity
being much less rapid in the first instance.
A more detailed report of the effects on animals than
it was possible to give in the above tables is not without
interest. For example, in Table IV, No. 2, which received
60 milligrams immediately after the solution had been
rendered decidedly alkaline, was very sick indeed, whereas
No. 3, which received the same amount after two hours
in the incubator, was only slightly affected. In the case
of Nos. 5 and 6 the effects observed corresponded more
closely to those obtained with the crude bacterial cell
substance. It is unfortunate that the time of death was
not ascertained in the case of No. 5. No. 6 did not suc-
cumb until seven hours after the injection. On autopsy
9
130 PROTEIN POISONS
there was considerable fluid in the peritoneal cavity, and
the vessels of the mesentery were markedly congested.
The omentum was particularly injected and a few minute
hemorrhages could be made out. The most plausible explan-
ation of the above facts is found in the theory that the
poison has not been destroyed in the alkaline solution, but
rather has entered into chemical combination with the
alkali and that we are again dealing with it in a combined
instead of in a free state. The fact that the same amount
will not cause a fatal result is thus readily explained, since
the outcome depends largely on the rapidity with which
the poison acts. If it is present in a state of combination
which must be broken up before it can exert its deleterious
action on the body, and if this combination is only slowly
decomposed, the nerve cells, for which it apparently has a
special affinity, are not subjected to an overwhelming dose
at one time, as in the case of the intraperitoneal injection
of the free poison.
The* results obtained in animals Nos. 5 and 6, Table V,
are very interesting. In these instances there were two
distinct illnesses, the first becoming manifest within from
twenty to thirty minutes after the injection and corre-
sponding in all respects to that following a non-fatal dose
of what we have for convenience termed the free poison.
The animals were decidedly in better condition at the end
of an hour; however, they then began to show symptoms
similar to those noticed after the injection of the crude
cell substance, i. e., roughening of the coat, stupor, and
slight convulsive movements. Recovery from this state
did not occur until after the lapse of from five to six hours.
It is evident that here the first signs of illness were due
to some of the poison which had not as yet combined with
the alkali, and hence still existed in the free state, whereas
the later symptoms were due to the effects of the slow
liberation of the same poison from its combination. In this
connection it is interesting to note that the combination
between the poison and the alkali which apparently takes
place in decidedly alkaline solutions is not an immediate
ACTION ON ANIMALS 131
one, but occurs gradually and reaches a maximum only
after the lapse of a considerable interval of time. That
the rapidity with which this combination is effected depends
largely on temperature is shown by the fact that it occurs
much more rapidly in a solution kept in the incubator
than in one which is allowed to stand at room temperature.
The results which follow the injection of a fatal dose of
the soluble poison intraperitoneally have already been
described. When a non-fatal dose has been injected the
symptoms first noticed are similar in all respects to those
following a fatal dose. The animal becomes restless, shows
signs of irritation of the peripheral nerves, incoordination,
and partial paralysis. The convulsive stage is not present,
as a rule, and when ,it is noticed is evidenced solely by
slight movements separated by considerable intervals of
time. We have never seen a case showing marked general-
ized convulsions which recovered.1 Recovery is apparently
rapid and complete, and within two hours after injection
the animal which has been desperately ill appears as well
as any untreated animal. Tne maximum effect is obtained
within from forty-five to sixty minutes in every instance.
The study of the changes in temperature in these animals
is particularly interesting. Within fifteen minutes the
rectal temperature has begun to fall and has reached a
minimum within one hour.
It remains stationary for a short time and then begins
to rise again, and at the end of three hours after the injection
has usually returned to normal or above.
On injecting the soluble poison subcutaneously, we find
that animals are able to withstand a much larger dose
than when the poison is given intraperitoneally. Thus, in
the case of a poison, 60 mg. of which invariably killed
when given intraperitoneally, it was found that 120 mg.
could be given subcutaneously without causing a fatal
result. However, the injection of a solution containing
180 mg. invariably caused death, the fatal issue occurring
« l This does rarely occur.
132
PROTEIN POISONS
in about the same length of time as in the case of animals
treated intraperitoneally. Thus a dose of 180 mgs. always
proved fatal in from one-half to three-quarters of an hour.
The symptoms are practically identical with those following
the intraperitoneal injection with the exception of the fact
that the various stages are much more sharply defined.
For example, the stage of peripheral irritation is much more
marked. The animal soon after injection becomes very
restless, runs around his cage, and scratche's his body.
This itching seems, however, to be general from the outset,
and is not, apparently, more pronounced in the immediate
FIG. 7
1U:>
/
-
k
/
'&
V
z
\
z
\
/
<*<;'
_
1 — .
— —
— —
— -"
Temperature curve of guinea-pig treated with 45 mgs. of the soluble
poison intraperitoneally.
neighborhood of the site of injection. If the animal has
been injected under the skin of the abdomen, its attention
is not necessarily first attracted to this spot, but it may
begin by scratching its nose or one of the extremities.
Another peculiar symptom, which is probably due to
peripheral irritation, and which is seldom seen in cases of
intraperitoneal injection, is the tendency which the animals
show to dig furiously in the shavings in the bottom of their
cages. This feature is quite characteristic, and is seldom
absent in pigs which have been treated subcutaneously.
The later stages are similar in all respects to those seen
following the intraperitoneal injection. The animal shows
ACTION ON ANIMALS 133
symptoms of incoordination, lies on one side, and finally
develops convulsions, with failure of respiration, the heart
continuing to beat regularly for some time after the com-
plete cessation of respiration. Here also the symptoms are
accompanied by a decided fall in the body temperature.
The results following the intravenous injection of the
soluble poison are given in the following table:
TABLE VI
Time of cessa- Time of cessa-
No. of animal. Amount of tion of respira- tion of heart-
poison injected tion after beat after
intravenously. injection. injection.
1 ... 10 mg. 4 minutes 7 minutes
2 ... 10 mg. Recovered
3 ... 10 mg. 3 minutes 6 minutes
4 ... 10 mg. Recovered
5 ... 15 mg. 4 minutes 6 minutes
6 ... 15 mg. 3 minutes 5 minutes
7 ... 15 mg. 4 minutes 7 minutes
8 ... 15 mg. 4 minutes 6 minutes
9 ... 20 mg. 3 minutes 5 minutes
10 ... 20 mg. 4 minutes 6 minutes
11 ... 20 mg. 3 minutes 6 minutes
12 ... 20 mg. 3 minutes 7 minutes
From the above table it will be seen that in all cases
respiration ceased within four minutes after injection.
Indeed, the respiratory embarrassment becomes pronounced
immediately following the injection. The animal struggles
for breath and there is violent retraction of the sternum.
No convulsions are seen following the intravenous injection,
this being probably due to the inhibitory influence of the
anesthetic which has been used during the preparation of
the animal for the operation. The failure of respiration in
the absence of convulsions would seem to be conclusive
evidence that the cessation of this function is due not to
mechanical interference during a convulsive attack, but to
a direct paralysis of the respiratory centre itself. Further-
more, the fact that the heart continues to beat in a perfectly
normal manner for from two to four minutes after respira-
tion has entirely ceased, would tend to show that the
134 PROTEIN POISONS
immediate cause of death is asphyxia brought about by
paralysis of respiration through the action of the poison.1
This action of the heart after the cessation of respiration
is exceedingly interesting, and is entirely analogous to that
mentioned as following the intraperitoneal injection. The
rate of the beat is decidedly lessened and at first the indi-
vidual beats are stronger. They gradually become more
and more feeble, however, until finally the heart stops in
diastole, the rate after the preliminary slowing remaining
unchanged until the end. It is worthy of note that in the
case of intravenous injections the fall of temperature, which
is so marked a feature after the intraperitoneal and
subcutaneous injections of the poison does not occur.
The explanation of this fact is doubtless to be found in the
very short interval of time which elapses between the
injection and a fatal outcome. As regards the size of the
lethal dose when given intravenously, we see that 10 mg.
often, and 15 mg. invariably, proved fatal.2
For purposes of comparison, we have always made use
of the poison obtained from the same extraction in our
intravenous, subcutaneous, and intraperitoneal injections,
and have therefore been able to ascertain with a fair degree
of accuracy the differences in dose required to bring about
a fatal result in the three cases. Thus, whereas 60 mg.
represents the fatal dose when given intraperitoneally, it
requires between 120 and 180 mgs. subcutaneously, and
only from 10 to 15 mgs. of the same poison to cause death
when given intravenously. These differences are un-
doubtedly due to the rapidity of absorption in the various
cases, and a fatal issue depends entirely on whether suffi-
cient poison reaches the sensitive area at one time to cause
cessation of respiration OP not.
It has now been shown that a very powerful intracellular
poison can be obtained from the colon bacillus. As has
been previously stated, the results given in the foregoing
experiments are those obtained with the poison from one
1 The physiological action of this protein poison is discussed on page 315.
2 The fatal dose of the purest form of the poison which we have
obtained is 0.5 mg. intracardially.
ACTION ON ANIMALS 135
extraction only. It must be understood that the poison
is not in a pure state and when it is stated that 60 mgs.
causes death when injected intraperitoneally we refer simply
to the powder obtained from a given extraction. We have
been able to procure powders which kill in doses of 15 and
even as low as 8 mgs. when given intraperitoneally. This
difference is in large amount due to the presence of sodium
chloride, since no attempt has been made to remove this
salt in the case of the less toxic powders by redissolving
in absolute alcohol.
There are several facts which lead us to believe that
this poison is the one which causes the symptoms of illness
and death in animals infected with the colon germ. Most
of these facts have already been brought out, but it may
not be out of place to briefly recapitulate at this point.
As has been previously seen, the results obtained with the
living germ, the dead bacterial substance, and the soluble
poison can best be explained on the ground that the poi-
sonous body in each case is the same. The differences in
action are not differences in symptoms, but simply in the
rapidity with which these symptoms become manifest.
While it is undoubtedly true that in animals dying with
the minimum fatal dose of the living germ, the convulsive
stage is not present or is only slightly marked, it is rarely
absent in cases where from three to four times the fatal
amount has been given. The sole difference between the
living germ and the soluble poison which would appear to
demand an explanation is the lack of evidence of a perito-
nitis in the latter case. This, we think, is best explained by
the fact that in the case of the soluble poison the poisonous
substance exists in an uncombined form, which, of course, is
not true in the case of either the living or the dead bacte-
rial cell. The uncombined poison is rapidly absorbed from
the peritoneal cavity, and hence the irritative effects which
would result from its retention in this place are absent.
As has been stated, one of the first signs of the action of
the poison is a lowering of the body temperature. This
hypothermia is usually present to a marked degree, and
is noticeable before any visible symptoms occur. It is,
136 PROTEIN POISONS
therefore, the best index which we have as to the exact
time at which the poison begins to exert its effect. In the
same manner the rise of temperature after the development
of hypothermia is the first indication of recovery. More-
over, if in an animal with a subnormal temperature a rise
occurs, it is an infallible sign of ultimate recovery, no matter
how grave the general condition may appear to be at the
time. We have laid great stress, therefore, on the changes
in body temperature as furnishing the most delicate test
of the action of the poison. It may be here stated that
the body temperature of guinea-pigs is ordinarily fairly
constant within certain narrow limits, and they are much
more satisfactory animals to work with in this respect
than are rabbits. Moreover, their temperature does not
seem to be materially altered by the injections of sterile
salt solution or such inert substances as a suspension of
pumice stone into the peritoneal cavity. As has been seen
in the case of the living germ, it is only after the lapse of
several hours that a fall in temperature occurs. This
would indicate that it is not until this time that sufficient
poison is liberated to cause notable toxic effects in the
animal. That it takes an appreciable time for the poison
to be liberated from the bodies of the bacilli is well illus-
trated in the instance of the dead bacterial substance.
Here it is only a question of dissolution of the bacilli and
the setting free of the contained poison, and yet it will
be noticed that an interval of at least two hours and usually
longer elapses before there is any noticeable fall in temper-
ature. The maximum effect in this case is reached between
four and six hours, and if the dose has been a non-fatal
one, recovery begins at the end of from eight to ten hours,
as is indicated by the upward trend of the temperature
curve at this point. In the case of the soluble poison, the
toxic effect begins at once. Within fifteen minutes the
temperature has begun to drop and within an hour has
reached a minimum. Recovery then begins and within
three hours the effect of the poison has worn off, as is best
evidenced by the return of the body temperature to normal
or above bv this time.
CHAPTER VII
THE PRODUCTION OF ACTIVE IMMUNITY
WITH THE SPLIT PRODUCTS OF THE
COLON BACILLUS1
IT may be stated that the work which we have done
with the colon bacillus up to the present time has in every
instance upheld the belief that the substances which give
rise to the phenomena occurring in animals infected with
the living colon germ exist as essential groups within
the bacterial cell and can be liberated from the latter
only by its disruption. Moreover, until these substances
have been separated from the other constituents of the
bacterial cell with which they are normally combined they
are unable to exert any deleterious action upon the body
cells. If the belief that the phenomena which result from
infection with the colon bacillus are due to the action of
the intracellular constituents of this organism is correct,
we would expect that it might be possible by chemical
means to split up this bacillus into different groups, the
injection of some of which into animals would be followed
by some of the results which are seen after inoculation with
the living germ. In a previous chapter we have shown that
it is possible to split off a toxic group which causes death
in animals with symptoms similar to those observed after
the injection of a fatal dose of living bacilli. However,
death is by no means the sole phenomenon which results
from the inoculation of animals with the colon bacillus.
The results which follow the injection of non-fatal doses of
the living germ are of equal if not of greater importance.
1 This chapter is taken, with but few changes, from an article by Victor
C. Vaughan, Jr., in the Journal of Medical Research, 1905, xiv, 67.
138 PROTEIN POISONS
It is a well-known fact that animals which have been
treated with non-fatal doses of either the living or dead
colon germ acquire a certain degree of immunity toward
subsequent infection with this bacillus. If, now, our theory
as regards the action of this bacillus is correct, one would
suppose that among the groups which we have been able
to split off there exist certain ones which possess the power
of producing immunity when injected into susceptible
animals. To ascertain whether such an active immunity
can be established in animals through treatment with the
split products of the colon bacillus is the aim of this chapter,
and we shall find it convenient to take up (1) the active
immunity obtained with the toxic portion, and (2) the
immunity obtained with the residue which remains after
the separation of the poisonous portion from the cellular
substance.
1. Immunization with the Poisonous Portion of the Cellular
Substance of the Colon Bacillus. — From the description of
the action of the "crude soluble poison" of the colon bacillus
given in the preceding chapter, it can be readily seen that
the poison with which we are working is one which exerts
its action with great rapidity. The difficulties of immunizing
animals with such a poison can be readily appreciated, and
it is inevitable that during the course of treatment a large
number should be lost. Up to the present time our attempts
to produce immunity with the toxic portion have been
largely confined to intraperitoneal and subcutaneous injec-
tions with what we have termed the free or uncombined
poison. However, as has been shown in a previous chapter,
it is possible to make use of this poison in a combined state
by rendering the solution of the toxic part decidedly alka-
line with sodium bicarbonate and allowing it to stand for
some time, preferably at incubator temperature. With
this combined poison, one is able to give much larger doses
without producing a fatal result, and, moreover, the effect
of the poison is in this instance manifested over a much
longer period of time. These two factors are, of course, of
primary importance in the production of immunity and it
THE PRODUCTION OF ACTIVE IMMUNITY 139
is quite possible that with the employment of the combined
poison a higher degree of immunization may be obtained.
That it is possible by means of repeated doses to induce a
certain amount of tolerance in animals for the poisonous
portion is illustrated in the following tables:
TABLE VII
20 MG. OF THIS POISON INVARIABLY CAUSED DEATH IN UNTREATED PIGS
WITHIN ONE HOUR
Total
Weight
Dose of poison in mg.
and when
given.
amount
Guinea-
pig No.
in
grams.
3/2 3/7
3/15
3/28
3/28
4/1
4/7
4/14
4/22
poison
in mg.
1
815
15.0 20
25.0
30.0
30
30
35
0
40.0
45
270
2
705
12.5 ..
15.0
17.5
20
25
30
.0
35.0
40
195
3
575
10.0 ..
15.0
20.0
20
25 1
30
,0
35.0
40
195
4
700
10.0 ..
15.0
20.0
20
25
30
0
35.0
40
195
5
580
9.0 ..
12.5
15.0
20
25
27
5
30.0
35
174
6
790
10.0 ..
15.0
20.0
25
25
30
0
35,0
35
195
7
545
7,5 ..
15.0
20.0
20
25
30,
0
32.5
35
185
8
655
10.0 ..
15.0
20.0
20
25
30
0
35.0
40
195
9
640
15.0 ..
20.0
25.0
25
30
35
0
40.0
45
235
The following table furnishes an index to the degree of
tolerance established in rabbits through the administration
of gradually increasing doses of the poison:
TABLE VIII
350 MG. OF THIS POISON CAUSED DEATH IN UNTREATED ANIMALS
WITHIN ONE HOUR
Weight Dose of poison in mgs. and when given.
Rabbit
No. grams. 12/12 l2/^ ^2/2» 1/9 1/14 1/19 1/26
1 1850 200 200 400 500 700 1000 1400 Died 20 min. after
last injection.
2 2800 250 300 400 500 700 900 1000
3 1950 300 250 400 600 700 ........ Died 30 min. after
last injection.
4 2100 200 250 400 500 700 1000 1200 Died 35 min. after
last injection.
From a study of the above tables it can be seen that in
the case of both guinea-pigs and rabbits after the adminis-
tration of several doses of gradually increasing strength,
a point is reached at which the animal is able to withstand
140 PROTEIN POISONS
the injection of from two to three times the amount which
would surely have proved fatal for an untreated control.
This would indicate that during the course of the treatment
the animal had developed either a slight degree of immunity,
or had established a certain amount of tolerance for the
poison. Which of these explanations is the correct one
can only be determined after a careful study of the subject
of the possible production of passive immunity and the
demonstration of a possible antibody in the blood of treated
animals. At present, owing to the slight amount of increased
resistance which the animals exhibit to the action of the
poison, we are inclined to believe that the question is one
of tolerance. Although the degree or tolerance thus far
secured has been limited, we do not feel justified in con-
cluding that greater resistance to the poison may not be
obtained. There are many factors of primary importance
in this work, all of which must be carefully studied before
definite conclusions can be drawn. For example, the
interval of time which is allowed to elapse between the
injections is a matter of first importance. Since the length
of time over which the poison acts is apparently so short,
it seemed quite probable that any reaction which might
occur on the part of the body would develop in a compara-
tively short time after the injection. With the object of
ascertaining whether this was true or not, animals were
treated daily with gradually increasing doses with the
following results:
TABLE IX
60 MG. OF THIS POISON INVARIABLY CAUSED DEATH IN UNTREATED
ANIMALS WITHIN ONE HOUR
Pig No. 1. Pig No. 2. Pig No. 3. Pig No. 4. Pig No. 5.
Day. Wt., Dose, Wt., Dose, Wt., Dose, Wt., Dose. Wt., Dose,
gm. mg. gm. mg. gin. mg. gm. mg. gm. mg.
1
425
45
385
45
460
45
390
45
405
45
2
380
50
360
50
415
50
355
50
385
50
3
385
60
375
60
420
60
365
60
385
60
4
385
80
375
80
450
80
370
80
380
80
5
405
100
370
100
450
100
375
100
390
100
6
405
112
385
112
460
112
385
112
7
420
125
400
125
470
100
410
112
Died in 30
Died in 30
Died in 30
minutes.
minutes.
minutes.
THE PRODUCTION OF ACTIVE IMMUNITY 141
From this we see that it is possible to establish a certain
amount of tolerance by means of daily injections of the
poisonous portion. Here again we find that it is compara-
tively easy to reach a dose which corresponds to about
twice the fatal amount, but above this the animal cannot
be carried. When death does result from a dose of the
poison which is too large to be borne by the treated pig, the
symptoms are identical in all respects with those which
occur in the case of an untreated animal, and a fatal result
follows in the same length of time.
The question now arose as to whether these animals
which had acquired a tolerance for the poisonous portion of
the colon bacillus were more resistant to inoculation with
the living germ than were untreated animals. In order to
ascertain this point, guinea-pigs which had received from
174 to 235 mg. of the toxic portion were inoculated intra-
peritoneally with doses of the living germ with the following
results :
TABLE X
1 c.c. OF A 16-HOUR CULTURE OF THE COLON BACILLUS USED IN THESE
EXPERIMENTS INVARIABLY KILLED A CONTROL WITHIN
TWENTY-FOUR HOURS.
No. of in-
Total
amount
Interval be-
tween last
injection and
Guinea- jections
of poison
inoculation
Amount and age of
pig No.
of poison.
received.
with germ.
culture.
Result.
1
9
270 mg.
11 days
1 c.c.
24-hour culture
Recovery.
2
8
195 mg.
11 days
1 c.c.
24-hour culture
Recovery.
3
8
195 mg.
11 days
1 c.c.
4-day culture
Died in 22 hrs.
4
8
195 mg.
11 days
2 c.c.
4-day culture
Dead in 24 hrs.
5
8
174 mg.
15 days
2 c.c.
24-hour culture
Recovery
6
8
195 mg.
25 days
2 c.c.
24-hour culture
Recovery
7
8
185 mg.
33 days
2 c.c.
24-hour culture
Recovery
8
8
195 mg.
25 days
2 c.c.
24-hour culture
Recovery
9
8
235 mg.
8 days
2 c.c
24-hour culture
Recovery
That an active immunity to the living colon bacillus is
also developed in rabbits which have been treated with
repeated injections of the toxic portion is illustrated by the
following experiments:
142 PROTEIN POISONS •
Rabbit No. 1 received between May 25 and July 5 eight
injections of the toxic part, the total amount of poison
injected being 855 mg. On July 8 this animal received
5 c.c. of a twenty-four-hour culture of the living germ
without apparent effect. The control inoculated at the
same time was found dead in eight hours.
Rabbit No. 2 received between May 28 and July 18
eleven injections of the toxic portion, the total amount of
poison given being 2475 mg. On July 27 this animal was
inoculated with 5 c.c. of a sixteen-hour culture of the colon
bacillus without effect. The control was found dead in
ten hours.
Rabbit No. 3 received between June 27 and July 18 seven
injections of the poison, the total amount given being 2100
mg. On July 28 this animal was inoculated with 6 c.c. of
a twenty-four-hour colon culture. Recovered.
Rabbit No. 4 received the same amount of poison as the
preceding one. Twelve days after his last treatment this
animal was given 6 c.c. of a forty-hour colon culture and
recovered. The control was found dead in eight hours.
From the foregoing experiments it becomes evident that
animals which have been treated with the toxic portion of
the colon bacillus acquire a certain degree of immunity to
the living germ. We are as yet unable to state whether it
is possible to obtain a high degree of immunity with the
poisonous portion or not. Thus far we have had animals
which have withstood inoculation with from two to four
times the fatal dose of the living germ. It is worthy of
note in this connection that animals which have received
one injection of a non-fatal dose of the poison are able to
withstand inoculation with twice the lethal dose of the
living germ on the following day. The immunity which
follows a single injection is, however, exceedingly transitory,
and has usually disappeared on the second day following
the treatment. This would seem to be the most marked
difference between the immunity which results from a
single injection of the toxic part and that which follows a
series of injections extending over a considerable interval
THE PRODUCTION OF ACTIVE IMMUNITY 143
of time. In the first instance the protection afforded is
very temporary, while in the second it is still present even
after the lapse of from twenty-five to thirty days. It may
be possible that in the case of the injections extending over
a long period of time the immunity obtained is of a higher
degree. This is a point which will require further study.
In the case of an animal which has been treated with-
the poisonous portion and has subsequently received a dose
of the living germ which would surely have proved fatal
for a normal animal, the symptoms noticed are identical
— Curve of normal animal inoculated with non-fatal dose of living bacillus.
• — Curve of immune animal inoculated with living bacillus.
in every respect with those which follow the injection of a
non-lethal dose in an untreated animal. This is a very
important fact and one on which we have laid much stress.
Moreover, as can be seen from Fig. 8, the temperature
curve corresponds very closely with that obtained in the
case of a normal animal inoculated with a non-fatal dose
of the living colon bacillus.
We see that in both instances there is no appreciable fall
in body temperature until from six to eight hours after
inoculation. At this time the minimum temperature has
144 PROTEIN POISONS
been reached in each case, and within from ten to twelve
hours it has again returned to normal. The similarity of
the symptoms in the two instances leads us to believe that
in all probability we are here dealing with an immunity
which is identical in character with that which is usually
spoken of as natural immunity. This idea has been further
upheld by the fact that we have been able to obtain from
egg albumen and peptone poisonous substances which
resemble the toxic portion of the colon bacillus in their
action, and by the injection of single non-fatal doses of
which it is possible to obtain the same transitory immunity
to the living colon germ as is observed after the injection
of the colon poison. That this toxic group is common to
certain bacteria and other protein bodies is not improbable,
and this would furnish an explanation not only of the
increased resistance to certain bacterial infections occurring
in animals treated with albumin and peptone, but of some
phases of natural immunity as well. However, this subject
will be more fully considered in a future paper on a compari-
son of these various poisons. It may be well to reiterate
at this point that we have conclusively shown that the
poison which we have been able to obtain from the colon
bacillus, and to which death is due in colon infection,
does not come directly from the peptone or albumen in
the culture medium, since we have obtained the same
poison from the bacillus when grown upon a protein-free
medium.
2. Immunization with the Residue Remaining after the
Extraction of the Poison from the Colon Bacillus. — The residue
remaining after the extraction of the poison from the
colon bacillus, which is insoluble in alkaline absolute alcohol,
is soluble in water. The resulting solution is, however,
quite decidedly alkaline in reaction, owing to the presence
of free alkali. Since it is essential that we should avoid
the irritative effects which would follow the injection of
this free alkali into the peritoneal cavity, the solution is
first rendered slightly acid with hydrochloric acid, and
then neutralized with sodium bicarbonate before injection.
THE PRODUCTION OF ACTIVE IMMUNITY 145
The solution of the residue thus obtained after sufficient
extraction with alkaline alcohol is non-toxic in the ordinary
sense of the term. However, the toxicity of a substance
for the body as a whole depends largely upon whether the
cells which it attacks are of fundamental importance in
maintaining the life of the animal or not. Thus a poison
which possesses a special affinity for the cells of the respira-
tory centre will inevitably lead to the production of marked
symptoms of poisoning on the part of the animal, while
one which exerts its effect upon the blood or connective-
tissue cells would not necessarily do so. Of course, in the
latter case, treatment over a prolonged period of time would
undoubtedly result in symptoms of chronic poisoning.
The residue is as potent a cell poison as is the toxic portion,
but the cells which it poisons are not directly concerned in
the carrying on of a function, the cessation of which would
prove immediately fatal to the organism as a whole. That
the residue is possessed of but slight toxicity is seen from
the fact that the injection of from 300 to 400 mg. into the
peritoneal cavity of guinea-pigs at a single dose has appar-
ently no effect upon the animal. There is no fall of tem-
perature such as is observed after the injection of the
poisonous portion, nor, on the other hand, is there any
appreciable rise. It may be well to emphasize at this point
that in order to study this portion of the colon bacillus and
its action it is absolutely essential that the toxic portion
should have been completely removed. In order to accom-
plish this it is necessary to extract the bacterial cell sub-
stance at least three times with the alkaline alcohol, and
frequently a fourth extraction is required. If all of the
poisonous portion has not been removed, the treated animal
begins to show evidences of poisoning, as lowering of tem-
perature, stupor, and, provided the extraction has been
very imperfect, death. These symptoms do not, however,
become manifest to a marked degree until from two to
four hours after the injection. This is in marked contrast
to the rapidity with which the free poison acts, and would
indicate that the poison in the imperfectly extracted residue
10
146
PROTEIN POISONS
still exists in combination with other constituents of -the
bacterial cell.
The question now arose as to whether the animals treated
with increasing doses of the residue had acquired any
immunity to infection with the living colon bacillus. To
ascertain this point, guinea-pigs were treated with this
portion and subsequently inoculated with the living germ
with the following results:
TABLE XI
Nos. 1, 2, 3, AND 4 RECEIVED A CULTURE, 1 c.c. OF A 12-nouR CULTURE
OF WHICH CAUSED DEATH IN UNTREATED PIGS WITHIN TWENTY-
FOUR HOURS. THE REST RECEIVED A CULTURE, \ c.c. OF WHICH
INVARIABLY PROVED FATAL TO UNTREATED ANIMALS.
No. of
Total
Time between
Guinea-
injections
amount.
last injection
pig
of
received
and inocula-
Amount of culture
No.
residue.
in gm.
tion.
injection.
Result.
1
. 9
0.29
16 days
1 c.c. 16-hour culture
Recovery
2
. 9
0.26
15 days
1 c.c. 24-hour culture
Recovery
3
. 9
0.3
16 days
2 c.c. 16-hour culture
Recovery
4
. 2
0.3
4 days
2 c.c. 16-hour culture
Death
5
. 4
0.9
3 days
2 c.c. 18-hour culture
Recovery
6
. 3
1.0
5 days
2 c.c. 24-hour cultu e
Recovery
7
. 3
1.0
5 days
3 c.c. 24-hour cultu e
Recovery
8
. 4
0.8
7 days
3 c.c. 20-hour cultu e
Recovery
9
. 4
0.9
14 days
3 c.c. 20-hour cultu e
Recovery
10
. 5
1.1
17 days
3 c.c. 16-hour cultu e
Death
11
. 4
0.8
30 days
3 c.c. 16-hour culture
Death
12
. 4
0.9
4 days
4 c.c. 16-hour culture
Recovery
13
. 4
0.8
7 days
4 c.c. 16-hour culture
Recovery
14
. 4
0.9
14 days
4 c.c. 18-hour culture
Death
15
. 4
0.9
5 days
5 c.c. 16-hour culture
Death
16
. 4
0.8
7 days
6 c.c. 16-hour culture
Death
From the above table it will be seen that guinea-pigs
which have been treated with that portion of the colon
bacillus which is represented by the residue have acquired
an active immunity to at least eight times the fatal dose of
the living germ. The degree of immunity produced does
not depend so much upon the amount of residue which
has been injected as upon the number of treatments and
the interval of time over which they have been continued.
THE PRODUCTION OF ACTIVE IMMUNITY 147
For example, No. 3, which received a total amount of 0.3
gram in nine doses, was able to withstand 2 c.c. of a sixteen-
hour culture after an interval of sixteen days, while No. 4,
which received the same total amount in two doses, suc-
cumbed to the injection of 2 c.c. of a sixteen-hour culture
given at an interval of four days after the last dose of
residue. Again we notice that the length of time over which
the immunity lasts is rather short in the case of animals
which have received a large amount of the substance in a
few doses continued over a short period. Thus the immunity
to 3 c.c. of a culture, 0.5 c.c. of which proved fatal for
untreated pigs, was lost between the fourteenth and the
seventeenth day following the injection of the last dose of
residue, and that to 4 c.c. of the same culture disappeared
between the seventh and the fourteenth day.
That it is possible to secure active immunity to the living
colon bacillus in rabbits by the injection of the colon residue
is shown in the following experiments:
Rabbit No. 1 received on August 13 a solution which
contained 1 gram of the residue. On August 20 a second
injection of 2 grams was given. Eighteen days later this
animal was inoculated with 5 c.c. of an eighteen-hour culture
of the colon bacillus. From two to four hours after injec-
tion this rabbit was apparently very sick, the symptoms
resembling those which are seen following the injection
of the toxic portion. However, it then began to improve,
and eventually completely recovered. A control inoculated
at the same time with an equal amount of the same culture
died in five hours.
Rabbit No. 2 received doses of 1 gram of the residue on
October 5, 10, and 20. Four days after the last injection
this animal was inoculated with 5 c.c. of an eighteen-hour
culture from which he recovered. A control given the
same dose at the same time died within five hours.
Rabbit No. 3 had the same treatment as did No. 2, and
six days after the last dose of residue withstood 5 c.c. of
an eighteen-hour culture. Eight days later this animal
received 10 c.c. of a sixteen-hour culture and did not die.
148 PROTEIN POISONS
Rabbits No. 4 and 5 had the same treatment as did Nos.
2 and 3, and seven days after the last dose each received
5 c.c. of a twenty-four-hour culture of the living colon
bacillus without a fatal result. A control which received
the same amount of the same culture died within five hours.
When we turn our attention to the symptoms which
follow the injections of living cultures of the colon bacillus
into animals which have been actively immunized with the
split products, we find that the clinical picture differs
materially according as to whether they have been treated
with the toxic portion or with the residue. As has been
previously mentioned, the symptoms which are observed
in a pig immunized with the toxic part are apparently
identical with those which one sees in a normal animal
after inoculation with a non-lethal dose of the living bacillus.
The picture which is obtained on inoculation of animals
rendered immune by previous treatment with the residue
is, however, quite different. In this case the animals
become apparently very ill within an hour after inoculation
with the living germ. Indeed, so noticeable was this fact
and the treated pigs appeared so much sicker than did the
controls that our first thought was that we had in some
manner increased their susceptibility to subsequent infec-
tion by treatment with the residue. However, after from
six to eight hours the treated animals appeared in much
better condition and eventually recovered, whereas the
controls invariably died. The temperature of the treated
animals runs a course entirely in accord with the symptoms.
Thus in a pig which had received 290 mg. of the residue,
and subsequently was inoculated with 1 c.c. of a twenty-four-
hour culture, the temperature fell from 100° to 95° F. within
four hours, and by six hours had once more begun to rise,
as is illustrated in Fig. 9.
The difference between the behavior of animals treated
with the toxic part and those which have been treated with
the residue toward cultures of the living germ is easily
explained if we consider the fact that in the first case we
are dealing with an animal which has acquired a certain
THE PRODUCTION OF ACTIVE IMMUNITY 149
amount of tolerance for the intracellular poison of the
colon bacillus as represented by the toxic part. In the
case of animals treated with the residue, however, no
tolerance for the poison contained within the colon bacillus
has been developed. If now the process which takes place
in both instances is a bacteriolytic one, it results that in the
case of the animal immunized with the toxic group the
effects of the poison contained within the bacterial cell and
liberated upon its disintegration will not become manifest
FIG. 9
103
102°
101
100°
99°
98
97°
96°
95°
The temperature curve of an animal immunized with the colon residue and
afterward inoculated with twice the lethal dose of the living culture.
until a sufficient amount of poison has been set free to
overcome the tolerance which the animal has attained
during the process of immunization. In the case of the
animal immunized with the residue there is no tolerance
to be overcome other than that which is present in all
animals, and the effects of the poison liberated through
bacteriolysis become apparent sooner and to a more marked
extent. Again, the fact that bacteriolysis may occur more
rapidly in the case of residue pigs than in those immunized
with the toxic group might explain in part the difference
150 PROTEIN POISONS
in behavior in the two cases. This is a point on which we
are as yet unable to give any definite results.
In order to study the differences in reaction to the living
germ in animals treated with the toxic part and those
immunized with the residue it is not only essential that
they should receive the same amount of the same culture,
but the dose given should not exceed twice that which would
prove fatal for a control. When a larger amount of the
living culture is given the differences are by no means so
clearly defined, although even in this case the animal which
has been treated with the residue shows symptoms of
severity at a much earlier time. As can be seen from Fig.
9, the temperature of a residue pig which had been inocu-
lated with twice the fatal dose of a living colon culture had
begun to rise at an interval of six hours after injection.
However, if an animal which has been rendered immune
by treatment with the residue is inoculated with six to
eight times the fatal dose of the living culture we find that
the temperature curve obtained is somewhat different in
character. The temperature falls with the same initial
rapidity, but instead of showing an early rise it continues
for some time at a low point and it is only at the end of
from eight to ten hours that any appreciable rise is mani-
fest. This we think is due to the fact that there has not
been enough of the bacteriolytic substance directly avail-
able to destroy all the bacilli contained in the large
amount of culture injected. The remainder of the germs
are destroyed by the same factors which are operative in
normal animals after the injection of a non-fatal dose of
the living bacillus. As can be seen from Fig. 8, it is only
after an interval of six to eight hours that there is any
appreciable fall in temperature in the case of a normal
animal inoculated with a non-fatal dose of the living culture.
This, we think, indicates that it is not until this time that
any appreciable amount of poison is liberated by bac-
teriolysis since, as we have seen in a previous chapter,
one of the first signs of the action of the intracellular poison
is a fall in body temperature. In a pig which has been
THE PRODUCTION OF ACTIVE IMMUNITY 151
immunized with the residue and subsequently inoculated
with a large amount of the living germ, we obtain evidence
of hypothermia at a much earlier period, owing to the
fact that bacteriolysis takes place very rapidly since the
bacteriolytic substance is present in a form available
for immediate use. If, however, the amount of this sub-
stance directly available is not sufficient to cause death
and bacteriolysis of all germs present, those bacilli which
remain are still capable of further reproduction. The same
mechanism which causes destruction of the bacteria in
normal animals, and which is probably connected with
the phenomenon of phagocytosis is, however, still operative
in the immune animal. Thus we shall have two influences
at work in the immune animal to cause bacteriolysis, one
acting rapidly, and the other manifesting its action only
after a considerable interval of time. We should there-
fore expect theoretically that we would find in the im-
munized animal a marked fall in temperature at an early
time, due to the setting free of the poison from the bodies
of the bacteria disintegrated by the directly available
bacteriolytic substance followed by a secondary rise, and
a succeeding fall due to the liberation of the poison by
means of the factors present in the normal animal. How-
ever, this is not actually the case, since the effect of the
poison liberated at first has not worn off before the second
period of bacteriolysis becomes well established. Conse-
quently, the intermediate rise of temperature is absent.
The results which follow the injection of the dead bac-
terial substance into animals immunized with the residue
are very interesting. As has been previously mentioned,
whereas animals treated with the residue develop an active
immunity to colon infection, they do not possess any greater
degree of tolerance for the colon poison than do untreated
animals. This is shown by the fact that the fatal dose of
the soluble poison is the same for the treated pig as for the
untreated control. This would lead to the belief that the
immunity obtained to the living colon bacillus is, in the
case of the residue animals, purely a bacteriolytic one. If
152 PROTEIN POISONS
this is true, one would suppose that on the injection of a
fatal amount of the dead bacterial substance death would
occur more rapidly in the immunized than in a normal
pig, provided the immune animal possesses a sufficient
amount of bacteriolytic substance directly available to
cause disintegration of all bacteria present. If, now, a
pig which has been immunized with the residue receives
5 c.c. of a twenty-four-hour culture of the colon bacillus
which has been deprived of life by means of heat, the
animal is very sick within from fifteen to twenty minutes.
The symptoms noted are similar in all respects to those
which are observed after the injection of the soluble poison.
The pig runs about the cage, scratches itself, and shows
the same evidence of lack of coordination and partial
paralysis of the hind extremities. This behavior is in
marked contrast to that seen in the case of a normal animal
which has received an injection of 5 c.c. of a twenty-four-
hour culture which has been rendered sterile by means
of heat. In this instance the animal appears perfectly
well until after the lapse of about an hour, when it begins
to show signs of illness such as roughening of the coat,
stupor, and indications of a beginning peritonitis. The
latter symptoms are those which we have described as
being due to the slow liberation of the poison from a com-
bined state. The same symptoms are observed in the case
of the immune pig, and are noticed at the same length of
time after the injection. The difference in the behavior
of the immune and the normal pig is seen to consist in the
fact that in the first instance we have symptoms of the
action of the free poison shortly after the injection of the
dead culture, which are entirely lacking in the second case.
This shows beyond doubt that in the immune pig there is
marked bacteriolysis of the dead bacilli and a consequent
liberation of the contained poison shortly after the injec-
tion of the dead bacterial cell into the peritoneal cavity.
Although we have as yet been unable to actually cause
death in an immune pig at an early period, the animals
are in every instance very ill within thirty minutes after
THE PRODUCTION OF ACTIVE IMMUNITY 153
the injection of the dead culture. In fact, several of them
have shown signs of the commencement of the convulsive
stage as evidenced by slight convulsive movements of the
head separated by considerable intervals of time. We have
been unable to secure a fatal result in these animals up to
the present time simply because we have worked with pigs
which did not possess a sufficient amount of bacteriolytic
substance directly available to cause disintegration of
enough bacilli to liberate a fatal amount of poison at one
time. It is worthy of note that this behavior of animals
immunized with the residue toward the dead bacterial sub-
stance furnishes additional proof of the fact that the poison
of the colon bacillus is an intracellular one. If the poison
existed, free in the culture medium we should expect that
the control would show evidences of its action at as early
a period as does the treated animal. However, as has
been stated above, this is not the case. The fact that the
treated animal shows symptoms of poisoning to a much
greater degree and at an earlier time than does the control
can be explained only on the ground that the poison with
which we are dealing is an intracellular one and is set free
only after the disintegration of the bacillus by bacteriolysis.
The question now arose as to whether the immunity
induced through the residue is specific for the colon bacillus
or not. In order to test this point, animals which had been
treated with this portion were inoculated with living cul-
tures of the typhoid bacillus with the following results:
TABLE XII
1 c.c. OF A 16-HOUR CULTURE OF THIS TYPHOID BACILLUS KILLED
CONTROLS WITHIN TWENTY-FOUR HOURS. 5 c.c.
DID NOT CAUSE DEATH
No. Total Time between
Guinea- injections amount last injection
pig
of
received
and inocula-
Amountof typhoid
No.
residue.
in gm.
tion.
culture injection.
Result.
1
. 4
0.8
6 days
1 c.c. 16-hour culture
Death
2
. 4
0.8
6 days
2 c c. 16-hour culture
Death
3
. 4
0.8
2 days
2 c.c. 16-hour culture
Death
4
. 4
0.8
6 days
3 c.c. 16-hour culture
Death
5
. 4
0.8
5 days
4 c.c. 16-hour culture
Death
154 PROTEIN POISONS
Although these experiments are by no means sufficient
in extent to warrant the conclusion that the injection of
the residue obtained from the colon bacillus furnishes no
increased resistance to typhoid infection, it can be seen
that the degree of immunity established must be very
slight. As far as the typhoid bacillus is concerned, the
immunity produced by the colon residue would appear to
be specific. If the immunity induced by the colon bacillus
is indeed specific, one would suppose that the immunizing
group is one which is found only in the residue obtained
from the colon germ.
As has been previously mentioned, we have found it
possible by treatment similar to that which we have used
in splitting up the colon bacillus to secure toxic substances
from egg albumen and peptone, which possess a similar
action when injected into the animal body, to that observed
after the injection of the colon poison. We have also
stated that the same transitory immunity to colon infec-
tion followed the injection of the albumin and peptone
poison as was obtained with the colon poison itself. The
albumen and peptone bear a further resemblance to the
bacterial cell substance in that the residue which remains
after alcoholic extraction is non-toxic. The question now
arose as to whether the injection of the albumen and pep-
tone residue afford any immunity to the living colon germ
or not. In order to ascertain this point animals were
treated with gradually increasing doses of these residues,
and • subsequently inoculated with the colon bacillus with
the following results:
THE PRODUCTION OF ACTIVE IMMUNITY 155
TABLE XIII
1 C.C. OF A
16-HOUR CULTURE OF COLON BACILLUS KILLED CONTROLS
IN TWENTY-FOUR HOURS. 5 c.c. DID NOT KILL
Peptone Residue
Total
Time between
Guinea-
No. injec-
amount
last injection
pig
tions of
received
and inocu-
Amount of colon
No.
residue.
in gm.
lation.
culture injection.
Result.
1
. 4
0.9
3 days
2 c.c. 16-hour culture
Death
2
. 4
0.9
4 days
2 c.c. 16-hour culture
Death
3
. 4
0.9
6 days
2 c.c. 16-hour culture
Death
4
. 4
0.9
3 days
2 c.c. 18-hour culture
Recovery
5
. 4
0.9
4 days
2 c.c. 18-hour culture
Recovery
6
. 4
0.9
5 days
2 c.c. 18-hour culture
Death
7
. 4
0.9
4 days
2 c.c. 16-hour culture
Death
8
. 4
0.9
6 days
3 c.c. 16-hour culture
Death
Guinea-
pig
Total
No. injec- amount
tions of received
TABLE XIV
Albumin Residue
Time between
last injection
and inocu- Amount of colon
No.
residue.
in gm.
lation.
culture injection.
Result,
1
. 4
0.9
3 days
1 c.c. 16-hou
culture
Death
2
. 4
0.9
5 days
1 c.c. 16-hou
culture
Death
3
. 4
0.9
3 days
2 c.c. 16-hou
culture
Death
4
. 4
0.9
4 days
2 c.c. 18-hou
culture
Death
5
. 4
0.9
5 days
2 c.c. 16-hou
culture
Death
6
. 4
0.9
5 days
3 c.c. 16-hou
culture
Death
7
. 4
0.9
4 days
4 c.c. 16-hou
culture
Death
From the above tables it can be readily seen that the
residues obtained from the peptone and albumin possess
little if any immunizing properties against infection with
the colon bacillus. In the case of the animals treated with
the peptone residue, the first three pigs received a different
residue from that given to the remainder. This residue
had been thoroughly extracted with alkaline alcohol, and
was evidently possessed of no immunizing properties what-
ever. On the other hand, the residue which was received
by pigs No. 4 to No. 8 inclusive had not been subjected to
so thorough an extraction. It is therefore highly probable
that the slight degree of immunity apparently obtained
in some of the latter animals was due to the presence of
156 PROTEIN POISONS
some of the toxic portion which had been left in the residue
as the result of incomplete extraction. In the animals
treated with the albumen residue we were unable to obtain
any evidence whatever of increased resistance to colon
infection.
It has now been shown that active immunity to the
colon germ can be produced in animals by treatment with
the split products obtained from this bacillus. This would
seem to furnish conclusive evidence that there exist within
the colon bacillus certain immunizing groups wrhich are
capable of being separated more or less completely from
the other constituents of the bacterial cell by means which
bring about a chemical cleavage of the latter. Furthermore,
it has been seen that the colon bacillus contains at least
two different groups, each of which when injected into the
animal body is capable of establishing a certain degree of
immunity toward subsequent infection with the living
germ. One of these groups is contained within the toxic
portion, and probably represents a group which is common
to many protein bodies, since, as has been shown, it is
contained in the poisons secured through the chemical
cleavage of egg albumen and peptone, as well as from the
colon bacillus. The degree of immunity thus far obtained
through the agency of this group is not great. The fact
that this group is apparently not specific to the colon
bacillus, but can also be obtained from other protein bodies,
furnishes an explanation of the increased resistance to
infection observed in animals previously treated with
solutions of egg albumen and peptone. Again it has been
shown that the residue, which, as has been previously
stated, is for all practical purposes non-toxic, also contains
an immunizing group. The immunizing group contained
in the residue differs from that found in the toxic portion
in one very important respect. It represents a group
which up to the present time we have been able to find
only in the colon bacillus, and which when injected into
animals affords protection against this bacillus alone. In
other wrords, the immunity produced with the residue is
strictly specific in character. Moreover, the degree of
THE PRODUCTION OF ACTIVE IMMUNITY 157
immunity to the living germ obtained through the employ-
ment of the residue is apparently much higher than that
which follows treatment with the toxic portion. It does
not seem improbable that the specific immunizing group
which is contained in the residue represents that group of
the colon bacillus which is of primary importance in the
development of specific acquired immunity to this germ.
The work just detailed may be summed up as follows:
1. Guinea-pigs treated at intervals of from three to
four days with intra-abdominal injections of the colon
residue acquire an active immunity to at least eight times
the ordinary fatal dose of the living bacterium.
2. The degree of immunity secured does not depend so
much upon the amount of the residue or non-poisonous
portion that has been injected as upon the number of
treatments and the interval of time over which they have
been continued.
3. The length of time over which the immunity continues
is rather short in the case of animals that have received a
large amount- of the residue in a few doses continued over a
short period.
4. Rabbits that received from two to three injections of
0.5 gram each of the residue acquire an immunity to
quantities of the living bacillus that kill the controls within
five hours.
5. The immunity induced by the colon residue or non-
poisonous part is specific, and previous treatments of
animals with the residues of egg-white, peptone, and the
typhoid bacillus give no immunity to the colon bacillus.
In continuing this work Vaughan and Wheeler1 decided
in the first place to ascertain whether or not a single dose
of the residue gives any immunity; if so, what degree of
immunity does it afford and how long does it continue?
The number of immunizing doses given in the work
already reported ran from three to nine, and the special
object of the work here reported is to ascertain the effects
of a smaller number of immunizing doses.
1 New York Med. Jour. June, 29, 1907.
158 PROTEIN POISONS
TABLE XV
These animals had one dose of 50 mg. of the residue,
as the non-poisonous portion is designated. The protocol
number, the weight of the animal, the interval in days
between the administration of the residue and the inocula-
tion with the twenty-four hours beef-tea culture of the
bacillus, the amount of the culture and the result are shown :
Protocol, Weight, Interval, Amount given,
No. gm. days. c.c. Result.
324 400 1 5 Recovery
325 420 3 5 Recovery
269 505 4 3 Recovery
275 300 4 5 Recovery
326 315 7 5 Death
270 600 9 3 Recovery
328 345 11 2 Recovery
207 455 21 2 Recovery
209 495 27 4 Death
210 625 27 4 Death
211 485 28 3 Death
139 240 36 2 Recovery
140 250 40 3 Death
141 250 42 2 Death
142 255 42 2 Death
In studying these results it will be well to consider the
minimum fatal dose as the unit, which in case of the cultures
used in these experiments is 1 c.c. of the twenty-four hour
beef -tea growth, and we will regard the animal that suc-
cumbs to 2 c.c. as having practically lost its immunity.
With this measure it will be seen that a single dose of 50
mg. of the colon residue gives to the animal a temporary
immunity of at least 5 units, which is in force twenty-four
hours after the treatment, and continues for at least four
days, but has begun to disappear by the seventh day.
However, some slight degree of immunity continues up to
the thirty-sixth day, but practically all is lost by the
fortieth day.
TABLE XVI
These animals had a single dose of 25 mg. of the residue.
The data are the same as given in the preceding table.
THE PRODUCTION OF ACTIVE IMMUNITY 159
Protocol, Weight, Interval, Amount given,
No. gm. days. c.c. Result.
330 460 1 5.0 Recovery
331 300 3 5.0 Death
332 225 7 5.0 Death
333 290 9 2.5 Recovery
334 235 11 2.0 Death
335 360 14 1.0 Recovery
Comparing Tables XV and XVI, it will be seen that the
immunity given by 25 mg. of the residue, although it may
be as great as that given by 50 mg. at the end of the first
twenty-four hours, declines more rapidly and is less at the
end of three days, and continues to be less at eleven days.
The only element of doubt that we can see in these con-
clusions lies in the small size of all the animals, save No. 1
used in Table XVI. However, we have not found that size
or weight of guinea-pigs are important factors, provided,
of course, that the animals are in good condition, in influ-
encing the result after inoculation with the colon or the
typhoid bacillus.
TABLE XVII
These animals had from two to three treatments, receiving
each time 50 mg. of the residue. These treatments were
at intervals of three days. The protocol number, the weight
of the animals, the number of treatments, the total amount
of residue given, the interval in days between the last
treatment and the inoculation, the amount of the culture
twenty-four hours old, and the result are given:
Amount of Amount of
Protocol, Weight, No. of residue, Interval, culture,
No. gm. treatments. mg. days. c.c. Result.
270 600 2 100 33 Recovery
271 600 2 100 34 Recovery
212 520 2 100 33 Recovery
213 530 2 100 3 4 Recovery
272 530 3 150 5 4 Recovery
273 540 3 150 5 5 Recovery
274 475 3 150 7 6 Recovery
278 580 3 150 7 6 Recovery
280 615 3 150 7 6 Recovery
214 585 3 150 12 5 Recovery
215 600 3 150 12 6 Recovery
216 485 3 150 12 6 Recovery
217 530 3 150 12 6 Recovery
218 475 3 150 12 7 Recovery
160 PROTEIN POISONS
Comparing Table XVII with Tables XV and XVI, it is
plainly evident that this immunity induced by two and three
doses at intervals of three to four days is greater in degree
and more lasting in its effects than that produced by a
single injection. This confirms the conclusion already
stated, but at the same time this additional work shows
that a single dose may serve to furnish protection against
at least five times the ordinary fatal dose for a few days
and against twice the fatal dose for one month.
TABLE XVIII
These animals received a single dose of 100 mg. of the
typhoid residue. The protocol number, weight of animal,
interval between treatment and inoculation, amount of
twenty-four-hour culture given, and the result are shown:
Amount of
Protocol, Weight, Interval, culture,
No. .. . gm. days. c.c. Result.
354 250 1 3 Recovery
355 320 1 4 Death on second day
356 280 3 3 Death
357 265 3 4 Death
358 270 6 2 Death
359 230 6 1 Recovery
219 570 28 2 Death
TABLE XIX
These animals received a single dose of 50 mg. of the
typhoid residue.
Amount of
Protocol, Weight, Interval, culture,
No. gm. days. c.c. Result.
336 360 1 3 Recovery
337 285 3 3 Death
338 245 7 3 Death
339 350 9 2 Death on second day
340 290 11 1 Recovery
341 255 14 1 Recovery
THE PRODUCTION OF ACTIVE IMMUNITY 161
TABLE XX
These animals received two and three immunizing doses
of the typhoid residue. The protocol number, the weight,
the number of immunizing doses, the interval in days between
the last treatment and the inoculation, the amount of
culture twenty-four hours old, and the result are given:
Protocol, Weight, No. of residue, Interval, culture,
No. gm. treatments. mg. days. c.c. Result.
220 375 2 100 32 Recovery
221 495 2 100 33 Death
222 650 3 150 52 Recovery
223 480 3 150 5 3 Recovery
224 435 3 150 53 Recovery
144 605 3 150 74 Recovery
145 665 3 150 13 4 Recovery
The minimum fatal dose of the twenty-four-hour cul-
ture of the typhoid bacillus employed in this case is 0.5
c.c. It will be seen from Table XVIII that a single dose of
100 mg. of the residue gives the animal an immunity of
six units at the end of twenty-four hours, and that the
immunity was less than eight units at that time. On the
third day the immunity was found to be diminishing, but
on the sixth day the animal bore twice the fatal dose. The
animal which received eight units at the end of the first
day evidently was close to the borderline, because it did
not die until the second day, and the normal time for an
untreated guinea-pig to live after receiving the minimum
fatal dose is less than twelve hours. Table XIX shows that
a single dose of 50 mg. is quite as efficient as one of 100 mg.
Table XX indicates that multiple immunizing doses give
a higher degree and a more lasting immunity than that
secured by a single dose.
Theoretical Considerations and Conclusions. — We wish to
offer certain theories that we have reached after making
these experiments. In order to save space we will condense
our views as follows:
11
162 PROTEIN POISONS
1. All the proteins with which we have worked contain
a poisonous group, and the probabilities are that this is
true of all proteins, be they bacterial, vegetable, or animal.
2. Proteins may be split into poisonous and non-poisonous
groups, either artificially in the retort or in the animal
body.
3. The splitting up of the protein in the animal body is
due to a proteolytic ferment which is the product of certain
cells.
4. This ferment is specific for the protein which calls it
into existence.
5. Our conception of the origin and nature of these specific
ferments is as follows: The cell is made up of molecules;
the molecules consist of atoms, and the atoms of electrons.
The molecule may be likened to the universe, composed
of suns, planets, and satellites. These are in harmonious
and rhythmic motion. The molecule of the foreign protein
introduced into the body has a structure similar to that
of the cell molecule, and when one is brought within the
attractive range of the other, one or the other, or both,
must undergo certain disturbances. Suppose that an
atomic group is split off from the animal cell and enters
the attraction sphere of the molecule of the foreign protein,
then the harmonious arrangement of the atoms and elec-
trons of the latter will be affected ; indeed, the molecule may
be disrupted as completely as it is in the retort under the
influence of dilute alkali.
Our residues evidently have the same effect on the cells
of the body that the proteins from which they come do,
and .in this way we may explain the specific action of the
residues. The special group broken off from the cell mole-
cule depends upon the composition of the protein which
comes within its attractive sphere, and it seems that our
residues contain that portion of the molecule of the foreign
protein which possesses this property of bringing into
existence, or rather of activating, its own specific ferment.
The ferment is, according to our conception, a portion of
the animal cell, an atomic group within the cell molecule,
THE PRODUCTION OF ACTIVE IMMUNITY 163
and does not become a real active ferment, or it is not
activated until the foreign protein comes within its sphere
of attraction. It occurred to us that if this theory has
much of truth in it we might test it. We thought that the
introduction of a small portion of the residue might give
some immunity immediately. We therefore injected doses
of 25 mg. of the colon and 12.5 mg. of the typhoid residue
into the abdominal cavity of guinea-pigs, and thirty min-
utes later inoculated these animals intra-abdominally with
living cultures, and found that the colon animals had in
that short time acquired an immunity of five units and
the typhoid one of six units. However, we found that
larger immunizing doses did not give us results so good,
and this is easily explained by supposing that this ferment
set free or activated by the residue is in part used up
in its reaction with the residue itself. The reason multiple
doses repeated at intervals give us a higher degree of im-
munity than single doses may be due to more cells being
acted upon or to the accumulation of the ferment in the
blood. The theory of the modus operandi of the residue
which we have offered is tentative, and we hope to be able
to investigate it further.
It will undoubtedly occur to the reader, as it has to us,
to ask the question how it is that the residue sensitizes or
activates, while the bacillus itself, living or dead, has no such
effect or at least is not nearly so effective. The only answer
that we can suggest to this question is that in order to be
effective in its action the sensitizer must be in solution,
and that being in this state it reaches every part of the
circulatory system in a few seconds. Possibly cell permea-
tion may be necessary to the most perfect sensitization.
CHAPTER VIII
THE SPLIT PRODUCTS OF THE TUBERCLE
BACILLUS AND THEIR EFFECTS
UPON ANIMALS1
The Organism. — The tubercle bacillus employed in the
experimental work herewith reported is one which, after
having been grown for many years on artificial culture-
media, has lost its virulence for rabbits and guinea-pigs.
We have repeatedly demonstrated this fact during the
past six years, but in order to have renewed evidence we
have, at the beginning of this research, inoculated 4 rabbits
and 5 guinea-pigs intra-abdominally with loops of the
glycerin beef-tea culture, and these animals having been
killed from three to six months after inoculation, have in
no instance showed any evidence of infection. On artificial
culture-media this bacillus grows abundantly, and in this
respect it has served our purpose in furnishing a large
amount of cellular substance. It has been grown in the
ordinary glycerin beef-tea and has been harvested after
periods of from one to six months. Growths obtained
after from one to two months at 37° have given us the
most satisfactory material.
The Cellular Substance. — The bacterial substance is
collected on hard filters, dried between folded filters, and
thoroughly extracted, first with alcohol and then with
ether, in large Soxhlets. In this paper we will say nothing
concerning the fats and waxes extracted with alcohol and
ether. The cellular substance is next rubbed up in a mortar
and passed through a fine-meshed sieve. As thus prepared,
1 The first part of this chapter is taken from a paper read by Vaughan
and Wheeler before the International Congress on Tuberculosis in 1908.
THE SPLIT PRODUCTS OF TUBERCLE BACILLUS 165
the powder shows the bacilli more or less broken into a
cellular debris when examined microscopically. The indi-
vidual bacilli take the carbolic stain, but this is now
washed out with dilute nitric acid.
The Cleavage of the Cell. — The cellular substance, pre-
pared as stated above, is placed in large flasks, fitted with
reflux condensers, covered with from fifteen to twenty
times its weight of absolute alcohol, in which 2 per cent,
of sodium hydroxide has been dissolved, and heated for
one hour at 78°, the boiling-point of absolute alcohol.
Three successive extractions are made, using a new portion
of alkaline alcohol each time. This treatment splits the
cellular substance into two portions — one soluble and the
other insoluble in absolute alcohol. These portions we will
designate as the "cell poison" and the "cell residue."
The Cell Poison. — This is soluble in the alcohol, which is
carefully neutralized with hydrochloric acid. The precipi-
tated sodium chloride is removed by filtration and the
filtrate containing the poison is evaporated in vacua at or
below 40°. This leaves the cell poison as a brownish mass
containing a small amount of sodium chloride, which can
be removed by repeated solutions in absolute alcohol and
evaporation. The poison resembles that obtained from other
protein bodies. It is freely soluble in absolute alcohol ; less
freely in water. Its aqueous solutions give all the color pro-
tein reactions, including that of Molisch, which is not given
by the poisonous groups that we have obtained from other
proteins. In powder form it is deliquescent and becomes
darker as it absorbs water. The tubercle protein apparently
contains much less poison than the cellular proteins of the
colon and typhoid bacilli. The latter are split up by our
method into about one-third poison and two-thirds residue,
while 25 grams of the cellular substance of the tubercle
bacillus yielded less than 3 grams of poison.
The Cell Residue. — This is the portion insoluble in the
alkaline alcohol. It is placed in Soxhlets and extracted for
many hours with absolute alcohol in order to remove
traces of the cell poison and free alkali. After this it is
166 PROTEIN POISONS
dried and powdered. It is partially soluble in water, and
the soluble part constitutes that which we have used in our
experiments.
The Bacterial Filtrate. — After the bacilli have been
removed, the culture medium is concentrated on a steam-
bath to one-sixth its volume. This concentrated fluid is
poured into five times its volume of absolute alcohol, which
throws down a heavy, sticky precipitate. This precipitate
is placed in Soxhlets and extracted, first with alcohol and
then with ether. Next, it is powdered and split up with
alkaline alcohol after the method used with the cellular
substance. This breaks it up into poisonous and non-
poisonous groups, which we distinguish from the corre-
sponding bodies obtained from the cellular substance by
designating them as "the precipitate poison" and "the
precipitate residue."
The Precipitate Poison. — This differs in none of its physical
or chemical properties, so far as we have investigated, from
the cell poison.
The Precipitate Residue. — This is freely and wholly soluble
in water. It gives all the protein reactions and is precipitated
by uranyl acetate and metaphosphoric acid.
The Final Filtrate.- — In this manner we have designated
that portion of the culture-medium that remains after the
concentrated medium has been precipitated by five times
its volume of absolute alcohol. The alcoholic filtrate gives
a voluminous precipitate with an alcoholic solution of
mercuric chloride, showing that all the protein material
has not been precipitated by the alcohol. This filtrate is
freed from alcohol by distillation and has been used in
some of the animal experiments described later.
It will be seen that we have split up the tubercle cell
into two portions: the cell poison and the cell residue.
The culture-medium has been concentrated and then pre-
cipitated with five times its volume of absolute alcohol,
and this precipitate has been broken up into two portions:
the precipitate poison and the precipitate residue, and the
portion of the culture-medium left after the removal of
THE SPLIT PRODUCTS OF TUBERCLE BACILLUS 167
the alcoholic precipitate we have designated as the final
filtrate.
The Effect of the Cellular Substance on Animals. — It must
be borne in mind that the cellular substance with which
we are now dealing is that of a tubercle bacillus that is
avirulent to rabbits and guinea-pigs and that it has been
thoroughly extracted with alcohol and ether. There remains,
as it were, only the protein skeleton of the bacillus.
We have injected into the abdominal cavities of twenty-
four guinea-pigs single doses, varying in amount from 5
to 200 mg. of the cellular substance, and from these experi-
ments we make the following statements:
1. In no case was death caused directly by the injection.
One pig that received 20 mg. was found dead six days
later. There were several caseous nodules in the omentum
and one on the under surface of the liver. Microscopic
examination showed that these consisted of masses of
leukocytes and the debris of the injected bacilli. Another
that received 5 mg. was found dead nine days later, but
careful search failed to reveal any traces or effect of the
injection. Animals that received from 100 to 200 mg.
remain apparently well four months after the injection.
2. It gives in guinea-pigs no immunity to a subsequent
inoculation with a virulent bacillus. Six pigs that had
received single intra-abdominal injections of the cellular
substance in amounts varying from 15 to 200 mg. were
inoculated one month later with a loop of a virulent culture
of bacillus tuberculosis and all developed tuberculosis and
died from it within from nineteen to one hundred days.
3. It does, for a short time at least, sensitize guinea-pigs
to the tuberculosis bacillus. This is an interesting and,
in our opinion, a hopeful point. The following are illus-
trations of this action: Pig No. 159, weight 530 grams,
received, December 18, 25 mg. of the cellular substance.
Thirteen days later it was given intra-abdominally a large
loop of the avirulent culture suspended in salt solution.
The animal was sick within a few minutes. Within half
an hour it developed the first and second stages of anaphyl-
168 PROTEIN POISONS
axis. Within forty-five minutes its rectal temperature
had fallen to 96° F. and it was found dead the next morning.
Postmortem showed a hemorrhagic peritonitis. Pig No. 163,
weight 535 grams, received 45 mg. of the cellular substance
December 18. Its subsequent treatment and its results
were the same as recorded of the preceding animal. Pig.
No. 151, weight 555 grams, received, December 18, 125
mg. of the cellular substance. Twenty-three days later
it had intra-abdominally a large loop of the avirulent
culture. It died within sixteen hours and showed a
hemorrhagic peritonitis.
If we interpret these results correctly we infer that the
cellular substance had sensitized these animals and that
the bacilli of the second dose were broken up so rapidly and
their poisonous constituents set free so speedily that the
animals died. If this interpretation be correct, there
remains at least the possibility that there may be found
in the bacillary substance some constituent that may
stimulate the cells of the animal body to split up and
destroy tubercle bacilli. We will return to this before
we close.
The Effect of the Cell Poison on Animals. — This body,
obtained by splitting up the cellular substance with alkali
in absolute alcohol, is, like all similar bodies that we have
obtained from bacterial, vegetable, and animal proteins,
a poison. It develops the three stages of peripheral
irritation, partial paralysis, and terminal convulsions.
When given in sufficient quantity it kills within an hour
both healthy and tuberculous animals. When given to
healthy animals in very small repeated doses it has no
visible effect. In larger repeated doses it causes in healthy
animals a condition of chronic intoxication characterized
by loss of flesh and general marasmus. When given even
in very small repeated doses to tuberculous animals it
intensifies the tuberculous process, and in all cases the
treated animals die before the controls. There is no evi-
dence that it elaborates any antitoxin, and it is harmful, so
far as our experiments show, and we have made many with
THE SPLIT PRODUCTS OF TUBERCLE BACILLUS 169
this body upon both normal and tuberculous animals; it
has nothing to recommend it. What is true of the cell
poison is equally true of the precipitate poison and the
final filtrate. The effects of these poisons on animals are
harmful — only harmful.
The Effects of the Cell Residue on Animals. — This is the
non-poisonous group obtained by splitting up the cellular
substance with alkali in absolute alcohol. On healthy
animals it has no recognizable ill effect, either in single or
repeated doses, either large or small. In this product we
see the one small ray of hope of finding, among the split
products, a body that may possibly be of service in the
treatment of incipient and localized tuberculosis. Before
giving the basis of this slight hope we will tell what we have
done with this product.
In the first place, it sensitizes guinea-pigs to the tubercle
bacillus. The following are illustrations: Guinea-pig No.
199, weight 530 grams, received, December 20, 50 mg. of
the residue. Thirteen days later it had intra-abdominally a
large loop of the avirulent culture (the same amount given
to pigs 159, 163, and 151, see p. 167). This injection was
made at 11.30 A.M. At 12 M. the temperature had fallen
to 96° F. At 2.30 P.M. it was 97.1°, and at 5 P.M. it was
99°, and the animal was apparently well.
Pig No. 200, weight 530 grams, received 50 mg. of the
residue intra-abdominally. Thirteen days later it had
intra-abdominally one large loop of the avirulent culture.
The rectal temperature before the injection was 101° F.
Within half an hour it had fallen one degree, but went no
lower and the animal seemed to be but little disturbed.
We have similar records of other animals.
We compare these results with those recorded of animals
159, 163, and 151, and tentatively conclude that the sensi-
tizing agent in the cellular substance is the portion that
we have designated as the residue. But the avirulent
bacilli killed the animals sensitized with the cellular sub-
stance because when the bacteriolytic ferment was set
free or activated by the second injection, it split up not
170 PROTEIN POISONS
only the bacilli introduced by the second injection, but also
those remaining in the body from the first injection and
these together supplied enough free poison to kill.
It will require much experimentation to show what degree
of sensitization can be secured by the residue, what size
doses should be used, and how long the condition of sensi-
tization continues. If men, as well as guinea-pigs, can
be sensitized with the residue, there is the possibility that
it may be of service in the treatment of initial and localized
tuberculosis, because it may be used to bring into existence
and activate a specific bacteriolytic ferment which will
split up and destroy the few bacilli that are in the body,
but we can readily see that this might be harmful rather
than beneficial when the number of bacilli in the body is
large enough to furnish a dangerous amount of the poison
when set free. In this case the old adage that it is not wise
to disturb sleeping dogs might be remembered.
The Effect of the Precipitate Residue on Animals. — This
is the most interesting of the split products of the tubercle
bacillus, and it deserves much more study than we have
as yet been able to give it. On healthy animals it has no
recognizable ill effects either in single or repeated doses,
large or small. We took 6 half-grown pigs and injected
into the abdominal cavity every third or fourth day 50
mg. of this residue. These injections were begun April 20
and continued until June 11. During this time each animal
received sixteen injections, a total of 800 mg. each. All
increased normally in weight, and five days after the last
injection all were killed and carefully examined and found
to be perfectly normal. We were led to do this because of
the following experience: Three sets of pigs were inocu-
lated intra-abdominally with the living a virulent culture.
These inoculations were made December 14, 1905. The
animals of the first set had no treatment, and when killed
April 11, 1906, were found to be perfectly normal. Those
of the second set had, December 18, 50 mg., December 22, 60
mg., and December 26, 70 mg. of the cell residue. All in
this set were killed April 11, and were also found to be
THE SPLIT PRODUCTS OF TUBERCLE BACILLUS 171
wholly free from infection. Those of the third set had,
December 18, 30 mg., December 22, 75 mg., and December
26, 100 mg. of the precipitate residue. Half of this set died
before April 11 of tuberculosis, and the other half were
found to be tuberculous when killed on that date. To us
this indicates that the precipitate residue has some specific
effect upon tuberculous animals. We suspect that the ill
effect in these instances was due to the size of the doses,
because in reality the doses of the precipitate residue were
much larger than those of the cell residue — much larger,
indeed, than the figures indicate, because, as we have
stated, the cell residue is not freely soluble in water, while
the precipitate residue is wholly soluble. In making up our
solutions we weighed out each residue and, in fact, the
animals received only the soluble parts of the amounts
stated in the figures. We can easily understand how exces-
sive doses given soon after inoculation with the avirulent
culture might induce such a result. This culture is aviru-
lent because it makes only an ineffectual attempt to grow
in the animal body. The feeble effort is resisted and over-
come by the natural defences of the healthy body. Now,
if these natural defences were wholly occupied in disposing
of the material injected, which should have been only
sufficient to awaken these defences, then the bacilli would
meet with no resistance and would multiply.
The precipitate residue sensitizes guinea-pigs to the
tubercle bacillus just as the cell residue does. Evidently
our so-called residues are much alike, and it is more than
probable that they contain the same active constituent.
In cultures from three to six months old many of the bacilli
have undergone autolytic changes and the cellular sub-
stance has in part passed into solution. This is true of
both the poisonous and the non-poisonous groups of the
protein that makes up the cell substance. Of one thing
we have satisfied ourselves at least, and that is that no
preparation from the tubercle bacillus should be used in
the treatment of tuberculosis until the poisonous group of
the tuberculous protein and other proteins in the culture-
172 PROTEIN POISONS
medium be removed. This is too powerful a poison to be
injected repeatedly even in small doses into the animal
body.
One of us has for the past two years used solutions of
the cell residue in the treatment of tuberculosis in man.
The most suitable preparation is a 1 per cent, solution
filtered through porcelain. The cell residue in weighed
quantity is placed in a bottle with the proper volume of a
0.5 per cent, solution of carbolic acid, and the bottle is
carried on a mechanical shaker for twenty-four hours,
after which the content is passed through a porcelain filter.
Such a solution will keep indefinitely. We have used this
solution sufficiently to justify the following statements:
(1) It is of no value in advanced cases of pulmonary tuber-
culosis. (2) It may prove harmful even in initial cases if
the dose be too large or if small doses be too frequently
repeated. (3) When properly used in initial cases or in
localized tuberculosis, its action is apparently prompt and
specific. If the tubercle bacilli wholly disappear from the
sputum, as they may, the injections should be repeated
at intervals of from two to four weeks for some months.
We .wish it clearly understood that in well-established cases
of pulmonary tuberculosis no benefit from this treatment
can be expected. We believe that in initial cases this pre-
paration is preferable to any form of tuberculin.
Toxophor Group. — White and Avery1 have reported an
interesting research on the split products of the cellular
substance of the tubercle bacillus, especially of the toxo-
phor group. They used a strain virulent to guinea-pigs.
This was grown on glycerin broth cultures for six weeks,
and the cellular substance was washed with alcohol and
ether, ground in a ball mill and split up by our method.
The toxophor obtained by them agreed with that which
we have prepared. It is a yellowish-brown powder of
characteristic pungent odor, readily soluble in alcohol. Its
aqueous solutions are faintly turbid, and give the biuret,
1 Jour. Med. Research, 1912, xxvi, 317.
THE SPLIT PRODUCTS OF TUBERCLE BACILLUS 173
xanthoproteic, Adamkiewicz, Liebermann, Millon, and
Molisch tests; the last two faintly. Bromine water pro-
duces a white flocculent precipitate, but no color, showing
the absence of tryptophan. Injections were made into the
right external jugular vein of guinea-pigs of about 200
grams weight. The poison, as prepared, killed in doses of
1 to 15,000 body weight. White and Avery give such an
excellent statement of the symptoms and gross pathology
that we are induced to make the following quotation : " When
a quantity approaching the minimum fatal dose is given,
the first symptoms appear immediately, or, at most, within
thirty seconds. The animal becomes restless, scratches
its nose, and frequently utters a sharp hiccough. The
movements become incoordinate, the gait is unsteady.
The eyes are fixed, and stare. Respiratory embarrassment,
with diaphragmatic spasm sets in and increases to a degree
which causes the animal to spring from its feet, to buck,
and finally to fall on its side with convulsive twitching of
its legs, intermittent, and both clonic and tonic in character.
Involuntary micturition and defecation frequently take
place. The dyspnea becomes more marked, and then
ensue successive periods of apnea, lasting as long as twenty
to thirty seconds. These are followed by violent inspira-
tory efforts, during which the chest wall becomes fixed
in maximum inspiration. Cyanosis is noticeable in the
lips and ears, and becomes more marked. The convulsive
gasps increase in frequency and decrease in depth, until
finally only the lips move, the feeble and rapid dilatations
of the alse nasi marking the onset of death. This sequence
of symptoms is accompanied by a rapid and progressive
fall of the body temperature. Death takes place in from
one and one-half to six or seven minutes. Immediate
autopsy reveals first a cyanotic hue of the subcutaneous
and muscular tissues. The blood is dark in color and does
not clot readily. Beyond an exaggerated peristaltic move-
ment of the intestines, the abdominal viscera appear to
be normal. On opening the chest the lungs are found to
be in a state of maximum inflation, overlapping the peri-
174 PROTEIN POISONS
cardium, and forming a cast of the thoracic cavity. They
are pale and often slightly bluish in color, and frequently
exhibit punctate hemorrhages on the surface. The heart
still beats. Not infrequently there is definite heart block,
with an auriculoventricular arrhythmia of three to one.
Often there are petechial hemorrhages in the epicardium,
Greater extravasations are also seen, and in two cases
actual rupture of the ventricle had apparently taken place.
On section the lungs do not collapse, and on pressure only a
little frothy serum exudes. They are not edematous. They
float on water. The excised heart continues to beat for
several minutes. The gross appearance of the brain is
normal. A study of the pathological changes in the his-
tology of the lungs, heart, and brain has been undertaken,
but has not yet progressed sufficiently to warrant any
conclusions. When the dose is larger the acute symptoms
appear instantaneously, and their sequence is more rapid.
With a sublethal dose the onset is slower and the manifes-
tations are less violent. The animal shows evidence of
weakness, drops its hind legs, and frequently lies on its
side in collapse. The apneic stage is never reached, its
appearance therefore signifies inevitable death. Recovery
from a non-fatal dose is comparatively prompt even when
near the lethal borderline. Recovered animals exhibit no
visible sequela? of the intoxication."
These investigators have compared the acute intoxi-
cation produced in animals by the tuberculopoison with
anaphylactic shock, and conclude that there are no appre-
ciable points of difference in the symptomatology and gross
pathology of the two conditions. "They would therefore
appear to be identical." The tuberculopoison, like that
obtained from other proteins, is thermostabile. It also
agrees with the like poison obtained by the cleavage of
other proteins in the following particulars: (1) It lowers
the temperature when given in doses sufficient to produce
recognizable effects. (2) It does not sensitize animals
to the unbroken tuberculoprotein, while the haptophor
group is not poisonous and does sensitize to the whole
THE SPLIT PRODUCTS OF TUBERCLE BACILLUS 175
protein. (3) The injection of non-fatal doses of the poison
renders animals at least temporarily refractory to subsequent
injections of what would normally be fatal doses. We
have always held that this is due to the establishment of a
tolerance. This is important and we will refer to it again
when we discuss the action of tuberculin. (4) The poison
is not absorbed in vitro by brain, lung, or liver tissue. "These
experiments seem to emphasize the absence of any possible
identity of this protein fragment with the true toxins.
The results, however, are in accord with the symptoms
produced by the cell poison. Recovery from a sub-lethal
dose is rapid and complete, and this would imply that the
contact between the body cells and the poison is transi-
tory, and non-destructive. It appears to be more like a
fulminating irritation, and may result in an arrest of func-
tion due to a disturbance in the physical equilibrium of the
cells affected." (5) The serum of normal guinea-pigs
incubated with the poison does not materially, at least,
decrease its action. (6) The poison does not induce any
local reaction when introduced intradermically in guinea-
pigs sensitized to tuberculoprotein. Four animals sensitized
nineteen days previously with the cell residue, and which
had been found to be sensitive to an extract emulsion of
tubercle bacilli, and four animals rendered and proved
sensitive by a watery extract of tubercle bacilli received
intradermal injections of the poison, and under close obser-
vation showed no reaction. This is as should be expected.
The skin reaction, like other tuberculin reactions, results
from the cleavage of the tuberculoprotein. The cleavage
products produce no such reactions. (7) Auer and Lewis1
showed that prophylactic treatments with atropine save
a large percentage of animals from death by anaphylactic
shock on the reinjection of the homologous protein. We
claim that our protein poison is the active agent in
anaphylaxis. Now, White and Avery show that atropine
protects 75 per cent, of guinea-pigs from death after the
1 Amer. Jour. Physiology, 1910, xxvi, 439.
176 PROTEIN POISONS
administration of lethal and slightly supralethal doses of
the poison. (8) Morphine sulphate has been shown by
White and Avery to antagonize the action of the poison."
Of the animals tested (19) only 3 showed typical symptoms,
and with two of these death was slightly delayed. The
three autopsies revealed typical inflation of the lungs, with
epicardial hemorrhages in two. Nine of the pigs had only
slight symptoms, and although the issue was fatal, death
was delayed from forty-two minutes to over six hours. On
section, however, six of the animals showed inflated lungs
with epicardial hemorrhages. Two animals recovered. It
will be noted that in five cases a dose of 1 to 12,000 failed
to produce typical immediate symptoms. Further investi-
gations of the effects of morphine might lead to a better
knowledge of the factors concerned in the sequelae of
parenteral administration. (9) Banzhaf and Steinhardt1
studied the effects of chloral hydrate upon the action of
our poison prepared from egg-white, and came to the
following conclusions: "Normal guinea-pigs under the
influence of chloral (by intracardiac and intramuscular
injections) were completely protected against one and
one-fourth fatal doses of the poison (given intracardiacly).
If two or more fatal doses were given death resulted. Chloral
mixed with the poison and then given caused irregular results
which were interpreted as meaning that there is no chemical
union of the chloral and poison in vitro. We assume that
the chloral protected by union with certain vital cells."
White and Avery used a 2.5 per cent, solution of chloral
in normal salt. The injections were made intravenously.
"With the exception of 2 animals displaying typical symp-
toms, both of which received amounts of the poison con-
siderably in excess of that required to kill, 6 of the 13
survived the injection with slight or no symptoms, while
5 succumbed in from two to thirteen hours without exhibiting
the classic respiratory spasms. Autopsy showed the typical
findings in 2, while in 2 others there was a partial inflation
1 Jour. Med. Research, 1910, xxiii, 1.
THE SPLIT PRODUCTS OF TUBERCLE BACILLUS 177
of the lungs with punctate hemorrhages beneath the peri-
cardium. The results, although not so strikingly positive
as those of Banzhaf and Steinhardt, at least tend to confirm
their conclusion." (10) Banzhaf and Steinhardt found
that lecithin given intraperitoneally in doses of from 250
to 500 mg. or more to serum-sensitized guinea-pigs pro-
tected them from a second injection of 5 c.c. of horse serum
given twenty-four hours later. When lecithin was emul-
sified with the Vaughan poison or given twenty-four hours
before the poison was injected, no protection was afforded.
From this Banzhaf and Steinhardt conclude that lecithin
prevents the cleavage of the protein in a sensitized animal
on reinjection and that it does not neutralize or modify
the action of the preformed poison. White and Avery,
from their experiments, come to the following conclusion:
"Lecithin emulsion injected simultaneously with the
poison seems to possess a slight and irregular prophylactic
action. Incubation of the poison with lecithin emulsion
for an hour at 37.5° increases this neutralizing property.
A dose of 1 to 12,000 was not affected. The preliminary
administration of lecithin protected some of the animals,
delayed death in others, and was without effect in the
remainder. The results were too inconstant to warrant
definite conclusions."
White and Avery are inclined to the opinion that our
crude protein poison contains a plurality of active substances,
and in this they are probably right. They say: "The
effects provoked by the parenteral administration of the
artificially obtained poisonous substance in non-fatal doses,
and as modified by atropine, morphine, chloral, and other
drugs, seem to suggest the plurality of its action. It is
conceivable that the poisonous fraction obtained by
Vaughan's method contains either an essential component
which is several in its physiological action, and which in
sufficient doses exerts its primary and dominating effect on
the respiratory mechanism, or that it contains groups or
individual constituents of different selective vital affinities,
the most eminent of which is for the peripheral or central
12
178 PROTEIN POISONS
cells functioning in respiration. It is not unreasonable to
hope that a further separation of the poisonous fraction
into its components and a more intimate study of their
various actions on the animal economy may furnish valuable
clues not only to the relation of these chemical substances
to true anaphylactic processes, but also to the physiological
nature of the varied phenomena of hypersensitiveness."
It seems to us, theoretically, that there must be a whole
spectrum of poisons in the protein molecule. We have
shown that at least one group in this molecule is poisonous.
The poisonous action of the protein molecule becomes
more marked as we proceed in stripping off certain side
chains. Peptone is more poisonous than the native protein
from which it is obtained. Our product is' more active
than peptone. Between the two there must be a group of
bodies, each of which is more active than the peptone and
less active than our split product. Indeed, we are con-
fident that we have discovered some of these intermediate
bodies. As has been stated, when the alcohol employed
in the cleavage of the protein molecule is not absolute we
obtain products that are quite unlike our poison in physical,
chemical, and physiological properties. They are sticky
and gummy. They contain some carbohydrate, responding
to the Molisch test, and yielding a reducing substance
after prolonged boiling with dilute mineral acid; while our
final product is not gummy and fails to show any evidence
of carbohydrate content, except in that from tubercle
bacilli. These other bodies kill much less promptly. The
paralytic symptoms are more marked, and the convulsive
stage is either only slightly in evidence or wholly wanting.
Interesting experiments on sensitization to tuberculo-
protein have been made by Baldwin1 and Krause.2 We
make the following extracts from this work: Animals may
be sensitized by any of the ordinary products of the tubercle
bacillus. Sensitization may be secured by introducing
the protein by any parenteral route, by the peritoneal
1 Jour. Med. Research, 1910, xxii, 189. 2 Ibid., xxii, 275; xxiv, 361.
THE SPLIT PRODUCTS OF TUBERCLE BACILLUS 179
cavity, subcutaneously, subdurally, intracerebrally, by post-
orbital injection, and probably by intravenous injection,
though the last-mentioned method was not tried. Sensi-
tization may be obtained by the injection of only 0.05 mg.
of the protein. The best preparations for sensitization
are those in which the protein is in solution. The shortest
period of incubation found was six days. This was when
the sensitizing dose was given postorbitally. Before the
twenty-first day sensitization is uneven and inconstant.
After this period it proceeds with great regularity, and the
longest duration noted was two hundred and eighty-six
days. It is likely that it continues in the guinea-pig through-
out life. The size of the sensitizing dose bears no relation
to the period of incubation. Acute anaphylactic shock
follows when the reinjection is given intravenously or post-
orbitally. The minimum toxic dose on reinjection was found
to be 0.99 mg. of the dry protein, and the minimum fatal
dose on reinjection 1.6 mg. Attempts to establish passive
anaphylaxis have been uniformly unsuccessful. Infected
animals become autosensitized and are killed by injections
of large amounts of the tuberculoprotein. This protein does
not act like a toxin, and when injected into animals does
not lead to the elaboration of an antitoxin. "If an animal
be infected experimentally it begins to react to tuberculin
about the fifteenth day; in like manner, the non-tuber-
culous but protein-treated animal will react to a second
injection about two weeks after the first. Again, both
the tuberculous and the sensitized non-tuberculous animals
react to exceedingly small doses of the protein; indeed, a
certain proportion of the tuberculous will undergo an
intoxication that is identical with acute anaphylaxis,
provided the toxic dose is applied postorbitally, while if
the sensitized animal receives its toxic injection by a route
that renders absorption less rapid — e. g., an intraperitoneal
injection— the resulting intoxication will tend to approxi-
mate what is generally observed as the tuberculin reaction
in the infected guinea-pig (without, of course, any focal
reaction). Therefore, while the facts will not at present
180 PROTEIN POISONS
warrant the flat declaration that the two phenomena result
from the same fundamental causes, there are enough data
at hand to justify the elaboration of a working hypothesis
that such is the case." The important question of the
relation between sensitization and immunity to infection
has been tested by Baldwin and Krause. Series of guinea-
pigs were sensitized to tuberculoprotein. The fact that
they were in full sensitization was demonstrated by testing
some of each set. Those that recovered from anaphylactic
shock and known as "refractory" were inoculated an hour
after the reinjection. Lot A had received a total of 25 c.c.
of the watery extract in seventeen doses over a period of
thirty-nine days. Lot B had received a total of 13 c.c.
of the watery extract in ten doses over a period of thirty-
nine days. Lot C had received a total of 8 c.c. of the watery
extract in six doses over a period of thirty-nine days. Lot
D had not been sensitized. The last sensitizing doses were
given June 14, 1910. All of these animals were inoculated
with the same amount of a virulent culture of the tubercle
bacillus July 1, 1910. Sixty-two days after inoculation
all the animals were killed and examined. A summary of
the findings is stated as follows: "The refractory animals
suffered most. The disease was pretty well disseminated
in all of them, and they exhibited far more tuberculosis
than any of the animals that had not been intoxicated,
and than any of the controls. . . . The animals that
were sensitized in various ways all became diseased. As
a general thing, we may say that the more protein the
animal received during preliminary treatment, the less was
the resultant infection. So far as one could tell from the
toxic symptoms of the test animals there was very little
difference in the average degree of sensitization in the
several sets of guinea-pigs. The results of inoculation
were, however, different. It is most likely that the differ-
ences were altogether independent of any degree of raised
or lowered resistance conferred by the sensitive state, but
that they were due to the heightened immunity that followed
the protein injections."
THE SPLIT PRODUCTS OF TUBERCLE BACILLUS 181
Krause concludes the paper from which the above was
taken as follows:
"1. Sensitization of non-tuberculous guinea-pigs with
tuberculoprotein does not alter their resistance to experi-
mental tuberculous infection.
"2. Sensitization to tuberculoprotein and relative
immunity (increased resistance) to infection can occur
coincidently in the same animals.
"3. Resistance to infection is markedly lowered during
the period that a sensitized animal is suffering from symp-
toms of anaphy lactic shock."
The third conclusion is certainly justified from the results
of the research, and is what might have been predicted
at the start. Whether the first conclusion is in any way
contradictory to the previous statement "that the more
protein the animal received during preliminary treatment,
the less was the resultant infection," we leave the reader
to determine for himself. This line of experimentation
should be continued with all the tuberculoprotein prepara-
tions and with variations in size and frequency of dosage.
Thiele and Embleton1 have reviewed the literature of
Sensitization in tuberculosis, and have experimented with
reference to both active and passive hypersensitiveness to
tubercle bacilli, and the relation to the tuberculin reaction
in man. The conclusions reached are stated as follows:
(1) Guinea-pigs may be typically sensitized with pow-
dered tubercle bacilli. (2) Guinea-pigs may be passively
sensitized with the blood or tissues of animals actively
sensitized. (3) Guinea-pigs may be sensitized to tuberculin
with the blood of tuberculous patients who are highly
sensitive to tuberculin. (4) Likewise, guinea-pigs may
be sensitized with tuberculous tissue from man, or with
that from tuberculous guinea-pigs. (5) By regulating the
dose one can induce fever or cause the temperature to
fall below the normal in actively sensitized guinea-pigs
with tuberculin. (6) The same results can be obtained in*
1 Zeitsch. f. Immunitatsforschung, 1913, xvi, 411.
182 PROTEIN POISONS
animals sensitized with heterologous or homologous tissue.
(7) A cutaneous reaction has not been obtained.
This work is confirmed and supplemented by that of
Sata,1 who sensitizes guinea-pigs with a single injection of
tuberculous serum, in doses of 0.1, 0.5, or 1.0 c.c. subcu-
taneously, intraperitoneally, or intravenously, and uses a
reinjection of old-tuberculin intravenously. When the
dose of the reinjection is as much as 0.5 c.c. acute ana-
phy lactic death results. With smaller doses there is
elevation of temperature.
Many investigators have failed to sensitize animals
with tuberculin, while most have succeeded with dead
bacilli and with aqueous extracts. This is not surprising;
indeed it is what should have been expected. Tuberculin
consists of digested, denatured proteins of relatively simple
composition. It is well known that peptones and poly-
peptids do not sensitize. The protein poison when detached
from other groups in the protein molecule sensitizes neither
to itself, nor to the unbroken protein. The fact that tuber-
culin does not sensitize or does so imperfectly raises a
serious question as to its employment as a therapeutic
agent. It is undoubtedly an excellent diagnostic agent
because its relatively simple structure may favor its prompt
cleavage when injected into an animal already sensitized
by the disease. But if it is not a sensitizer its therapeutic
good effect, if it has any such effect, must be confined to
the possible establishment of a tolerance to the tuberculo-
poison. Sensitization to tuberculoprotein can be induced
by bacillary emulsions, with watery extracts, and with
the non-poisonous residue. If the sensitization secured by
the last-mentioned agent is as good as that produced
by the others, it has the advantage of not containing any
poison. On the other hand, if the therapeutic effect desired
consists in the development of a tolerance to the poison,
tuberculin must be preferred unless wre should use the more
•completely isolated poison.
> Zeitsch. f. Immunitatsforschung, 1913, xvii, 62.
THE SPLIT PRODUCTS OF TUBERCLE BACILLUS 183
There are those who, while admitting that animals can
be sensitized to tuberculoprotein, hold that the tuberculin
reaction is not an anaphylactic one. We think that it is,
and that the fact that tuberculin does not sensitize or does
so imperfectly does not contradict this. It is probable that
when tuberculin does sensitize at all it is due to the fact
that it contains traces of but little altered or unaltered
tuberculoprotein. The tuberculin reaction should be
regarded as a phenomenon resulting from a reinjection.
The animal is already sensitized by the disease.
Koch in his early work pointed out two facts, which in
a way seemed to be contradictory, but which have been
found to be true. First, he showed that a tuberculous
animal behaves toward a second infection differently from
a normal animal, the former resisting the second infection
by forming an inflammatory area about the point of the
second inoculation, this leading to necrosis, and recovery
without extension of the infection. Second, Koch stated
that the injection of dead tubercle bacilli into tuberculous
guinea-pigs killed them within from six to forty-eight
hours, while like injections into normal guinea-pigs had
no such effect. These apparently contradictory statements
have not only been confirmed, but have been found not in
any way in conflict. The studies of Romer, Hamburger,
and others have shown that the following conditions must
prevail in order to fully demonstrate the first statement of
Koch: (a) The first injection must be a weak one, per-
mitting the disease to run a chronic course. (6) The time
interval between the first and second inoculations must
be relatively long, the resistance to the second infection
increasing with time, (c) The dose of the second injection
must not exceed a certain limit. What happens to the
bacilli of the second inoculation? Why should these fail
to develop, while those of the first inoculation continue to
grow? Kraus and Hofer1 have found that tubercle bacilli
injected into the peritoneum of a tuberculous animal are
» Deutsch. med. Woch., 1912.
184 PROTEIN POISONS
destroyed by lysis within an hour. There is some lytic
destruction of tubercle bacilli in the peritoneum of a
healthy guinea-pig, but this does not compare in rapidity
and completeness with that occurring in the tuberculous
animal. But why is the action of this lytic agent manifested
so effectively against the bacilli of the second inoculation
while those of the first apparently proceed in uninterrupted
growth? One of Hamburger's experiments1 may throw
some light on this question. He made his reinoculation
subcutaneously on each side. On the left he injected a
small dose, on the right a large one. On the left there was
no development; on the right a tuberculous nodule developed.
We infer from this and similar observations made by others
that the lytic agent which destroys the tubercle bacillus
and which is produced in larger amount in tuberculous
than in normal animals, because the cells of the former have
been sensitized, is stored in the cells as a zymogen, and
is activated only when tuberculoprotein is brought into
contact with the cell, and possibly is active only, or is
most active, in statu nascendi. The ferment is capable of
destroying only a given amount of bacilli or is wholly
inactive in the presence of a great excess of substrate,
or its action is soon interrupted by the accumulation of
fermentative products. However, it is quite certain that
all the bacilli of the second inoculation are not always
killed because it may happen according to observations
of Hamburger that some months after the second inoc-
ulation, during which time there may have been no
evidence of infection, tubercular processes appear and
develop like a primary infection. Whatever the true
explanation, it is a fact that the tuberculous animal is more
resistant to additional infection than the normal animal is
to primary infection. This led Lowenstein2 to say: "Only
the tuberculous organism is tuberculosis-immune." Ham-
1 Beitrage z. klin. d. Tuberculose, xii.
2 Handbuch d. path. Mikroorganismen, Kolle u. Wassermann, zweite
Auflage.
185
burger and Toyosuku1 infected guinea-pigs subcutaneously,
and after the disease had become chronic, they submitted
these animals along with normal ones to a dust rich in
tubercle bacilli. The normal animals developed pulmonary
tuberculosis, while the tuberculous ones failed to do so.
Romer2 developed chronic subcutaneous tuberculosis, and
then inoculated intracutaneously and intravenously, and
in this way demonstrated the immunity of the tissues of
the tuberculous animal to infection with tuberculosis. The
submental and cervical glands of normal guinea-pigs become
tuberculous on feeding with as small an amount as 0.1
mg. of living bacilli, but these glands are not affected when
tuberculous guinea-pigs are fed with living bacilli. Many
other investigators have experimented along the same
line, with like results. That the unaffected tissues and
organs of tuberculous men are largely immune to infection
with the tubercle bacillus is a matter of every-day observa-
tion. In pulmonary tuberculosis the sputum laden with
bacilli passes through the upper air passages without, as
a rule, infecting them. Besides, there are cases of healed
tuberculosis with virulent tubercle bacilli in their expec-
toration. There are tuberculosis carriers just as there are
typhoid carriers.
Koch tried various methods in his attempts to immunize
animals to tuberculosis. Early in his investigations he
tried feeding animals with both living and dead cultures.
For two months he fed rats exclusively on the bodies of
animals dead from tuberculosis. From time to time a
rat was killed, and most of them were found normal. In
a few, small nodules were detected in the lungs. But these
animals after feeding for weeks upon tubercular tissue,
developed tuberculosis promptly when inoculated intra-
peritoneally. Later, Koch made the following statement:
"All attempts to cause absorption of living or dead bacilli,
by administration subcutaneously, intraperitoneally, or
intravenously, have failed me and also other investiga-
tors. When dead bacilli are injected subcutaneously they
1 Beitrage z. klin. d. Tuberkulose, xviii 2 Ibid., xiii.
186 PROTEIN POISONS
uniformly cause suppuration, and they can be easily stained
and detected in large numbers in the abscesses thus formed
after months. When injected into the peritoneal cavity
they are better absorbed, and I have obtained some immunity
in this way, but they generally cause local inflammations,
which lead to adhesions with stenosis and occlusion of the
intestine, so that a large percentage of the animals is lost.
When injected intravenously into rabbits, dead bacilli
cause tubercular nodules, similar to those observed after
infection, in the lungs, and in these nodules the unaltered
bacilli can be found after a long time. By this method
absorption does not proceed in the desired way. Having
been convinced that the unaltered bacilli could not be
used, I attempted to render them absorbable through the
action of chemical agents on them. The only method of
this kind which I have found effective consists in boiling
the bacilli with dilute mineral acid or with 'strong alkali.
In this way tubercle bacilli may be so changed that they are
absorbed in toto, and in large amount, though slowly, when
administered subcutaneously. But marked immunity has
not been reached in this way, and it seems that these chemi-
cal agents cause such thorough alteration in the bacillary
substance that its immunizing property is destroyed."
This conclusion reached by Koch has been justified by
all subsequent investigators. Levy1 has tried to prepare a
vaccine by such chemically indifferent substances as glycerin,
25 per cent, solution of milk sugar, and 10 to 25 per cent,
solutions of urea. The object in these experiments was to
kill the bacilli by the withdrawal of water and without
changing their immunizing properties. Levy stated that
these vaccines contain no living bacilli, and with them he
apparently increased the resistance of guinea-pigs to infec-
tion with tubercle bacilli, but Romer doubts the complete
killing of the bacilli by these agents. Heating the bacilli
to 70° or 80° has failed to furnish an effective vaccine.
Lowenstein2 tried to prepare a vaccine by exposing tubercle
1 Med. Klinik, 1905, 1906; Centralbl. f. Bak., xlii, xlvi, and xlvii.
2 Zeitsch. f. Tuberkulose, 1905, vii.
THE SPLIT PRODUCTS OF TUBERCLE BACILLUS 187
cultures to daylight for a year, but this failed. The same
investigator tried formalin, with a like result. Bartel1
made a brie of tubercle bacilli and lymph glands, and
Schroder tried a like experiment with spleen pulp, but
neither of these preparations proved effective. Calmette
and his students have tried preparations obtained by the
action of iodine and its salts upon tubercle bacilli, but
without success. Similar preparations with chlorine have
been tried by Mossu and Goupil.2 Noguchi3 stated: "The
inoculation of guinea-pigs with tubercle bacilli, which have
been killed by soaps (sodium oleate), develops in these
animals a complete or partial resistance to subsequent
inoculation with a virulent culture of the same strains
of bacilli. In short, a condition of immunity to tuberculosis
can be induced in guinea-pigs by the injection of an emulsion
of tubercle bacilli in oleic soaps." Zeuner,1 following the
work of Noguchi, has used an extract of tubercle bacilli in
solution of sodium oleate in the treatment of tuberculosis.
Broil5 found that guinea-pigs treated with this preparation
survived only a few weeks. Marxner6 tried it on goats, and
found no evidence of infection in two cases on postmortem,
but Lowenstein7 states that this is explained by the fact
that the animals were sectioned too soon after inoculation,
and adds that he sectioned goats three years after this
treatment, followed by inoculation, and found tubercular
cavities, the size of a man's head, in the lungs. Deycke
and Much8 found that, "One part of tubercle bacilli is
dissolved in two parts of a 25 per cent, solution of neurin
when kept at 52° for twenty-four hours. This forms a
perfectly clear syrup which becomes cloudy on cooling.
We attribute this phenomenon to the presence in the bacilli
of fatty bodies with high melting-points." Later it was
1 Wien. klin. Wochenschrift, 1905.
2 Compt. Rend, de 1'Acad., 1907.
3 Centralbl. f. Bak., 1909, lii, 85.
4 Zeitschf. f. Tuberkulose, xv.
6 Berlin tierarztliche Wochenschft., No. 47.
i • Zeitsch. f. Immunitatsforschung, x, xi, xii. 7 Loc. cit.
8 Beitrage z. klinik d. Tuberkulose, xv, Heft 2.
188 PROTEIN POISONS
found that solutions of cholin have a similar, but less
marked, solvent action on tubercle bacilli; also that these
solutions injected into men are without harmful effect.
Attempts to immunize animals with these solutions have
been made, without success. Aronson1 has extracted fat
from tubercle bacilli with trichlorethylen, and has attempted
to immunize with the residue, but has not reported any
success. Calmette and Guerin2 have tested the protective
action of tubercle bacilli grown on media containing bile
acids. They stated in 1910 that by the tenth generation,
these cultures of bovine tubercle bacilli become so attenu-
ated that they can be used as a vaccine. Since these
investigators have made no later report it is fair to assume
that their expectations have not been realized. Attempts
have been made to immunize animals to tuberculosis with
the virulent bacillus by beginning with so small an amount
as one bacillus (Webb and Williams) and increasing the
dose; with cultures attenuated in varying degrees; with
living and dead cultures of various varieties of the tubercle
bacillus, such as human, bovine, avian, chicken, from cold-
blooded animals, etc.; with strains of other acid-fast bacilli,
as those of timothy, butter, manure, etc.; with normal
and specific sera; but up to the present time no satisfactory
results have been obtained. This is an interesting subject,
and we would like to. go into some detail, but it lies outside
the scope of this book.
Lowenstein3 has shown that the tubercle bacillus will
grow, though not abundantly, on a medium of the following
simple composition:
Ammonium phosphate 6 parts
Glycerin 40 parts
Distilled water 1000 parts
Although the growth is slow in developing and sparse,
it elaborates an active tuberculin. This is additional
evidence that growth of bacteria consists essentially of
synthetical processes.
1 Berl. klin. Woch., 1910 No. 35. * Comp. rend, de 1'Acad., cli, 1.
» Centralbl. f. Bak., 1913, Ixviii, 591.
CHAPTER IX
THE ANTHRAX PROTEIN1
Literature. — Since anthrax is the most typically infec-
tious of all diseases, and since so many theories have been
evolved concerning it, we may be pardoned for briefly
reviewing the literature. As early as 1805 Kausch2 wrote
a monograph on this disease in which he held that it is
due to paralysis of the nerves of respiration; but he offered
no explanation of the paralysis. Delafond3 held that
anthrax has its origin in the influence of the chemical
composition of the soil on the food, thus inducing patho-
logical changes from malnutrition. The contagious nature
of the disease was clearly established in 1845 by Gerlach.4
This was confirmed by the studies of Heuzinger,5 and was
endorsed by Virchow in 1855, since which time it has never
been questioned. However, as early as 1849 the bacilli
had been seen by Pollender.6 Pollender did not publish his
observations until 1855, but he states that they were made
in the fall of 1849. First, he examined the blood of five
cows dead from anthrax, and compared this with material
taken from the spleen of a healthy animal. The examina-
tions were not made until from eighteen to twenty-four
hours after death, and he states that the blood was stinking,
thus indicating that it had become contaminated with
putrefactive organisms, but the description which he gives
1 The first part of this chapter is abstracted from a paper by J. Walter
Vaughan, published in the Trans. Assoc. Amer. Phys., 1902, xvii, 313.
2 Ueber der Milzbrand des Rindviehes.
3 Traite sur la Maladie du Sang des Betes a laine, 1843.
4 Magazin f. Thierheilkunde.
6 Die Milzbrandkrankheiten der Thieren und der Menschen.
6 Vierteljahresschrift f. gerichtliche Medicin, 1855, viii, 103.
190 PROTEIN POISONS
shows that he actually saw anthrax bacilli. He used a
crude compound microscope made by Plossl, and he gave
his attention to the blood corpuscles, chyle globules, and
the bacilli. His description of the microorganisms may be
condensed as follows: The third and most interesting
microscopic bodies seen in anthrax blood are innumerable
masses of rod-like, solid, opaque bodies, the length of which
varies from -j^-g- to 2~ro °f a nne> and the breadth averages
ToVo °f a nne- They resemble the "vibrio bacillus" or
"vibrio ambiguosus." They are non-motile and neither
water nor dilute acids, nor strong alkalies have any effect
upon them, and for this reason he concluded that they must
be regarded as vegetable organisms. He questioned whether
they existed in the blood of the living animal or resulted from
putrefaction, but was inclined to believe the former, and
thought they might represent the infecting organism, or at
least the bearer of the infection. It will be seen that Pol-
lender presented no positive proof that these rod-like bodies
had any causal relation to the disease. In 1856 Brauell1
inoculated sheep, horses, and dogs with blood taken from
animals sick with anthrax, and in this way demonstrated
that the disease could be transmitted to sheep and
horses, but not to dogs. He found sheep highly sus-
ceptible, horses less so, and dogs quite immune. He also
demonstrated the presence of the bacilli in the blood of
sick animals before death. It is interesting to note that
he fell into an error concerning the motility of the bacilli.
He states that when seen in fresh blood they are non-motile,
but later they become highly motile. This was, of course,
due to contamination. It should be noted that Brauell
also made examinations of the blood of various domestic
animals suffering from other diseases, and demonstrated
the absence of the bacillus in these. In 1863 Davaine2
published three valuable papers on anthrax. In the first
he states that in 1850 Rayer inoculated sheep with the
blood of others dead from anthrax, and in this way trans-
1 Virchow's Archiv, 1857, xi, 132.
» Compt. Rend, de 1'Academie des Sciences, Ivii, 220, 351, 386.
THE ANTHRAX PROTEIN 191
mitted the disease. It appears that Rayer published a short
note of this work in the Bull, de la Soc. de Biologie in 1850,
but we have not had access to this publication. Davaine's
own work was of the greatest value and shows great skill
for that time. Probably the most important experiments
that he made were those in which he demonstrated that
the blood of an animal sick with anthrax is not capable of
transmitting the disease to others unless it contains the
bacillus. It may be of interest to describe briefly the
experiments which led to the establishment of this fact.
Rabbit A was inoculated with anthrax blood. Forty-six
hours later, examination showed no bacilli in the blood of
A. At that time twelve or fifteen drops of blood were
taken from the ear of this animal and injected into rabbit
B. Nine hours later the blood of A was reexamined and
found to contain a large number of bacilli. This blood was
injected subcutaneously into rabbit C. One hour later
rabbit A died, and twenty hours later C died, while B
remained free from infection. Space will not permit us
to follow the literature of anthrax further, save those parts
that bear on the presence of a chemical poison. Pasteur,
De Barry, Koch, and others studied the morphology, life
history, and cultural characteristics of the bacillus, and in
this way founded the science of bacteriology. The reader
is referred for the theories of the action of this bacillus
to the works of Bollinger,1 Szpilman,2 Joffroy,3 Touisant,4
and Nencki.5
Investigations. — In 1877 Pasteur and Joubert6 filtered
anthrax cultures and the blood of animals sick of this
disease through porcelain, and injected the germ-free
filtrate into animals without inducing the disease, and
concluded, quite properly, that this bacillus does not
1 Zur Path, des Milzbrandes, 1872.
2 Zeitsch. f. phys. Chem., 1880, iv, 350.
3 Compt. Rend. Soc. de Biol., 1873 and 1874. •
4 Comp. Rend, de 1'Acad., 1879, xci, 195; xciii, 163.
6 Berichte d. deutsch. Chem. Gesellschaft, 1884, 2605.
6 Comp. Rend, de 1'Acad., Ixxxiv, 905.
192 PROTEIN POISONS
produce a soluble poison. Subsequent investigations, in
our opinion, have established the correctness of this con-
clusion. However, there have been several claims to the
discovery of soluble poisons in cultures of the anthrax
bacillus, and in the bodies of animals dead with this disease,
and we will now briefly review some of these claims which
are of historical interest.
Hoffa1 obtained from pure cultures of the anthrax bacillus
small quantities of a substance which he believed to be a
ptomain, and the specific poison of this disease. When in-
jected under the skin of certain animals it at first increased
the respiration and the action of the heart. After a short
period the respirations became deep, slow, and irregular.
Later the temperature fell below the normal, the pupils were
dilated, and a bloody diarrhea set in. Autopsy showed
the heart to be in systole, the blood dark, and ecchymoses
were found on the pericardium and peritoneum. Subse-
quently, Hoffa believed that he had succeeded in isolating
this substance from the bodies of animals dead of anthrax.
He named it anthracin, and undoubtedly convinced himself
for the time at least that he had discovered the specific
poison of this disease. No subsequent investigator — and
several have repeated the experiments — has been able to
confirm Hoffa's results, and it is now more than probable
that his "anthracin" resulted from the action of the agents
used in its detection and separation upon the constituents
of the fluids with which he worked.
In 1889 Hankin2 grew the anthrax bacillus in Liebig's
meat extract, to which fibrin had been added, and from
this filtered culture he precipitated with ammonium sul-
phate an albumose, which, while not directly poisonous to
animals when injected simultaneously with an inoculation
of the anthrax bacillus, caused more speedy death than
when the bacillus only was used. He concluded that the
albumose destroys or lessens the natural resistance of the
1 Ueber die Natur des Milzbrandgiftes, 1886.
^ British Med. Jour., 1890, ii, 65; Proc. Royal Soc., xlviii, 93.
THE ANTHRAX PROTEIN 193
animal to the disease, after which the bacillus is able to
continue the elaboration of its poison in the body.
Petermann1 repeated Hankin's experiments, and obtained
an albumose which elevates the temperature from one to
two degrees, but is otherwise without poisonous effects and
without protective influence against the anthrax bacillus.
Hankin and Wesbrook2 repeated and modified the experi-
ments of the former, and reached the following conclusions:
(1) The anthrax bacillus elaborates a proteolytic ferment
by means of which albumose may be formed from proteins,
but these have no immunizing action. (2) The bacillus
produces another albumose which is not due to the soluble
ferment, but to an intracellular ferment. (3) This albumose
was obtained in a relatively pure condition. This was done
by growing the bacillus in a solution of pure peptone. It
confers partial immunity against anthrax, when given in
small doses to mice. (4) To animals susceptible to anthrax
this albumose, in ordinary doses at least, is not poisonous.
(5) Those animals which are relatively immune to anthrax,
such as the rat and frog, are easily poisoned by this albumose.
(6) On the contrary, young rats which are susceptible to
anthrax are not poisoned by this substance.
Klemperer3 obtained from cultures of the anthrax bacillus
a substance which caused elevation of temperature when
injected subcutaneously, but which was not submitted to
further investigation. Brieger and Frankel4 endeavored
to prepare a tox-albumin from the organs of animals dead
of anthrax. They cut the tissue into fine pieces, rubbed
up with water, allowed to stand for twelve hours in an ice-
box, and filtered through porcelain. The filtrate was
concentrated in vacuo at 30° to one-third its volume, and
after being acidified with a few drops of acetic acid, was
treated with ten times its volume of absolute alcohol. The
mixture was then allowed to stand twelve hours longer in
an ice-box, after which the precipitate was collected on a
1 Ann. de 1'Institut Pasteur, 1892, vi, 32. 2 Ibid., vi, 633.
3 Zeitsch. f. klin. Med., 1892, xx, 165.
« Berl. klin. Woch., 1890, xxvii, 241, 268, 1133.
13
194 PROTEIN POISONS
filter, dissolved in a small volume of water, refiltered,
reprecipitated with alcohol, this being repeated until a
perfectly clear aqueous solution was obtained. The albu-
mose was further purified by dialysis, and as thus obtained,
it was found to be freely soluble in water and to give the
ordinary reactions for proteins. These investigators failed
to make any satisfactory study of this product, and the
repetition of their work by others has led to negative results.
Marmier1 attempted to isolate a poison from cultures
grown in a medium of the following composition:
Water 1000.0
Peptone 40.0
Sodium chloride 15.0
Sodium phosphate 0.5
Potassium phosphate 0.2
Glycerin 40.0
Before inoculation this fluid was filtered through porcelain,
and then sterilized at 110°. The peptone used was obtained
from the commercial preparation by the precipitation of
the albumoses with ammonium sulphate, and the salts
were removed by dialysis. In this medium the anthrax
bacillus, especially the sporeless form, grew abundantly.
In order to obtain the poison the culture was filtered and
saturated at room temperature with ammonium sulphate,
which produced a more or less abundant precipitate. This
was allowed to stand for some hours and filtered, after which
the precipitate was washed with a saturated solution of
ammonium sulphate. Subsequently the precipitate was
dissolved in water, freed from salts by dialysis, concen-
trated, feebly acidified with sulphuric acid, and precipitated
with alcohol. The substance thus obtained was found to
be soluble in water and in a 1 per cent, solution of phenol.
It was said not to give any of the reactions for albumoses
or alkaloids, but this can scarcely be true. This work has
had no confirmation, and is mentioned here simply because
of its historical interest.
1 Ann. de 1'Institut Pasteur, 1895, ix, 533.
THE ANTHRAX .PROTEIN 195
Heim and Geiger1 grew anthrax bacilli in eggs after the
method of Hueppe, extracted with alcohol, precipitated
the extract with mercuric chloride, filtered, treated the
filtrate with platinum chloride, decomposed the precipitate
thus formed with hydrogen sulphide, filtered, rendered
alkaline with potassium hydrate, and divided into two
portions, one of which was extracted with ether and the
other with benzol. The amount of material removed with
ether was small, but that obtained in the benzol extract
was large. When either of these residues was taken up in
a few cubic centimeters of feebly acidified water and injected
intra-abdominally into mice, it caused salivation and
lacrymation, followed by muscular convulsions and death.
The smallest dose of the benzol extract which proved to
be fatal was 0.5 c.c., while a similar amount of the ether
extract caused only transient symptoms. Apparently no
controls were employed by these investigators, and the
evidence that they obtained any poison from the anthrax
bacillus is too slight to deserve serious attention.
Ivanow2 has demonstrated the presence of certain volatile
acids, formic, acetic, and caproic, in anthrax cultures, but
there is no proof that the bacilli had anything to do with
the production of these bodies, or that they are concerned
in any way in the symptomatology or pathology of the
disease; certainly these same volatile acids are found in
the cultures of many bacteria, both pathogenic and non-
pathogenic.
Petri and Massen3 detected hydrogen sulphide in anthrax
cultures, but inasmuch as at the same time they found this
gas in every one of the thirty-six other bacteria examined,
it cannot be said to be of any specific importance. More-
over, spectroscopic examination of anthrax blood fails to
show the presence of hydrogen sulphide or any of its com-
pounds, and there is no evidence that this gas has any
connection with the disease.
1 Lehrbuch der Bakteriolog. Untersuchungen u. Diagnostik, 1894, 229.
2 Ann. de 1'Institut Pasteur, 1892, vi, 131.
3 Arbeiten aus d. kaiserlich. Gesundheitsamte, 1893, viii, 318.
196 PROTEIN POISONS
Fermi1 has shown the presence of both diastatic and
proteolytic ferments in anthrax cultures, but as all living
cells, including bacteria, elaborate such ferments, this
discovery fails to make us acquainted with the poison of
the anthrax protein. Maumus2 found that by its growth
on potato the anthrax bacillus converts some starch into
sugar, and Reyer3 showed the presence in anthrax cultures
of a ferment which coagulates casein.
Klein4 removed anthrax bacilli from agar cultures of
forty-eight hours' growth, placed them in 5 c.c. of bouillon,
and after the tube had been held for five minutes in boiling
water, injected the contents into the peritoneal cavity of
a guinea-pig, without results. After a few days the injec-
tion was repeated, and four or five days later these animals
were inoculated subcutaneously and intra-abdominally,
with small doses of a living culture. All died within forty-
eight hours, of typical anthrax. From these experiments
Klein concluded that the anthrax bacillus contains no
intracellular poison, and that treatment with the cellular
substance confers no immunity on guinea-pigs.
Conradi5 attempted to solve the question of the existence
of an anthrax poison by the following methods:
1. Guinea-pigs were inoculated intraperitoneally with
anthrax, and immediately after death the peritoneal fluid,
which varied in amount in different individuals, but aver-
aged from 10 to 15 c.c., and contained in each field from ten
to twenty microorganisms, was filtered through porcelain.
In some of the experiments the filter of Chamberland was
employed while in others that of Kitasato was used. The
filtered peritoneal exudate was injected into mice, rats,
and guinea-pigs subcutaneously, intravenously, and intra-
peritoneally, and always without effect. The amounts of the
filtered exudate injected into mice varied from 2 to 4 c.c.;
1 Arch. f. Hygiene, 1890, x, 1.
2 Compt. Rend, de la Soc. de Biologie, 1893, v, 1071.
3 Ibid., 309.
4 Centralbl. -f . Bakteriologie, 1894, xv, 598.
5 Zeitsch. f. Hygiene, 1899, xxxi, 287. .
THE ANTHRAX PROTEIN 197
that into rats from 5 to 12 c.c.; in guinea-pigs from 4 to 15
c.c.; in rabbits from 10 to 20 c.c.; and in one dog, 25 c.c.
was injected subcutaneously. In proportion to body weight
the mice received by far the larger injections. The experi-
ments indicate that in the peritoneal exudate of guinea-
pigs inoculated with anthrax there is no appreciable amount
of soluble toxin.
2. Many guinea-pigs were inoculated with anthrax, and
directly after death their livers and spleens were removed
under aseptic conditions and rubbed up in a sterilized
mortar with sterile sand to which a little physiological salt
solution had been added. After thorough rubbing the
mixture was diluted with physiological salt solution and
filtered through a Chamberland tube under four atmos-
pheres of pressure. The filtrate was injected subcutaneously,
intravenously, and intraperitoneally into mice, rats, guinea-
pigs, and rabbits, and in every case without effect.
3. Conradi, finding the preparation of collodion sacs
difficult, substituted for them the vegetable membranes
from phragmites communis, first used by Podbelsky.1
These sacs, after sterilization, were filled with bouillon
cultures of the anthrax bacillus, and after laparotomy
under ether were placed in the abdominal cavities of animals.
The animals used were guinea-pigs, rabbits, and dogs, all
of which remained well, notwithstanding the presence of
these tubes containing virulent cultures of the anthrax
bacillus in their abdominal cavities. These experiments
satisfied Conradi that the anthrax bacillus does not produce
any soluble toxin, and he next turned his attention to
determining whether or not this organism possesses any
intracellular poison.
4. The anthrax exudates obtained as described in the first
series, in quantities of from 5 to 6 c.c., were placed in test-
tubes, 0.5 c.c. of toluol added to each, the tube closed with
a sterilized cork, thoroughly shaken, and then allowed to
1 To one experienced in the preparation of collodion sacs this must be
regarded as a clumsy substitution.
198 PROTEIN POISONS
stand for ten days in the dark at room temperature. At
the expiration of this time the contents of many tubes were
placed in a separator and the toluol removed. Having shown
by inoculation that the germ contained in these tubes was
dead, the exudate thus sterilized was injected into suscep-
tible animals subcutaneously, without effect.
5. Asporogenous cultures were sterilized by exposure
for one hundred and ten hours to — 16°. After exposure
to this temperature, the tubes were kept in an incubator
at 20°, long enough to see that they remained sterile, after
which they were injected subcutaneously into susceptible
animals, and always without effect.
6. A number of rabbits and guinea-pigs were simul-
taneously infected with anthrax, and after death the livers
and spleens were subjected to a hydraulic pressure of 500
atmospheres. The fluid thus obtained, which on micro-
scopic examination showed the presence of bacteria, was
passed through a Chamberland filter, and then injected
subcutaneously, intraperitoneally, and intravenously into
mice, rats, guinea-pigs, and rabbits, and always without effect.
7. The experiment of Brieger and Frankel in which they
prepared their anthrax toxalbumin was repeated with
negative results.
From these experiments Conradi reaches the following
conclusions: "By no method known at present can it be
shown that the anthrax bacillus forms either an extra-
cellular or an intracellular poison within the bodies of either
susceptible or insusceptible animals. Indeed, these experi-
ments increase the probability that the anthrax bacillus
does not form any poisonous substance, therefore the solu-
tion of the manner in which anthrax infection results remains
unknown. Whether improved chemical methods will
lead to its detection or not cannot be determined, but for
the present the anthrax bacillus at least must be regarded
as a purely infectious microorganism." If this conclusion
reached by Conradi be true, the mechanical interference
theory is the best that can at present be offered so far as
anthrax is concerned.
THE ANTHRAX PROTEIN 199
In all our work the anthrax bacillus has been grown in
lloux flasks, as we have not dared try it in the large tanks,
consequently the amount of cellular substance obtained
has been small. This work was begun in 1900 and carried
on intermittently. We have felt it desirable to exercise
great care in handling quantities of the anthrax bacillus.
We have always opened the flasks and removed the growth
over a shallow tray containing some powerful germicide.
We will make a few extracts from our protocols. In May,
1900, the growth was removed from twelve Roux flasks,
placed in one liter of 1 per cent, sulphuric acid, and kept
in the incubator at 37° for twenty-four hours. At the
expiration of this time cultures showed the bacillus still
alive. The suspension was then placed in the autoclave
and heated at 100° for thirty minutes. Cultures now showed
that the bacillus had been killed. The suspension was
passed through a Chamberland filter with the aid of a
pump. The clear filtrate was poured drop by drop into
twice its volume of absolute alcohol. A finely flocculent,
white precipitate formed, was collected on a hard paper,
washed with alcohol until the filtrate was no longer acid,
and dried in vacuo over potash. This substance is not
colored with nitric acid and heat, but on the addition of
ammonia the characteristic orange of the xanthoproteic
test is developed beautifully. It does not give the biuret
and other protein tests. Later an unweighed portion of
this substance was dissolved in water and injected intra-
abdominally in a guinea-pig. Twelve hours later the
animal was found dead. The heart was in partial diastole
with clots in both auricles and ventricles. The peritoneal
cavity contained a few cubic centimeters of a clear fluid.
Two days later a guinea-pig received intra-abdominally
60 mg. of the alcoholic precipitate. When last seen that
night, five hours after the injection, the animal was dying.
The next morning it was posted and the condition described
above was found. The heart's blood was found to be sterile.
This experiment was repeated a number of times, with
similar results.
200 PROTEIN POISONS
This work was not resumed until 1902, and on repeating
the work the alcoholic precipitate failed to manifest any
poisonous action, but in this instance we heated the acid
extract for two hours. Inasmuch as the prolonged heating
seemed to be the only difference in the methods of pro-
cedure, another trial was made in which the acid extract
was heated in the autoclave at 110° for exactly ten minutes.
The alcoholic precipitate freed from acid as before was
ground to a fine powder in an agate mortar. An unweighed
portion of this powder was dissolved in 5 c.c. of water and
injected intra-abdominally. When last seen that night,
nine hours after the injection, the breathing was difficult
and irregular. The animal was found dead the next morning.
Autopsy showed extreme subcutaneous edema over the
abdomen. The peritoneal cavity contained a few cubic
centimeters of a clear fluid, and a smaller amount of bloody
exudate was found in the pleural cavity. The heart was
in diastole and the most marked changes were found in
the lungs. These were greatly congested and the left upper
lobe seemed to be consolidated. Closer examination showed
portions of the lungs to be completely hepatized. Many
of the air cells were filled with exudate and blood corpuscles.
The kidneys were highly congested and the liver seemed
pale and friable. Further experiments with weighed por-
tions of the powder showed the minimum fatal dose for
a guinea-pig when given intra-abdominally to be about
50 mg. Smaller doses down to 15 mg. made the animals
very sick, but failed to kill.
The poisonous group obtained from the anthrax protein
by cleavage with 1 per cent, sulphuric acid is destroyed,
at least greatly weakened, by prolonged boiling in aqueous
solution.
The cellular substance of the anthrax bacillus, prepared
by our method, is the least toxic of all the bacterial proteins
with which we have worked. This is true whether the cell
protein is derived from a pathogenic or a non-pathogenic
organism. It requires not less than JO mg. of the cell
substance, after extraction with ohol and ether, to
THE ANTHRAX PROTEIN 201
kill a guinea-pig after intra-abdominal injection. Smaller
doses make the animals sick, but do not kill, and confer
no marked immunity to inoculation with living cultures.
The cell substance has been split up by our method into
poisonous and non-poisonous portions. The former differs
in no recognizable way from the poisonous group obtained
from other proteins, and treatment of animals with the
latter fails to establish any noticeable immunity to inocula-
tion with the bacillus.
It has been shown that the poisonous group in the cellular
protein of a non-pathogenic bacillus may be more effective
than that in the anthrax bacillus. The minimum lethal
dose of the air-dried cell of the prodigiosus for a guinea-pig
of from 200 to 300 grams body weight, when injected intra-
abdominally, is less than 3 mg., while that of the anthrax
bacillus for the same animal is about 200 mg. Even the
lemon sarcine, the least toxic of the non-pathogenic organ-
isms examined, surpasses the anthrax bacillus in toxic
action. These facts convince us that the pathogenicity of
a bacterium is not measured by its capability of furnishing
a poisonous group, but by its ability to grow and multiply
in the animal body. The high degree of infectivity shown
by the anthrax bacillus in some animals is due to the fact
that it grows without hindrance on the part of the secre-
tions of those animals. On the other hand, its failure to
infect other species is due to the inhibiting action of certain
secretions of these animals.
Like other proteins, that of the anthrax bacillus contains
a poisonous group. The chief constituent of the anthrax
bacillus is a glyconucleoprotein, and by this we do not
mean a physical mixture of carbohydrate, nuclein, and
protein, but a molecule containing these constituents as
atomic groups. The intracellular poisons contained in
bacterial cells are not preformed toxins, as supposed by
Pfeiffer, but are atomic groups in a complex molecule.
The poison can be obtained from the protein only by
processes which disrupt the molecule. Mere solvents,
such as water, alcohol, ether, saline solution, and glycerin,
202 PROTEIN POISONS
do not detach the poisonous group. We have obtained
poisonous substances from the anthrax bacillus by two
methods. The substances thus obtained differ physically,
chemically, and physiologically. The one obtained by the
action of 1 per cent, sulphuric acid is insoluble in alcohol
and doos not give the distinctive protein reactions. The
one obtained by cleavage of the bacterial cell with a 2 per
cent, solution of sodium hydroxide in absolute alcohol is
soluble in alcohol and does give the biuret and Millon
reactions. The former kills only after some hours, and
leaves marked pathological changes. The other kills in a
few minutes and leaves no gross alterations. This, however,
does not prove that the poisonous group in the two prepara-
tions is not the same. It may be that in the one the poisonous
group is still closely attached to other groups, and energetic
measures may be necessary to tear it off, and as a result
of this the injury done to the body cells and recognized
at autopsy may be due. In the other preparation the
poisonous group is already detached, and consequently
its effects are manifest immediately. On the other hand,
our work does not show that the poisonous group in the
two preparations is the same. It leaves this question quite
undetermined. As we have stated elsewhere, there is
probably in the protein molecule a whole spectrum of
poisons, one derivable from the other, a chain of poisonous
groups, one differing from the one next it by having one
more or one less link. There is at present no more impor-
tant and no more difficult subject than that of the chemistry
of the protein molecule. The researches of Fischer have
done much to show some features of the structure of the
protein molecule. We know, as a result of Fischer's work,
that proteins are to be regarded as polymers or condensa-
tion products of the amino acids, but between the native
protein and the amino acids into which it may be split,
there is a long list of intermediary products about which
we know practically nothing. The ordinary, native proteins
are not primarily poisons. The amino-acids which result
from their ultimate cleavage are not poisonous, but between
THE ANTHRAX PROTEIN 203
the two there are many split products, of varied sizes,
which are poisonous. Besides, some of the amino-acids
may be converted into highly poisonous substances.
Whether a given protein molecule, on being disrupted, sup-
plies an active poison or not is determined by the lines of
cleavage and these are dependent upon the cleavage agent
and the conditions under which it acts. The work of Fischer,
as valuable as it is, has been and is of little or no service in
elucidating the processes of parenteral digestion, which must
be better understood before we can read the first line in the
true history of disease, either exogenous or endogenous.
Rosenau and Anderson1 sensitized animals by subcu-
taneous injections of extracts of the anthrax bacillus.
Sobernheim2 was not able to confirm this work, and made
some statements that deserve attention. He said that the
cell substance of the anthrax bacillus is quite different
chemically and biologically from that of other bacteria,
and that it is wholly devoid of poisonous properties, what-
ever the amount and method of administration may be.
Our own work, as already stated, shows that this is not
true. Busson3 has reinvestigated this question of sensiti-
zation with anthrax protein. Preisz4 has shown that
when the anthrax bacillus is grown at a high temperature
(42.5° C.) after the manner used by Pasteur in preparing
his vaccine, the membrane becomes mucilaginous and
more permeable. With bacilli thus prepared Busson suc-
ceeded in inducing a mild form of sensitization by intra-
peritoneal injections. The sensitized state was recognized
by a more marked elevation of temperature over the con-
trols on reinjection. It is undoubtedly true that the anthrax
bacillus is protected by its capsule against the action of
ferments produced in the bodies of infected animals, but
that anthrax protein is so radically different from other
1 Hygienic Lab. Bull., 1907, No. 36.
2 Kraus und Levaditi, Handbuch d. Technik u. Methodik d. Immuni-
tatsforschung, ii.
3 Zeitsch. f. Immunitatsforschung, 1912, xii, 671
4 Centralbl. f. Bak., 1911, Iviii.
204
PROTEIN POISONS
bacterial proteins is an unwarranted assumption, and that
it is not poisonous in any dose we have shown not to be
true. When the anthrax protein is obtained in solution
without alteration of its constitution, and when this solu-
tion is properly administered we dare say that it will be
found to sensitize animals as well as any other protein.
As we have had occasion to point out more than once,
it is necessary to have a protein in solution in order to
develop exquisite sensitization, and it must be in solution on
reinjection in order to induce the most striking form of
anaphylactic shock. Permeation of the body cells seems to
be essential to the most complete sensitization, also to its
demonstration on reinjection.
Roos1 has shown that salvarsan is an effecient germicide
for the anthrax bacillus, both in vitro and in vivo, and
Becker2 and Bettman3 have successfully treated anthrax
in man with this preparation.
1 Zeitsch. f. Immunitatsforschung, 1912. xv, 487.
z Deutsch. med. Woch., 1911.
3 Ibid., 1912.
CHAPTER X
THE CELLULAR SUBSTANCE OF THE
PNEUMOCOCCUS *
The Strain. — The strain of the pneumococcus with which
this work was done was presented us by Dr. J. J. Kinyoun,
of Washington. When the culture was received, a mouse,
guinea-pig, and rabbit received intraperitoneal inoculations.
The mouse died in twenty-three hours, the guinea-pig in
twenty-four, and the rabbit in twenty-seven. Cultures
were made from the heart blood of each of these animals,
and all found to be pure. Our growths were made in 5 per
cent, glycerin bouillon, and with these, Roux flasks and
the tanks containing a medium made of 3 per cent, of agar
and 1 per cent, of ox serum in the 5 per cent, glycerin
bouillon were inoculated. The flasks and tanks were kept
at 38° for four days, when the growth was removed. The
growth seemed to reach maturity in this time at the tem-
perature mentioned. When kept longer it began to dry
and contract. In some instances the growth was not
harvested until the sixth day. The cellular substance thus
obtained was handled in the usual manner, i. e., it was
thoroughly extracted with alcohol and ether.
The strain was found to be highly virulent and remained
so throughout the year of work with it. Fig. 10 shows the
effects of the living organism on guinea-pigs after intra-
abdominal inoculation of 0.00001, 0.000001, and 0.0000001
c.c. of a twenty-four-hour bouillon culture.
Fig. 11 shows the relative effects of the living organ-
ism and 5 mg. of the cellular substance. It will be seen
1 The first part of this chapter is founded on work done in the Hygienic
Laboratory of the University of Michigan in 1905-06 by Dr. J. F. Munson.
206
PROTEIN POISONS
that with the living organism there was no marked fall
in temperature until about the tenth hour, when the fall
became evident and continued until death. This held
FIG. 10
\
FIG. 11
TIME
TEMP. I
TEMPI!
v •
.
3
1
0 1
1 1
.' 1
., 1
4 1
• 1
6 1
; 1
8 1
j :
c :
1
N.T.
99.9
/
\
1 HR.
101.4
98.2
v
/
S
*"
•--.
/
2 HR8.
99.7
100.
\
/
^N
/
\
3 MRS.
99.4
01.3
\
/
\
\
I
5 MRS.
98.2
00.1
A
/
\
\
6 MRS.
96.8
00.6
\
\
7 HR8.
96.2
01.
— \
'
—
s
VI
a H s.
00.7
^
j
k
\
96
»Jn s.
91.8
\
I
S
r
\
10 H S.
96.4
01.9
\
— f
—
\
as
1Z H S.
95
02.2
\
\
14 H S.
55
99.5
\
21 H 8
/
93
HF-
COVEREC
\l
\
\
\
91
\
-90-
good with some variations in the exact time when the fall
began in all our experiments with the living organism when
twenty-four-hour cultures were employed. During the
time before the fall begins, the organism is growing and
multiplying. The fall indicates dissolution of the cell and
the liberation of its poisonous constituents. When older
cultures are used, especially when the amount is large, the
fall in temperature may appear much earlier, and is due to
the presence of autolyzed cells in the cultures. This is
true not only of the pneumococcus, but of the cholera,
typhoid, and other bacteria.
The cellular substance of the pneumococcus, prepared
by our method, is a white powder, in which the individual
cells readily take the stains, and it is found to be quite free
from debris. We administered it in suspension in sterile
salt solution, and generally intraperitoneally. It seems
to be more irritating than other bacterial cellular sub-
stances with which we have worked, and even 5 mg. sus-
pended in 5 c.c. of salt solution and injected into the peri-
toneal cavity seems to cause pain. The animal soon becomes
quite normal in appearance, and remains so for an hour
or two, when the fur behind the ears begins to roughen and
gradually the whole coat takes on this state. The posterior
extremities become weak and the animal is unable to main-
tain the erect posture. The weakness intensifies into a
paralytic state, and finally the animal lies stretched out
on its side, and seems quite unable to make a struggle.
Rarely there are convulsions, but, as a rule, respiration
slowly and quietly fails, and it is often difficult to tell
just when it stops. Following these symptoms, the tem-
perature which at first may be slightly elevated grad-
ually falls, and has been frequently at 85°, rarely at
75°, before respiration wholly ceases. When the dose
is a non-fatal one the lowest point is usually reached
about the seventh hour, when the temperature rises as
gradually as it fell. The cellular substance of the pneu-
mococcus is not highly poisonous compared with similar
preparations from other bacteria. It is rather interesting
208 PROTEIN POISONS
to make some comparisons here. As has been stated, the
strain with which our work was done was highly virulent,
killing half-grown guinea-pigs in doses of 0.0000001 c.c.
of a twenty-four-hour culture given intraperitoneally. At
the same time our old stock culture of the pneumococcus
did not kill in doses of less than 1 c.c., and yet the cellular
substances of the two, measured by toxicity, were prac-
tically the same. This and similar observations wTith other
bacteria lead us to conclude that virulence is measured
by rate of multiplication and not by chemical differences
in cellular poison content. Moreover, when two animals
were killed with the two strains the cells seemed to be as
abundant in one as in the other. The more virulent strain
multiplies the faster. Virulence may depend upon several
factors, but rate of multiplication is certainly one of them,
and on a common medium as the animal body this must
depend upon the effectiveness of the ferments whose func-
tion it is to prepare and utilize the pabulum on which
the organism feeds. Our highly virulent strain furnished
a cellular substance which killed guinea-pigs in doses of
1 to 10,000. Occasionally smaller doses killed. The smallest
fatal dose on first injection of which we have a record was
1 to 19,000, but the surely fatal minimum was 1 to 10,000,
and as we have stated, the less virulent strain of pneumo-
coccus furnished cellular substance of the same degree of
toxicity. A comparison of the virulent strain of the pneu-
mococcus with our strain of colon is also of interest. With
our colon bacillus the minimum constantly fatal dose was
1 c.c. of a twenty-four-hour bouillon culture. Sometimes
a dose of 0.5 c.c. killed, and the smallest fatal dose, as we
found it, was 0.25 c.c. This organism yielded a cellular
substance, which as a coarsely ground powder always killed
1 to 50,000; when finely ground it killed 1 to 75,000, and
sometimes as high as 1 to 2,000,000. Our virulent pneumo-
coccus killed in 0.0000001 c.c. doses, and yielded a cellular
substance which, when ground to the finest possible powder,
killed only 1 to 10,000. Surely these are strong arguments
for our belief that the pathogenicity of a microorganism is
not measured by its poisonous cell content, but by the rate
with which it multiplies in the animal body or the intensity
and rapidity with which it converts body proteins into its
own proteins.
It must be borne in mind in considering what we are
about to say in this paragraph that at the time these experi-
ments were conducted we knew but little about protein
sensitization, and they were not conducted with the phe-
nomena of sensitization in view. Had we known then what
we now know the lines of investigation would have been
drawn somewhat differently. However, this makes a review
of our old protocols all the more interesting and valuable.
We tried to immunize animals with the cellular substance.
It will be worth while to follow one set of these experiments
through. We take the three tables on p. 210 just as they
stand in the protocol.
It will be observed that in the second and third injections
made at intervals of five and six days we killed one-third
of our animals. Now we know that this was due to the
fact that we partially sensitized the animals.
Failing absolutely to even render our animals tolerant
to the dead germ substance, we tried to weaken it by heat,
but in this we were equally unsuccessful. However, we
did prove that heating the cellular substance of the pneu-
mococcus to 144° for five minutes in the autoclave does
not destroy its intracellular poison. We also found that
by heating the cells some of the poison passes into solution,
and may be filtered through porcelain.
We split up the cellular substance with a 2 per cent,
solution of sodium hydroxide in absolute alcohol, and
obtained a non-poisonous and a poisonous portion. In
both small and large doses the former had no visible effect
on animals, but it gave no immunity to subsequent inocu-
lations.
The poisonous fraction kills animals in about the same
doses as are required by similar preparations from other
proteins. The symptoms are not wholly identical with
those induced by poisons obtained from other proteins.
14
TABLE XXI. — MARCH 9, 1905. CAGE VI. PNEUMOCOCCUS GERM SUBSTANCE;
DOSES SUSPENDED IN 5 c.c. SALT SOLUTION.
"§
o.
g
a
a
a
a
3
a
ft
d
d
ft
a
|
a
a
>
si
41
I
§
a
i
i
a
8
I
§
•i
»
|
9
e
M~
i*
1
S
1
1
s
i
8
•p
h
-iS
I*'
r^
-^>
L
00
O
P
3
S
'3
£
O
P
PQ
tH
1
XJ
1-1
jg
<N
XI
CO
J3
•*
xi
>o
A
t-
O
<N
o
IN
12.20
Black
340
2.5
136,000 100.0
102.9
104.3
99.0
98.9
98.0
96.7
99.0
99.7
99.5
12.18
Gray
350 5.0 70,000 100.2
101.5
103.4
100.4
97.1 +
96.6
94.2— 95.0
94.8
98.2
12.17
Gray curly
370 10.0 37,000 100.3
100.1
99.9
97.6
97.6
98.8
97.4 194.2—
99.2
100.8
12.15
Wh., yel.,
black
390 15.0 26,000 100.8
101.3
101.2
96.4
98.1
PS 6
95.8
94.2—
976
101 4
12.10
Yel., white
39520.0 19.750 101.0
101.7
102.4
99.2
98.9
P76
94.6
96.8
P66
1009 +
12.05
Yellow
025
25.0 25.000 101.2
102.0
102.7
101.0
98.1
P73
97.6
97.4
PP3
10">4
TABLE XXII. — MARCH 14, 1905. THIS is, so FAR AS POSSIBLE, AN EXACT
REPETITION OF THE WORK OF MARCH 9, 1905, USING THE SAME PIGS. DOSES
GIVEN AT 11.45 A.M.
Black
Gray
Gray, curly
Yel., white,
black
Yel., white
Yellow
1
521
508
545
542
545
P
2.5
50
PQ
-S
o
£
a
I
a
!
a
I
1
ft
a
1
XI
00
i
i
o
101.0
100.6
ft
i
XI
(N
XI
XI
co
X)
98.4
100.2
XI
XI
96.4
P87
208,400
101,600
100.9
100.2
102.6
105.4
105.0
103.0
103.0
103.8
98.8
100.1
102.3
101.2
101.9
100.8
Died
Died
10.0
150
54,550
36,133
101.7
101.2
103.3
102.7
104.1
100.5
102.4
100.0
99.3
99.8
98.4
98.1
95.3
98.1
100.2
90.7
101.0
91.4
101.8
95.4
20.0
27,250
100.4
100.6
101.2
97.9
96.6
95.2
87.8 86.7
88.7
90.3
601
25.0
24,040
102.0
103.0
103.7
100.1
100.0
97.0
91.4
92.5
97.0
97.3
TABLE XXIII. — MARCH 20, 1905. CAGE VI. THE USUAL DOSES WERE GIVEN
(THIRD TIME). ANIMALS REACTED MORE SEVERELY THAN BEFORE. AN IMPRESSION
DRAWN FROM THEIR BEHAVIOR. Nos. 3 AND 6 DIED. PLACING IN INCUBATOR DID
NOT SAVE THEM
Sf
£
4!
•
•si
a
p
PQ
1
i
|
B
o
a
S
t-I
a
c
CM
a
ri
CO
ft
a
i
a
XI
co
f
£
H«
a
J3
0?
PI 4
S
S
o
1
M
>D
Black
Gray
Gray, curly
Yellow
550 2.5
542 5.0
552 10.0
60925.0
99.6
103.4
104.3
102.6
P6 4
90.5
PI 4
PO 5
96 7
Dead
Dead
101.0
100.6
103.6
100.0
103.2
100.3
101.7
96.0
97.1
86 9
92.392.7
86.9 84.2
95.0
P6 8
93.2
98.2
101.0
98.2
97.8
93.2
PO 5
86.086.0
P5 P
The temperatures of Nos. 3 and 6 at nine and one-half hours rose because they were placed
in the incubator.
THE CELLULAR SUBSTANCE OF PNEUMOCOCCUS 211
The difference lies in the less marked convulsive character
of the third stage. When injected intraperitoneally in
guinea-pigs, the coat soon roughens, and for some minutes
the animal seems quiet. Weakness of the hind limbs
develops and the upper part of the body is shaken by
spasms resembling severe hiccough. The paralysis rapidly
develops and spreads, and the animal lies on one side.
It dies in most instances without general convulsions.
The respiration becomes slower and the heart continues to
pulsate for some minutes after respiration has stopped.
Death occurs in from twenty minutes to an hour. The
temperature curve begins to fall soon after the injection,
and continues until death. In cases of recovery the first
sign of improvement is a rise in temperature, and it comes
up more slowly than it went down.
It is worthy of note that in case of inoculation with the
living organism there is an incubation period of about ten
hours. This is followed by the complete triumph of the
infection, and is shown by the even and constant fall in
temperature. With the dead cellular substance the incu-
bation period is shortened, but the character of the fall is
the same. With the free poison there is no period of incu-
bation, and the temperature begins to fall in a few minutes.
In all, the temperature curve is the same in general char-
acter. Certainly, it must be true that the poison which
affects and kills the animal must be the same in the living
and dead cell, and in the split product.
We demonstrated that the poisonous portion, like that
obtained from other proteins, establishes on repeated
injections of non-fatal doses a certain degree of tolerance,
but gives no immunity against infection.
The recent work of Rosenow1 on the autolytic cleavage
products of the pneumococcus, and certain other bacteria
is of great interest and value. The pneumococcus readily
undergoes autolysis, and Rosenow has studied the products
resulting in this way. The following are some of the more
1 Jour. Infect. Dis., ix, 190; x, 113; xi, 286, 480.
212 PROTEIN POISONS
important facts demonstrated by this investigator: (1)
Animals may be sensitized with dead pneumococci or
with extracts from the same. The sensitizing dose may be
given subcutaneously, intraperitoneally, intravenously, or
intrapleurally. In order to induce anaphylactic shock the
reinjection must be made intravenously or intracardiacly.
In the sensitized animal both dead and living pneumococci
are dissolved more rapidly than in normal animals. This
explains the slight but definite immunity to virulent cultures
manifested by sensitized animals. (2) Fresh pneumococci
suspended in salt solution and kept at 37° for forty-eight
hours, under ether or over chloroform, undergo autolysis
by which a poison is liberated. This poison injected
intravenously or intracardiacly in normal animals causes
anaphylactic shock. In guinea-pigs this poison induces
death by spasm of the bronchioles and consequent arrest
of respiration. In dogs it causes marked fall in blood-
pressure and delays the coagulation of the blood. This
poison is split off from the pneumococcus protein not only
in autolysis, but also by normal and immune sera and by
leukocytic extracts. (3) The cleavage of pneumococcus
cell substance by autolysis or the other agents mentioned,
is accomplished by proteolytic ferments, as is shown by
increased production of amino bodies as the poison is set
free. Finally, the digestive process reaches a point when
the poison itself is digested and rendered inert. "The fact
that virulent pneumococci have within themselves a proteo-
lytic enzyme which splits their protein into a highly toxic
substance, is strong indication that certain strains of pneu-
mococci may cause infection forthwith without first rendering
the host allergic. This is quite in keeping with the fact
that in pneumococcus infections an incubation period is
not an invariable rule. On the other hand, in certain
instances, a previous sensitization before symptoms set
in, probably occurs. This might well be the case in lobar
pneumonia when the chill occurs a week or ten days after
the patient contracted a severe cold or bronchitis. The
distribution by lobes in typical cases may be related to
THE CELLULAR SUBSTANCE OF PNEUMOCOCCUS 213
the bronchial spasm which this toxic substance produces.
That early dyspnea and increased respiration before con-
solidation is demonstrable is in keeping with this idea.
(4) Morphine, ether, urethane, atropine, and adrenalin,
protect normal guinea-pigs against the toxic material
obtained in vitro from pneumococci, and also sensitized
guinea-pigs on reinjection."
Recently (December, 1912) we found a small bottle of
the powdered pneumococcus cellular substance prepared
by Munson nearly seven years before (March, 1906). It
is a fine, yellowish-white powder, looking very much like
wheat flour. It has stood during these years in a cupboard,
kept closed except when momentarily opened to put some-
thing in or take something out. Microscopic examination
showed the pneumococci as clearly and in as perfect form
as in a fresh preparation. It kills guinea-pigs on intra-
abdominal injection in the same doses (1 to 10,000 of body
weight), and just as promptly as it did more than six years
ago. Five hundred milligrams of this was weighed, sus-
pended in 500 c.c. of salt solution, 10 c.c. of chloroform
added, and the whole allowed to stand at 37°. After twenty-
four hours, 10 c.c. of the opalescent supernatant fluid was
injected into the external jugular vein of a guinea-pig.
Within two hours the rectal temperature had fallen below
94°, and the animal remained sick for some hours, but
gradually recovered. The same experiment repeated at
the end of forty-eight and seventy-two hours killed the
guinea-pigs within two hours. These animals died with
the symptoms of a subacute anaphylactic shock. We
conclude from this that the intracellular autolytic ferment
had remained intact during the years that had elapsed
since the preparation of the cellular protein. Six days
after the suspension had been prepared and placed in the
incubator a like injection killed the guinea-pig within three
minutes. This animal died with the symptoms of acute
anaphylactic shock, and autopsy showed the lungs distended
and minute petechial hemorrhages in the pericardium.
CHAPTER XI
PROTEIN SENSITIZATION OR ANAPHYLAXIS
Introduction. — The older medical literature occasionally
records facts which in the light of more recent and extended
knowledge are known as the phenomena of protein sensi-
tization. Such were some of the experiences recorded in
the early attempts at the transfusion of blood. Many of the
untoward results reached in this procedure and beyond
the ken of that time are now fully explained. Behring and
Kitashima1 found on immunizing an animal to tetanus
toxin that it died in convulsions notwithstanding the fact
that the blood serum was richly charged with antitoxin.
They explained this by assuming the existence of a con-
dition of " hypersensitiveness" to the toxin. With our
present knowledge we see no reason for ascribing this to
the toxin. There is, so far as we know, no evidence that
animals can be rendered hypersensitive to either toxin or
antitoxin. Neither has ever been obtained free from pro-
teins, and since all true proteins, so far as we know, sensitize,
there seems no sufficient justification in ascribing a sensi-
tization induced by a protein solution containing a toxin
to the latter. Buchner2 repeatedly injected bacterial pro-
teins into men and noticed that the cardinal indications
of local inflammation, tumor, rubor, dolor, and calor
resulted. Furthermore, he noted that fever increased with
repeated injections. Krehl and Matthes3 induced fever in
animals by repeated injections of albumose and peptone.
Weichardt4 made an advanced study in the domain which
1 Berl. klin. Woch., 1901, No. 6.
2 Berl. klin. Woch., 1890, 216; Munch, med. Woch., 1891, No. 3.
3 Arch. f. exper. Path. u. Pharm., 1895, xxxv, 232; ibid., 1896, xxxvi, 437.
4 Berl. klin. Woch., 1903, No. 1.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 215
we now designate as anaphylaxis. He repeatedly treated
rabbits with protein expressed from placental cells, and
found that some of these died promptly on subsequent
injections. Furthermore, he mixed the serum of animals
thus treated with placental cells and obtained a soluble
poison which he named synzytiotoxin. Later, he showed
that hay fever results from the parenteral digestion of the
proteins of pollen. Both of these points will be discussed
in more detail later. Wolff-Eisner1 discussed the theory
of endotoxins and their application to various diseased
conditions, in a very suggestive manner, but added little
to our exact knowledge. Richet2 has made many valuable
contributions on this subject. In his first report made with
Portier in 1902, he worked with an extract from the ten-
tacles of a muscle and showed that an injection of this made
the animal much more susceptible to a second one. Unfor-
tunately, he coined the word anaphylaxis as most suitable
to cover this condition of increased susceptibility. He used
this word understanding it to mean "without protection,"
and indicating that the first injection destroyed any natural
resistance that the animal might possess against the poison.
Now, we know that the condition of sensitization is essential
to certain forms of immunity, as was first indicated by
Vaughan and Wheeler,3 and the inappropriateness of the
term anaphylaxis is self-evident. However, the word has
come into general use, and with this explanation we will
continue it. V. Pirquet4 proposed and has continued the
use of the word "allergic," meaning altered energy. This
is much more suitable, inasmuch as it simply expresses a
fact and binds no one to any theory. However, "allergic"
has not been usually employed, and we will use "protein
sensitization," "hypersensitiveness," "anaphylaxis," and
"allergic" as synonyms.
1 Zentralbl. f. Bakt., 1904, xxxvii; Munch, med. Woch., 1906; Derm.
Zentralbl., 1906; Berl. klin. Woch., 1907.
2 Compt. rend, de la Soc. biol., 1902; Ann. de 1'Institut Pasteur, 1907, xxi,
497; ibid., "1908, xxiii; ibid., 1909.
3 Jour. Infect. Dis., 1907.
4 Munch, med. Woch., 1906.
216 PROTEIN POISONS
The fact that animals which have once received an
injection of protein are liable to sudden death after a second
injection of the same kind has been known for many years.
Ever since the opening of the Hygienic Laboratory of the
University of Michigan (1888), animals once used have
been segregated and kept in cages marked "used animals,"
which indicated that conclusions could not be safely drawn
from results obtained when these animals were employed a
second time. In the standardization of diphtheria antitoxin
it soon became evident that the guinea-pigs that survived
one test could not be relied upon in a second one. In the
late nineties, Parke, Davis & Co., large manufacturers of
antitoxins, ascertained this fact and offered to supply the
Hygienic Laboratory of the University of Michigan with
" used" guinea-pigs at a small price. The offer was accepted,
but the animals were found dear at any price, as they
suddenly and unexplainably died when treated with horse
serum.
This condition evidently was observed by others, and
Theobald Smith mentioned it to Ehrlich, who set Otto to
work to find the explanation. Otto1 published his results
under the title "Das Theobald Smithsche Phanomen der
Serumiiberempfindlichkeit." However, simultaneously with
these observations on animals used in the standardization
of antitoxin, the profession had occasion to observe the
effects of injections of antitoxin in human beings. As
early as 1903, v. Pirquet2 wrote concerning certain clinical
effects following antitoxin treatment, and in 1905 he and
Shick published a monograph "the serum disease," "Die
Serumkrankheit. ' '
Definition. — Friedemann3 offers the following definition:
"We speak of anaphylaxis when the organism, in conse-
quence of a previous treatment with an antigen, after a
period of incubation becomes hypersensitive to the same
or to a closely related substance, and when this condition
1 V. Leuthold Gedenkschrift, 1906.
a Wien. klin. Woch.
3 Jahresb. u. d. Ergeb. d. Immunitatsforschung, 1910, vi.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 217
can be passively transferred to fresh animals by the serum
or organ extracts of the sensitized animal." Biedl and
Kraus,1 omitting passive anaphylaxis, give the following:
"By anaphylaxis we mean that state of specific hypersen-
sitiveness induced in animals by protein injections, and in
which symptoms of poisoning follow subsequent injections
of the same protein in doses which would have no effect
upon untreated animals." With some explanation to be
given later we accept these definitions as quite satisfactory.
In the meantime it is desirable to have a clear understanding
of the meaning of the terms employed in discussing this
subject. The substance which induces the anaphylactic
state is generally known as the "antigen." This implies
that it gives rise to the production of an antibody, and the
selection of this word has been determined by an attempt
to correlate the phenomena of anaphylaxis with the theory
evolved by Ehrlich in explanation of the production of
antitoxins by treatment with toxins. In truth the " antigen"
of anaphylaxis is not a toxin, nor is the new substance
generated in the body of the treated animal an antitoxin.
The term " anaphylactogen" is unobjectionable, since it is
applicable to any substance which induces the anaphylactic
state. Sensitizer is a good word, and commits one to no
theory. The same is true of the term " sensibilisinogen"
used by our French confreres. The sensitizer causes the
body cells of the treated animal to elaborate a specific
proteolytic ferment which digests or splits up the sensitizer.
Again, following the nomenclature of Ehrlich, this ferment
elaborated as a consequence of the introduction of the
sensitizer is generally designated as the "antibody." It
would be equally rational to speak of pepsin as an antibody
to beefsteak, because the former digests the latter. The
theory evolved by Ehrlich in his studies on toxin immunity
is the product of a genius of the highest order. It has
stimulated research, which has resulted in discoveries of
the greatest importance, but the attempt to explain all
1 Kraus and Levaditi's Handbuch d. Technik u. Methodik d. Immunitats-
forschung. Erganzungsband.
218 PROTEIN POISONS
physiological and pathological processes by this theory,
and to describe them in the nomenclature of this theory is
unscientific. To say that anaphylaxis is the result of
protein — antiprotein reaction — is to talk jargon. When
foreign proteins are taken into the alimentary canal they
must be digested before they are absorbed. This means
that their large molecules must be split into smaller ones,
and this must be continued until there are no more protein
molecules left. Every protein molecule contains a poisonous
group, and in normal, alimentary digestion this group is
rendered non-poisonous by further cleavage before absorp-
tion takes place. When foreign proteins find their way into
the blood and tissues they must be digested. This is accom-
plished, as it is in the alimentary canal, by proteolytic
ferments, but the danger from the poisonous group in the
protein molecule is evidently greater in parenteral than in
enteral digestion. Both enteral and parenteral digestion
are physiological processes. Every living cell has its own
proteolytic ferments, otherwise it could not live. When
stimulated it pours out this ferment, and it does so only
when stimulated. The function of a cell ferment depends
upon the kind of cell elaborating it, and to a certain extent
upon the stimulating substance. The proteins are the
normal stimulants to cell secretion. When a foreign pro-
tein is introduced into the blood or tissue it stimulates
certain body cells to elaborate that specific ferment which
will digest that specific protein. When such a protein first
comes in contact with the body cells the latter are unpre-
pared to digest the former, but this function is gradually
acquired. The protein contained in the first injection is
slowly digested, and no ill effects are observable. When
subsequent injections of the same protein are made, the
cells, prepared by the first injection, pour out the specific
ferment more promptly and the effects are determined by
the rapidity with which the digestion takes place. The
poisonous group in the protein molecule may be set free
so rapidly and in amount sufficient to kill the animal. This
in brief is an explanation of the phenomena of anaphylaxis.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 219
The Sensitizer. — The sensitizing agent most thoroughly
studied is blood serum. When a small dose of blood serum
is injected into a guinea-pig intravenously, subcutaneously,
intracranially, or intra-abdominally, and, after a period
of twelve days or longer has elapsed, a second injection is
made, the animal develops the symptoms of anaphylactic
shock, which, in the majority of instances, terminate
fatally. This reaction is specific. The animal is sensitized
to the blood serum of the species of' animal from which the
blood was taken and not to the sera of other species. The
amount of serum necessary to sensitize a guinea-pig is
surprisingly small. Rosenau and Anderson found 0.000001
c.c. of horse serum sufficient. Besredka places the minimum
amount necessary to secure uniform results at 0.001 c.c.
while one-tenth of this proved sufficient in a considerable
percentage of the animals. The sensitizing dose of horse
serum ordinarily employed in experiments upon guinea-pigs
is 0.01 c.c. Large doses sensitize, but a longer time is
required. When 5 c.c. is given the time which elapses
before complete sensitization results may be as long as
three months. The larger the dose the longer the time
essential to sensitization. Besredka is inclined to the
opinion that when large doses are given there is no sensitiza-
tion until the greater part of the injected protein is elimin-
ated. If he means that it is eliminated unchanged, he is
certainly wrong. The protein of the first injection is slowly
digested, and the larger the amount the longer the time
required for the digestion, and complete sensitization does
not occur until all the protein of the first injection has been
disposed of and the cells have had time to accumulate a
reserve of the preferment. At least this is our explanation
of this point. The second dose, in order to produce a fatal
result, must be considerably larger than the minimum
sensitizing dose. The proportion between the minimum
sensitizing and minimum fatal dose has been placed by
Doerr and Russ at 1 to 1000. The second dose, in order
to kill the animal promptly, must contain at least a fatal
dose of the protein poison, but it may contain many times
220 PROTEIN POISONS
this amount and not kill. Whether the second dose kills
or not depends not only upon the amount of poison it
contains but upon the rapidity with which the poison is
set free.
There has been some difference of statement concerning
the effect of heat on the sensitizing properties of blood
serum. Rosenau and Anderson1 found that animals could
not be sensitized with serum which had been heated at 100°.
Doerr and Russ2 placed the point at which loss of sensitizing
properties occurs at 80°. Kraus and Volk3 raised it to 90°.
Besredka has straightened out this matter and has correctly
shown that the sensitizing properties of a protein are in part
at least dependent upon its physical state, and that diluted
serum may be heated even to 120° without losing its
capability of sensitizing. It is probable that no protein
completely sensitizes the body cells unless it be in at least
partial solution. Heating undiluted blood coagulates the
protein, and in this way leads to a decrease of its stimu-
lating effects upon the body cells. Besredka has shown that
the sensitizing property of blood serum is thermostabile.
Wells4 has very properly pointed out that it is the physical
change induced in the protein by coagulation and not
chemical alteration, which decreases its efficiency as a
sensitizer, and he calls attention to the fact first shown
by Besredka that proteins not coagulated by heat, do not
decrease in their sensitizing effects when their solutions are
boiled. This is true of casein, for instance, but when milk
sours and coagulation of the casein results it is not so ready
a sensitizer. Wells, furthermore, shows that other methods
of coagulation, as by precipitation with alcohol, lessen the
sensitizing properties. He suggests that the finely coagu-
lated particles of protein may be seized upon by phagocytes
and destroyed. In confirmation of this we have found that
proteins insoluble in water, such as edestin, sensitize more
efficiently when dissolved in salt solution than when sus-
1 Hygienic Laboratory, Bulletin No. 45.
2 Zeitsch. f. Immunitatsforschung, i, 110.
3 Ibid., 731. " Jour. Infect. Dis., v.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 221
pended in water. Furthermore, we have found that bacterial
proteins suspended in normal salt solution and heated to
154° in the autoclave under 2 kilos of pressure are more
efficient sensitizers than the unheated suspensions. All
these facts support the theory that body cells are best
sensitized when the protein comes in intimate contact with
them. Possibly cell permeation is necessary for the most
complete sensitization.
Besredka finds that when the protein of the second
injection is heated it is less likely to kill, and he concludes
that proteins contain a thermostabile, sensitizing, and a
thermolabile toxic component. We fail to see how such a
conclusion follows his findings. If the physical condition
of a protein affects its sensitizing properties, why should
it not affect its toxic action on reinjection? The poison is
set free by the digestive action of the specific ferment
elaborated as a result of the first injection. Why should
not the physical state of the protein affect the rapidity
and thoroughness with which it is digested, and consequently
the amount of the protein poison set free or activated at
one time? Doerr and Russ have apparently answered this
question in a satisfactory manner. By carefully conducted
experiments they show that heat affects the sensitizing
and toxic properties of proteins in the same ratio.
It should be understood that temperatures high enough
to disrupt and destroy proteins are destructive to their
sensitizing properties. According to Rosenau and Anderson
a temperature of 200° removes every trace of the sensitizing
property of proteins.
The influence of the digestive ferments of the alimentary
canal on the sensitizing properties of proteins is an interesting
and important subject, since it bears upon the possibility
of sensitization by administration through the digestive
tract. This point has been especially studied by Wells1
and Pick and Yamanouchi.2 The former submitted egg
1 Loc. cit.
2 Zeitsch. f. Immunitatsforschung, i, 676; Wien. klin. Woch., 1908, 1513.
222 PROTEIN POISONS
albumen to tryptic digestion and found that as the digestive
action advanced the sensitizing property receded. Some
have claimed to sensitize animals with peptone and even
with amino-acids, but since the minutest quantity of
protein suffices to sensitize, it is more reasonable to suppose
that the peptone and amino-acid preparations were not
absolutely free from protein. Vaughan and Wheeler have
shown that the poisonous portion of the protein molecule
does not sensitize in either small or large doses. Frances-
chelli1 found that when tissue is autolyzed for months, and
until every trace of the biuret reaction is lost, the fluid
shows no diminution in its sensitizing properties. This
agrees with the finding of Vaughan and Wheeler, that their
non-poisonous portion of the protein molecule sensitizes
even when it does not respond to the biuret reaction. All
this suggests that the sensitizing group in the protein
molecule is not itself a protein, or at least not a biuret, body.
However, the sensitizing group is destroyed in normal
digestion, and it is only under abnormal conditions that
protein sensitization results through the alimentary canal.
We will return to this subject later.
Whether or not all proteins contain the sensitizing group
cannot as yet be answered with certainty. According to
Doerr and Russ the globulin of the blood serum is the only
protein in that fluid which sensitizes, while Wells concludes
that in egg-white the albumen is the only active agent.
Wells purified the albumen of egg-white by recrystallization
after the method of Hopkins and Pinkus, and he found
that the purer his albumen, the smaller the amount neces-
sary to sensitize. Gay and Adler2 attempted by fractional
precipitation of blood serum with ammonium sulphate to
separate the anaphylactogenic from the other protein
constituents, and they obtained an euglobulin which sensi-
tizes but does not prove toxic on the second injection.
Quite naturally it seemed to them that they had succeeded
in isolating the sensitizing constituent of blood serum, and
i Archiv f. Hygiene, 1909, Ixx, 163. 2 Jour. Med. Research, xviii, 433.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 223
they proposed to call it "anaphylactin." But ammonium
sulphate alters the chemical nature of proteins, and Armit1
has shown that after precipitation with this reagent the
poisonous group in it cannot be extracted by the method
of Vaughan and Wheeler. It is probable therefore that
the "anaphylactin" of Gay and Adler contains the poisonous
group, but so combined that it is not set free either in vitro
or in vivo.
Besredka has pointed out that in anaphylactic experiments
with milk the fluid should be boiled for about twenty
minutes, the injections should be made into the peritoneal
cavity, larger guinea-pigs of from 300 to 400 grams' weight
should be used, and a period of from sixteen to twenty
days allowed to elapse between the sensitizing and test
injections. When these conditions are complied with, an
exquisite sensitization with uniform results is secured.
Besredka was not able to sensitize guinea-pigs with milk
given by mouth or rectum. We have found rabbits, espe-
cially young ones, easily sensitized by either of these avenues.
In our work we have observed that hungry rabbits will
eat milk when mixed with other food and seldom are sensi-
tized, but when the milk is introduced into the stomach of
a fasting rabbit through a tube, or injected into the rectum,
the milk can soon be detected in the heart's blood, and the
animal becomes sensitized. Evidently when the milk is
taken normally into the stomach, it is digested; when
forcibly fed through a tube, it is in part, at least, absorbed
undigested. Like the blood, milk as between different
species of animals shows a strictly specific action. Animals
sensitized to woman's milk do not react when treated with
cows' milk, and vice versa. By this method we have iden-
tified the source of milk stains deposited on wood for
months. The evidence concerning the difference between
the proteins of the milk and those of the blood of the same
animal is somewhat conflicting. Besredka2 found that
1 Zeitsch. f. Immunitatsforschung, 1910, v. 703.
2 Compt. rend. Soc. biol., Ixiv, 888; Ann. de 1' Institut Pasteur, 1909, xxiii.
224 PROTEIN POISONS
animals sensitized to cows' serum were not affected on the
subsequent injection of cows' milk, and vice versa. Wells1
obtained results that were not constant. Uhlenhuth and
Handel2 and Thomsen3 did sensitize to serum with milk and
vice versa; the latter used woman's milk and human serum.
Bauer4 by fixation of the complement method seems to have
shown that the albumin and globulin of milk are closely
related to the same constituents of the blood, while the
casein of the milk is a protein unlike any in the blood.
This is probably correct and explains the inconstancy in
the experiments.
The differences between the proteins of the blood serum
and those of the erythrocytes have been demonstrated
by the anaphylactic reaction. This has been shown uni-
formly by the experiments of H. Pfeiffer,5 Pfeiffer and
Mita,6 Thomsen,7 Doerr and Moldovan,8 and Uhlenhuth
and Handel.9 These investigators have found it impossible
to sensitize guinea-pigs against blood serum with erythro-
cytes and vice versa. In demonstrating this fact it is necessary
to fully separate the corpuscles and serum. The corpuscles
should be well washed in order to accomplish this, and
provision must be made against solution of the corpuscles
in the serum. The corpuscles of each species contain
specific proteins and therefore those of one species do not
sensitize to those of another.
According to Dunbar,10 the sexual cells are as specific as
blood sera. The proteins of organ extracts are specific as
between different species of animals, with some exceptions
to be noted later, but as between different organs from the
same species, and between the blood serum and organ
Loc. cit.
Zeitsch. f. Immunitatsforschung, iv, 761. 3 Ibid., iii, 539.
Munch, med. Woch., 1908, No. 16; Zeitschr. f. exper. Path. u. Ther.,
1909, vii.
Zeitsch. f. Immunitatsforschung, 1910, viii.
Ibid., vi; ibid., v. ' Ibid., i; ibid., iii.
>> Zeitsch. f. Bakt. Ref., v.
» Zeitsch. f. Immunitatsforschung, iv, 761.
10 Ibid., 740; vii, 454.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 225
extracts they are not strictly specific. It will be readily
understood that it is difficult to obtain organ extracts
wholly free from the blood of the same animal.
That the crystalline lens contains proteins different from
those found in any other part of the body was demonstrated
some years ago by Uhlenhuth, by the precipitin reaction,
and it was believed that the proteins of the crystalline lens
are identical in all animals. This was apparently confirmed
by anaphylactic tests, as shown by the work of Andrejew1
and that of Kraus and Sohma.2 Guinea-pigs can be sensi-
tized with the proteins of their own lenses, or with those of
other animals, and sensitization with the proteins of the
lens from any animal responds to the same from any animal.
The question whether the proteins of the crystalline lens
are identical in all animals is of the greatest biological
interest. The precipitin and the first sensitizing tests
seemed to establish this belief, but quantitative experi-
ments, such as those of Kapsenberg,3 indicate that there are
slight differences in the proteins of the crystalline lens from
different species. Kapsenberg, after reviewing the literature
and detailing his own work, concludes: (1) Guinea-pigs are
easily and uniformly sensitized with the lens substance of
other animals. The dose on reinjection necessary to induce
fatal anaphylactic shock is small.4 (2) Guinea-pigs may be
sensitized to the protein of their own lenses, but this is
done with difficulty and the fatal dose on reinjection is
large (40 mg.). (3) The proteins of the crystalline lens are
specific, but not markedly so. Dunbar finds that the
specificity of fish proteins is not so marked as those from
mammals.
Rosenau and Anderson5 succeeded in sensitizing guinea-
pigs to placental tissue from the same animal, and Locke-
mann and Thies6 sensitized rabbits to the serum of the
1 Arb. aus d. Kaiserl. Gesundheitsamte, xxx, 450.
2 Wien. klin. Woch., 1908, 1084.
3 Zeitsch. f. Immunitatsforschung, 1912, xv, 518.
4 In one instance as low as 2.5 mg. of protein but usually 6 mg.
5 Hygienic Laboratory Bulletin, No. 45.
e Biochem. Zeitsch., xxv.
15
226 PROTEIN POISONS
rabbit fetus. Gozony and Wiesinger1 passively sensitized
rabbits with the blood serum to amniotic fluid in two cases
of eclampsia.
Some years ago Obermayer and Pick2 found that the
serum of rabbits treated with proteins radically changed
by being iodized, nitrified, or diazoized, did not precipitate
the native protein, but did act upon the altered protein
with which the animal had been treated, and this occurred
without reference to the original sources of the protein.
Wells3 and Pick and Yamanouchi4 were not able to sensitize
animals with iodized protein to an iodized protein obtained
from another species.
Sensitization to egg-white has been studied by Vaughan
and Wheeler5 also by Wells.6 The former used in most of
their experiments egg-white diluted with an equal volume
of salt solution. Guinea-pigs sensitized to egg-white from
chickens responded to test injections of egg-white from
tame ducks, though less energetically and less constantly,
and still less to egg-white from robbins. Vaughan and
Wheeler by the method already described (p. 98) split
egg-white into a non-poisonous, sensitizing portion and a
poisonous, non-sensitizing portion. They believe that a
similar cleavage occurs as the result of ferment action on
the second injection in sensitized animals. This work
forms the basis of their theory of anaphylaxis, which will
be discussed later.
All bacterial proteins are anaphylactogens. Indeed, the
Koch reaction with tuberculin is an anaphylactic test, but
this will be discussed later. Bacterial proteins act as ana-
phylactogens, whether living or dead, formed or in solution.
On account of the physical state of the protein the reaction
is generally less pronounced, and strong than with proteins
in solution. Bacterial anaphylaxis has been studied by
» Orvosi hetilap, liii, 418. 2 Wien. klin. Woch., 1906.
3 Loc. cit.
4 Zeitsch. f. Immunitatsforschung, i, 676.
6 Jour. Infect. Dis., June, 1907.
6 Loc. cit.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 227
Rosenau and Anderson,1 Kraus and his students,2 and
others. The first mentioned sensitized animals with sub-
tilis, colon, typhoid, anthrax, and tubercle bacilli. The
second dose given after eleven days or longer induced
anaphylactic symptoms. In some instances repeated
injections seem to be necessary in order to induce a high
degree of sensitization. The evidence concerning the
specificity of bacterial anaphylaxis is somewhat conflicting.
Kraus and Doerr sensitized guinea-pigs with intraperitoneal
injections of one loop or less of cultures of typhus, dysentery,
cholera, v. Nasik and v. El-Tor. A second injection of a
maceration of homologous cultures given intravenously
after from twenty to twenty-five days was followed by
marked dyspnea, discharge of urine and feces, and coma.
Some recovered, but others died within ten minutes. These
investigators found this reaction strictly specific. In
another experiment they found that animals sensitized with
a maceration of the dysentery bacillus did not respond to
the toxin of this bacillus, but did to a second treatment
with the maceration. There is no proof that toxins sensi-
tize. Delanoe sensitized guinea-pigs to the typhoid bacillus.
He secured the most marked effects when sensitization was
induced by repeated injections, and one month or longer
elapsed before the test injection. He did not find the
reaction markedly specific. Vaughan and Wheeler sensi-
tized guinea-pigs to colon, typhoid, and tubercle proteins,
and in this way secured a certain degree of immunity to
living cultures. They also sensitized animals with the
non-poisonous proteins of the colon and typhoid bacilli, and
secured the same degree of immunity to living cultures.
This subject will be enlarged when we discuss the relation
of anaphylaxis to the infectious diseases.
The purest known proteins act as sensitizers. Even the
crystallized proteins as hemoglobin, crystalline egg-white,
and such pure vegetable proteins as edestin and excelsin
1 Loc. cit.
2 Wien. klin. Woch., 1908, Nos. 18, 28, and 30.
228 PROTEIN POISONS
act as exquisite sensitizers. Moreover, it has been found
that the more thoroughly a protein is purified the more
perfectly it sensitizes and the smaller the dose necessary
to sensitize or to kill on reinjection. Wells found that
purified casein acts more perfectly and in smaller doses
than a corresponding quantity of milk, and the sensitizing
dose of crystallized egg-white is less than one one-hundredth
that of native egg-white, and the killing dose on reinjec-
tion less than one-fifth. These facts have led Wells to
suggest that the mixed albumins may contain substances
which antagonize the anaphylactic reactions. Since pure
proteins sensitize and kill on reinjection, it seems reasonable
to conclude that the sensitizing and poisonous groups are
constituents of the same molecule. Edestin in its most
highly purified form is believed to be a chemical unit, and
not a mixture of proteins. This can be split by our method
into sensitizing and poisonous portions. It is true that the
amount of the non-poisonous portion necessary to sensitize
is larger than that of the unbroken molecule necessary to
accomplish the same purpose, and it is possible that sensi-
tization with this product is due to the fact that it contains
a trace of the unbroken molecule, but the fact that no
amount of this portion induces the slightest anaphylactic
symptoms on reinjection is not in harmony with this view.
It seems more reasonable to assume that in the process of
cleavage, which is crude, a large part of the sensitizing
group is destroyed. It is certain that the poisonous portion
does not sensitize to either itself or the unbroken molecule.
By our method the molecule is disrupted and in so doing
both portions are largely destroyed. The final word on
this matter cannot be spoken until we know that we have
absolutely pure proteins to start with, and we have more
perfect methods for the cleavage of the protein molecule.
However, it seems certain that the sensitizing properties
of the protein molecule reside in a group or in groups which
are destroyed by digestion long before the poisonous group
is markedly impaired. The sensitizing group seems more
labile than the poisonous one.
PROTEIN SEN SIT IZ AT ION OR ANAPHYLAXIS 229
Through the courtesy of White and Avery we have been
permitted to read an unpublished research of theirs on
"Some Immunity Reactions of Edestin." From this we
excerpt the following findings: (1) The smallest sensitizing
dose of pure crystallized edestin given intraperitoneally
is 0.0001 mg. Guinea-pigs sensitized with this dose react
fatally when the reinjection intravenously is not less than
50 mg. When the sensitizing dose is from 0.1 to 5 mg.,
0.5 mg. causes a fatal dose on intravenous reinjection.
(2) Guinea-pigs sensitized to edestin do not react on intra-
venous injection of gliadin, or globulins from squash seed,
castor bean, and the hazel-nut. Two animals reacted,
one fatally, to intravenous injections of flaxseed globulin.
The fatal dose of flaxseed globulin was, however, from
forty to one hundred and twenty times the minimum fatal
intoxicating dose of edestin. (3) Guinea-pigs from a sensi-
tized mother inherit sensitization, though in a lessened
degree. (4) The intraperitoneal injection of from 0.05 to
0.1 c.c. of edestin immune serum into a guinea-pig sensitizes
the latter to such an extent that it reacts fatally to an
intravenous injection of edestin on the following day. (5)
"When edestin is hydrolyzed by an alcoholic solution of
sodium hydrate by the method of Vaughan, a substance
is formed which produces a fatal intoxication in the guinea-
pig apparently identical with true anaphylactic shock. The
intravenous injection of one part of this poison to forty
thousand parts of guinea-pig by weight constitutes the
minimum fatal dose." It should be stated that this is the
crude poison. (6) "When suitable amounts of edestin and
edestin-immune serum are allowed to remain in contact
for a given length of time, a precipitate is formed which,
when washed with salt solution and mixed with fresh
guinea-pig complement and incubated at body temperature,
yields a substance or substances which when injected into
a guinea-pig intravenously produces a fatal intoxication,
apparently identical in every way with the anaphylactic
reaction. Fresh complement, when allowed to act under
similar conditions with edestin alone, yields no poisonous
230 PROTEIN POISONS
substance. From edestin, therefore, by the action of
immune serum and complement, under the experimental
conditions noted, a toxic product is obtained which seems
to correspond with the anaphylatoxin of Friedberger."
It seems most probable that anaphylactogens, agglu-
tinogens, precipitinogens, and lysinogens are identical.
In other words, one group in the protein molecule causes
the animal cells to develop a substance which under certain
conditions may act as an agglutinin, a precipitin, or a
lysin. We are inclined to the belief — not yet positively
demonstrated — that the same ferment may under varied
conditions act as an agglutinin, a precipitin, a lysin, or it
may cause a deeper cleavage in the protein molecule,
resulting in the liberation of the protein poison. Through
the researches of Friedberger, Doerr and Russ, and others,
it has been made quite certain that anaphylactogens and
precipitinogens are identical, and that these properties
reside in the same intramolecular group. As proteins are
altered by heat or digestion, their properties as anaphylac-
togens and precipitinogens are decreased in the same
ratio. The protein obtained by one-third to one-half
saturation of serum with ammonia sulphate is strongly
active both as a precipitinogen and as an anaphylactogen,
while that obtained by full saturation is inactive in either
direction. Whether this is due to physical or chemical
alteration has not been determined.
We may condense our statements concerning anaphylac-
togens as follows : They are proteins which when introduced
parenterally into animals stimulate the body cells to elaborate
specific ferments for the purpose of their digestion. When
introduced into a sensitized animal they are digested so
rapidly that the split products, some of which are poisonous,
produce certain more or less violent and characteristic
symptoms which may terminate in death. All anaphylac-
togens are proteins, and all proteins contain a certain
poisonous intramolecular group. This group is physio-
logically the same in all proteins, hence the identity of
the symptoms of anaphylactic shock whatever the protein
PROTEIN SENSITIZATION OR ANAPHYLAXIS 231
by which it is induced. All anaphylactogens contain a
sensitizing intramolecular group which is not the same in
any two kinds of proteins, hence the specificity of sensiti-
zation. We have succeeded in splitting some proteins into
non-poisonous, sensitizing, and into poisonous, non-sensi-
tizing portions. Whether all proteins contain a sensitizing
group or not has not been determined. Our views con-
cerning anaphylactogens differ from those held by others.
They think that in mixed proteins, such as blood-serum,
corpuscles, organ cells, egg-white, etc., there is some one
protein which sensitizes and some other one which is toxic.
We hold that the sensitizing and toxic proteins are groups
in the same molecule. We think that we have demon-
strated this by obtaining both groups from such pure proteins
as edestin. Artificially crystallized proteins, such as egg
albumen prepared by the method of Hopkins and Pinkus,
are not suitable for this work because they are changed
chemically by the ammonium sulphate, and are not split
up by our method. From our researches we conclude that
the sensitizing group of the protein molecule is much more
complicated in its chemical structure than the toxic group.
Further discussion along this line will be indulged in when
we take up the poisonous portion.
We are aware of the claims made by Bogomoletz1 and
by Pick and Samanouchi,2 that lipoids may act as anaphyl-
actogens, but they have not convinced us that their prepara-
tions were wholly free from proteins.3 Besides, it is possible
that a non-protein may act indirectly as an anaphylactogen.
This may be due to the substance causing some cleavage
in the proteins of the body and these products may sensitize.
This question will arise again when we discuss hypersensi-
tiveness to certain medical agents.
Volatile Sensitizers. — Rosenau and Amos4 have demon-
strated that the exhaled air contains a substance which
1 Zeitsch. f. Immunitatsforschung, v and vi. 2 Ibid., i, 676.
s Thiele and Embleton, Zeitschr. f. Immunitatsforschung, 1913, xvi, 160,
have investigated the claim of Bogomoletz that lipoids act as anaphylacto-
gens and have been unable to confirm his work.
4 Jour. Med. Research, 1911, xxv, 35.
232 PROTEIN POISONS
sensitizes animals to the blood serum. The exhaled breath
of men condensed and injected into guinea-pigs sensitizes
these animals to subsequent injections of man's serum. Of
99 guinea-pigs submitted to this test, 26 manifested recog-
nizable symptoms of anaphylactic shock, and 4 of these
died on injection of human serum. " The fact that a number
of our experiments resulted negatively may mean either
that the organic matter is present in the expired air in
exceedingly small amounts, or that the guinea-pigs with
which we worked did not come from a very sensitive race.
There are indications in our work which suggest that the
expired breath from certain persons contains more organic
matter than from other persons; also that the amount
varies with conditions. We obtained a greater percentage
of reactions in the guinea-pigs injected with the liquid
condensed from the expired breath of females than in those
injected with the liquid condensed from the expired breath
of males. Whether this is a mere coincidence or not may
be determined only by collecting more extensive data.
" The logical conclusion from our results is that protein
substances under certain circumstances may be volatile.
It seems unlikely that such a complex molecule should
possess the power of passing into the air in a gaseous form.
The volatility, however, now in question, may resemble
that solubility which deals with particles in suspension hi
a physicochemical state (colloidal suspension). The protein
may simply be carried over in 'solution' in the water vapor.
"A comparatively large number of the guinea-pigs
inoculated subcutaneously with the condensed liquid from
the expired breath developed sloughs at the site of the
injection. It is not certain whether this was due to the
pressure of the relatively large amount of liquid injected,
or to some irritating principle contained in the liquid.
Occasionally the local effects may have been due to the
fact that the liquid was cold when injected. The injection
of the condensed liquid caused no other untoward symp-
toms upon the animals, which is quite contrary to the
observation on rabbits of Brown-Sequard and others."
PROTEIN SENSITIZATION OR ANAPHYLAXIS 233
Later, Rosenau has announced that guinea-pigs kept in
stables with horses become sensitized to horse serum.
Wells and Osborne1 have studied the anaphylactic reac-
tions of some pure vegetable proteins, such as the globulins
from castor bean, flax-seed, and squash-seed, edestin from
hemp-seed, excelsin from Brazil nuts, legumins from peas
and vetch, vignin from cow peas, glycinin from soy beans,
gliadin from wheat and rye flour, hordein from barley, and
zein from maize. " It has been found that all these proteins
cause typical anaphylactic reactions in sensitized animals,
with all features essentially the same as when serum and
other animal materials containing proteins are used. The
minimum doses which produce sensitization and the time
of incubation are about the same as with animal proteins
but as a rule the symptoms are of somewhat slower onset
and less stormy course than are those obtained with foreign
sera, and the minimum intoxicating doses are larger. There
are also considerable differences in the toxicity of the several
vegetable proteins to sensitized animals, but the reasons for
these differences have not yet been investigated. The most
toxic proteins as measured by the frequency of severe and
fatal reactions, were the globulin of squash-seed, vignin,
excelsin, and castor-bean globulin, which usually caused
death when given in 0.1 gram doses to properly sensitized
animals. Edestin caused the least severe reactions of any
of the proteins, while hordein and glycinin seldom caused
fatal reactions. Nevertheless the minimum sensitizing
and intoxicating doses of edestin and squash-seed globulin
are essentially the same. The influence of the food of the
guinea-pig upon the anaphylactic reaction is of particular
importance in experiments with vegetable proteins, since
the natural food of the guinea-pig is vegetable. Experiments
showed that continuous feeding with a vegetable protein
rendered guinea-pigs immune to this protein, so that they
could not be sensitized to it. Although brief feeding with
animal proteins (cows' milk, foreign sera, egg albumen)
1 Jour. Infect. Dis., 1911, viii, 66.
234 PROTEIN POISONS
renders the animal sensitive to the corresponding animal
protein, probably sufficiently protracted feeding with
animal proteins will likewise confer immunity. The sensi-
tization through feeding is specific for the protein food,
showing that during the processes preceding and including
absorption of the food protein no change takes place which
robs it entirely of its biological specificity. The close
similarity, if not identity, of the legumins of the pea and
vetch was shown by the interreaction of these proteins,
and the close relation to vignin from the pea was also
indicated. The near relation or probable identity of the
gliadins from wheat and rye was also shown."
This is in accord with our findings of some years ago,
when we demonstrated that vegetable, bacterial, and animal
proteins contain the same poisonous group.
The Sensitizing Group in the Protein Molecule. — As has
been stated (Chapter V) we have split proteins into
poisonous and non-poisonous portions. This has been done
with proteins of most diverse origin, bacterial, vegetable,
and animal, and we have found no true protein which has
failed to undergo this cleavage. Certain pseudoproteins,
like gelatin, do not respond to this test, but all true proteins,
so far as tested, have been split into poisonous and non-
poisonous portions. This is the foundation stone of our
theory of protein sensitization. All true proteins are
sensitizers, and so far it has not been shown that sensiti-
zation can be established by any non-protein substance.
All sensitizers develop symptoms of poisoning on reinjec-
tion. These symptoms induced by reinjection are identical
in manifestation and sequence with those induced in the
fresh animal by the injection of the poison split off from
the protein molecule by chemical agents, or by the ferments
in the serum or organ extracts of sensitized animals. There-
fore, we have concluded that anaphylactic shock is due to
the cleavage of the molecule of the protein sensitizer on
reinjection, and the liberation of the protein poison, and
this cleavage is due to a specific proteolytic enzyme developed
in the cells of the animal body as a result of the first injec-
PROTEIN SENSITIZATION OR ANAPHYLAXIS 235
tion. We have repeatedly shown that the poisonous group
obtained from the protein molecule by cleavage with
chemicals or with ferments does not sensitize animals.
This is contrary to the generally accepted view, and our
claim on this point has met with either silence or denial,
but we have tested this matter so often and with poisons
obtained from so many and such a variety of proteins that
we have no hesitancy in affirming that the poisonous group
in the protein molecule does not sensitize animals. But it
is said that toxins are necessary to elaborate antitoxins,
and that the latter can be produced in no other way. This
is true, but the protein poisons are not toxins, and they
lead to the elaboration of no antibodies. The toxins are
specific; the protein poisons are not. The blood serum of
an animal treated properly with a toxin neutralizes the
toxin both in vitro and in vivo, while the blood serum of a
sensitized animal renders the protein with which the animal
has been treated, when brought in contact with it under
proper conditions, either in vitro or in vivo, poisonous. It
seems to us that it has been positively demonstrated that
the sensitizing and toxic groups in the protein molecule
are not the same. It might be argued that in ordinary
protein mixtures, such as blood serum and egg-white, one
protein may contain the sensitizing group and another the
toxic group. This may be true, but when pure proteins,
such as edestin, are used the two groups must exist in the
same molecule. The specificity of proteins is demonstrated
in sensitization. The toxic group shows no specificity.
This property characterizes the sensitizing group, and it
is in these groups that the fundamental and characteristic
property of each protein resides. The exact structure and
chemical nature of neither the sensitizing nor the poisonous
groups have been determined. The latter seems to be
physiologically the same in all proteins, the former is specific
in every protein. By our method, detailed in Chapter V,
the poisonous group is easily obtained; not in a chemically
pure condition, but so that its presence can be demon-
strated. The poisonous group, being the same in all proteins,
236 PROTEIN POISONS
is obtained from all by the same or by like methods. The
sensitizing group, being the same in no two proteins, cannot
be isolated from all by the same method. We have been
able to obtain specific sensitizing groups from colon, typhoid,
and tubercle protein quite uniformly. From the pneumo-
coccus and related organisms we have never succeeded in
obtaining a sensitizing group. From egg-white we have
rarely succeeded, generally failed. It seems evident to
us that the sensitizing groups in many proteins are highly
labile bodies, probably of such delicate structure that they
easily fall to pieces.
If sensitizers are ever to have a legitimate place in the
treatment of disease, it will be of the highest importance
to obtain them free from the poisonous group. Every
time an unbroken protein is introduced into the body it
carries with it, and as a part of it, a poison. From the very
careless, rash, and unwarranted way in which "vaccines"
of most diverse origin and composition are now used in
the treatment of disease, this matter certainly cannot be
understood or its danger appreciated by those who subject
their patients to such risks. It should be clearly understood
that all proteins contain a poisonous group — a substance
which in a dose of 0.5 mg. injected intravenously kills a
guinea-pig. This poison is present in all the so-called
"vaccines" now so largely used, and it is not strange that
death occasionally follows the use of "phylacogen" or
similar preparations. Not only do these proteins contain
a poison, but when introduced parenterally the poison
is set free, not in the stomach, from which it may be removed,
but in the blood and tissues. It is possible that vaccine
therapy may become of great service in the treatment of
disease. Even now there are occasional brilliant results
which are reported while the failures and disasters are not
so widely advertised. But before sensitization can be of
great service in a therapeutical way we must secure sensi-
tizers free from poisonous constituents. Until recently
the existence of, or the possibility of preparing non-toxic
sensitizers has been made evident only by our work.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 237
Recently, confirmation of our studies along this line have
come: (1) From White and Avery,1 who have prepared
by our method a sensitizing group from tubercle cell sub-
stance. (2) From Zunz,2 who, as the result of a most
exhaustive research, has shown that one of the primary
albumoses (the synalbumose of Pick) sensitizes, but does
not induce anaphylactic shock on reinjection. Zunz states:
Both active and passive anaphylaxis can be induced by
the three so-called primary proteoses (hetero-, proto-, and
synalbumose), but not by thioalbumose, nor the other
so-called secondary proteoses, nor by Siegfried's pepsin-
fibrin-peptone-j8, nor by any of the abiuret products of
peptic, tryptic, or ereptic digestion.
Animals sensitized with hetero-, proto-, or synalbumose
develop anaphylactic shock on reinjection with the original
serum, acid albumin, hetero- or proto-albumose, but not
after reinjection with synalbumose, thio-albumose, the other
secondary proteoses, pepsin-fibrin-peptone-|8, or any of
the abiuret products of peptic, tryptic, or ereptic digestion.
The hetero- and proto-albumoses both sensitize and induce
anaphylactic shock, while synalbumose sensitizes only.
It follows, therefore, that sensitization and the production
of anaphylactic shock are due to different groups in the
protein molecule.
Wells and Osborne,3 working with the purest vegetable
proteins known, hordein from barley, glutinin from wheat,
and gliadin from both wheat and rye, find that: "Guinea-
pigs sensitized with gliadin from wheat or rye give strong
anaphylactic reactions with hordein from barley, but these
are not so strong as the reactions obtained with the homolo-
gous protein. Similar results are obtained if the sensitizing
protein is hordein, and the second injection is gliadin. We
have here a common anaphylaxis reaction developed by
two chemically distinct but similar proteins of different
biological origin, thus indicating that the specificity of
1 Jour. Med. Research, 1912, xxvi, 317.
2 Zeitsch. f. Immunitatsforschung, 1913, xvi, 580.
3 Jour. Infect. Dis., 1913, xii, 341.
238 PROTEIN POISONS
the reaction is determined by the chemical constitution of
the protein rather than by its biological origin. This is
in harmony with the fact that chemically closely related
proteins have, as yet, been found only in tissues biologically
nearly related.
"From the results of these experiments it seems probable
that the entire protein molecule is not involved in the
specific character of the anaphylaxis reaction, but this is
developed by certain groups contained therein, and that
one and the same protein molecule may contain two or
more such groups."
Evidently the view that the protein molecule contains a
sensitizing group, one or more, is finding strong experi-
mental support. In our opinion this view was demonstrated
by Vaughan and Wheeler1 as early as 1907, but recent
work, such as that by Zunz, Gay, Wells and Osborne, and
others, strengthens the evidence then offered. According
to our theory every protein molecule contains a chemical
nucleus, key-stone or archon. This is the protein poison,
and is physiologically much the same in all proteins. One
protein differs from another in its secondary or tertiary
groups. In these resides the biological specificity of proteins.
Biologically related proteins contain chemically related
groups, and in these are found the sensitizing agents. The
chemical structure of the protein molecule determines its
biological differentiation and development. It is not,
therefore, surprising to find that a pure protein from wheat
sensitizes to another closely related protein from such a
biologically closely related grain as rye. This, however,
does not indicate that the proteins from the two grains are
wholly identical in chemical structure. It only shows that
the two protein molecules contain among their secondary
groups identical or closely related atomic combinations.
The same can be said of the fact that certain non-pathogenic
acid-fast bacteria may, at least partially, sensitize animals
to the tubercle bacillus. Biological relationship is deter-
>Jour. Infect. Dis., iv, 476.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 239
mined by the chemical structure of the protein molecule.
We hold this to be true of all specific biological tests for
proteins, whether they be agglutination, precipitin, lytic,
complement deviation, or anaphylactic tests. The chemical
structure of the protein molecule determines all of these.
The form and function of every cell is determined by the
chemical structure of its constituent proteins. That the
sensitizing agent in the protein molecule resides in its
secondary groups is shown by: (a) The fact that sensitiza-
tion is within limits specific; (6) The fact that the residues
left after stripping off these secondary groups by proteo-
lytic digestion or by the action of dilute bases and acids,
do not sensitize. Peptones, polypeptids, amino-acids, and
the protein poison do not sensitize to either themselves
or to the unbroken proteins from which they have been
derived.
The Animal. — Guinea-pigs give the most striking results.
They are easily sensitized and anaphylactic shock develops
promptly and violently. It is worthy of note that the
work of Besredka in France and of Rosenau and Anderson
showed great difference in the reaction in guinea-pigs
in the two countries. In this country practically 100 per
cent, of the animals sensitized to horse serum die on the
second injection, made intraperitoneally ; while in France
the highest percentage of fatality following the same pro-
cedure is 25. It was at first supposed that this difference
is due to the race of horse supplying the serum, but Rosenau
and Anderson, using Besredka's serum, obtained the same
results as with the American serum. The same investi-
gators say that the difference is not due to races of guinea-
pigs. In our work with egg-white we noted a much higher
percentage of mortality with short-haired, smooth-coated
animals than with the long, curly-haired ones.
Doerr and Russ,1 using ox serum, with the second dose
constant at 0.2 c.c., found the following comparative
results by varying the sensitizing dose, both doses being
1 Zeitsch. f. Immunitatsforschung, ii, 109; ibid., iii, 181.
240 PROTEIN POISONS
administered intravenously: (1) With a sensitizing dose
of from 0.01 to 0.001 c.c., the second dose was followed
uniformly by sudden death. (2) With the sensitizing dose
reduced to from 0.0001 to 0.00001, the period of incubation
was prolonged, but after this the results of the second
dose were the same as in the former instance. (3) When
the doses were further reduced to 0.000001 c.c., the animals
were not sensitized. Pfeiffer and Mita found that 0.1
c.c. of a 10 per cent, suspension of red corpuscles uniformly
sensitized guinea-pigs. Vaughan and Wheeler found that
1 c.c. of a solution of egg-white in an equal volume of salt
solution sensitized all guinea-pigs, and the result was death
within thirty minutes or less when the second dose con-
sisted of from 2 to 5 c.c. of the same solution. Wells found
the minimum sensitizing dose of the purest crystalline
egg albumen which he obtained to be 0.05 mg. Kraus
and Doerr employed from \ to 1 loop of agar cultures of
bacteria for the sensitization of guinea-pigs, but Holobut
and Delanoe found repeated injections more efficient.
In guinea-pigs subcutaneous, intraperitoneal, and intra-
venous injections of soluble proteins are practically alike
in sensitization. Rosenau and Anderson sensitized guinea-
pigs by feeding them horse serum. The question of sensi-
tization by way of the alimentary canal will be discussed
more fully later.
Rabbits are not so easily and uniformly sensitized as
guinea-pigs. Friedemann1 recommends the following method
for the complete sensitization of rabbits: The intravenous
injection of 1 c.c. of a heterologous serum per kilo; repetition
of the same after one month, and the giving of same dose
in the same way eight days later. When this is done the
animals are found to be highly anaphylactized. We have
found rabbits highly anaphylactized when treated daily
with very small intravenous injections for a week, and
then after from three to six months later given a like
intravenous injection.
1 Zeitsch. f. Immunitatsforschung, ii.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 241
Doerr and Russ were unable to sensitize mice, but Braun
succeeded after repeated injections and Ritz after a single
treatment, when the reinjection was made intravenously.
Goats and sheep have been sensitized by Friedemann
and Isaac,1 and horses and birds by Doerr.
Dogs are not easily anaphylactized. Friedemann and
Isaac, also Remlinger,2 failed, and the former suggested
that it was due to the fact that this animal is largely car-
nivorous, but Biedl and Kraus,3 using the finer method
of measuring anaphylactic shock by fall in blood pressure,
has made this animal of great value in studying the phe-
nomena of anaphylaxis. The dog has also been employed
by Arthus,4 by Manwaring, Pearce, Edmunds, and others.
Vaughan and Wheeler failed to sensitize cats by a single
intra-abdominal injection of egg-white, but repeated injec-
tions were not tried. More recently it has been shown
that the cat can be easily sensitized. Up to the present
time no animal thoroughly tested has failed to respond
to protein sensitization.
We are without sufficient data to determine with any
certainty the relative susceptibility of man to protein
sensitization. As with other animals, man's susceptibility
evidently varies within wide limits with the protein supplying
the anaphylactogen. Pirquet and Shick have shown that
a high degree of sensitization may result from a single
relatively small dose of a heterologous serum. The high
degree of sensitization shown by many in hay fever and
in susceptibility to certain foods raises questions which
will be discussed later.
Period of Incubation. — By the period of incubation we
indicate the interval of time between the introduction of
the anaphylactogen and that time when the body is recog-
nizably disturbed by a reinjection of the same or a closely
related protein. It is quite properly designated as the
1 Zeitsch. f. Exp. Path. u. Ther., i, 513.
2 Compt. rend, de la Soc. biol., Ixii, 23.
3 Wien. klin. Woch., 1909, 363.
4 Compt. rend, de 1'Acad. Sci., cxlviii, 1002.
16
242 PROTEIN POISONS
pre-anaphy lactic state. It covers the time necessary for
the development of anaphylaxis. The first injection of
the foreign protein is without manifest effect upon the
animal, but in reality it has a most profound effect. It
induces changes which may continue throughout life, and
may be transmitted from mother to offspring. The limits
of the pre-anaphylactic state have been studied only in the
guinea-pig sensitized to horse serum. In these studies
it appears that the shortest time required for the develop-
ment of the anaphylactic state is from six to nine days,
and the usual time from ten to twelve days. Otto, Rosenau
and Anderson, and Gay and Southard uniformly found
that large doses of the anaphylactogen (5 c.c. or more of
horse serum) prolonged the pre-anaphylactic state or
delayed the full development of sensitization. Friedberger
and Burkhard1 have apparently contradicted this finding,
but since the maximum dose employed by the latter was
only 1 c.c., we fail to see that there is any conflict. Evi-
dently there is a maximum amount of anaphylactogen
which the body cells can dispose of within six or eight days,
and that this for horse serum in the guinea-pig is something
more than 1 c.c., and something less than 5 c.c. As has
been stated, Doerr and Russ found that when the sensi-
tizing dose was less than 0.001 c.c. of ox serum the pre-
anaphylactic stage is also prolonged. It seems rational
to conclude from all the evidence at hand that with sensi-
tizing doses of 0.001 to 1 c.c. of serum the average duration
of the pre-anaphylactic state is from ten to twelve days,
with a minimum period of six days. With sensitizing doses
above or below these limits the period of incubation may be
prolonged.
The Anaphylactic State. — Rosenau and Anderson, also
Gay and Southard, found that guinea-pigs sensitized to
horse serum remain in this condition for at least two years.
It may possibly continue throughout life. Vaughan and
Wheeler found that guinea-pigs lose their anaphylactic
1 Zeitsch. f. Immunitatsforschung, iv, 690.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 243
state to egg-white after about one year, and after this
time they can be resensitized. The same investigators
found that guinea-pigs sensitized to colon or typhoid
proteins begin to lose their sensitization after thirty
days. Pirquet and Shick report the continuance of the
sensitized state in man after treatment with diphtheria
antitoxin up to more than seven years, and Currie1 up to
five years.
The Reinjection. — This term has come to have in this
connection a restricted and definite meaning. Repeated
injections may be employed in inducing anaphylaxis, but
by " reinjection" we mean the one made after the anaphyl-
actic state has been established. While subcutaneous,
intra-abdominal, and intravenous methods of adminis-
tration are alike suitable and effective in inducing anaphyl-
axis, the intravenous reinjection is much the most effective.
Besredka has been partial to the intracerebral introduction
of the "reinjection," and he claims that it has advantages
over the intravenous, although the latter is the most effec-
tive. With a set of highly sensitized guinea-pigs and with
the same serum he obtained the following comparative
results with the different methods of administration: Fatal
dose, intravenously, TV to ^V c-c-5 intracranially, y1^ to -§-;
intraperitoneally, only about one-half the animals responded
to 5 c.c., and subcutaneously this amount had scarcely any
effect. Besredka prefers intracranial injections because
they are not so delicate as the intravenous, and more easily
measured. In conjunction with Steinhardt2 he has developed
a method of standardizing sera. He has found that sera
differ widely in toxicity, as tested on sensitized guinea-pigs
by intracerebral injections, the fatal dose varying from \
c.c. in a sample thirteen years old, to y^-g- c.c. in some fresh
samples. This variation is in part due to age and in part
to the horses from which they are taken. His studies on
the effects of age on the toxicity of sera as tested upon
1 Jour, of Hygiene, viii. 35.
2 Ann. d. 1'Institut Pasteur, 1907, 157.
244 PROTEIN POISONS
anaphylactized animals are interesting and important.
He finds that sera from the different horses of the Pasteur
Institute, all of the same race, and on the same food, show
some, but not marked, variations. On the day that it is
drawn, horse serum is highly toxic tested by this method.
During the first ten days the toxicity rapidly decreases,
after that time more slowly. In making these tests he
uses guinea-pigs already used in the standardization of
diphtheria antitoxin, and thus saves expense. The standards
for therapeutic sera, established by the Frankfurt Institute
are as follows: (1) It must be clear and contain no marked
deposit. (2) It must not contain bacilli. (3) The highest
phenol content must be 0.5 per cent. (4) It must contain
no free toxin, especially tetanus toxin. Besredka thinks
a fifth requirement should be made, namely, that the D. L.
should be less than ^o c-c-> as tested intracerebrally on
sensitized guinea-pigs. He makes the following statement
concerning the average serum of the Pasteur Institute:
The first day it is hypertoxic (D. L. is -^ c.c.); on the
eleventh day it has fallen to one-half (D. L. is -£$ c.c.);
by the forty-fifth day the last-mentioned dose induces
severe symptoms, but does not kill; after two months it
has fallen to ^ c.c., and after this the decrease is very slow.
He thinks it wise not to use a serum less than two months
old. In France all therapeutic sera are heated to 50° before
distribution, and Besredka states that cases of serum
disease are less frequent, and when they do occur, less
serious than in countries in which unheated sera are em-
ployed. The temperature cannot be raised above 60°
without weakening the antitoxin. Rosenau and Anderson
have tried many chemicals and ferments with the hope of
destroying the anaphylactic toxicity of therapeutic sera
without injuring the antibody, but with wholly negative
results. Other methods of averting the dangers of serum
disease will be discussed elsewhere.
Besredka finds that milk may be heated to 100° for twenty
minutes, or to 120° for fifteen minutes without appreciable
loss in its anaphylactic toxicity. It begins to lose, however,
PROTEIN SENSITIZATION OR ANAPHYLAXIS 245
when the temperature reaches 130°; at 135° to 140° it
becomes gelatinous and is no longer toxic when tested on
sensitized guinea-pigs.
Symptoms. — The symptoms induced by the reinjection
of a homologous or closely related protein into an anaphyl-
actized animal vary within certain limits in different species
of animal, but in the same species are constant, whatever
the protein used. This in itself is a strong argument in
favor of the claim made by us that the anaphylactic poison
is the same, in its physiological action at least, whatever
the protein be. In other words, it is strongly in favor of the
view that all proteins, at least all which possess the capa-
bility of sensitizing animals, contain the same poisonous
group and the symptoms are due to the liberation or
activation of this poison.
When a sensitized guinea-pig receives a reinjection of
the same protein to which it has been sensitized after a
proper interval of time, certain characteristic and prac-
tically invariable symptoms develop; generally within
five or ten minutes, sometimes as late as thirty or forty
minutes. These symptoms develop in three stages, which
are best studied when they do not proceed too rapidly.
For this reason the reinjection should be made intra-
peritoneally. When given intravenously the symptoms
develop so rapidly that a study of the different stages may
be difficult or quite impossible. The first stage is that of
peripheral irritation. The animal is excited and evidently
itches intensely, as is shown by its attempts to scratch
every part of its body that it can reach with its feet. The
second stage is one of partial paralysis. The animal lies
upon its side or belly, with rapid, shallow, and difficult
breathing. It is disinclined to move, and when urged to
do so shows more or less incoordination of movement, and
muscular weakness, with partial paralysis, especially obser-
vable in the posterior extremities, which it drags. Rarely
the animal dies in this stage. The third, or convulsive
stage, begins with throwing the head back at short intervals.
The convulsions become general, more frequent and violent,
246 PROTEIN POISONS
and the animal having reached this stage, usually dies in a
convulsion or immediately following one. Expulsion of
urine and feces is frequent in the convulsive stage. Recovery
after reaching the convulsive stage is exceedingly rare.
When this stage is not reached, recovery usually occurs,
and is so prompt and complete that after a few hours, or
at most by the next day, the animal cannot be distin-
guished from its perfectly healthy fellows.
This is an exact reproduction of the picture of poisoning
an untreated guinea-pig with the protein poison of Vaughan
a"nd Wheeler, and another indication that this and the
anaphylactic poison are one and the same.
In dogs the first two stages, as seen in the guinea-pig,
occur with some variations. The first stage is one of excite-
ment. The animal moves about uneasily and cries. He
retches and sometimes vomits. Expulsion of urine and
feces frequently occurs. In the second stage, one of
great muscular weakness, he lies flat on his side or belly,
with his head on the table. When placed on his feet he
stumbles, falls, and lies with extended legs, as if paralyzed.
There may be marked expiratory spasms with retching,
and repeated expulsion of feces. There is finally suppres-
sion of urine. The animal remains in this state of depression
for many hours, and then dies or slowly and completely
recovers, so that the next day it seems as well as ever.
Again, this is a duplication of the poisoning produced in an
untreated dog with the protein poison.
Acute anaphylactic shock is seen in men being treated
with sera or other albuminous fluids. We saw it repeatedly
some years ago when we treated tuberculosis with yeast
nuclein, and the tuberculin reaction is one of sensitization.
It is not our purpose to go into detail concerning the ana-
phylactic phases seen in man under various conditions.
That part of our subject will be dealt with later. At
present we are to speak only of acute anaphylactic shock in
man. When the homologous protein is injected into a man
sensitized by disease or by previous treatments, symptoms
develop promptly, often within a few minutes, usually
PROTEIN SENSITIZATION OR ANAPHYLAXIS 247
within a few hours. The stage of peripheral irritation is
characterized by the sudden appearance of a rash. The
rashes that occur most promptly are urticarial or erythema-
tous. We have seen such a rash rapidly spread over the
surface like a blush in every direction from the point of
injection, and soon involve the entire surface. The lips and
tongue seem swollen, and often the backs of the hands are
swollen. The individual becomes apprehensive, says that
he cannot breathe, and falls into a state of more or less
marked collapse. In extreme instances there is retching,
and occasionally vomiting. The second stage, that of
great muscular weakness, continues for a variable time and
usually rapidly passes away. In rare instances speedy
death results.
The Mechanism of Anaphylaxis. — Gay and Southard1
were the first to study the pathological changes induced
by anaphy lactic shock in guinea-pigs. They reported
minute hemorrhages in the pleura and in the mucous mem-
brane of the stomach, and showed that the lungs are inflated
after death. Auer and Lewis2 made plain that death in
guinea-pigs from anaphylaxis is not due to effects on the
central nervous system, but is due to tetanic contraction of
the smooth muscles of the bronchioles. They also demon-
strated that these spasms could be averted and life saved
by preventive injections of atropine. These findings have
been fully confirmed by subsequent researches, especially
those of Biedl and Kraus. It is to the last-named investi-
gators that we owe the most complete and satisfactory
demonstration of the mechanism of anaphylactic shock.
Biedl and Kraus3 have summed up their own and the
researches of others on this point up to the time of their
writing (1910). We will first follow the summary and then
review the work done since that time.
In dogs, fall in blood pressure is a characteristic and
1 Jour. Med. Research, 1908.
2 Jour. Arner. Med. Assoc., 1909.
3 Kraus and Levaditi, Handbuch d. Technik u. Methodik d. Immuni-
tatsforschung, Erganzungsband, i.
248 PROTEIN POISONS
constant result of the reinjection. When sensitization is
not complete, fall in .pressure may be the only symptom.
In all cases there is complete parallelism between the
clinical symptoms and the fall in blood pressure. As the
latter proceeds the former increase in intensity, and in
recovery rise in pressure accompanies the return to the
normal. With a normal pressure of from 120 to 150 mm.
of mercury, soon after the reinjection, and as the pulse
grows smaller and faster and the general depression deepens,
the blood-pressure in the femoral artery falls to 80 or 60,
sometimes to 40 or even less. The character of the curve
changes, the effects of respiratory movements become less
marked, and cease altogether as the pressure approaches
the lowyest point. Now, only the movements of retching
and expiratory spasms cause transitory rises in the curve.
When the lowest point is reached the depression is greatest
and recovery is indicated by and accompanies rise in
pressure. The corneal and cutaneous reflexes remain
intact throughout, and exclude both a central narcosis and
peripheral muscular paralysis. The absence of marked
respiratory disturbances is an additional indication in the
same direction, and, furthermore, shows that the respiratory
function of the red corpuscles is not at fault.
The genesis of the fall in blood pressure becomes an
interesting question. The type of the fall and the accom-
panying condition of the pulse show that it is not due to
weakness of the heart's action. More than fifty years ago
it was shown by Marey that a fall in blood-pressure accom-
panied by increased frequency of the pulse is not due to a
decrease in the strength of the heart, but in all probability
to. decreased peripheral resistance. Both in the course of
the fall and after it has reached the lowest point there is
no irregularity in the action of the heart. On the contrary,
while in non-narcotized dogs immediately after the reinjec-
tion the heart beat becomes slower and sometimes irregular,
as the pressure falls the heart becomes and remains regular.
That the heart is not injured is furthermore shown by the
fact that with spasmodic expirations in which the abdominal
PROTEIN SENSITIZATION OR ANAPHYLAXIS 249
viscera are compressed, the pressure invariably rises. It
follows that the low blood pressure in anaphylactic shock
is due to decreased peripheral resistance from marked
peripheral vasodilatation.
The next question is to determine whether the dilata-
tion is due to paralysis of the vasomotor centre or that of
the periphery. At first it seemed that the trouble might
be central, because stimulation of the vasomotor centre
failed to increase the blood pressure. But this is negatived
by the fact that stimulation of the peripheral vasomotor
apparatus otherwise than through the centre also failed
to increase the pressure. Not only did irritation of the
terminal end of the splanchnic fail to raise the pressure,
but the intravenous injection of from 0.1 to 0.2 mg. of
adrenalin, which in normal animals is followed by marked
increase in pressure, in anaphylactic shock is either wholly
without effect or has but slight influence.
It is generally held that the capability of increasing blood
pressure possessed by adrenalin is due to its action on the
nervous apparatus in the vessel walls, and possibly in part
on the vessel muscle. It follows that the vasodilatation
of anaphylactic shock is due to paralysis of the peripheral
vasomotor apparatus. Stimulation of the vasomotor centre
naturally fails to raise the pressure because the end appa-
ratus does not work. It should be stated that this failure
of adrenalin to raise the pressure occurs only in the stage of
deep depression when the pressure is low. If the pressure
begins to rise, as recovery begins, then the administration
of adrenalin carries it up rapidly. That the fall in blood
pressure in anaphylactic shock is due to a transitory paralysis
of the peripheral vasomotor apparatus seems to be quite
conclusively demonstrated.
As has been shown by Boehm, barium chloride causes a
marked and fairly persistent increase in blood pressure,
which is due to its stimulating effect upon the smooth
muscles of the vessel walls. In anaphylactic shock even
when the pressure has fallen to the lowest point, the adminis-
tration of barium chloride causes a marked rise. Moreover,
250 PROTEIN POISONS
when barium chloride is given before the reinjection, the
latter does not cause a fall in blood pressure. Still more
striking is the fact that when barium chloride is given in
anaphylactic shock, as the pressure rises the symptoms
disappear; also, when this substance is given in doses which
cause in normal animals a marked and peristent increase
in pressure, before the reinjection, the latter induces no
anaphylactic symptoms. That the animals upon which
these observations were made were in a sensitized state
was proved by inducing passive anaphylaxis in normal
animals with their sera. It will be seen from the above
that in experimental anaphylaxis in dogs, barium chloride
is efficient both as a preventive and a curative agent.
The antagonistic action of barium chloride demonstrates
the peripheral genesis of anaphylactic vasodilatation, but
it does not wholly settle the question as to whether the
dilatation is due to the effect of the poison on the nerves
or on the smooth muscle. The failure of adrenalin and
the success of barium chloride in raising the pressure in
anaphylactic shock render it highly probable that the
anaphylactic poison lowers the blood pressure by paralysis
of the smooth muscle of the vessel walls. It seems quite
certain that barium chloride and the anaphylactic poison
act upon the same peripheral apparatus, that the action of
the former is stimulating, and that of the latter is paralyzing,
and the former is the stronger and able to prevent or replace
the latter.
Having established the fact that fall in blood pressure
is a marked and constant result of the anaphylactic poison,
the symptoms become easily explainable. The resulting
anemia of the brain explains the disturbances of respiration,
the retching, the expulsion of urine and feces, the great
depression and muscular weakness, and the speedy recovery,
when death does not result.
Biedl and Kraus give as additional phenomena of
anaphylactic shock the following: (1) On reinjection
the coagulability of the blood falls markedly or wholly
disappears. Before the reinjection is made the blood of a
• PROTEIN SENSITIZATION OR ANAPHYLAXIS 251
sensitized dog coagulates like that of a normal animal, while
that in anaphylactic shock remains fluid for hours and
even days. (2) During anaphylactic shock the polynuclear
leukocytes wholly disappear from the blood, while the
lymphocytes and platelets remain. (3) A second reinjec-
tion made in the depression phase or some hours later, or
on the next day after recovery, is wholly without effect.
This is true whether the amount of serum employed in the
second reinjection is small or large. Furthermore, the
animal is anti-anaphylactic after the shock has been either
prevented or relieved, by the administration of barium
chloride. However, Biedl and Kraus did find one dog
which rapidly recovered under barium chloride responsive
to a second reinjection made the next day.
Biedl and Kraus compare anaphylactic shock with the
poisoning of normal dogs with Witte's peptone, and find
that even in the minutest details they are not only similar,
but identical. The intravenous injection of Witte's peptone
in normal dogs in doses from 0.3 to 0.03 grams per kilo
causes fall in blood pressure, loss of coagulability of the
blood, the disappearance of polynuclear leukocytes, and
peptone immunity. Moreover, peptone poisoning can be
prevented or relieved by injections of barium chloride.
They conclude that anaphylactic intoxication is caused by
a poison which is physiologically identical with the active
constituent of Witte's peptone. This is of the highest
importance to us because we hold that the protein poison
of Vaughan and Wheeler is the active principle of Witte's
peptone, and in fact of all proteins which contain ana-
phylactogens. We have prepared this poison from Witte's
peptone as well as from other proteins, bacterial, vegetable,
and animal. We will return to this point.
It should be understood that the above extracts from the
researches of Biedl and Kraus refer only to serum anaphyl-
actic intoxication, as observed in dogs. As has been stated,
the cause of death from anaphylactic shock in guinea-pigs
was discovered by Gay and Southard and more fully studied
by Auer and Lewis, and has been confirmed by subsequent
252 PROTEIN POISONS
investigators. There is spasmodic contraction of the muscles
of the bronchioles. This is independent of central injury,
or, in other words, is due to peripheral action, and is pre-
vented or relieved by the intravenous administration of
atropine in doses of from 1 to 10 nig., provided the drug is
given before the heart stops. In anaphylactic shock in
guinea-pigs there is a primary rise in blood pressure,
which after an intravenous reinjection lasts from thirty
seconds to two minutes. This is followed by a sudden
fall which may go as low as 20 or even 10 mm. of mercury.
But the fall in blood pressure is not the cause of death.
It is on account of the difference in action of the anaphyl-
actic poison in guinea-pigs and dogs that the symptoms in
the two species vary. The sudden onset, the stormy progress,
and the fatal ending of the symptoms in guinea-pigs are
seldom or never seen in dogs. In the former, spasmodic
contraction of the bronchioles prevents the expiration of
the air, and when the lungs are laid bare they are seen to
fill the thoracic cavity; they do not collapse, and are pale
and bloodless. Biedl and Kraus have shown that these
conditions, characteristic of anaphylactic intoxication in
guinea-pigs, result also when guinea-pigs are poisoned with
peptone. Besides, fatal poisoning with peptone may be
prevented by the intravenous injection of atropine. Thus,
it is shown that in these animals also the anaphylactic
poison is identical, physiologically at least, with the active
constitutent of Witte's peptone. In dogs this poison par-
alyzes the vessel muscles of the splanchnic region, while
in guinea-pigs it stimulates the constrictor muscles of the
bronchioles.
Biedl and Kraus, having come to the conclusion that
anaphylactic intoxication and peptone poisoning are iden-
tical, discuss the poisonous property of peptone. Pick and
Spiro state that there are peptones which do not lower the
blood pressure or lessen the coagulability of the blood, and
that there are digestive products containing no albumose
or peptone, or only traces of either, which do induce these
poisonous effects. They conclude that they are peptones
PROTEIN SENSITIZATION OR ANAPHYLAXIS 253
devoid of peptone action, and there may be peptone action
without peptone. They think that in the peptic digestion
of proteins there is formed in small amount a highly poison-
ous body for which they propose the name peptozym.
Popielski,1 who has made a chemical and physiological
study of Witte's peptone, states that the albumose con-
tained in it is without effect, and that peptone prepared
from it by the method of Pick has the same, but less marked,
action as the original Witte's peptone. From this he con-
cludes that the active agent is not peptone. He also con-
cludes that, in peptic digestion, a highly poisonous sub-
stance is formed along with the peptone, and on account
of its action he proposes the name "vasodilatin." This
he obtained in an impure state by fractional precipitation
of aqueous solutions of Witte's peptone with hot, absolute
alcohol. This substance is highly active and contains
relatively small amounts of albumose and peptone, and no
cholin. This agrees well with our own work. As has been
stated, we have prepared our protein poison from Witte's
peptone, but Nicolle and Abt2 could not obtain it by our
method from Defresne's peptone, and we subsequently
confirmed this. Gastric digestion is a progressive process,
and it progresses through its successive stages at widely
differing rates. When it is arrested as in the manufacture
of peptones, the product may contain the poisonous group,
either in combination or free, or the digestion may have
continued to the destruction of the protein poison. This
seems a simple and rational explanation of the above-
mentioned findings, and reconciles their apparent contra-
dictions. If this be the correct explanation, one batch of
peptone may contain the poison, while another from the
same manufacturer may contain no trace of it. The protein
poison is a group in the protein molecule; at each successive
step in the digestive process it exists in a smaller and more
labile molecule, and finally it itself is broken up and rendered
inert.
1 Arch. f. Exp. Path., Ivi; Pfliiger's Archiv, cxxxvi.
2 Ann. d. 1'Institut Pasteur, 1907.
254 PROTEIN POISONS
We have tried to extract the protein poison from Witte's
peptone by long-continued shaking with absolute alcohol,
but with only negative results. It is probable that the
poison as it exists in peptone is in the form of a larger
molecule than is split off by our method, and consequently
is not soluble in absolute alcohol.
Passive Anaphylaxis. — The serum of a sensitized animal
introduced into a fresh animal renders the latter suscep-
tible. The second animal may be of the same or another
species. In the former case the condition induced by the
transference of the serum is known as homologous, and in
the second as heterologous passive anaphylaxis. On account
of the ease and completeness with which it is sensitized the
guinea-pig is most frequently the recipient, whatever be
the species of the donor. The transfer of the condition of
sensitization from the mother to her offspring is an illus-
tration of homologous passive anaphylaxis. This has been
studied especially by Rosenau and Anderson, Gay and
Southard, and Otto. The latter has found the young
sensitive at forty-four days after birth. Gay and Southard1
were the first to demonstrate experimental passive anaphyl-
axis. These investigators found their recipients first sensi-
tive on the fourteenth day. This indicated a somewhat
long period of incubation for the development of the ana-
phylactic state in the recipient, and this was not easily
explainable. Otto and Friedemann2 injected the anaphyl-
actic serum subcutaneously and the antigen intraperi-
toneally twenty-four hours later. With shorter intervals
they failed to obtain any response. Braun3 was the first
to inject the anaphylactic serum intravenously, but even
with this method a short period of incubation seemed to be
necessary. By injecting both sera intravenously, Doerr
and Russ4 cut down the supposed period of incubation to
four hours, but their most striking and constant results
1 Jour. Med. Research, 1907, xvi, 143.
2 Munch, med. Woch., 1907.
3 Zeitsch. f. Immunitatsforschung, 1910, iv.
« Ibid., iii, 181.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 255
at that time were obtained by injecting the anaphylactic
serum intraperitoneally and the antigen intravenously,
twenty-four hours later. Later it was shown by Doerr
and Russ, also by Biedl and Kraus, that acute symptoms,
with death, may result from the simultaneous intravenous
injection of anaphylactic serum and antigen. There is
therefore no period of incubation in passive anaphylaxis,
and this condition loses much of the theoretical importance
which was attached to it so long as it seemed to require a
period of incubation for its development. As we shall see
later, the serum and organ extracts of sensitized animals
mixed in vitro in proper proportions with the antigen produce
a poison which kills fresh animals in anaphylactic shock.
It has been found by Gay and Southard that 0.1 c.c. of
serum suffices to render the recipient passively anaphylactic.
In reference to passive homologous anaphylaxis in rabbits,
Friedemann1 makes a statement of which the following is
a summary: Among the old authorities, Weichard with
a mixture of placental cells and the antiserum, Batelli
with laked blood and the corresponding hemolysin, and
Nicolle with horse serum and anti-horse serum, succeeded.
More recent authorities2 have failed to secure passive
homologous anaphylaxis in rabbits. The divergencies are
probably explained by Friedemann, who recommends the
following: (1) Antigen and antiserum should be injected
intravenously and simultaneously. When the antigen is
introduced twenty-four hours after the serum there is no
marked reaction. (2) There is an optimum proportion
between anti-serum and antigen. By employing 2.5 c.c.
of anti-serum, Friedemann obtained no results when the
antigen varied from 2.5 to 0.25 c.c., but did obtain positive
effects when the amount of antigen was reduced to from
0.025 to 0.0025 c.c.
This corresponds exactly with our researches on anaphyl-
axis in vitro (see p. 274).
1 Jahresbericht u. d. Ergeb. d. Immunitatsforschung, 1910, vi, 67.
2 Braun, Kraus, and others.
256 PROTEIN POISONS
Doerr and Russ1 attempted to measure the antibody
in sera by the following methods: (1) A series of guinea-
pigs received intraperitoneally 1 c.c. of antiserum and
twenty-four hours later decreasing amounts of the antigen
intravenously. By this method they determined the
smallest amount of antigen necessary to induce sudden
death. (2) A series of guinea-pigs received intraperitoneally
decreasing doses of the antiserum, and twenty-four hours
later a constant dose (0.01 to 1 c.c.) of the antigen intra-
venously. In this way they determined the smallest amount
of serum necessary to induce sudden death. In our opinion
these experiments, while of great value, do not give results
which justify standardization of the so-called antibodies.
Anaphylactic shock is not determined, at least wholly, by
the amount of antigen given, nor yet by the amount of
antibody in the animal, but by the proportion between
the two.
We have stated that Gay and Southard were the first
to discover passive anaphylaxis, and that their work was
done with guinea-pigs as both donors and recipients. Our
French confreres generally credit Maurice Nicolle2 with
this discovery, and in his work rabbits served both as
donors and recipients. It seems that the work of Nicolle
was done before that of Gay and Southard, but not pub-
lished until after the work of the Americans appeared in
print. This is the statement made by Levaditi.3 However,
the work done in one country was quite independent of that
done in the other, and Nicolle showed that rabbits which
were being treated daily by intraperitoneal or intravenous
injections of 1 c.c. of horse serum furnished a serum which,
when injected into the peritoneal cavity of a fresh rabbit,
sensitized the latter, as was demonstrated by the subcu-
taneous injection twenty-four hours later of 1 c.c. of horse
serum, this injection giving rise to an inflammatory edema —
the Arthus phenomenon (see p. 262). When the reinjection
1 Zeitsch. f. Immunitatsforschung, iii, 181.
2 Ann. de 1'Institut Pasteur, 1907, xxi, 128.
3 Jahresb. u. d. Ergeb. d. Immunitatsforschung, 1907, iii, 40.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 257
was made into the brain, sudden death resulted. It will
be seen from the above that passive anaphylaxis may be
demonstrated in the recipient not only by general reac-
tion, as anaphylactic shock, but by local reaction also.
Even before the work of either Gay and Southard, or that
of Nicolle, v. Pirquet and Shick1 had, by a reversed method,
demonstrated passive anaphylaxis. To each of three rabbits
they administered 10 c.c. of horse serum; twenty-four hours
later, two received each 2 c.c. of rabbit-antihorse serum,
and 1 c.c. of rabbit serum, all subcutaneously in the ear.
In the first two, edema resulted, while in the third there
was no effect. By this reversed method anaphylactic
shock may be induced. Pick and Yamanouchi2 injected
2 c.c. of ox-serum subcutaneously in young rabbits of about
700 grams weight, and 14 days later 5 c.c. of rabbit anti-ox
serum intravenously, causing anaphylactic shock and death.
Weil-Halle and Lemain3 have observed both local and
general symptoms of anaphylaxis in both guinea-pigs and
rabbits when simultaneously on one side rabbit-antihorse
serum and on the other normal horse serum, were injected.
A few hours after such injections, the tissue around the
point of injection of the antiserum becomes edematous,
infiltrates, and becomes necrotic. Either speedy death
follows, or the animal becomes cachectic and dies after
two or three weeks.
Kraus, Doerr, and Sohma4 found that the blood-serum
of rabbits sensitized to the proteins of the crystalline lens,
renders the recipients anaphylactic to the same proteins,
and that this sensitization is strictly specific, and Doerr and
Kraus5 have made a similar showing in bacterial anaphyl-
axis. Furthermore, Richet6 has made like demonstrations
with the serum of animals sensitized to mytilocongestion
and crepitin.
1 Die Serumkrankheit, 1906.
2 Zeitsch. f. Immunitatsforschung, i, 676.
3 Compt. rend, de la Soc. biol., 1907.
« Wien. klin. Woch., 1908, No. 30.
5 Ibid., No. 28.
• Ann. de 1'Institut Pasteur, 1907-08.
17
258 PROTEIN POISONS
The following additional facts concerning passive ana-
phylaxis are of importance and will be referred to again
when we come to discuss the theories of anaphylaxis:
1. Otto1 has shown that blood-serum taken from an
animal in the pre-anaphylactic stage (eight days after the
first injection of horse serum) sensitizes recipients.
2. Gay and Southard and others have shown that the
blood-serum taken from animals in the so-called anti-
anaphylactic state sensitizes the recipients.
3. Otto has shown that a third species of animal may
be used to demonstrate anaphylaxis; thus, a guinea-pig
which has been treated with rabbit-antihorse serum mani-
fests anaphylactic symptoms when subsequently treated
with horse serum. We2 have shown that under proper
conditions the anaphylactic poison may be generated
in vitro.
Besredka3 in discussing passive anaphylaxis states that
results are inconstant. In our opinion this is due to the
difficulty in securing the proper proportion between the
antigen and the anaphylactic serum introduced into the
recipient.
Anti-anaphylaxis. — We wish to join Friedemann4 in a
protest against the use of this term. However, as Friede-
mann states, it is so deeply embedded in the literature of
anaphylaxis that it cannot be omitted. German and
American authors are not the only ones who object to the
term anti-anaphylaxis, and the explanation of it given by
Besredka and Steinhardt. Levaditi5 has produced potent
arguments against the views of his confreres. The work
done by Besredka and Steinhardt is of the highest value,
but we differ from them in the explanation of their results.
As Levaditi states, we owe the discovery of this condition
to Otto and to Rosenau and Anderson, but the name and
1 Munch, med. Woch., 1907, No. 39.
2 Zeitschr. f. Immunitatsforschung, xi, 673.
3 Kraus and Levaditi, Handbuch d. Technik u. Methodik d. Immuni-
tatsforschung, Erganzungsband i, 240.
4 Jahresbericht u. d. Ergeb. d. Immunitatsforschung, 1910, vi, 54.
* Ibid., iii, 37.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 259
the most thorough study of it have come from the researches
of Besredka and Steinhardt.
Theobald Smith1 stated that the guinea-pigs which had
received the largest doses of diphtheria toxin-antitoxin
mixture more frequently survived the second dose than
those that received smaller doses. Rosenau and Anderson2
found that animals to which they gave reinjections before
the end of the period of incubation (before twelve days)
were not responsive when another reinjection was made
at the end of twelve days, and that a longer period than
another twelve days had to pass before they became respon-
sive. All who have tested this point in serum anaphylaxis
have found that the larger the sensitizing doses employed,
the longer must be the period before typical anaphylactic
shock appears on reinjection. Moreover, all who have
experimented with anaphylactic shock, whatever the antigen
employed, and whatever the avenue of administration,
have found that animals which survive the first reinjection
are for a time thereafter, which is variable, refractory to a
second reinjection. Besredka and Steinhardt observed
that it was easy to develop the refractory state in guinea-pigs
by either of the following methods: (1) The intracerebral
injection of 0.25 c.c. of horse serum before the expiration
of the period of incubation (twelve days). (2) The intra-
cerebral injection of less than the fatal dose (^5- to -f^ c.c.)
after the period of incubation. (3) The intraperitoneal
injection of 5 c.c., after the period of incubation. The last
method seems to apply only to French guinea-pigs, which,
as we have already stated, are not so easily sensitized as
those of other countries. (4) Rectal injections. The rectum
is cleansed with a glycerin-enema and then 10 c.c. of a
dilution of serum with an equal volume of normal salt
solution is injected. This never affects sensitized guinea-
pigs, and after twelve hours they generally prove refractory
to poisonous doses given intracerebrally. Besredka3 secures
1 Jour. med. Research, 1904, xxii, 3.
2 Hygienic Laboratory Bulletin, 1906. No. 29.
3 Kraus and Levaditi, Handbuch d. Ergeb. d. Immunitatsforschung,
Erganzungsband, i.
260 PROTEIN POISONS
the refractory state in sensitized guinea-pigs by these
methods. He accomplishes a similar result by giving the
reinjection while the animal is deeply narcotized with either
ether or alcohol. He finds that in this state many animals
survive the reinjection made at any time and by any
method, and that after recovery they are completely, but
only temporarily, refractory. Of these methods he prefers
the rectal injection, or better still, the subcutaneous injec-
tion of a less than fatal dose. For the reinjection he prefers
the cerebral method. Since these investigations, made by
Besredka, are of the highest importance both theoretically
and practically, we must study them more in detail. They
are of theoretical importance because he holds that mixed
proteins contain not only sensitizing and toxic substances,
but also a vaccinating body and the last mentioned of
these he claims vaccinates sensitized animals, thus ren-
dering them immune to reinjections. This part of his
study he has carried out most thoroughly with milk, and
along this line we will follow him. Milk heated for fifteen
minutes at«120° still sensitizes and kills on reinjection, but
when heated for fifteen minutes at 130° it neither sensitizes
nor kills sensitized animals on reinjection, but does render
sensitized animals refractory to reinjections of milk heated
to only 120°. From these results he concludes: (1) That
the sensitizing and toxic components of milk behave alike
under the influence of the temperatures mentioned. (2) That
the vaccinating component can be separated from the other
two. In other words, the milk heated to 130° vaccinates or
immunizes against anaphylaxis, while it can neither sensi-
tize fresh animals nor kill sensitized ones. In order to
determine the nature of the vaccinating substance he coag-
ulates the milk with Bulgarian lactoferment and separates
the coagulum from the whey by the centrifuge or by filtra-
tion through paper. The whey immunizes sensitized
animals, but is without toxic action. The whey is neutral-
ized with soda and the flocculent precipitate which forms
is separated from the supernatant fluid by decantation of
the latter, and then made into a gelatinous mixture with
PROTEIN SENSITIZATION OR ANAPHYLAXIS 261
salt solution. This mixture renders sensitized animals
refractory, and Besredka finally concludes that the vacci-
nating constituent of milk is lactoprotein, which is not
destroyed by heating to 130° nor removed from solution
by coagulation with the Bulgarian ferment. Against this
conclusion we must call attention to the following facts
demonstrated by Besredka himself: (1) Milk or serum
introduced into the stomach, rectum, or peritoneal cavity
of sensitized animals, renders them refractory to a cerebral
reinjection. (2) Small, non-fatal, doses of milk or serum
given subcutaneously or in any other way has a like effect.
Why, therefore, is it not reasonable to say that the vacci-
nating property of the whey or of the precipitate obtained
from it is not due to the small amount of casein which it
undeniably contains? Especially is this query pertinent,
since as Besredka states, certain food authorities hold
that casein is the only protein in milk. We must conclude
that while Besredka's work along this line is most interest-
ing and valuable, he has failed to prove the existence of a
vaccinating component in milk.
Besredka's work on the refractory state, which he calls
anti-anaphylaxis, is of practical value in pointing out a
possible way by which sensitized individuals may be saved
from anaphylactic shock in the therapeutic administration
of sera. We will return to this point later (see Chapter XV).
We wish now to inquire whether there is any con-
dition that may properly be designated as anti-anaphyl-
axis. In discussing passive anaphylaxis we have seen that
serum taken from animals while in the refractory state,
whether it be before the complete development of sensiti-
zation or after recovery from a non-fatal reinjection, and
transferred to normal animals renders the recipient sus-
ceptible. It hardly seems proper to say that an animal
has been desensitized when its blood-serum has this effect.
If the blood-serum of refractory animals rendered sensitized
animals refractory, the term anti-anaphylaxis might be
proper, but this is in no case true. The refractory animal
is still sensitized, but the degree of sensitization has been
262 PROTEIN POISONS
so lowered that it no longer manifests itself in anaphylactic
shock when a reinjection is made. Moreover, in all instances
in which it has been tested, the refractory state is only
temporary, and sooner or later the sensitized condition is
sufficiently restored to be recognizable by anaphylactic
shock. We will take up this point again when we discuss
the theories of anaphylaxis.
The Arthus Phenomenon. — This phenomenon, already
referred to, deserves more detailed study. Arthus1 observed
that a single injection of horse serum into rabbits, whether
the amount was small or large, whether the injection was
made subcutaneously, intraperitoneally, or intravenously,
whether the serum was unheated or heated to 57°, had no
recognizable effect at the time or later, but that repeated
injections were followed by certain constant results. When
daily subcutaneous injections of 5 c.c. were made, the
following effects were observed: After the first three injec-
tions the serum was readily absorbed in a few hours; after
the fourth there appeared a soft infiltration about the
point and this disappeared after two or three days; after
the fifth the infiltration, which appeared, was hard, edema-
tous, and required five or six days for its absorption; after
the sixth there appeared a hard, compact, aseptic mass
which remained unchanged for weeks; after the seventh,
there was the same condition much accentuated. The
skin over the swelling became red, then whitish and dry;
the tissue became gangrenous and finally dropped out,
leaving a deep wound which slowly contracted into a scar.
This local reaction became more marked, and more extended,
with further repetitions, of the injections. This reaction is
strictly specific, inasmuch as animals first treated with
horse serum do not react to subsequent treatment with
other sera or with milk. Subsequent studies show that
this reaction can be obtained in the same way in guinea-
pigs, rats, and pigeons. It is not necessary that the sensi-
tizing and the reinjections be made in the same way. If
1 Compt. rend, de la Soc. biol., 1903, 817.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 263
the former be subcutaneous, the latter may be intraperi-
toneal or intravenous, or vice versa. However, when the
animal has been sensitized by several (six to eight) subcu-
taneous or intraperitoneal injections, and the reinjection
consisting of 2 c.c. is made into the vein, the animal, in
many instances, dies from anaphylactic shock. Imme-
diately after the reinjection it shakes its head, and evidently
becomes anxious. Breathing is frequent, polypnea reaching
sometimes as high a,s 200 to 250 respirations per minute.
There is expulsion of stool. Immobility, apnea, and exoph-
thalmia appear and the animal may die within three minutes
after receiving the reinjection. Obduction shows the heart
in systole and the blood fluid.
The rabbits which recover after manifesting disturbances
of respiration pass into a long-continued cachectic state.
Rabbits which have received daily intravenous injections
and then, after a period of time receive the reinjections
by the same route, generally, but not invariably, die in
anaphylactic shock. Sensitization by repeated intravenous
injections and subcutaneous reinjections, invariably result
in the production of the Arthus phenomenon.
Nicolle1 found that rabbits in which Arthus' phenomenon
has been developed become highly receptive to, and readily
succumb to, infection. The same investigator produced
the Arthus reaction in guinea-pigs, though he found these
animals less suitable than rabbits for the demonstration
of this form of anaphylaxis. Lewis2 also demonstrated
in guinea-pigs the phenomenon of Arthus. Remlinger3 has
done similar work. He injected massive doses (10 c.c.)
of horse serum, at intervals of one week, from six to eight
times, and one month after the last, gave a reinjection.
On doing this his animals either developed general symp-
toms or passed into a cachectic condition and died. Vaughan
and Wheeler4 killed guinea-pigs by daily intraperitoneal
injections of egg-white.
1 Ann. d. 1'Institut Pasteur, 1907, xxi, 128.
2 Jour. Exper. Med., 1908.
3 Compt. rend, de la Soc. biol., 1907, Ixii, 23.
4 Jour. Infect. Dis., June, 1907.
264 PROTEIN POISONS
Arthus and Brim1 studied the microscopic changes in
the tissues in this reaction. A fluid containing a few poly-
nuclear leukocytes first infiltrates the subcutaneous tissue.
Later, the infiltration approaches the surface, and forms a
line of cleavage between the stratum corneum and the
stratum lucidum. The subcutaneous connective tissue is
converted into a homogeneous mass, and there are extrava-
sations of blood. Finally, the process leads to necrosis with
a sharp line of demarcation. "It is an aseptic necrosis
which first involves the connective tissue and vessels, and
finally the epidermis."
From a study of the Arthus phenomenon we draw two
conclusions: (1) A prolonged period of incubation is not
necessary in order to induce the anaphylactic state. Such
a period is necessary in order to secure the explosive mani-
festation of anaphylaxis, but the development of the
specific "antibody" begins soon after the first injection of
the anaphylactic protein. (2) There is no such thing as a
condition of antianaphylaxis. If there were, certainly
animals which are receiving daily injections should manifest
it, but the only effect is to suppress the explosive character
of anaphylaxis. We will return to these questions when we
take up the theories.
Anaphylaxis and Toxic Sera. — It is well known that a
single injection, even in very small amount, of certain
sera into animals of another species proves fatal. One of
the most highly poisonous sera is that of the eel, a very
minute quantity of which injected into a guinea-pig causes
death. Doerr and Raubitschek2 have studied the toxic
and anaphylactic effects of eel serum on guinea-pigs. They
find that heating this serum to 58° destroys its toxic action.
A single dose of this heated serum has no apparent effect
upon guinea-pigs, but does sensitize them so that a second
dose of the same is followed by anaphylactic shock. This
demonstrates that the toxin and the anaphylactogen of eel
1 Compt. rend, de la Soc. biol., 1903, 1478.
2 Berl. klin. Woch., 1908, No. 33.
PROTEIN SENSITIZATION OR AN APR YL AXIS 265
serum are distinct substances, the former being thermo-
labile and the latter thermostabile. The same investigators
demonstrated the same thing in another way. The toxin
of eel serum is destroyed by acidifying the serum with
hydrochloric acid, and is not restored on neutralization
(differing in this last respect from certain other toxins,
such as those of cobra poison and of diphtheria). Eel
serum when acidified with from 0.4 to 1 per cent, of hydro-
chloric acid and then neutralized has no poisonous action
in a single dose on guinea-pigs, but does sensitize them to a
second dose of the serum treated in the same way. Pre-
cipitation of eel serum by saturation with ammonium
sulphate carries down both the toxin and the anaphylac-
togen. Ox serum behaves in a similar way with eel serum
on guinea-pigs, and it also is robbed of its toxic property
when heated to 60°. The blood serum of guinea-pigs
treated with unheated eel serum contains both antitoxin and
the substance produced by the anaphylactogen, and with
this serum fresh animals may be protected against unheated
eel serum and anaphylactized to heated serum.
It follows from these researches that the substance
elaborated in the organism by an anaphylactogen is not
an antitoxin. This does not mean that the animal which
dies from the first dose of a toxic serum and the one that
dies from the second dose of a heated serum do not die
from the effects of the same poison. The toxic serum owes
its toxicity to a ferment which splits up the proteins of
the animal's body, setting a poison free. The unheated
serum leads to the elaboration in the animal's body of a
ferment which splits up the protein of the heated serum
on the second injection, setting a poison free. With the
toxic serum the ferment is introduced into the guinea-pig,
and splits up the proteins of the body. When the heated
serum is injected the cells of the guinea-pig elaborate a
ferment which splits up the proteins of the heated serum
on its second injection. In both instances the poison is
generated by the parenteral digestion of proteins, and in
all probability is the same.
266 PROTEIN POISONS
The Toxogens. — There has been marked diversity of
opinion concerning the nature of the substance developed
in the body under the influence of the anaphylactogen.
Most German authorities, following the nomenclature
introduced by Ehrlich in his masterly studies of toxins and
antitoxins, have designated the sensitizing proteins as
antigen and the substance elaborated in the organism as
antibody. It is evident that if antigen be appropriate for
the sensitizing substance, the substance produced under
its influence must be the "antibody." We have already
expressed our opinion concerning the unfitness of the
word "antigen." The inappropriateness of the term
"antibody" is equally evident. The anaphylactogen,
instead of rendering the organism resistant to subsequent
injections, renders it more sensitive. It is true, as we shall
see later, that this increased sensitiveness may be a most
delicate and efficient means of subsequent protection.
It sharpens the agents of defence, but it does not blunt
the implements of attack. It prepares the body cells for
subsequent contests, but it does not disarm the invader.
It places in the hands of the defender more efficient means
of warfare, but it does not impair the equipment of the
attacking force. It is not a shield for protection, but a
sharpened sword for battle.
The blood serum of an animal which has been treated
with a toxin, mixed in vitro with the toxin in proper propor-
tion, may be injected into a fresh animal without effect.
The blood serum of an animal treated with an anaphylac-
togen, mixed in vitro in proper proportion with the anaphyl-
actogen and injected into a fresh animal kills it. Surely
there is no justification for the use of the terms "antigen"
and "antibody" in explaining the phenomena of sensiti-
zation. Moreover, their employment confuses and misleads,
and in our opinion they should be discarded. However, like
many other terms improperly used in scientific research,
they have become so deeply engrafted into the literature
that they cannot be eliminated, but their inappropriateness
should be clearly understood.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 267
Not all the German authorities have used the term
"antibody" in indicating the substance elaborated in the
organism in the development of the anaphylactic state.
Otto calls it the "reaction-body," and to this there can be
no objection. V. Pirquet, as we have seen, used "allergy,"
meaning altered reaction instead of anaphylaxis, and
"allergin" for the substance which reacts with the foreign
protein on reinjection. Besredka calls the sensitizing agent
sensibilisinogen, and the substance developed under its
influence, sensibilisin. Nicolle uses the term, albumino-
lysin and Richet the word toxogen. All of these are free
from the objections which we have urged to the term
"antibody." We have adopted Richet's term, but this
does not imply condemnation of the others. Indeed, the
word "toxogen" needs some explanation in order to prevent
error following its use. As we shall see later, the anaphyl-
actic poison is not a toxin. The word "toxogen" is used
by us as meaning a generator of poisons, and these poisons
are not toxins, inasmuch as they do not lead to the elabora-
tion of antitoxins when introduced into the animal body.
The toxogen is a ferment.
Pfeiffer1 long before the word anaphylaxis had been
coined really discovered the fundamental fact which later
research has confirmed. This is known as Pfeiffer's phe-
nomenon. He found that when cholera vibrios are injected
into the abdominal cavity of a guinea-pig, which has pre-
viously been immunized by repeated injections of non-fatal
doses of the living culture, they are dissolved like sugar
or salt in water. This destruction of the vibrios can be
demonstrated by microscopic study, but notwithstanding
the destruction of the bacteria, the animal is poisoned, and
dies. In fact the more powerful the lytic serum and the
more rapid and complete the destruction of the bacteria,
the more certain and prompt is death. Later, it was shown
by Bordet that with fresh lytic serum the vibrios may be
dissolved in vitro. Furthermore, it was shown by Pfeiffer
1 Zeitsch. f. Hygiene, 1903.
268 PROTEIN POISONS
that living cultures of the cholera, typhoid, colon, and
many other bacilli secrete no toxin, but that the cellular
proteins of these organisms are themselves poisonous.
By these experiments Pfeiffer laid the foundation of our
knowledge of lytic immunity, which, as we shall see, is
the chief protective function in the anaphylactic state.
The anaphylactogen which he used was the cellular protein
of the cholera bacillus. This caused the elaboration of
the toxogen, which is contained in his lytic serum, and this
digesting the anaphylactogen on the second injection, split
it up with the liberation of the poison. From these researches
Pfeiffer developed his theory of endotoxins, which we will
discuss later.
The next important work done along this line was that
of Weichardt.1 He extracted the proteins from placental
cells, and found that the blood serum of rabbits which had
received repeated injections of such extracts when mixed
with the anaphylactogen either in vitro or in vivo, produced
a poison which killed rabbits with the typical symptoms of
anaphylactic shock.
The toxogen exists in the blood serum and in the tissues
of sensitized animals, and with the former it may be trans-
ferred to normal animals, thus establishing passive anaphyl-
axis. As we have seen, passive anaphylaxis may be induced
in either homologous or heterologous animals. In the study
of anaphylactic sera one observation has, in our opinion,
led several authorities astray. It has been found that
passive anaphylaxis, in some instances at least, may be
induced with anaphylactic serum, either unheated or heated
(56°). From this it has been inferred that the toxogen is
thermostabile. In fact, the toxogen consists of amboceptor
and complement, and the latter is destroyed by a tempera-
ture of 56°; but when heated, anaphylactic serum is injected
into a fresh animal the recipient does or may furnish the
complement. Whether it does or does not, determines the
degree of success in inducing passive anaphylaxis, which,
as we have seen, is not constantly accomplished.
1 Berl. therap. Woch., 1903, No. 1.
Reference has already been made to the endotoxin theory
of Pfeiffer. This assumed the existence of a preformed
poisonous body in the cell, and cytolysis resulted in setting
it free. It was believed to be an intracellular toxin, an
independent and separate molecule, and not a group in a
more complex molecule. This theory was applicable only
to cellular proteins. An endotoxin, as understood by
Pfeiffer, could not exist in a soluble protein, and since soluble
proteins are most efficient as anaphylactogens, the theo-
retical endotoxin cannot be the anaphylactic poison. Indeed,
there is now no reason for believing in the existence of the
endotoxin. The brilliant work of Friedemann1 has shown
that red blood corpuscles may be dissolved without setting
free any active poison, and, on the other hand, the poisonous
group from the hemoglobin molecule may be extracted
without dissolving the corpuscles. Hemoglobin is not an
active poison.2 Animals are not affected by a large amount
of it given in a single dose, but it is an anaphylactogen
which means, according to our understanding at least,
that its molecule contains a poisonous group which is
liberated on reinjection through the cleavage action of the
toxogen. Friedemann showed 'that the poisonous group
can be extracted from the proteins of the red corpuscles
without dissolving them. He used 3 c'.c. of a heavy sus-
pension of washed ox-corpuscles. To this he added an
equal volume of a highly active anaphylactic serum, and
after a time separated the corpuscles in the centrifuge.
The corpuscles were again washed and incubated for
a short time with fresh rabbit serum, then placed in an
ice-box, then centrifuged, and the colorless fluid injected
into fresh animals induced anaphylactic shock. By this
method Friedemann was the first to prepare the anaphyl-
actic poison in vitro. This work has been confirmed, and
it has been shown fully that the formation of the anaphyl-
actic poison is quite independent of hemolysis. Thomsen3
1 Zeitsch. f. Immunitatsforschung, ii.
2 This is Friedemann' s statement, not ours. We have found hemoglobin
quite poisonous, even to the species supplying it.
3 Zeitsch. f. Immunitatsforschung, i, 741.
270 PROTEIN POISONS
has demonstrated that in guinea-pigs sensitized with
erythrocytes, there is no recognizable hemolysis on reinjec-
tion, although anaphylactic shock occurs. It is generally
believed, as first taught by Bordet, that in hemolysis the
stroma only is involved. When unbroken corpuscles are
used the anaphylactic poison may come from either the
hemoglobin or the stroma, or from both. We have anaphyl-
actized animals with hemoglobin and with stroma. The
former is more easily done on account, we presume, of its
more ready solubility. We have found the stroma difficult to
dissolve without using so much alkali that the preparation
is not suitable for animal injection, and suspensions, as a
rule, do not so readily sensitize as solutions. Friedberger
and Vallardi1 have found that only by having stroma,
amboceptor, and complement in proper portions can they
prepare the anaphylactic poison, an excess of any one giving
negative results. Moreover, while the poison is quickly
generated under proper conditions from the unbroken
corpuscles a much longer time is required when the stroma
only is used. It has been shown byNeufeld and Dold2
that the anaphylactic poison can be extracted from bacteria
without cytolysis. Furthermore, they have found that the
anaphylactic poison is more easily extracted from those
bacteria which are least susceptible to lytic influences.
For instance, the pneumococcus which is highly resistent
to lytic influences easily yields its anaphylactic poison,
even at 0°, while no poison is obtained at 37°, and the
cholera bacillus, which is highly labile, yields the poison
with more difficulty. Friedberger and Schiitze3 found that
the tubercle bacillus, which is highly resistent to lysis,
readily supplies the anaphylactic poison. We have shown
that the tubercle bacillus from which the protein poison
has been extracted, leaves a residue which is not only not
poisonous, but sensitizes fresh animals, and this has been
confirmed by the later researches of White and Avery.4
1 Zeitsch. f. Immunitatsforschung, vii, 94.
2 Ueber Bakterienempfindlichkeit u. ihre Bedeutung f. Infektion.
*.Berl. klin. Woch., 1911, No. 9.
4 Jour. Med. Research, 1912, xxvi, 317.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 271
Friedberger's excellent work on the extraction of the
anaphylactic poison from bacteria shows the necessity of
attending to the quantitative proportions between the
bacteria, amboceptor, and complement, also that the
poison may be destroyed by prolonged digestion. The
amount of cellular substance necessary to supply a fatal
dose of the poison is smaller than the lethal dose of the
living, unbroken cells. This confirms our work, for we
demonstrated some years ago that the protein poison is
only a part of the larger molecule which is a part of the
cell. The greater part of the bacterial cell is, after the
removal of the poisonous portion, wholly without toxic
action. Neufeld and Dold have shown that there is no
relation between the amount of poison in a given bacillus
and its pathogenic action, and Friedberger and Goldschmidt1
have obtained the anaphylactic poison from the prodigiosus
and other non-pathogenic bacteria. All of this is confirma-
tory of work we did many years ago. In 1902 we published
the following findings:
THE EFFECTS OF INTBAPEKITONEAL INJECTIONS OF THE AIR-DRIED CELLS
OF THE BACILLUS PRODIGIOSUS IN GUINEA-PIGS
No. Weight in gm. ' Dose in mg. Result.
1 260 50 +
2 305 50 +
3 287 20 +
4 ...... 272 10 +
5 260 5 +
6 270 3 +
7 252 2
8 252 1
THE EFFECTS OF INTRAPERITONEAL INJECTIONS OF THE AIR-DRIED
CELLS OF THE BACILLUS VIOLACEUS IN GUINEA-PIGS
No. Weight in gm. Dose in mg. Result.
1 220 30 +
2 230 20 +
3 255 15 +
4 265 10 +
5 210 5
1 Zeitsch. f. Immunitiitsforschung, vi, 299.
272 PROTEIN POISONS
THE EFFECTS OF INTRAPERITONEAL INJECTIONS OF THE AIR-DRIED CELLS
OF SARCINA AURANTIACA IN GUINEA-PIGS
No. Weight in gm. Dose in mg. Result.
1 240 25 +
2 300 15 -;-
3 305 10
THE EFFECTS OF INTRAPERITONEAL INJECTIONS OF THE FINELY-GROUND
CELLS OF THE COLON BACILLUS IN GUINEA-PIGS
No. Weight in gm. Dose in mg. Result.
1 172 4.09 +
2 170 4.05 +
3 165 3.66 +
4 195 2.60 +
5 135 1.80 +
6 145 1.45 +
7 165 0.825 +
8 200 0.10
9 175 0.085
10 162 0.081
As is well known, rabbits repeatedly treated with some
foreign protein, such as horse serum, furnish a serum
which precipitates the foreign protein in vitro. The rabbit
at the same time, and by the same treatment, is sensitized.
Quite naturally one suspects that toxogens and precipitins
are identical. Friedemann was the first to test this question
experimentally. He mixed the blood serum of sensitized
rabbits in vitro, in varying proportions with the homologous
anaphylactogen. The precipitate which formed was col-
lected and washed in the centrifuge; then it was digested
with fresh rabbit serum in order to supply the complement.
Such preparations after varying periods of digestion, were
injected intravenously into rabbits, but with negative results.
Friedberger,1 using guinea-pigs instead of rabbits, succeeded
fully in producing the anaphylactic poison by this method.
For two reasons the guinea-pig is better suited for this
work than the rabbit. The blood of the former is richer
in complement, and this animal is the more susceptible
to the action of the anaphylactic poison. However, in
1 Zeitsch. f. Immunitatsforschung, iv, 636.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 273
carrying out this work, even with the guinea-pig, the
results are not constant, and variations in the quantitative
relations of the solutions concerned in the reaction may
lead to failure. An excess of the anaphylactogen prevents,
apparently at least, the action of the ferment, and no
poison is formed. Here lies an important question which
we will briefly discuss. Some years ago when we were
testing the lethal doses of certain bacterial cellular pro-
teins,1 we frequently observed that a small dose killed,
while two or three or more times this amount did not; or
the smaller dose killed within a shorter time than the
larger. The proteins were administered intra-abdominally,
and in suspension. Finally, we demonstrated that the more
finely powdered cell substance was ground the more poisonous
it became. By this we mean that smaller doses killed.
Later we found, much to our surprise, that high tempera-
tures increased the toxicity of the suspensions of cellular
proteins. We came to the following conclusions: (1) The
toxicity of the bacterial suspensions is determined by the
rapidity and completeness with which the cells are split
up by the ferments of the body. (2) Other things being
equal, the rapidity and completeness with which the cells
are digested depend upon the proportion of surface exposed
to the action of the ferment. (3) Grinding the powder more
finely increases the surface exposure of a given weight, and
therefore leads to the liberation of a larger amount of the
poison in a given unit of time. (4) When the bacterial
suspensions are heated the proteins contained in the cells
are partly dissolved, or at least the molecular surface is
extended, digestion is more rapid and complete, and the
substance becomes more efficient as a poison, not because
more poison is generated, but because that contained in
the cell is made more available. Later in our work on
protein fever2 we came upon the same thing in a new guise.
We found that intra-abdominal and intravenous injections,
single or repeated, of egg-white in large doses in rabbits
1 Trans. Assoc. Amer. Phys., 1902.
2 Zeitsch. f. Immunitatsforschung, ix, 458,
18
274 PROTEIN POISONS
had but little or no effect on the temperature of the animal,
while small doses frequently repeated caused rapid eleva-
tion of temperature, and death within ten to twelve hours.
Here, again, small doses kill while larger ones are without
visible effect. The explanation is in our opinion the same
as that given for the bacterial suspensions. When 1 c.c.
of the egg-white dilution is injected into the ear vein of
the rabbit and diluted with all the blood in the animal
body, the molecular surface of the foreign protein is im-
mensely greater than when 10 c.c. of the egg-dilution is
injected. The egg-white has no poisonous action until it
is split up by ferments, and the rapidity and completeness
with which this is done is determined in part at least by
the extent of the molecular surface of the substrate. The
same thing is seen in the action of the precipitins. An
excess of the antigen prevents precipitation. We believe
the matter of molecular surface exposure to be of great
importance in the various phenomena of anaphylaxis. The
greater it is in the anaphylactogen the more potent is its
action both in sensitizing and on reinjection.
Friedberger suggests that in the preparation of the
anaphylactic poison in vitro, excess of the anaphylactic
serum or prolonged time exposure may carry digestion
beyond the formation of the poison; itself being split up.
This is in accord with our findings as reported below.
The formation of the anaphylactic poison from soluble
proteins in vitro was first done in our laboratory.1 The
importance of this matter leads us to reproduce the experi-
mental part of -our report:
Our method of procedure is as follows: Experimentation
has so far been confined to guinea-pigs. The chest of the
etherized animal is opened and the blood is drawn from
the heart into sterilized tubes and thus the serum is obtained.
The animal dies from bleeding. The organs are rubbed
up in a conical glass with sand, stirred with 30 c.c. of physio-
logical salt solution, and allowed to stand for subsidence
1 Zeitsch. f. Immunitatsforschung, xi, 673.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 275
one hour. The supernatant fluid is then removed and is
known as the organ extract. Egg-white and horse serum
are diluted with physiological salt solution until 0.1 c.c.
contains the amount of protein desired in the individual
experiment, but in case more than 10 mg. of protein is
used, a multiple of 0.1 c.c. constitutes the solvent. These
solutions are freshly prepared for each experiment, and
everything is done aseptically. The volume of serum or
organ extract used is 5 c.c., and the amount injected into
the heart of the animal is, unless otherwise noted, 4 c.c.
Further details will appear in the record of the experiments.
1. One milligram of egg-white incubated for thirty
minutes in 5 c.c. of the serum or organ extracts of a normal,
unsensitized guinea-pig is without marked effect when
injected into the heart of another unsensitized guinea-pig.
TABLE XXIV
No. Fluid. Effect.
1 .... Serum None
2 .... Liver None
3 .... Kidney Slight scratching
4 .... Spleen None
2. One milligram of egg-white incubated for thirty
minutes in 5 c.c. of the serum or organ extracts of a guinea-
pig killed three days after sensitization to egg-white is
without marked effect when injected into the heart of a
fresh guinea-pig.
TABLE XXV
No. Fluid. Effect.
1 .... Serum Slight scratching
2 .... Liver None
3 .... Kidney None
4 .... Spleen None
5 .... Brain None
3. With the conditions the same as in Table XXV, except
that the animal supplying the serum and organ extracts
had been sensitized to egg-white fourteen days before being
killed, the effects were marked as showed in Table XXVI.
276 PROTEIN POISONS
TABLE XXVI
No. Fluid. Effect.
1 . . . . Serum (3 c.c.) Dead in four minutes
2 .... Liver Dead in four minutes
3 .... Kidney Dead in four minutes
4 .... Spleen Convulsions, recovered
It should be remarked that in all instances in which
symptoms followed they were characteristically those of
the protein poison.
4. With the serum or extract of the same animal em-
ployed in Table XXVI, but with the incubation prolonged
to ninety minutes, the effects were less marked, as recorded
in Table XXVII.
TABLE XXVII
No. Fluid. Effect.
1 .... Serum (3 c.c.) First and second stages
2 .... Liver Convulsions, recovery
3 .... Spleen Slight
4 .... Kidney First and second stages
We infer from this that the digestion continued until
the poison was in part destroyed.
5. With the serum and extracts from the same animal
employed in Tables XXVI and XXVII, but the fluids after
the addition of the egg-white kept in the cold room for
twenty-four hours, the effects were not marked, as shown
in Table XXVIII.
TABLE XXVIII
No. Fluid. Effect.
1 .... Serum None
2 .... Liver None
3 .... Kidney Slight scratching
4 . . . . Spleen None
6. With the serum and extracts obtained from an animal
killed seventeen days after being sensitized to egg-white,
the effects varied with the time of incubation, as recorded
in Table XXIX.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 277
TABLE XXIX
Time of
No. Fluid. incubation. Effect.
1 ... Serum 15 min. Dead in 30 minutes
2 ... Serum 30 min. Dead in 6 minutes
3 ... Liver 15 min. First and second stgaes
4 ... Liver 30 min. Dead in 6 minutes
5 ... Kidney 15 min. First and second stages
6 ... Kidney 30 min. Dead in 5 minutes
7 ... Spleen 15 min. First stage
8 ... Spleen 30 min. First and second stages
It seems from this that fifteen minutes is too short a
time for the full development of the poison.
7. The ferment passes through hardened filter paper.
The serum and organ extracts from an animal killed twenty
days after sensitization to egg-white were filtered through
hardened paper. To each 5 c.c. of the filtrates 1 mg. of
egg-white protein was added, incubated for thirty minutes,
and then injections were made intracardiacly in fresh
guinea-pigs, with the results shown in Table XXX.
TABLE XXX
No. Fluid. Effect.
1 .... Serum (3 c.c.) Death delayed 12 hours
2 .... Liver First and second stages
3 .... Kidney Dead in 8 minutes
4 .... Spleen Dead in 6 minutes
8. The ferment passes through a Berkefeld filter. The
serum and organ extracts of an animal killed twenty-two
days after sensitization to egg-white were filtered through
a Berkefeld V. To each 5 c.c. of these filtrates 1 mg. of
egg-white protein was added, incubated for thirty minutes,
and then intracardiac injections were made in fresh guinea-
pigs with results shown in Table XXXI.
TABLE XXXI
No. Fluid. Effect.
1 .... Serum Dead in 6 minutes
2 .... Liver Dead in 9 minutes
3 .... Kidney Dead in 4 minutes
4 .... Spleen Dead in 6 minutes
278 PROTEIN POISONS
9. The poison formed by the action of the ferment on
the protein passes through hardened filter paper. One
milligram of egg-white protein was added to each 5 c.c.
of serum and organ extracts obtained from a guinea-pig
killed twenty days after sensitization to egg-white; then
these portions were incubated for thirty minutes, filtered
through hardened paper, and injected intracardiacly in
fresh guinea-pigs, with the results shown in Table XXXII.
TABLE XXXII
No. Fluid. Effect.
1 .... Serum Convulsions, recovery
2 .... Liver Dead in 12 minutes
3 .... Kidney Dead in 2 minutes
4 .... Spleen First and second stages
10. The poison passes through a Berkefeld V. The
serum and organ extracts of a guinea-pig, killed twenty-
three days after sensitization to egg-white, were treated
with 1 mg. of egg-white protein to each 5 c.c., incubated
for thirty minutes, filtered through a Berkefeld V, and
the filtrates injected intracardiacly in fresh guinea-pigs,
with the results shown in Table XXXIII.
TABLE XXXIII
No. Fluid. Effect.
1 .... Serum (3 c.c.) First and second stages
2 .... Liver Dead in 6 minutes
3 .... Spleen Dead in 10 minutes
4 .... Kidney Dead in 9 minutes
11. Serum and organ extracts from sensitized animals
are inactivated when heated to 56° for thirty minutes.
In our first experiment on this point the fluids were
placed in small Erlenmeyer flasks, and these were set in
water at 56° and allowed to stand for thirty minutes; then
5 c.c. portions, to each of which 1 mg. of egg-white protein
was added, were kept in the incubator for thirty minutes.
The result was that the inactivation was incomplete, as
shown by Table XXXIV.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 279
TABLE XXXIV
No. Fluid. Effect.
1 .... Serum First and second stages
2 .... Liver Dead in 4 minutes
3 .... Kidney First stage
4 .... Spleen None
As a control to the experiments of Table XXXIV, animals
were treated with the serum and organ extracts from the
same animal to which egg-white had been added and incu-
bated for thirty minutes without previous subjection to
heat. The results are shown in Table XXXIV A.
TABLE XXXIV A
No. Fluid. Effect.
1 .... Serum (3 c.c.) Dead in 6 minutes
2 .... Liver Dead in 6 minutes
3 .... Kidney Dead in 5 minutes
4 .... Spleen First and second stages
In a repetition of this test, the serum and extracts were
placed in thin sealed tubes and kept submerged for thirty
minutes in water at the temperature of 56°. After this
the fluids were treated with 1 nag. of egg-white protein,
incubated for thirty minutes, and injected into the hearts
of normal guinea-pigs, with the results shown in Table
XXXIV B.
TABLE XXXIV B
No. Fluid. , Effect
1 .... Serum None
2 .... Kidney None
3 .... Liver None
4 .... Spleen None
As a control to the experiments of Table XXXIV B, the
unheated extracts from the same animal were employed,
with results as shown in Table XXXIV c. There was not
enough serum to use in the control.
280 PROTEIN POISONS
TABLE XXXIV c
No. Fluid. Effect.
1 .... Kidney Dead in 8 minutes
2 .... Liver Dead in 5 minutes
3 .... Spleen Dead in 4 minutes
12. Serum and organ extracts inactivated by heating
to 56° may be reactivated by the addition of corresponding
fluids obtained from an unsensitized animal. The serum
and organ extracts obtained from an animal killed twenty-
seven days after being sensitized to egg-white were inacti-
vated by being heated to 56°, then treated with equal
volumes of serum and corresponding, organ extracts, egg-
white added, 1 mg. to each 5 c.c. of fluid, and incubated
for thirty minutes. These fluids when injected into the
hearts of unsensitized animals produced the results shown
in Table XXXV.
TABLE XXXV
No. Fluid. Effect.
1 .... Serum (3 c.c,) First and second stages
2 .... Liver Dead in 34 minutes
3 .... Kidney Dead in 56 minutes
4 .... Spleen Dead in 45 minutes
It will be noticed that when reactivated fluids were used,
death was not so speedy.
As a control to the experiments of Table XXXV, the
inactivated fluids from the same animal were used without
the addition of complement, with results shown in Table
XXXV A.
TABLE XXXV A
No. Fluid. Effect.
1 .... Serum (3 c.c.) None
2 .... Liver None
3 .... Kidney None
4 .... Spleen None
13. The serum or organ extracts of animals sensitized
to egg-white do not produce a poison when incubated with
PROTEIN SENSITIZATION OR ANAPHYLAXIS 281
horse serum for thirty minutes. The serum and organ
extracts of an animal killed twenty-six days after sensiti-
zation to egg-white were incubated with horse serum
protein, 1 mg. to each 5 c.c. of fluid, for thirty minutes,
and then injected into the hearts of unsensitized animals,
with the results shown in Table XXXVI.
TABLE XXXVI
No. Fluid. Effect.
1 .... Serum (2 c.c.) None
2 .... Liver None
3 . . . Kidney None
4 .... Spleen None
14. The serum and organ extracts of an animal sensitized
to horse serum do produce a poison when incubated with
horse serum. The serum and organ extracts of an animal
killed eleven days after sensitization with horse serum were
incubated for thirty minutes with 1 mg. of horse serum
protein to each 5 c.c. of fluid and then injected into the
hearts of fresh animals, with the results shown in Table
XXXVII.
TABLE XXXVII
No. Fluid. Effect.
1 Serum Dead in 24 minutes
2 .... Spleen Dead in 18 minutes
3 .... Liver Dead in 27 minutes
4 .... Kidney Dead in 60 minutes
5 .... Brain Dead in 4 minutes
15. The serum and organ extracts of an animal sensitized
to typhoid bacilli by the intensive method (described in
Zeitschrift /. Immunitatsforschnng, vol. ix, p. 458) when
incubated with living typhoid bacilli for thirty minutes
do produce a poison. This animal had been thus treated
twenty-seven days before. One-tenth of a loop from an
agar slant four days old was added to each 5 c.c. of fluid.
The results are shown in Table XXXVIII.
282 PROTEIN POISONS
TABLE XXXVIII
No. Fluid. Effect.
1 . Serum Dead in 4 minutes
Spleen Dead in 16 minutes
Brain Dead in 4 minutes
Liver Dead in 4 minutes
Kidney Dead in 6 minutes
16. A like experiment with the serum and organ extracts
of an animal treated twenty-eight days before with the
bacillus of cholera gave the results shown in Table
XXXIX.
TABLE XXXIX
No. Fluid. Effect.
1 .... Serum Dead in 14 minutes
2 .... Liver Dead in 6 minutes
3 .... Brain Dead in 8 minutes
4 .... Kidney Convulsions, recovery
5 .... Spleen First and second stages
17. Sensitization with egg-white in the guinea-pig does
not continue indefinitely. Vaughan and Wheeler found
that these animals, when the second injection was made
four hundred days or longer after the first, proved not to
be in a condition of sensitization, and that injections made
after this time resensitized. Apparently in cases of sensi-
tization with egg-white the serum first loses the specific
ferment. A guinea-pig killed thirty-four days after sensi-
tization to egg-white furnished a serum and organ extracts
which when incubated with egg-white and injected into the
hearts of fresh animals gave the results shown in Table XL.
TABLE XL
No. Fluid. Effect.
1 .... Serum (3 c.c.) None
2 .... Liver Dead in 7 minutes
3 .... Kidney Dead in 18 minutes
4 .... Brain Convulsions, recovery
5 .... Spleen Convulsions, recovery
PROTEIN SENSITIZATION OR ANAPHYLAXIS 283
18. When the amount of egg-white protein added to
5 c.c. of the fluid before incubation was varied, the results
as shown in the promptness and intensity of the effect of
the poison were found to vary. The serum and organ
extracts of two guinea-pigs, one killed twenty-eight days
and the other thirty days after sensitization to egg-white,
were divided into portions of 5 c.c., to which varying amounts
of egg-white were added, incubated for thirty minutes, and
injected into the hearts of fresh animals, with the results
recorded in Table XLI.
TABLE XLI
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
Amount of
Fluid.
egg-white.
Serum
0.5 mg.
Spleen
0.5 mg.
Liver
0.5 mg.
Kidney
0.5 mg.
Serum
5.0 mg.
Liver
5 . 0 mg.
Kidney
5 . 0 mg.
Spleen
5.0 mg.
Brain
5.0 mg.
Serum
10.0 mg.
Brain
10.0 mg.
Spleen
10.0 mg.
Kidney
10.0 mg.
Liver
10.0 mg.
Serum
20.0 mg.
Brain
20.0 mg.
Spleen
20.0 mg.
Kidney
20.0 mg.
Liver
20.0 mg.
Effect.
Slight
Dead in 6 minutes
Dead in 4 minutes
Dead in 8 minutes
Convulsions, recovery
Dead in 2 hours
Dead in 5 minutes
Dead in 70 minutes
Dead in 2 hours
Convulsions, recovery
First and second stages
First and second stages
Dead in 80 minutes
Dead in 60 minutes
None
Slight convulsions
Dead in 1 hour
Dead in 95 minutes
Dead in 65 minutes
19. The protein poison prepared from egg-white by
cleavage with a 2 per cent, solution of sodium hydroxide,
and still by no means pure, kills guinea-pigs when injected
into the heart in doses of 0.5 mg. The minimum fatal dose
of the pure poison must be much less than this. The symp-
toms are the same as those induced by injections of the
serum and organ extracts from guinea-pigs sensitized to egg-
white after incubation with egg-white for thirty minutes.
284 PROTEIN POISONS
It will be seen that the serum and organ extracts of
sensitized guinea-pigs contain an agent which when mixed
with homologous anaphylactogens in vitro in proper pro-
portions and incubated for the proper time produces
a poison which when injected intracardiacly into fresh
animals causes typical anaphylactic shock and sudden
death. This poison-producing agent is a ferment and is
inactivated by a temperature of 56° and reactivated on the
addition of serum or organ extracts from normal animals.
Like the toxins and many other ferments this one consists
of amboceptor and complement. The latter is destroyed
by a temperature of 56°, but being a constituent of normal
serum and organ extracts, its loss is made good on the
addition of these substances. The ferments formed in
anaphylaxis are strictly specific, their specificity being
determined by the anaphylactogen and residing in the
amboceptor. The ferment elaborated in anaphylaxis,
like the toxins, consists of amboceptor and complement.
The anaphylactogen is not a toxin and the substance
produced in the body under its influence is not an antitoxin.
The anaphylactogen does not even contain a toxin group;
it contains a poison, and it is this that is set free on reinjec-
tion. As we have stated, there is no more justification in
calling the anaphylactic ferment an antibody than there
would be in designating the proteolytic ferments of the
alimentary canal antibodies.
The Poison. — While anaphylactogens and anaphylactic
ferments are specific the poison is not specific. It is one
and the same thing whatever the anaphylactogen, and in
our opinion it is the poisonous group in the protein mole-
cule. Our studies on the protein poison, done before the
phenomena of anaphylaxis were known, demonstrated the
presence of a poisonous group in widely diversified proteins,
and it probably exists in all true proteins. We found it in
bacteria, both saprophytic and pathogenic, and, as has
been stated, we then were convinced that the pathogenicity
of bacteria bears no relation to the poison content of the
molecule of its cellular protein. The pathogenicity of a
PROTEIN SENSITIZATION OR ANAPHYLAXIS 285
bacterium is determined by its capability of growing in
and ultimately sensitizing the animal body. Furthermore,
we demonstrated, to our own satisfaction at least, that the
symptoms and lesions of the infections are not directly
due to the multiplication of the bacteria in the body, but
to their destruction by the sensitized cells of the animal
body, because at the time when the growth and multipli-
cation of the bacteria proceed most rapidly — in the period
of incubation — there are no symptoms and no lesions. The
onset of the disease marks the time when sensitization
becomes manifest.
We found the protein poison in diverse animal and vege-
table proteins, and with the few substances in which we
did not find the protein poison, such as gelatin and some
peptones, we were not able to sensitize guinea-pigs.
The reasons we have for holding that our protein poison
is identical with the anaphylactic poison may be stated
as follows:
1. The protein poison exists in all true proteins, so far
as they have been tested, consequently it exists in all
anaphylactogens.
2. Whatever the protein from which the poison is obtained,
its physiological action is the same. While there may be
and probably are chemical differences in the protein poison
as obtained from diverse proteins, physiologically there is
no difference. Likewise the symptoms in anaphylaxis are
the same whatever the anaphylactogen.
3. The symptoms induced in fresh animals by the protein
poison are identical in every detail with those observed in
sensitized animals after reinjection. They come on in the
same time, proceed in the same order, and terminate alike.
4. Friedberger has shown that guinea-pigs killed with
the protein poison show the Auer-Lewis phenomenon in
the lungs.
5. Edmunds has shown that dogs killed with the protein
poison manifest the same symptoms as those studied in
anaphylactic shock. The lowered blood pressure found
in anaphylactic shock and in peptone poisoning in dogs is
286 PROTEIN POISONS
just as marked in those under the influence of the protein
poison.
6. Our poison is the active principle in peptone, and
when it has been extracted from peptone the residue is no
longer poisonous.
7. When the poison has been removed from an anaphyl-
actogen the residue may or may not sensitize, but in
no case does it induce the symptoms of anaphylaxis on
reinjection.
8. The activity of the protein poison is progressively
increased to a certain point in proteolytic digestion. Peptone
is more poisonous than the protein from which it is formed,
and the same is true of some of the products of tryptic
digestion. The protein poison in ordinary proteins is not
active because it is combined with other groups, and as
these groups are detached it becomes more and more
poisonous. The protein molecule is a highly complex
organic compound made up of many groups, some of which
are basic, and some acid in character, and at least one which,
when detached from the others, is highly poisonous, and it
is poisonous because of the avidity with which it disrupts
the proteins of the body. To make it simpler we may say
that the protein molecule is a neutral or basic salt, and as
the basic elements are split off it becomes an acid salt, and
finally a free acid, and with each step its poisonous action
increases because its capability of depriving other salts of
their basic elements increases. Finally the acid, itself a
complex body, becomes disrupted and looses its poisonous
properties.
9. Since proteolysis is a progression in which complex
molecules are broken into simpler and still simpler ones, in
all proteolytic digestion there is an increase in the activity
of the protein poison up to a given point, when it ceases to
be a poison. It follows, therefore, that whatever the specifi-
city of the proteolytic ferment, at some stage in the process
the poison is more or less freed from the groups which tend
to prevent its action. The protein molecule has definite
lines of cleavage, and is disrupted only along these lines,
PROTEIN SENSITIZATION OF ANAPHYLAXIS 287
and in all cases its poisonous group is at some stage of the
process activated as it were. If it were not for the fact that
the poisonous group is not readily diffusible through animal
membranes, and especially through the walls of the alimen-
tary canal, all proteins would be poisonous to us even when
taken by the mouth, because the protein poison is set free
in alimentary digestion, but not being readily diffusible, it is
split up and rendered inert as digestion proceeds. When,
however, digestion is parenteral, escape from the effect
of the protein poison is impossible, and the ultimate effect
upon the organism is determined wholly by the amount ren-
dered active at one time. When it is set free with explosive
rapidity and in relatively large amount it induces anaphyl-
actic shock, and possibly death. When set free slowly and
in small amount, we have fever or fall in temperature,
according to the amount of the poison liberated. When set
free either in the circulating fluid or when it passes into
this fluid immediately we have systemic effects. When
set free locally we have inflammation in the adjacent
tissue. Narrowly used the term anaphylaxis refers to the
symptoms of anaphylactic shock. In a wider sense it
covers all the phenomena of parenteral protein digestion.
Some think that parenteral digestion is always abnormal,
either artificially induced or due to pathological conditions.
We doubt the truth of this assumption. By inhalation,
through abrasions and possibly through the alimentary
canal, man must be frequently, almost constantly, taking
into his blood and tissues very minute traces of undigested
proteins, but ordinarily the amounts thus taken in are so
infinitesimally small that the body cells are not sensi-
tized, and no harm comes. While, as we have seen, some
anaphylactogens sensitize in very small doses, these are not
infinitesimal, and there are measurable doses which do not
sensitize. The limits vary with the protein and the animal.
Friedemann and Isaac,1 also Pfeiffer and Mita,2 think
1 Zeitsch. f. exp. Path., 1905, i, 513; 1906, ii; 1908, iv, 830,
2 Zeitsch. f. Immunitatsforschung, 1909, iv.
288 PROTEIN POISONS
that the poison does not come, solely at least, from the
protein used in the reinjection. Friedemann says that it is
generally held that the poison comes from the antigen,
but that this is pure hypothesis. He holds that the poison
may come from any one of the factors in the reaction —
the anaphylactogen, the amboceptor, and the complement.
He holds that the minimum killing dose of the protein
on reinjection is so small that it cannot be supposed to
furnish a fatal quantity of the poison, and he thinks that the
ferment once set in action by the reinjection may go on and
digest the proteins of the animal body. Friedemann has
done most valuable work on metabolism in anaphylaxis, and
he holds that the increase in the nitrogen output is greater
than all of this element contained in the reinjection, and
therefore he thinks that the evidence that the whole of
the poison at least does not come from the foreign protein
is incontrovertible. He is undoubtedly right in his finding
that nitrogen metabolism in anaphylaxis is far beyond
that which can be accounted for by the nitrogen in the
foreign protein, and in this he has been confirmed by others.
Our own work1 proves the same thing, but in our opinion
this does not show that the poison itself has any other
source than the protein of the reinjection. In the first
place, as we have seen, the minimum of the protein neces-
sary to produce anaphylactic shock is much greater than
that necessary to sensitize. Rosenau and Anderson sensi-
tized one guinea-pig with 0.000001 c.c. of horse serum,
and Besredka found sensitizing doses under 0.001 c.c.
uncertain, and he found the smallest killing dose to be
^5- c.c. even when given intravenously. It will be seen
from these figures that there is a big difference between
the sensitizing and the killing dose. One-fortieth of a cubic
centimeter of horse serum contains about 2 mg. of protein.
We found in our work that serum albumin yields about
one-third its weight of poison, then 2 mg. would yield
0.66 mg., and the protein poison obtained by us, in a crude
1 Jour. Amer. Med. Assoc., 1909.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 289
way and far from pure, kills guinea-pigs when injected
intracardiacly in doses of 0.5 mg. The minimum fatal dose
of the pure poison as split off by the ferment in the body
must be much less than this. It will be seen from this
that the proteins of the reinjection, even when the smallest
fatal dose is used, probably contain enough of the poison
to kill. In the second place the ferment developed in
anaphylaxis is specific. It splits up its own anaphylactogen
and no other protein. There is no reason for supposing
that it can digest the body proteins. It seems to us that
this supposition is wholly untenable. It is contrary to all
we know about the specificity of the anaphylactic ferment.
How then may we account for the greatly increased nitrogen
metabolism? When the foreign protein is split up the
split products chemically react with the protein molecules
of the animal's body. The liberated poison tears off the
basic groups from the body molecules, and this goes on
to the extent and during the time that the cleavage con-
tinues. We see this in its most marked form in the Arthus
phenomenon, for in this the process is more localized. We
have shown that foreign proteins injected intravenously
in rabbits soon disappear from the circulating blood, and
after this they may be detected in the skin and in other
tissues. We have already seen (p. 262) to what extent
destruction of tissue may occur in the Arthus phenomenon.
There is an additional explanation of the augmented
nitrogen metabolism. The protein matter resulting from
the disruption of the molecules of the body by the split
products from the anaphylactogen is digested by the normal
non-specific parenteral ferments, and in this way nitrogen
elimination is increased. As has been stated, Friedberger
has obtained the poison by digesting precipitates and
bacterial cells with homologous anaphylactic sera. As
thus obtained and injected intravenously it induces ana-
phylactic shock and death. He has obtained like results
by digesting bacterial cells with the normal serum of guinea-
pigs. It might be assumed from this that Friedberger's
poison is not the true anaphylactic poison, but in our
19
290 PROTEIN POISONS
opinion, it is the poisonous group in the protein molecule,
and this is the anaphylactic poison, it matters not what
the agent be which has detached it from the other groups.
This agent may be wholly chemical, such as we have used
in the retort, and it may be any proteolytic ferment, the
ferment of the gastric juice, that of a specific or a non-
specific serum. As we have stated, since proteolysis consists
in the successive and progressive disruption of the protein
molecule, in at least one stage of this process, whatever
causes it, the protein poison must be released from com-
bination with those groups which in the original molecule
neutralize it. The sera of many animals, possibly of all,
contain proteolytic ferments; some are more active than
others; some act upon certain while some act upon other
proteins. The products of proteolysis resulting from different
ferments certainly differ, and even the poisonous group as
detached from the non-poisonous groups by different
ferments probably differs in its molecular structure, but
the poisonous principle is the same in all cases. Even its
physiological action may be slightly modified, though there
is no evidence of this, by variation in the lines of cleavage
along which the protein molecule is disrupted. The pieces
into which the large protein molecule is split depend upon
the shape, weight, and force of the hammer that strikes
it, and the point where the blow falls. Some of the pieces
are large and some small, and when the blow is especially
effective the pieces may be so small that the poisonous
group is broken and rendered inert; but even when protein
is fused with caustic alkali the cyanogen group is still in
evidence.
It seems to us that Friedberger's work has only confirmed
our contention, first published in 1907, that the anaphyl-
actic poison is the poisonous group in the protein molecule.
Friedberger calls his poison "anaphylatoxin." We join
Friedemann in protesting against this name. The substance
is not a toxin, as we now understand that word. Fried-
berger has demonstrated this fact himself, inasmuch as he
has shown that his poison does not induce immunity, nor
PROTEIN SENSITIZATION OR ANAPHYLAXIS 291
does it cause animals treated with it to elaborate an anti-
toxin.
That the anaphylactic poison is a protein derivative is
certain; whether it is still a biuret body has not been deter-
mined. The chemistry of the protein poison has been
discussed (p. 101).
/3-iminazolylethylamin. — This amin is produced by split-
ting off carbon dioxide from histidin, and this may be done
by either chemical or bacterial agencies. It was first
prepared synthetically by Windam and Vogt,1 and then
by Ackermann2 by the action of putrefactive bacteria on
histidin. About the same time it was detected in ergot,
and its physiological action investigated by Barger and
Dale.3 In the same year Kutscher4 isolated from ergot a
substance which chemically could not be distinguished
from this amin, but which was believed to have a some-
what different physiological action. /3-iminazolylethylamin,
hereafter designated by the abbreviation /3-i, was suggested
as a possible agent in inducing anaphylactic shock by Dale
and Laidlaw,5 who made a thorough study of its physio-
logical action. It is highly poisonous, 0.5 mg. being sufficient
to kill a guinea-pig, with all the symptoms of anaphylactic
shock, when administered intravenously. Dale and Laidlaw
describe its action on guinea-pigs as follows: "In large
guinea-pigs, weighing 800 to 1000 grams, injection of 0.5
mg. into the external saphenous vein caused death in a few
minutes. The immediate effect was a marked respiratory
impediment, resulting in violent but largely ineffective
inspiratory efforts, during which the lower ribs were drawn
in. After a time the respiratory convulsions ceased, and
the animal lay comatose, though the heart continued to
beat for some time longer. Post mortem: The lungs were
found permanently distended. If the fatal amount were
1 Berichte, 1907, xl, 3691.
2 Zeitsch. f. physiol. Chem., 1910, xlv, 504.
3 Proc. Chem. Soc., 1910, xxvi, 128.
4 Zentralbl. f. Physiol., 1910, xxiv, 163.
6 Journal of Physiology, 1911, Ixi, 318.
292 PROTEIN POISONS
given more slowly, as in two doses of 0.25 mg., after the
second of which death ensued rapidly, the final condition
of pulmonary distention was extreme. Death was clearly
due to asphyxia, evidently resulting from progressive
obstruction to the respiration, sufficient in its early stages
to prevent the exit of air sucked into the lungs by the
violent inspiratory spasms, and later becoming complete.
The larger the initial dose, and, therefore, the earlier the
obstruction became complete, the less pronounced the
distention of the lungs. Such an effect could only be due
to constriction of the bronchioles by spasm of their muscular
coats, though the effect would be aided by increased bron-
chial secretion. Preliminary injection of atropine, though
it did not abolish the action, had decided protective value.
After 5 mg. of atropine a dose of 1 mg. of /3-i intravenously
had the normal effect, but another guinea-pig, which received
a preliminary injection of 5 mg. of atropine, recovered from
subsequent intravenous injections of 0.5 mg., 0.25 mg., and
again 0.5 mg. of /3-i given in fairly rapid succession; whereas
one dose of 0.5 mg. was, in our experience, invariably
fatal when given intravenously to a guinea-pig untreated
with atropine. Whether atropine actually weakens the
bronchial spasm, or merely modifies the effect by preventing
secretion, must remain uncertain. We were unable to
remove the obstruction when once developed by a sub-
sequent injection of atropine."
In dogs and cats /3-i causes a marked fall in blood-pressure,
and in this respect also agrees with the anaphylactic poison.
On the smooth muscle, notably on that of the virgin uterus,
it has a markedly stimulating effect. It does not, according
to the findings of the English investigators, affect the
coagulability of the blood.
In a later paper Barger and Dale1 make a further com-
parison between the physiological action of /3-i and peptone
poisoning, especially with the action of the " vasodilatin"
of Popielski, and they state their conclusions as follows:
1 Journal of Physiology, 1911, Ixi, 499.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 293
"The hypothetical vasodilatine must, therefore, be
regarded as consisting of at least two substances:
"1. /3-iminazolylethalamin, causing fall of blood-pressure,
and the other characteristic effects on plain muscle and
gland cells, but not affecting coagulation of the blood.
"2. Another substance, or other substances, which renders
the blood incoagulable, and which may or may not play
some part in the other effects."
Friedberger and Moreschi1 conclude from its behavior
toward alkalies that it is not the true anaphylactic poison
which they believe to be the anaphylatoxin of Friedberger.
Biedl and Kraus2 hold, contrary to Barger and Dale,
that /3-i does delay the coagulation of blood in dogs. They
say: "In dogs 3 mg. of this substance causes immediate
fall in blood-pressure, retards the coagulation of the blood,
and induces the phenomena of anaphylaxis."
With our poison, /3-i seems to agree closely. Both induce
bronchial spasm and distention of the lungs in guinea-pigs,
and cause prompt and marked fall in blood pressure in
dogs. Neither destroys the coagulability of the blood.
In the purest form in which we have obtained it our poison
kills guinea-pigs intravenously in doses of 0.5 mg., and this
is the fatal dose of /3-i. When the active agents in our
crude poison are isolated we shall not be surprised if /3-i
or some closely allied body is among them.
/3-i has been prepared by Barger and Dale from the
mucosa of the small intestine of the ox by boiling with 0.1
per cent, of hydrochloric acid, and further treatment with
silver nitrate, and excess of baryta, according to the method
of Kutscher. In regard to this work they make the following
statement: "We have no evidence with regard to the origin
of the /3-i in the extract of intestinal mucosa. All possible
precautions were taken to avoid putrefaction before the
material was worked up. Moreover, a piece of intestine
removed immediately after death, or even during life, from
1 Berl. klin. Woch., 1912, No. 16.
2 Zeitsch. f. Immunitatsforschung, 1912, xv, 447.
294 PROTEIN POISONS
an anesthetized animal, washed, scraped, and worked up
immediately gives an extract with the characteristic physio-
logical action of /3-i. Bayliss and Starling showed that the
depressor substance could be extracted from fresh mucous
membrane of the dog's intestine by alcohol. It must
probably, then, be regarded as a normal product of intes-
tinal mucosa, though whether it is present in living cells,
or only formed when these are killed and disintegrated,
remains uncertain."
/3-i has recently become a commercial product under the
name "ergamin;" it is also known as "histamin."
Mellanby and Twort1 have confirmed Ackerman's findings2
that histidin is converted into ergamin by bacterial agencies,
and have demonstrated that it is formed in this way in
the alimentary canal. They have isolated a bacillus which
causes this conversion : " It is a small bacillus with rounded
ends, non-motile, and Gram-negative. It will grow aero-
bically or anaerobically on the ordinary laboratory media.
The optimum temperature is about 37°. The growth on
gelatin, agar, and broth is similar to that of bacillus coli.
Milk is clotted and no liquefaction of gelatin takes place.
Acid and gas are produced in media containing glucose,
lactose, or dulcite." "In the alimentary canal of a guinea-
pig, at least, and probably in that of most mammals, the
bacillus capable of producing /3-i from histidin is present
from the duodenum downward. It is legitimate, therefore,
to assume that the presence of the histidin base, described
by Barger and Dale, is due to bacterial decomposition
going on in the intestine." It grows and produces ergamin
in alkaline Ringer's solution containing 0.1 per cent, of
histidin. When the concentration of the histidin is greater,
the growth is not so prompt nor the conversion so com-
plete. "It is evident that the toxic symptoms produced
by the substance together with its presence in the alimentary
tract must bring it under consideration as a possible cause
1 Jour. Physiol., 1912, xlv, 53.
2 Zeitsch. f. physiol. Chemie, 1910, Ixv, 504.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 295
of pathological conditions. It is probable that under normal
conditions the liver can deal adequately with /3-i, as it
can with the amins of tyrosin and tryptophan, and render
it innocuous; but if this defensive mechanism of the liver
breaks down for any reason, then many toxic symptoms
will no doubt follow. For instance, one of us has elsewhere
suggested that the condition of cyclic vomiting in children
may be due to the excessive accumulation of such sub-
stances as j8-i in the intestine, causing, from time to time,
an exacerbation of symptoms. In any case a fact which
would appear to point to means of lessening the formation
of this substance in the alimentary canal is worth consid-
eration. This base is not produced in an acid medium, and
this fact is additional support to the medical treatment,
as advocated by Metchnikoff, involving the injection of
lactic acid producing bacilli. It is necessary, however,
to point out that the colon bacillus responsible for the
production of the toxic product is not killed by the acidity
of a medium, but its energies are only directed along other
lines, so that as soon as an alkaline reaction returns the
production of the histidin base continues."
The Kyrins.— These bodies have been studied and described
by Siegfried,1 who regards them as intermediate products
between the proteins and the amino acids. His method
of preparation consists in digesting the protein (fibrin,
casein, etc.) for three weeks at 38° to 39°, with from 12 to
16 per cent, hydrochloric acid. This mixture is filtered
and the filtrate precipitated with phosphomolybdic acid.
The precipitate is extracted with dilute sulphuric acid
and precipitated with alcohol. Solution and precipitation
with these reagents are repeated about fifteen times, but
after the ninth a substance of constant composition is
secured, and this is a kyrin. With the cleavage process
carried one step farther these bodies are converted into
amino acids. It will be seen from the method of preparation
that they are closely related to the diamino acids (arginin
1 Zeitsch. f. physiol. Chem., 1906, xlviii, 54.
296 PROTEIN POISONS
and lysin). The precipitate with phosphomolybdic acid
may be crystallized, and the picrate differs from the corre-
sponding salts of arginin and lysin by its solubility in alcohol.
The kyrins give the biuret reaction, the color differing
from that given by peptone in being more distinctly a
Bordeaux red. Siegfried states that the kyrin formed
from fibrin splits on further cleavage into lysin, arginin,
and glutamic acid.
The kyrins are said to be highly poisonous, and Kam-
mann1 has suggested that they, or similar bodies, may be
the active agents in the production of anaphylactic shock.
However, we have been unable to find any record of thor-
ough studies of the poisonous action of these cleavage
products. Schittenhelm and Weichardt2 have studied
two kyrins, one prepared from hemoglobin and the other
from gelatin. The latter is not poisonous, and this agrees
with our work in which we failed to obtain the poisonous
group from gelatin. The globinokyrin is moderately
active, but not so poisonous as the protamins.
Anaphylatoxin. — Friedberger3 treated rabbits with lambs'
serum until he obtained abundant precipitates with the sera
of these animals. These precipitates were deposited in
a centrifuge, washed with salt solution, and then digested
with normal guinea-pig serum for some hours in the incu-
bator. When this was done and the serum decanted and
injected into normal guinea-pigs, the animals promptly
died with all the symptoms of anaphylactic shock. A
poison had already been obtained in a similar manner by
Weichardt from placental tissue, and by Friedemann from
blood corpuscles (see p. 269). Friedberger named the poison-
ous substance which he obtained by the digestion of specific
precipitates with normal serum, anaphylatoxin. Desig-
nating this substance as a toxin might be criticized, but
at that time Friedberger believed in his theory of sessile
receptors, and it is plain that he regarded the poison which
1 Zeitsch. f. Immunitatsforschung, 1911, xi, 659.
2 Ibid., 1912, xiv, 609. 3 Ibid., 1910, iv, 636.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 297
he had prepared as a toxin. One of the conclusions stated
in the paper in which he reported this work is as follows:
"Die Bildung eines Antitoxin gegen das Anaphylatoxin
ist mir bischer noch nicht einwandsfrei gelungen, jedoch
ist sie wahrscheinlich." He also concluded that the poi-
sonous action of this substance is destroyed by a temperature
of 65°. So far he has not announced the successful prepara-
tion of an antitoxin, and further research has convinced
him that anaphylatoxin is thermostable. In these respects,
therefore, anaphylatoxin does not differ from the poison
obtained by the cleavage of proteins with chemical agents
or ferments. Later, Friedberger1 became convinced that
anaphylatoxin is not a specific body; it is a common product
of the cleavage of diverse proteins. This, also, distinguishes
it from toxins, one invariable characteristic of which is their
specificity.
One of the most important contributions made to the
literature of anaphylaxis is the paper by Friedberger and
Vallardi.2 In this contribution Friedemann is properly
credited with having been the first to produce the anaphyl-
actic poison by ferment action in vitro. The following
statement is made: "Concerning the nature of anaphyl-
atoxin we know nothing, but we are justified in assuming
its close relationship to the split product obtained by
Vaughan and Wheeler through the action of alkaline
alcohol on proteins, also the similarity of its action with
that developed by poisoning with peptone, as shown by
Biedl and Kraus, and Pfeiffer and Mita, is evident."
With specific precipitates, the stroma of blood corpuscles,
and with whole corpuscles, under the action of ambo-
ceptor and complement, anaphylatoxin was developed,
and its effect on fresh animals was demonstrated. The
specific precipitates were obtained by treating rabbits with
lambs' blood and then mixing the sera from these animals.
Such precipitates were collected in a centrifuge, washed
1 Zeitsch. f. Immunitatsforschung, 1910. vi. 179.
2 Ibid., 1910 vii, 94.
298 PROTEIN POISONS
twice with salt solution, and then incubated with the
serum of a normal guinea-pig. The ferment in the serum
split up the precipitate with the liberation of the poison,
and when the serum containing the poison was injected
into a fresh guinea-pig the animal promptly died, with all
the symptoms of anaphylactic shock. Washed stroma
treated in a similar way yielded the same poison. With
whole red corpuscles the results are complicated by the
poisonous action of the liberated hemoglobin, which is an
active poison without further cleavage. Friedberger desig-
nates the precipitates, stroma, and corpuscles subjected to
the action of the normal serum as antigens. We have a
marked antipathy to the use of this term in discussing
the phenomena of anaphylaxis, and would designate the
substances submitted to the action of the normal serum as
substrates, and regard the serum as containing the ferment.
It might be said that the terms we use are of but little
importance, and the meaning is the one important thing.
This is true, but we use words to express ideas, and we
hold that the term "antigen" in this connection confuses
and tends to lead to gross misconception. As has been
shown by Friedberger and others, anaphylatoxin may be
obtained by incubating various proteins with normal
serum, and why should we call the substances thus split
up by the ferment in the serum an antigen? Would it not
be just as proper to denominate starch which is converted
into sugar by amylase an "antigen?" But this is a digres-
sion, and we will return to the work of Friedberger and
Vallardi. They found that when specific precipitates and
stroma were used the amount of the poison obtained was
not in proportion to the amount of substrate used. With
a larger amount of substrate, that of the serum (ferment)
being constant, they obtained no poison. With the amount
of substrate very small, they obtained either no poison or
at least not enough to demonstrate its presence by its
effect on the animal. They obtained positive results, as
shown by anaphylactic death only when the amount of
substrate but slightly exceeded that necessary to kill a
PROTEIN SENSITIZATION OR ANAPHYLAXIS 299
sensitized animal on reinjection. When the amount of
substrate employed was very small they obtained either
too little of the poison to affect the animal, or, what is more
probable, the digestion was so active that the poison itself
was destroyed. When whole blood corpuscles were used
the amount of poison obtained did increase with an increase
in the substrate, but the poison thus formed in greater
abundance was hemoglobin. At least this is our explana-
tion of their results. On the theory of Friedberger, his own
results are difficult, or, as we think, impossible of explana-
tion, while on our theory they explain themselves, and,
in fact, are exactly what might have been expected.
Friedberger and Vallardi find that both the subjective
and objective symptoms of poisoning with anaphylatoxin
are identical with those of both active and passive anaphyl-
axis, and in this we quite agree with them. Biedl and
Kraus have held, and still hold, that anaphylatoxin cannot
be the true anaphylactic poison because, as they claim,
the condition of the lungs after death in guinea-pigs from
this poison is not the same. We agree with Friedberger,
who holds that the distention of the lungs after anaphyl-
actic death, first mentioned by Gay and Southard, and
more fully emphasized by Auer and Lewis, is found after
death from naturally poisonous sera, from poisonous anti-
sera, from peptone as shown by Biedl and Kraus, from the
poison of Vaughan and Wheeler, from the /3-i compound of
Barger and Dale, and possibly after poisoning from other
substances as well. Lung distention due to constriction
of the bronchioles is, in guinea-pigs at least, a constant
result of the protein poison, but should not be considered
as pathognomonic of this poison. Friedberger and Jeru-
salem1 attempted to isolate and study the physical and
chemical properties of anaphylatoxin. It should be under-
stood that the poison as they had it is in guinea-pig serum.
They state their conclusions as follows: (1) The solution
can be evaporated to dryness (in vacuo) without loss of
1 Zeitsch. f. Immunitatsforschung, 1910, vii, 748.
300 PROTEIN POISONS
toxicity. (2) By evaporation and resolution in smaller
volume it can be concentrated. (3) It cannot be extracted
from the serum by ether or chloroform. (4) It can be
precipitated without loss of toxicity by alcohol.1 (5) Ana-
phylatoxin is not a globulin. (6) It can be obtained by the
action of complement upon heated (as well as unheated)
precipitates.
Biedl and Kraus claim that anaphylatoxin cannot be
the true anaphylactic poison because it does not induce
anaphylactic shock when injected into the brain, and
Besredka has shown that with serum the reinjection of a
very small dose into the brain causes the shock. To this
Friedberger very properly replies that the only form in
which he has obtained the poison is in solution in guinea-
pig serum, and he cannot introduce a large enough quan-
tity of this into the brain without doing mechanical injury,
and, moreover, the minimum reinjection dose in the brain
is not smaller than that required intravenously. It should
always be borne in mind that Friedberger's anaphylatoxin
is a solution of a poison whose physical and chemical
properties are not known in blood-serum.
Friedberger2 states that after death from anaphylatoxin
the blood does not coagulate. This seems to complete the
identity of the action of this poison with that formed in
anaphylaxis. The symptoms and all the postmortem
findings seem to be identical.
Friedberger3 and his students showed that various bacteria,
such as the vibrio of Metchnikoff, the prodigiosus, the
typhoid and tubercle bacillus, when incubated with normal
guinea-pig serum, furnish a soluble poison which, when
injected into normal animals intravenously, causes anaphyl-
actic shock and death. The poison obtained from these
diverse bacterial proteins as well as that obtained from
1 It will be understood that the proteins of the serum were precipitated
by the alcohol and the poison carried down with the precipitate. It does
not mean that the poison would necessarily be precipitated from aqueous
solution by alcohol.
2 Zeitsch. f. Immunitatsforschung, 1910, viii, 239.
3 Ibid., 1911, ix, 369.
PROTEIN SENSITIZATION OR .ANAPHYLAXIS 301
specific precipitates, blood corpuscles, stroma, and other
proteins, is in all instances the same in its physiological
action. It matters not whether the bacteria submitted
to the action of the serum be heated or unheated, the
result is the same. In other words, Friedberger and his
students accomplished with the proteolytic ferments in
blood-serum by the cleavage of proteins just what we did
nearly ten years earlier by chemical agents. We demon-
strated that all proteins, living or dead, formed or without
form, contain a poisonous group, and that the physiological
action of this group is the same whatever the protein from
which it is obtained. We split up the pathogenic and non-
pathogenic bacteria, vegetable and animal proteins of the
most diverse kind, and obtained from each and every one
the same poison, and now the same has been accomplished
by ferments. This we regard as a confirmation of our
statement made many years ago that the protein molecule
contains at least one poisonous group. Besides, the poison
obtained by us, when we split up proteins with chemical
agents, is the same or very closely related to that now
obtained by the cleavage of the same proteins by the more
delicate agency of ferment action. As we stated at the
time, our method was crude, and the poison was obtained
only at great loss, but the principle is the same. From our
work we developed the theory of the relation of the split
protein products to immunity and disease, which was
formulated in 1907, and which, in our opinion, is confirmed
in every particular by the work of Friedberger and others.
We fail to see why he and our German confreres in general
still use the Ehrlich nomenclature in discussing the protein
split products. Bacterial cellular substance is submitted
to a ferment in vitro, and broken up, and why should
this substance be called "an antigen" and the ferment
an "antibody?" The theory of sessile receptors, the only
theory, so far as we know that Friedberger ever originated,
was long ago demonstrated to be false by his own work.
He has adopted another theory, one which he certainly
did not originate, but for the establishment of which he
302 PROTEIN POISONS
has done much, and still he employs the language of his
own theory long since discarded by himself.
Friedberger and Nathan1 showed that by the action of
normal guinea-pigs' serum on normal horse serum, or vice
versa, in proper proportions, a poison is set free. The serum
that is to serve as substrate is inactivated by being heated
to 56°, and this is then acted upon by the ferment in the
unheated serum. It will be understood that the amount
of the substrate must be small. In fact, it was found that
the poison is produced when the substrate contained not
more than 1 mg. of protein. When guinea-pig serum was
used as the ferment, it was found to act best in quantities
of about 6 c.c. For instance, when inactivated horse serum
in quantities of from 0.01 to 0.0005 c.c. was incubated with
6 c.c. of normal guinea-pig serum for eighteen hours, and
then 4 c.c. of this injected intravenously into guinea-pigs
of about 200 grams, the animal promptly died from anaphyl-
actic shock. On the other hand, when guinea-pig serum
was used as the substrate and horse serum as the ferment,
somewhat larger quantities of each were needed. For
instance, with 0.1 c.c. of inactivated guinea-pig serum
incubated with 8 c.c. of horse serum for twenty-four hours
a fatal amount of the poison was obtained, while with the
substrate reduced to 0.01 c.c. no symptoms were induced.
Friedberger, in reporting this work, expressed astonishment
that from a very small amount of protein, enormous quan-
tities of which, in its unbroken state, could be injected
into animals without recognizable effect, there could be
obtained a potent poison, and still years before we had
split up these proteins by chemical means and obtained
the same poison. What we had done with chemical agents,
Friedberger did with ferments. We claimed ten years ago
that we had split the protein molecule along definite lines
of cleavage, and that our product was not a mere degre-
dation body. This demonstration that ferments split the
protein molecule along the same lines is a justification of
our claim.
1 Zeitsch. f. Immunitatsforschung, 1911, ix, 567.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 303
This work of Friedberger, in our opinion, confirms another
claim which we put forth some years ago. There are two
kinds of parenteral proteolytic enzymes in the body, or
capable of being developed in the body. One of these is
non-specific and the other specific. The former is found
in the normal blood serum, and the latter is developed by
protein sensitization. The normal serum of this guinea-
pig under proper conditions splits up most diverse proteins
with the liberation of the poisonous group. The blood-
serum and organ extracts of the sensitized animal contain
specific ferments which cleave the special protein to which
the animal has been sensitized. The non-specific protein
takes care of the small amounts of foreign protein which are
constantly finding their way into the blood without having
undergone digestion. Ordinarily, these enter the blood
in such small amounts that they are rapidly and fully
digested beyond the poisonous stage by these non-specific
proteolytic ferments. Among its other functions the blood
is a digestive fluid, and it exercises this function not only
on the unbroken proteins which find their way into it from
the outer world, but also upon certain substances which are
constantly coming into it as a result of tissue metabolism.
Another important research from Friedberger's laboratory
is reported by him and Girgolaff.1 Rabbits and guinea-pigs
were treated with homologous proteins, bacteria, and sera,
and after the animal had developed the specific ferment
(antibody) it was bled to death by opening the aorta and
transfused with salt solution until all the blood was washed
out. Then a portion of some organ from this animal was
implanted in the abdominal cavity of another, and after
recovery from the operation this animal was found to be
sensitized. A few illustrations will best explain this work. A
guinea-pig of 200 grams received 1 c.c. of lambs' serum intra-
venously. Fourteen days later this animal was exsanguinated
and washed out with salt solution. Then two pieces of its
spleen — about half of this organ— were implanted in the
1 Zeitsch. f. Immunitatsforschung, 1911, ix, 575
304 PROTEIN POISONS
abdominal cavity of a fresh guinea-pig of about the same
weight. Fourteen days later this animal received intra-
venously 1.5 c.c. of lamb serum, and promptly died of
anaphylactic shock. Autopsy showed the lungs distended,
the heart still beating, and the blood of the right heart had
not coagulated at the expiration of ten minutes.
The organ may be implanted into another species, as is
shown by the following: A rabbit was treated with lamb
serum until the precipitation titer of the serum was 1 to
1000; then the rabbit was exsanguinated and perfused with
salt solution, and two pieces of its spleen implanted in
the abdominal cavity of a guinea-pig. Fourteen days
later this guinea-pig received intravenously 1.5 c.c. of lamb
serum, and died of anaphylactic shock. Guinea-pigs thus
treated were found to be sensitized not only to lambs, but
also to rabbit serum, thus proving that the implanted
organs continued to secrete both their normal and their
specifically developed proteolytic ferments (antibodies).
Like results were obtained when pieces of kidney were
transplanted. Moreover, the animals thus sensitized by
the receipt of transplanted organs retained their sensitized
condition for a time at least after the removal of the im-
planted tissue. The following is an illustration. A guinea-
pig received intravenously 1 c.c. of lamb serum. Five
days later it had two subcutaneous injections of 1 c.c. of
lamb serum. Eight days later it was killed and transfused,
and portions of its spleen and kidney implanted in a fresh
guinea-pig. Six days later the implanted tissues were
wholly removed, and after the guinea-pig had fully recovered
from this operation it was found to be still sensitized to
lamb serum. Evidently the implanted tissue had not only
continued to develop its specific ferment, but had discharged
it in part at least into the blood. These experiments are
of the highest value for two reasons: (1) They caused
Friedberger to wholly abandon his theory of sessile receptors,
and (2) they show that the specific ferment developed in
protein sensitization is a cellular product, and that the
cells of the spleen and kidney, possibly of other organs as
well, elaborate it.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 305
Later this work was continued by Girgolaff,1 who discusses
the possible explanation of these findings. (1) It might
be possible that the recipient is passively anaphylactized
by the' transfer of some serum from the donor. This suppo-
sition is held untenable for two reasons. First, the washing
out is so thoroughly done, and second, the volume of the
piece of organ transferred is too small. Moreover, if it were
passive anaphylaxis, the recipient should be most respon-
sive to reinjection within a day or two, while in fact it is
not responsive until after seven or eight days. (2) It might
be suggested that some of the protein used in sensitizing
the first animal is carried over to the second and actively
sensitizes it. This is highly improbable on account of the
small amount of protein used in sensitizing the first animal;
the length of time (in some cases fourteen days) elapsing
before the transfer of the tissue, and the thorough washing
out given the first animal. Besides, this was shown to be
impossible because in one instance a rabbit was killed and
its organs transferred to another rabbit three hours after
the former had received a bacterial suspension, just at the
time when the bacilli should have been most abundant in
the tissue, and these animals were not sensitized. (3) The
only conclusion which seems to have any justification is
that the cells of the tissue removed, having acquired a new
function while in their normal location, continue to exercise
this function in their new location. We regard this as a most
complete verification of our theory of anaphylaxis, in which
we hold that a new function is developed in certain cells of
the animal body by the sensitizer. Moreover, it does seem
that this work should lead to the discarding of all theories
involving an "antigenrest," about which much has been
said.
Vaughan, Vaughan, Jr., and Wright2 demonstrated that
the serum and organ extracts of normal guinea-pigs do
not form a poison when incubated with egg-white, but
1 Zeitsch. f. Immunitatsforschung, 1912, xii, 401.
2 Ibid., 1911, xi, 673.
20
306 PROTEIN POISONS
that corresponding preparations from guinea-pigs sensitized
to egg-white do (p. 274).
Friedberger and Mita1 show that summer frogs can be
anaphylactized. From 0.1 to 0.5 c.c. of lamb serum was
given by the abdominal vein, or into the dorsal lymph sac
as a sensitizing dose. From one to four weeks later a rein-
jection causes characteristic symptoms. The animal soon
becomes stupid and lies with extended limbs. When placed
on its back it does not regain its normal position. Sudden
death does not follow, and the animal usually survives
from twelve to twenty-four hours. When the heart is
watched through a fenestrated chest, the pulse is seen to
grow slow and irregular, and finally the heart stops in
diastole. Anaphylatoxin was found to have a similar
action on the isolated heart.
Friedberger and Scymanowski2 show that the presence
of leukocytes lessens the formation of anaphylatoxin, and
apparently destroys it when abundantly formed. They
question whether this is due to an activity of the leukocyte
or to its absorption of the poison. We suggest that the
leukocytes destroy the poison by digesting it and converting
it into a harmless body.
More than ten years ago (see p. 46) we showed that the
poison contained in the cellular substance of the diphtheria
bacillus is a wholly different thing from the toxin elaborated
by the same organism. This convinced us that the protein
poisons — substances obtained by the cleavage of the protein
molecule — are not toxins. When the diphtheria bacillus grows
it elaborates and excretes a soluble ferment known as diph-
theria toxin. When injected into animals in sufficient doses
this toxin kills after from two to five days. When repeatedly
injected in smaller doses the body elaborates an antibody —
an antitoxin. When the cellular substance of the diphtheria
bacillus is split up by our method a poison is obtained.
This is not a toxin, but a poison. When injected into
1 Zeitsch. f. Immunitatsforschung, 1911, x, 362.
* Ibid., 1911, xi, 485.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 307
animals in sufficient doses it kills within a few minutes —
not after days. When repeatedly injected in non-lethal
doses the animal body does not elaborate an antibody —
an antitoxin. Diphtheria toxin is specific; it is an exclusive
product of the diphtheria bacillus, and animals treated
with it in the proper way produce a specific antibody.
The cellular poison of the diphtheria bacillus is not specific;
the same poison is contained in other proteins, and it gives
rise to no specific antibody. We convinced ourselves
many years ago that the protein poisons and the toxins
are not related bodies, and we demonstrated that diphtheria
toxin gives no protection against poisoning with the active
substance contained in the protein molecules making up the
cell substance of the diphtheria bacillus. For this reason
we have never discussed the phenomena of sensitization
due to protein poisons in terms applicable only to toxin.
Years after we did this work Friedberger and Reiter1 con-
firmed it by showing that the protein poison obtainable
from the cellular substance of the dysentery bacillus is a
wholly different substance from the toxin of the same
bacillus, but they make no mention of our work in this
connection, and they still call the protein poison a toxin
and speak of the antibody.
It has been fully demonstrated that the protein poison
can be detached from its combination in the molecules of
specific precipitates, blood corpuscles, stroma, many bac-
teria, etc., by incubation with normal guinea-pig serum.
Other proteins require for their disruption and for the
liberation of the poisonous group, sera in which specific
ferments have been developed by sensitization. The
proteolytic ferment in normal guinea-pig serum is not
specific. It is capable of digesting many, but not all,
proteins. When a guinea-pig has been sensitized to a
given protein, its serum contains not only the general,
non-specific proteolytic ferment normal to it, but in addition,
the specific ferment. Whether the latter is a wholly new
1 Zeitsch. f. Immunitatsforschung, 1911, xi, 493.
308 PROTEIN POISONS
product or is due to a modification in. the former we cannot
say. For the present we will confine our attention to the
general, non-specific proteolytic ferment in normal guinea-
pig serum. Like all ferments, it is supposed to consist of
two parts. (1) A thermostabile part, known as the ambo-
ceptor, and (2) a thermolabile part, known as complement
or alexin. The latter is destroyed by a temperature of 56°,
and serum heated to this point is inactivated. It has been
found by most observers that inactivated serum from the
guinea-pig, or, in fact, any inactivated ferment solution,
does not function. Seitz1 found in a few instances that
by incubating certain bacteria with inactivated serum he
obtained a free poison, but this is contradicted by the
experience of so many investigators that we must conclude
that his technique was defective.2 There are, however,
some points of real interest in connection with this reaction.
Since bacteria and certain other proteins, when incubated
with normal serum, yield a soluble and active poison, why
does this reaction not occur when these proteins in the
unbroken condition are injected directly into the blood?
The only answer to this question seems to be that the ferment
is in a more readily available form in the serum than it is
in blood. The ferment in the serum probably comes largely
from the breaking down of leukocytes. When an unbroken
protein is injected into an animal usually the first effect
upon the blood is a leukopenia. Certainly there is for a
time a diminution of the leukocytes in the peripheral blood.
After a time there is a hyperleukocytosis, and this is generally
believed to be for the purpose of breaking up the foreign
protein. There is, however, another possible explanation.
It may be that in the circulatory blood the disruption is
carried beyond the point of setting free the poison. It may
itself be disrupted and converted into relatively harmless
substances. There is also the possibility that the inclusion
of the foreign protein by the phagocyte may delay the
disruption of the former.
1 Zeitsch. f. Immunitatsforschung, 1911, xi, 588.
2 See article by Lura, ibid., 1912, xii, 467.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 309
It seems that all bacteria, both pathogenic and non-
pathogenic, at least so far as tested, yield a poison when
incubated with normal serum of the guinea-pig. Some are
acted upon or disrupted more promptly and quickly than
others, but with prolonged incubation all yield enough
of the poison to affect animals to a recognizable extent.
This confirms our work in which it was shown that all
proteins, so far as tested, yield a poisonous fraction when
properly disrupted by chemical agencies. It was this
work that led us to conclude that every protein molecule
contains a poisonous group. Bold and Aoki1 have obtained
the poison by incubation with serum from streptococci,
meningococci, gonococci, b. mallei, pestis, pneumonise,
paratyphus, chicken cholera, swine erysipelas, yeast cells,
actinomyces, the spirochetes of chicken spirillosis, and
Russian relapsing fever. They did fail to obtain it from
the spores of certain molds, but this does not prove that
these spores do not contain a poison. It simply shows that
the proteins of these spores are resistant to the cleavage
action of the ferment contained in the normal serum of
the guinea-pig.
Boehneke and Bierbaum2 find that repeated, alternate
freezing and thawing have no effect upon either the substrate
or the ferment in the production of the poison, and no
effect on the poison itself.
Bessau3 sensitized animals simultaneously with ox and
horse sera. The injections were made on each side of the
thorax subcutaneously. After full sensitization had been
developed, as was demonstrated by reinjection of controls
which had been sensitized to only one serum, those doubly
sensitized were given sublethal reinjections of one of the
sera, and after recovery from the effect induced they were
found to be less susceptible to reinjections of the second
serum. He also determined the minimum fatal dose of
anaphylatoxin prepared from typhoid bacilli on fresh
1 Zeitsch. f. Immunitatsforschung, 1912, xii, 200.
2 Ibid., 1912, xiv, 130.
3 Centralbl. f. Bakteriol., 1911, Ix, 637.
310 PROTEIN POISONS
guinea-pigs, and found that animals sensitized to a serum,
and given a non-fatal reinjection of the homologous serum,
after recovery survived the minimum fatal dose of anaphyl-
atoxin. From these experiments Bessau reached the
following conclusions: (1) The condition or state of anti-
anaphylaxis is not specific. (2) It is not due to absorption
of the ferment (antibody). (3) It is due to increased toler-
ance of, or lessened susceptibility to, the poison. Friedberger
and his students1 have taken up these points, and by exact
quantitative experiments have demonstrated that the
state of anti-anaphylaxis, like that of anaphylaxis, is
strictly specific, but that it is true that increased tolerance
does play a part in the experiments as made by Bessau.
When an animal is simultaneously sensitized to two sera,
and after the condition of sensitization has been fully
developed, a non-fatal reinjection of one of these sera
renders the animal after recovery absolutely insusceptible
to any dose of the serum which has been employed in the
reinjection, but leaves it still susceptible to the second serum
in doses only slightly larger than those required to kill
control animals sensitized to that serum only. We can
make this plainer by the following statement. When an
animal is sensitized to two sera, two specific proteolytic
ferments are developed. When an animal in full sensitiza-
tion to both sera is treated with a non-lethal reinjection of
one of them, the specific ferment for this serum is exhausted,
and a certain amount of the poison is set free, not enough
to kill the animal, but enough to give the animal increased
tolerance to the poison. Consequently the fatal dose of the
other serum necessary to kill on reinjection, say, twenty-four
hours later, is larger than the minimum fatal dose when the
animal has been sensitized to only one serum. We demon-
strated (see page 139) many years ago not only that tolerance
to the protein poison can be increased, but that resistance
to living cultures of pathogenic bacteria may be increased
by repeated doses of the poison. Furthermore, we showed
1 Zeitsch. f. Immunitatsforschung, 1912, xiv, 371.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 311
that in neither of these instances is the action specific,
nor does the poison have the action of a toxin, nor is the
increased tolerance of it due to the production of antitoxin.
Repeated treatment of animals with the poison, beginning
with a sublethal dose and gradually increasing the dose,
may enable the animal to bear three or four times the
minimum lethal dose, as tested on fresh animals, but the
effect induced is never quantitatively comparable to that
obtained by similar treatments with increasing doses of toxin.
Besides, we were never able to find any evidence of the
presence of an antitoxin in the blood serum of the treated
animal. For these reasons we decided years ago that the
protein poison is not a toxin. Moreover, we found that the
increased resistance to typhoid infection came just as
promptly and was as marked when the animal was treated
with poison obtained from egg-white as that obtained by
repeated treatments with the poison split off from the
cellular substance of the typhoid bacillus. This demon-
strated that the tolerance obtained to the protein poison
is not specific. This is another clear proof that the poisonous
group contained in the protein molecule is not a toxin.
Years after our work had been done and reported, Fried-
berger1 found that after a guinea-pig had recovered from
severe poisoning with his anaphylatoxin, it would bear a
certainly fatal dose of the same, and at that time he thought
that he had secured a toxin-antitoxin immunity. Later
still H. Pfeiffer2 found that the urine of an anaphylactized
guinea-pig is highly poisonous; also, that treating a sensi-
tized guinea-pig with such urine made it more resistant
on reinjection; also, that after recovery from anaphylactic
shock, guinea-pigs are more resistant to the poison in the
urine of anaphylactized animals. He also thought that
he had established a toxin-antitoxin immunity, but if we
read their later works with correct interpretation, neither
Friedberger nor H. Pfeiffer now believe that the protein
1 Zeitsch. f. Immunitatsforschung, 1910, iv, 636.
2 Ibid., 1911, x, 550.
312 PROTEIN POISONS
poison is a toxin, in the sense of diphtheria or tetanus
toxin, though both continue to call it a toxin. The work of
Bessau and Friedberger confirms ours and establishes
beyond any doubt that the increased tolerance brought
about by repeated administrations of the protein poison
by direct injection or by recovery from anaphylactic shock
is not of the nature of a toxin-antitoxin immunity.
It seems to us that there is one point about anti-anaphyl-
axis which both Bessau and Friedberger fail to see. When
a sensitized animal is reinjected with the homologous
protein and recovers, it immediately loses, for a time at
least, its responsiveness to the same protein. As Fried-
berger has shown, injections of two hundred times the
amount necessary to kill the sensitized animal is without
effect. Indeed, the animal seems to be returned suddenly
to the condition of a fresh animal, one which has never
received a protein injection. The usual explanation is
that all the specific ferment in the sensitized animal has
been exhausted by the non-fatal reinjection. Bessau
thinks it due, as we have seen, to a decreased susceptibility
to the poison, or, as we say, to an increased tolerance of
the poison. Both of these are undoubtedly factors, and
they may be the most important factors in the sudden
development of the anti-anaphylactic state, but they are
not the only factors, and we are inclined to the opinion
that they are not the most important. That the specific
ferment (the antibody of other writers) is not wholly
exhausted is shown by the fact that the blood serum of
an animal in the anti-anaphylactic state, when transferred
to a fresh animal, sensitizes the recipient. This could not
be if the ferment had been wholly used up. There must
still be active ferment in the portion of blood serum trans-
ferred. The factor which we suspect of being of greatest
importance in the production of the anti-anaphylactic
state is the changed relation between the amount of ferment
and the substrate. This is one of the most important and
interesting problems connected with protein sensitization.
In our earliest work with the cellular bacterial poisons,
PROTEIN SENSIT1ZATION OR ANAPHYLAXIS 313
when suspensions were injected into the abdominal cavity
we found that large doses often failed to kill, or killed
slowly, while smaller doses killed more certainly and more
promptly; then we found that grinding our cellular sub-
stances more finely increased their toxicity; later we found
that the introduction of large quantities of egg-white into
fresh animals was without visible effect, while the repeated
injection of very small doses produced prompt effects and
speedily killed. Later still we ascertained that when a
small amount of the blood serum of a guinea-pig sensitized
to egg-white was incubated with from 1 to 5 mg. of egg-white
in vitro we obtained an active poison, but when the amount
of egg-white present was greatly increased there was no
evidence of the production of a poison. Friedberger has
repeatedly met with the same thing in preparing his ana-
phylatoxin. With a small amount of ferment and an
excessive amount of substrate the reaction is impeded.
Then the presence and accumulation of the products of
fermentation retard the fermentative process. The con-
centration of the ferment, the substrate, and the products
of fermentation all influence the rapidity wTith which the
fermentative process proceeds, and all of these are altered
when a few cubic centimeters of the blood serum of an
animal in the anti-anaphylactic state is transferred to a
fresh animal and the latter receives an injection of the
proper protein. Besides, it is possible that in the prepara-
tion of the serum the amount of available ferment is increased
by disruption of the leukocytes.
Friedberger at first stated that his anaphylatoxin is
thermolabile. If this be true it cannot be identical with or
very closely related to our protein poison, which is ther-
mostabile. Later, Friedberger found that in acid solution
anaphylatoxin is thermostabile. It is well to see how these
differences can be reconciled. It must be understood that
anaphylatoxin has never been isolated, not even partially,
from the serum in which it is formed, and of course the
serum is alkaline. Years ago we showed that our poison
in alkaline solution decreases in toxicity, and that this
314 PROTEIN POISONS
decrease takes place rapidly at high temperature. We
showed this by making aqueous solutions of the poison
alkaline with sodium bicarbonate and keeping them in the
incubator for varying periods. Friedberger finds that
when he heats the alkaline serum containing anaphylatoxin
to 65°, its toxicity decreases, but when the serum is made
acid it may be heated to 100° without appreciable loss in
toxicity. It will be seen, therefore, that the two substances
behave in the same manner when heated in alkaline solution.
We never supposed that the heat destroyed our poison,
but that on combination with alkali, which combination is
hastened by heat, it becomes less poisonous, and Friedberger
has failed to show that this is not true of anaphylatoxin.
There is another striking point of similarity between our
poison and Friedberger's anaphylatoxin, and in this par-
ticular both substances show a close relationship to peptone.
It has long been known that when an animal is quite fully
under the influence of peptone the further administration
of peptone has but little effect. This is true of both our
poison and anaphylatoxin. We have designated this as
tolerance, but it must be admitted that it is an unusual
form of tolerance and it needs further investigation.
Besredka, Strobel, and Jupilli1 refuse to accept anaphyl-
atoxin as the true anaphylactic poison because its adminis-
tration to sensitized animals does not induce the so-called
anti-anaphylactic state. Of course it does not and should
not be expected to do so. The anti-anaphylactic state is
due to the partial exhaustion of the specific proteolytic
ferment, and the retarding effects of the products of digestion
on the remaining ferment. Administration of the poison
itself, already formed, uses up none of the ferment, and the
other products of the cleavage action, besides itself, are
not present. Another reason the French investigators
give for concluding that anaphylatoxin is not the true
anaphylactic poison is that the former is not specific in
origin and may be obtained equally from diverse proteins;
1 Zeitsch. f. Immunitatsforschung, 1913, xvi, 250.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 315
certainly it can be obtained from all true proteins, as we
demonstrated many years ago. The specificity of an infec-
tious disease does not lie in the poison which is formed, but
in the ferment by which it is formed. The same poison is
contained in all bacteria, pathogenic and non-pathogenic,
indeed, in all proteins, but there are specific ferments which
break up one protein more readily and more completely
than other ferments. The specificity lies in neither the
substrate, except that it must be a protein, nor in the
cleavage product, but in the agent that effects the cleavage.
Physiological Action of the Protein Poison. — Edmunds1
has made the most thorough study of the protein poison,
as prepared by Vaughan and Wheeler, reported up to the
present time. • His experiments were made on dogs and
with the "crude soluble poison" made from casein. This
preparation contains something less than 10 per cent, of
the poison in the purest form in which, so far, we have been
able to obtain it, and this is not chemically pure. On
account of its importance we make the following, somewhat
lengthy, abstract from the paper by Edmunds.
Intravenous injections in intact dogs are reported as
follows: "The most prominent symptoms were a marked
depression, disturbance of the alimentary canal, and some
respiratory disturbances, the latter consisting of slight
acceleration with a slightly labored expiration. In some
animals the respiratory symptoms, with the exception of
the slight acceleration, were scarcely noticeable. A study
of these symptoms shows that they resemble closely those
exhibited by dogs which are suffering from anaphylactic
shock, although they are milder than those described by
Pearce and Eisenbrey and others."
The effect on the circulatory system was studied upon
dogs anesthetized with morphine and paraldehyde. Blood
pressure was measured from the carotid, and the respiration
recorded by a tambour resting against the chest wall and
connected with a second one by which the movements were
1 Zeitsch. f. Immunitatsforschung, 1913, xvii, 105.
316 PROTEIN POISONS
traced upon blackened paper. Injection into the external
jugular ,vein of the soluble portion of 100 mg. of the crude
poison was followed immediately by a slow decline in blood
pressure, amounting to from 6 to 8 mm. Hg.; after about
twenty seconds the fall became rapid, passing from a normal
of about 72 mm. to about 20 mm. Synchronously with
the fall in pressure the heart-beat was at first slightly
accelerated, passing from 138 or 140 to 144, but when the
pressure reached the low point the heart-rate dropped to
92. The respiration was but little changed in rate, becoming
slightly slower with the fall in pressure, but the strength
was considerably decreased, neither inspiration nor expira-
tion being as complete as normal.
The blood pressure was slow to recover. In some instances
there was an increase of only a few millimeters after thirty
minutes. When under the full influence of the poison,
stimulation of either the sciatic or the great splanchnic
nerve with the induced current elicited no response, showing
peripheral paralysis of the vasomotors. In pithed animals
(with the brain and cord destroyed) the effect upon blood-
pressure was the same as on whole animals, only that the
initial pressure being small, the fall did not measure so
many millimeters. "That this action was peripheral to
the ganglia along the course of the constrictor fibers was
proved by the use of large doses of nicotine, sufficient being
given to paralyze them. When this stage was reached an
injection of the poison still produced the characteristic
fall.
"The localization of the point of action of the poison
upon nerve ending, receptive substance, or muscle wall
was studied with the aid of nicotine, epinephrin, and digitalis.
The action of nicotine was greatly weakened by the previous
injection of the poison. Where before the poisoning, nico-
tine had given a marked increase in pressure in the charac-
teristic manner, following 300 mg. of the poison 5 mg. of
nicotine raised the pressure from 16 mm. to only 62 mm.
The heart-rate was increased by the nicotine in the usual
manner, from 120 per minute to 216. The pressure curve
PROTEIN SENSITIZATION OR ANAPHYLAXIS 317
from the nicotine was not only lower than that usually
seen with such doses, but was much altered in shape, there
being a very slow rise in place of the precipitous increase
commonly obtained. The injection of nicotine was followed
after a short time by a dose of epinephrin which raised the
pressure from 22 to 220 mm. Further injections of large
doses of the poison were followed by repeated injections
of epinephrin which raised the pressure from 18 to 225 mm.
These experiments seemed to point conclusively to the
nerve endings as being the structure primarily acted upon
by the poison, as evidently the receptive substance which
is stimulated by the epinephrin had not been paralyzed by
the poison. In addition to epinephrin, both digitalis and
barium chloride raised the lowered blood pressure very
satisfactorily, demonstrating that the muscle cell in these
cases was not affected by the doses of poison given. In
some animals, however, the use of large doses ,of the poison
was followed by a lessened response to epinephrin and
digitalis, thus showing that while the nerve ends are first
affected, the effect of large doses is not necessarily confined
to these structures, but may spread to the receptive
substance and contractile substance proper."
Small doses (25 mg.) of the poison have but little effect
upon the heart, beyond a temporary increase in systole and
diastole in both auricle and ventricle. This increase is
followed in about a minute by some weakening in both
chambers, most marked in the auricle, both systole and
diastole being decreased. With larger doses (100 to 150
mg.) the same changes are produced, the only difference
being one of degree. The increase in systole in both auricle
and ventricle is quite marked in some cases, while the
change in the extent of the dilatation is not so great, but
is present in most cases. These changes lead to an increased
amplitude of beat which lasts usually from one to two
minutes, until the blood-pressure has reached its lowest
limit. The fall in pressure coming on while the strength of
the beat is increased finds no explanation in the behavior
of this organ.
318 PROTEIN POISONS
"The changes which the heart undergoes, following the
increase in systole described, consist in a weakening in both
chambers, while the extent of dilatation may be still further
increased, or it may show little change. Two factors come
in to complicate the changes produced by the poison. First,
the great fall in blood pressure produced by it decreases
the resistance against which the heart has to contract; and
second, the changes in respiration, which at times produce
a mild degree of asphyxia. It happened in some cases,
before the poison was given, the artificial, respiration had
seemed entirely adequate, after its administration the
lungs did not inflate so well and the blood showed signs of
deficient aeration. The heart changes were therefore studied
further in animals in which all other organs, save the lungs,
were excluded from the circulation. This experiment was
carried out on a large bulldog anesthetized in the usual
way. Under artificial respiration the sternum was cut
lengthwise, the two halves pulled apart, and the heart
exposed. The large vessels at the base of the heart were
dissected free and a clamp placed in position on the aorta
just below the origin of the left subclavian artery, this and
the right subclavian being tied. A loose clamp was also
placed on the inferior vena cava just above the diaphragm.
Cannulas were placed in both common carotids and external
jugular veins. The cannulas in the left carotid and the
right jugular were paraffined and connected by a short
paraffined rubber tube, thus providing a channel for the
blood from the left to the right heart. The cannula in the
right carotid was connected with the mercury manometer to
record the blood pressure, and that in the left jugular was
used to inject the poison. The clamps on the aorta and
the vena cava were now closed, thus shutting off the circu-
lation in all parts of the body, except the heart and lungs.
The pericardium was opened and sewed to the sides of
the chest to form a sort of cradle for the heart. The myo-
cardiograph was attached to the right auricle and ventricle
in the usual manner, and arranged to record their move-
ments on the kymograph." After recovery from the opera-
PROTEIN SENSITIZATION OR ANAPHYLAXIS 319
tion the pressure stood at 105 mm. Injection of the poison
caused a fall of only 2 or 3 mm. Hg. In this way it was
shown that the heart and lungs were not responsible for
the fall in pressure observed in the intact animals.
Isolated organs were perfused with solutions of the
poison and a dilatation of the vessels, probably due to
paralysis of the vasomotor mechanism, was observed.
This paralysis did not disappear after subsequent washing
with Ringer's solution, but did so promptly on the use of
epinephrin. The perfusion experiments, therefore, indicate
a local paralyzing effect upon the vessel walls.
Edmunds demonstrated by careful experimentation1
that the liver dilates with the fall in blood-pressure. This
seems to settle the question of the distribution of the blood
as the pressure falls. "The fall is due primarily to a
peripheral paralysis of the vasomotors running in the
splanchnic nerves." The spleen, kidneys, and intestine
do not show increase in volume, as the blood is drained
from these organs into the capacious blood channels of the
liver. Other vascular areas besides those innervated by
the splanchnics are affected. This was shown by the fact
that when the poison was injected into white dogs the skin
over the thorax and abdomen and down on the legs became
bright pink. When the liver was excluded from the circu-
lation the fall in blood pressure occurred, but less promptly.
Respiratory changes in the dog, due to the poison, are
not marked. The usual effects are slight acceleration and
weakening. With the chest walls open and under artificial
respiration, there would be, at times, signs of asphyxiation
which were easily relieved by a slightly stronger pressure
on the bellows. The most marked change in the blood
picture observed was a diminution in the eosinophiles, both
relatively and absolutely.
It has been observed by all who have studied the action
of peptone and the protein poison, that after the blood-
1 The details can be found in Zeitsch. f. Immunitatsforschung, 1913,
xvii, 105.
320 PROTEIN POISONS
pressure has fallen to the lowest limit the further adminis-
tration of the peptone or poison is without effect.
Edmunds closes his studies with the following conclu-
sions: "The toxic portion of the split protein molecule as
described by Vaughan and Wheeler produces in dogs when
injected intravenously the same symptoms as are seen in
these animals when suffering from acute anaphylactic
shock. An analysis of the changes shows the same rapid
fall in blood pressure due mainly to paralysis of the vaso-
motor endings of the splanchnic nerves. The blood does
not accumulate at the time of the fall in pressure in the
intestines or kidneys, but is drained from them into the
liver, and probably into the large abdominal veins. There
is no evidence of a constriction of the pulmonary vessels,
nor of lack of blood to the left side of the heart. In these
points the action of the protein poison agrees with the
changes described in anaphylactic shock, but whereas with
the latter the ability of the blood to coagulate may be lost,
this is not affected by the split product."
General Physiological Action of Proteins. — Schittenhelm
and Weichardt1 conclude a study of this subject as follows:
The compound proteins, as such, are relatively inactive.
In the doses employed they give rise to no symptoms and
do not affect blood pressure. Their components (the
globulins, histons, and protamins) are highly poisonous
compared with the native simple proteins. They cause a
marked fall in blood-pressure, delay blood coagulation,
influence respiration and temperature, and in small doses
may cause death. This is true even when they are of
homologous origin. From the composition of the protamins
and histons it has been inferred that their poisonous action
is connected in some way with their large diamino acid
content, but the globins do not contain a large amount of
these acids. On the other hand, as has been stated, the
kyrins contain a large amount of diamino acids and are
not so poisonous as the protamins and histons. It should
1 Zeitsch. f. Immunitatsforschung, 1912, xiv, 609.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 321
be remarked that while the globins do not contain a large
amount of the diamino acids, they are rich in the closely
related body, histidin, and the relation of this to the highly
poisonous /3-i body of Barger and Dale has been mentioned.
Certainly there is reason for suspecting that the poisonous
group or groups in the protein molecule has some close
chemical relationship to the diamino acids.
Sensitization is Cellular. — Dale1 has shown that the plain
muscle of a sensitized guinea-pig contracts when touched
with a dilute solution of the homologous protein. This
demonstrates that sensitization is cellular. Dale states
his conclusions as follows: "(1) Plain muscle from an
anaphylactic guinea-pig, freed from all traces of blood and
serum, has a very high degree of sensitiveness to the specific
sensitizing protein. The plain muscle of the virgin uterus
is essentially suited to the demonstration of the condition,
and exhibits a definite rise of tonus in response to extreme
dilutions of the antigen. (2) The effect is practically imme-
diate, i. e., the delay is not obviously more than can be
attributed to the method of application of the antigen.
(3) The response is not a mere exaggeration of the reaction
which normal, plain muscle gives to fresh sera in general.
Preparations of purified protein can be obtained (e. g.,
serum globulin precipitated by Gibson's method, or egg
albumen crystallized by Hopkins' method) which have no
effect on the normal plain muscle, but are as toxic for the
anaphylactic plain muscle as the native proteins. (4)
One dose of the specific antigen, in sufficient concentration
to produce "maximal" response of the anaphylactic plain
muscle, completely desensitizes the latter to further doses
of any dimensions, provided that the experiment is not
complicated by the use of an antigen preparation of normal
toxicity. Either normal or anaphylactic plain muscle gives
repeated responses to successive large doses of a normally
toxic serum or other native protein, but this phenomenon
is not anaphylactic response. (5) When sensitizing doses
1 Jour, of Pharm., 1913, iv, 167.
21
322 PROTEIN POISONS
of several antigens are given, a multisensitization of the
plain muscle can be demonstrated. Desensitization of the
muscle to one antigen is not without effect on its sensitiveness
to the others. (6) The washed plain muscle from guinea-pigs
immunized to an antigen by a series of injections, is sensi-
tive to the antigen, like that from anaphylactic pigs. But
the sensitiveness is in this case less rigidly specific, e. g.,
plain muscle from a guinea-pig immunized to horse serum
showed a subsidiary sensitiveness to sheep serum. (7)
The sensitiveness of the washed, plain muscle is seen in
passive as in active anaphylaxis, whether the serum pro-
ducing passive sensitization is obtained from sensitive or
immune guinea-pigs. (8) The actively or passively sensi-
tized plain muscle after being desensitized in vitro can be
resensitized in vitro by mere contact for some hours with a
not too great amount of sensitive serum. It has not been
found possible to sensitize normal plain muscle in exactly
the same way; but perfusion of a normal uterus for five
hours, with diluted serum from sensitive guinea-pigs, pro-
duced a decided passive sensitization. (9) The response
to the specific antigen of the bronchioles of the anaphylactic
guinea-pig is not impaired by excluding the abdominal
viscera and the brain from the circulation, and is produced
with apparently undiminished vigor in the isolated lungs
perfused with Ringer's solution."
Dale suggests that the response of plain muscle to its
specific sensitizer might be used for medico-legal purposes.
The suspected material might be used to sensitize a guinea-
pig. After allowing time for full sensitization the uterus
could be excised and a suspended horn tested with various
sera until the one giving the typical response was detected.
With the other horn the limits of the response might
be determined. A second method might be as follows:
"Guinea-pigs could be sensitized with a small injection of
known serum from the suspected species, e. g., human
serum. After the usual incubation period one pig would
be killed, the uterus excised, and the degree of sensitiveness
of the first horn to human serum tested. If the sensitiveness
PROTEIN SENSITIZATION OR ANAPHYLAXIS 323
were not of a high order, another pig could be tried at once,
or after a few days longer incubation period. When a
uterus was found which gave a large and clear response
to, say, 1 in 100,000 human serum, the other horn, kept
meanwhile in warm oxygenated Ringer, could be suspended
in a small volume such as 10 c.c. of Ringer's solution. A
dose of the suspended material could then be added, and
if no reaction were produced it would be clear that the
dose contained less than 0.0001 c.c. of human serum, which
should be sufficient evidence, in any ordinary case, that
the blood under examination was not human. If, on the
other hand, a decided response were produced, it would
only be necessary to test further the action of the specimen
on a normal uterus, so as to exclude primary non-specific
toxicity."
It has been demonstrated by Manwaring,1 and confirmed
by Voegtlin and- Bertheim2 that dogs sensitized to horse
serum do not respond on reinjection when the liver is
excluded from the circulation.
Theories. — Hamburger and Moro at one time suggested
that the first injection leads to the formation of precipitins,
and that on reinjection precipitates are formed, and induce
the anaphylactic symptom-complex by the formation of
capillary emboli. The formation of specific precipitins is
a reaction which occurs in vitro, but not in vivo. Besides,
the symptoms of anaphylaxis are not those due to emboli,
and finally, no emboli are formed.
Gay and Southard thought that as a result of the first
injection there remains in the circulation a protein rest
which they named "anaphylactin," and that this continues
to stimulate the cells, creating an abnormal affinity for
the homologous protein which on reinjection leads to
anaphylactic shock. The transfer of this "anaphylactin"
to a fresh animal was supposed to explain passive anaphy-
laxis, a phenomenon first studied by these investigators.
1 Zeitsch. f. Immunitatsforschung, 1910, viii, 1.
2 Jour. Pharm., 1911, ii, 507.
324 PROTEIN POISONS
Richet held that sensitizers contain a substance which
he called "congestin," and that this develops in the animal
another substance known as "toxogenin." The reaction
between the latter and the homologous protein on reinjec-
tion sets free a poison "apotoxin," which on account of its
effect on the nervous system, develops the symptoms of
anaphylaxis. Considering Richet's toxogenin a ferment,
we can accept this theory as essentially correct.
Besredka taught that the sensitizer contains two sub-
stances— " sensibilisinogen" and "antisensibilisin." On the
first injection the former develops in the animal body a
substance, "sensibilisin," and on reinjection the sensibilisin
and the antisensibilisin combine to form a poison which
acts on the nervous system. Besredka has not been able
to produce satisfactory proof of the existence of antisensi-
bilisin. His work along this line has already been referred
to (p. 260).
For reasons which will become evident as we proceed, it
is desirable to go somewhat into detail in considering the
theory of Friedberger. This was first published in 1909,1
and in this publication Friedberger clearly and unequivocally
set forth his theory. It may be known as the theory of
sessile or fixed receptors. The followng is an abstract of
the statement: On the first injection the protein finds but
few groups with which it can combine, and for this reason
it is not poisonous, even in large doses, just as happens
when tetanus toxin is injected into a naturally immune
animal. During the period of incubation the animal cells
develop specific receptors for the homologous protein.
With frequent injections at short intervals, as when the
object is to obtain a highly active precipitating serum,
the newly formed receptors are in large part cast off into
the blood. When a single small dose is given, as in sensi-
tization, less receptors are cast off into the blood, and more
remain attached to the cells. In this way an organism
relatively insusceptible to a given foreign protein is made
1 Zeitsch. f. Immunitatsforschung, ii, 208.
PROTEIN SENSITIZATION OR AN APR YL AXIS 325
highly susceptible, and on the second injection the protein
is firmly anchored to the cell, just as the cells of an animal
susceptible to tetanus anchor the tetanus toxin. The only
difference is that in the latter instance the receptors are
preformed, while in the case of sensitization they are
developed as a result of the first injection. It is, as Ehrlich's
theory explains, the same substance, the receptor, so long as
it remains attached to the cell, that is the cause of the
poisoning, and which becomes the cause of cure when
detached from the cell, and cast off into the blood. The
only difference, as has been stated, is that the substance
attached to the cell (the receptor) is not, at least in sufficient
quantity, preformed, and must be developed by the first
injection. Protein (toxin) immunity and anaphylaxis,
therefore, are alike save in the proportion and location of
the antibodies. When the precipitin is already in the body
fluids the injection of the homologous protein is without
effect; when the precipitin is still attached to the cell in
sufficient quantity the reinjection of the homologous protein
is followed by the phenomena of anaphylaxis. The anti-
bodies exist in two places: (1) As free antibodies in the
serum (known as precipitins in test-tube experiments).
(2) As sessile antibodies attached to the cells. In cases of
local sensitization, as in Arthus phenomenon, the local
cells only are affected because they are the only ones which
bear the sessile receptors. The animal escapes anaphyl-
actic shock because the cells of the body as a whole, and
especially those of the nervous system, do not carry the
specific sessile receptors. Friedberger, in his theory, explains
antianaphylaxis as follows: An animal is rendered anti-
anaphylactic when it receives a large reinjection before
the period of incubation is complete. In this case the
reinjection uses up the sessile receptors already developed,
but these are not enough to lead to anaphylactic shock,
and at the end of the period of incubation the new crop of
sessile receptors is not sufficiently developed to give rise
to the symptoms of anaphylaxis. Again, an anaphylactized
animal may be rendered antianaphylactic by a small dose
326 PROTEIN POISONS
of the antigen. This is explained by Friedberger by sup-
posing that the small dose uses up a part of the sessile
receptors, and that there are not enough left to induce
anaphylactic shock when another injection is made. In
short, he concludes: "In every case antianaphylaxis is
nothing more than anaphylaxis refraeta dosi." Passive
anaphylaxis is explained by Friedberger by supposing that
the free receptors in the blood of an anaphylactized animal
become, on injection into a fresh animal, anchored to the
cells, thus forming fixed or sessile receptors. This is Fried-
berger's theory. It is clean cut and clearly stated by its
distinguished author, but at present it has no support,
and is clearly out of harmony with known facts, some of
the most important of which have been discovered by the
researches of its own author. It was an attempt to make
the facts of anaphylaxis fit Ehrlich's theory of the action
of toxins and the production of toxin immunity, while
the trend of later research is to show that the two sets of
phenomena have but little in common. Friedberger's
theory would make the action of sensitizers, such as serum
albumin, egg-white, edestin, bacterial proteins, etc., identical
with that of diphtheria and tetanus toxin, abrin, ricin,
the venoms, etc. There is nothing in the theory about the
development of the proteolytic ferments, and the liberation
of a protein poison by the parenteral digestion of the sensi-
tizer on reinjection. If we read his works with correct
interpretation, Friedberger has abandoned his own theory
largely, if not wholly. Indeed, in Contribution VI,1 Fried-
berger plainly discards his own theory.
According to Friedberger's theory all sensitizers act like
the toxins; although at first only mildly toxic, they become
more so by developing the receptors, and thus rendering
the animal more susceptible. It is in a way proper for
Friedberger to speak of anaphylaxis as a "protein-anti-
protein" reaction. Friedberger calls the sensitizer an
antigen and the substance developed in the animal an
1 Zeitsch. f. Immunitatsforschung, 1910, vi, 179.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 327
antibody. According to our theory these terms are not
only inappropriate, but are confusing and misleading.
The theory of Vaughan and Wheeler was first published
in 1907, * two years before that of Friedberger, and while
it has been developed and, in our opinion, confirmed by
later investigations, there has been no material alteration
in it. To one who has read this chapter thus far the essen-
tials of this theory must be already fairly understood, but
a concise statement of its fundamental points must be
made here even if some repetition be necessary. The
proteins taken into the alimentary canal are broken up
by the digestive ferments into non-protein split products,
mostly amino-acids. During or after absorption these
pieces are resynthesized, in part at least, to form the body
proteins peculiar to the species. The precipitin test shows
that with the exception of the proteins of the crystalline
lens, those of all the fluids and tissues of the body are
peculiar to the species. Those of one species differ from
those of all other species. Just where this synthesis occurs
in the animal body we are not sure, but that the species
proteins are formed from the split products of the proteins
of the food has been positively demonstrated. Every
protein molecule contains a poisonous group. In the whole
molecule this group is saturated with other non-poisonous
groups. As the whole molecule undergoes cleavage as a
result of enzyme action, the poisonous group is more or
less completely liberated, and in this process it becomes
activated. In alimentary digestion the poisonous group
becomes most active at or about the stage of the formation
of peptone. As the digestive process proceeds, the poison-
ous group is itself disrupted, and ceases to be a poison.
The protein poison is not readily diffusible, and for this
reason it is retained in the alimentary canal until it is broken
up and rendered inert. When an unbroken or undigested
protein finds its way into the blood or tissues it must be
digested. There are two kinds of proteolytic digestion:
1 Jour. Infect. Dis., June, 1907.
328 PROTEIN POISONS
(1) Enteral, (2) parenteral. Proteins that escape enteral
digestion and find their way into the body must be digested
by the body cells. When some foreign protein, like the
blood serum of another animal, milk, egg-white, bacterial
proteins, etc., are injected into the blood or tissues they
must be digested. There is only one way in which this
can be accomplished, and that is by the cells of the body,
and there is only one way in which these cells can do this,
and that is by elaborating a specific proteolytic ferment
which will digest and destroy the foreign protein.
In experimental anaphylaxis the first injection introduces
into the body a foreign protein. This must be digested
and the body cells slowly elaborate a specific proteolytic
ferment which slowly digests it. In doing this certain
body cells acquire a new function. The protein of the first
injection is slowly digested usually without the develop-
ment of recognizable effects, and consequently we conclude
that the animal has not been affected or had its functions
altered in any way. But this is a mistake. The animal
has been profoundly affected. It has developed a new-
function which it may retain quite indefinitely, and which
may be transmitted from mother to offspring. The foreign
protein is digested and its poisonous group set free, but
this has been done so slowly and gradually that the effects
have not come within the range of our powers of recog-
nition. After the protein of the first injection has been
disposed of, the new ferment in the form of a zymogen
continues to be formed in the cells, and on the second
injection after the proper interval, this zymogen is acti-
vated and splits up the protein so promptly and so abun-
dantly that the liberated poison induces the symptoms of
anaphylactic shock.
The following statements formulated in 1907, in our
opinion, still hold good:
1. Sensitization consists in developing in the animal a
specific proteolytic ferment which acts upon the protein
that brings it into existence, and on no other.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 329
2. This specific proteolytic ferment stored up in the
cells of the animal as a result of the first treatment with
the protein remains as a zymogen until activated by the
reinjection of the same protein.
3. Our conception of the development of a specific
zymogen supposes a rearrangement of the atomic groups
of the protein molecules of certain cells, or an alteration
of their molecular structure. In other words, we regard
the production of the specific zymogen not as the formation
of a new body, but as resulting from an alteration in the
atomic arrangement within the protein molecule, and a
consequent change in its chemism.
4. Some proteins in developing the specific zymogen
produce profound and lasting changes in molecular struc-
ture, while the alterations induced by others are slighter
and of temporary duration, the molecular structure soon
returning to its original condition.
5. Bacteria and protozoa are living, labile proteins, while
egg-white, casein, serum albumin, etc., are stabile proteins.
The proteins of one group are in an active, while those of
the other are in a resting state, but both are essentially
proteins made up of an acid or poisonous chemical nucleus,
and basic, non-poisonous groups. Bacterial immunity
and protein sensitization, apparently antipodal, are in
reality the same, and each consists in developing in the
animal body the capability of splitting up specific proteins.
If the living protein be split up before it has time to multiply
sufficiently to furnish a fatal quantity of the poison, the
animal lives and we say it has been immunized. If the
stabile protein be introduced into the animal body it
leads to the development of a specific proteolytic ferment,
and if enough of it to supply a fatal dose be reinjected after
this function has been developed, the animal dies.
6. We are compelled to change our ideas concerning
the causation of the lesions of the infectious diseases.
Formerly, we believed the structural changes to be due
wholly to the living, growing, feeding microorganisms.
For instance, we were sure that the intestinal ulcerations
330 PROTEIN POISONS
of typhoid fever are caused by the living bacilli. Now we
know that these lesions follow the intravenous injection
of dead proteins. As has been stated, each foreign protein
has its predilection tissue in which it is largely deposited,
whose cells it especially sensitizes, and where it is disrupted.
This explains the characteristic lesions and symptoms of
the different infectious diseases. Bacterial inflammation
is essentially a chemical process, or is due to the disruption
of cell molecules through the chemical affinity between
certain groups in the bacterial cell and certain groups in
the cell of the animal. So long as the bacterial cells are
alive the chemism that holds the living molecule together
tends to resist this process of disintegration. The patho-
genic bacterium assimilates the nutritious constituents
of the fluids of the animal body, builds them into its own
tissue, converts them into substances foreign to the host,
and finally, when the bacterial cell goes to pieces either
from spontaneous dissolution, or through the aggressive
action of some animal cell, these reconstructed chemical
groups are set free and poison the animal, inducing lesions
in various tissues, and, in many instances, so interrupting
the vital functions as to cause death. It is in harmony with
these statements that Friedberger has been able to induce
aseptic pneumonia by spraying horse serum into the lungs
of guinea-pigs sensitized with the same, and Schittenhelm
and Weichardt have established "enteritis anaphylactica"
by the reinjection of egg-white into sensitized dogs. It
is more than probable that cholera infantum and the kindred
summer diarrheas result from the absorption of undigested
milk and consequent sensitization. The designation " protein
diseases" might be used to cover the majority of bacterial
and protozoal diseases, and many of these hitherto regarded
as autogenous.
7. It seems to be a physiological law that the specific
ferments elaborated by living cells are determined by the
proteins brought into contact with them, but as yet we
know but little concerning these bodies which we call
ferments. That they are labile chemical bodies resulting
PROTEIN SENSITIZATION OF ANAPHYLAXIS 331
from intramolecular rearrangement in the protein molecules
of the cell seems a plausible theory, but at present it is
only a theory. We know but little of the action of these
so-called ferments upon their homologous proteins. Our
knowledge of the chemistry of protein sensitizers is exceed-
ingly limited, and as we have pointed out, it is highly desirable
that work in this direction should be prosecuted with
vigor, because we need sensitizers free from the poisonous
group. Furthermore, there is the question why small
doses of protein induce fever while large doses have no such
effect. At present we have no satisfactory answer to this
question. If it could be conclusively demonstrated that
the toxins are ferments, the subject of the etiology of disease
would be greatly simplified. We have elsewhere (see Chapter
XV) given our reasons for holding that the toxins are
ferments, and at this point we wish to formulate what we
believe to be two biological laws:
(a) When the body cells find themselves in contact with,
or permeated by, foreign proteins they tend to elaborate
specific ferments which digest and destroy the foreign
proteins.
(6) When body cells are attacked by destructive ferments
they tend to elaborate antiferments the function of which
is to neutralize the ferments and thus protect the cells.
Zunz1 finds the proteoclastic (protein-splitting) properties
of blood-serum, as tested on the sensitizing protein, increased
in the anaphylactic state. This increase becomes measurable
in the pre-anaphylactic stage, usually about the fifth day
after the injection, and continues to be measurable to from
the twentieth to the sixtieth day. It is not recognizable
in blood-serum taken during or soon after anaphylactic
shock. Zunz concludes that the increased proteoclastic
property of the blood serum is not sufficient to fully account
for the phenomena of anaphylaxis. In this conclusion we
fully agree with the distinguished Belgian investigator.
In our opinion the failure of the blood serum taken during
1 Zeitsch. f. Immunitiitsforschung, 1913, xvii, 241.
332 PROTEIN POISONS
or soon after anaphylactic shock to show measurable proteo-
clastic effect is due to the accumulation of the cleavage
products in the blood. That all the phenomena of anaphyl-
axis are not due to cleavage ferments in the blood-serum
must be evident to all who have followed us thus far. The
work of Pfeiffer and Mita, that of Zunz, our own, and that
of others agree in showing that the property of splitting up
the sensitizing protein, in measurable quantity, at least,
disappears from the blood-serum of the sensitized animal
long before the anaphylactic state disappears. When a
guinea-pig is sensitized to horse serum, the blood-serum of
this animal looses the power to split up horse serum in vitro
in appreciable amount after from twenty to sixty days, but
the animal remains sensitive to horse serum for at least two
years, as shown by Rosenau and Anderson, and probably
so long as the animal lives. We must therefore agree with
Zunz that the increased proteoclastic property of the blood-
ferum of the sensitized animal is not sufficient to account
sor all the phenomena of sensitization. Our theory of sensi-
tization takes this into account. We hold that sensitization
develops in certain body cells a new function— that of
elaborating a new specific, proteoclastic ferment. The
duration of this new function varies with the sensitizing
protein and with the cells in which it is developed. In a
given cell this function must be limited by the life of the
cell. We do not know just what cells develop this new func-
tion, but we do know that the animal may remain in a
sensitized state long after the blood-serum fails to show any
cleavage action on the sensitizing protein in vitro. It may
be that the specific ferment present in the blood-serum of
recently sensitized animals comes from the white corpusc'es,
or it may come in part, or altogether; from fixed cells. It
seems justifiable to conclude that the ferment which manifests
its action in animals long after it is absent from the blood
must come from fixed cells, and that these are stimulated to
elaborate this ferment only when the specific protein is
brought into contact with them, probably only when they
are permeated by the specific protein. All the facts which
PROTEIN SENSITIZATION OR ANAPHYLAXIS 333
have been ascertained in regard to this matter indicate that
sensitization is secured only by alteration in the cell and that
in some cells the newly developed function is more persistent
than in others.
The facts of cross-sensitization seem in harmony with our
view that the protein molecule contains one or more special
sensitizing groups. The white of hen's eggs sensitizes to itself,
less perfectly to the white of duck's eggs, and very imper-
fectly or not at all to the white of robin's eggs. The serum
of man's blood sensitizes to itself and less fully to that from
the ape. Horse serum sensitizes to itself, and less fully
to that of the donkey. Certain non-pathogenic acid-fast
bacteria sensitize in some degree to the tubercle bacillus.
In short, the phenomenon of sensitization, like that of pre-
cipitation, may be employed to show biological relationship.
All this seems in harmony with the view that the specificity of
sensitization depends upon the similarity or dissimilarity in the
chemical structure of protein molecules from different sources.
Doerr1 makes the following statement concerning Fried-
berger's anaphylatoxin : So far as the matrix of the poison
is concerned it is highly improbable that it comes from the
bacteria or other antigens. That the most diverse proteins
should yield the same poison seems improbable. The theory
that the poison comes from the amboceptor, as held by
Wassermann and Keysser, is still less probable. That the
anaphylatoxin comes from the blood-serum, the one constant
factor in all the experiments in its production, is most prob-
able. During its formation or in the process of blood coagu-
lation, a poison is formed, but the serum obtained, after
coagulation is complete, is not poisonous on account of the
presence of antibodies. When the serum is digested with
bacteria, precipitates, etc., the latter absorb the antibodies
and thus the serum again becomes poisonous. That ana-
phylatoxin can be obtained when there are no formed elements
present, as for instance when inactivated horse serum is
1 Handbuch d. path. Mikroorganismen by Kolle and Wassermann, second
edition, ii, 947.
334 PROTEIN POISONS
digested with fresh guinea-pig serum, does not contradict this
theory because the colloidal bodies may serve as absorption
agents.
This explanation of the easy production of anaphylatoxin,
given by Doerr, is worthy of consideration. His failure to
understand how the same poison can be obtained from the
most diverse proteins has no weight with, us since wTe have
prepared a poison which has, grossly at least, the same
physiological action, from the most diverse proteins, bacterial,
vegetable, and animal. The protein molecule, wherever
found, must have some common nucleus, and this we believe
to be the poison. But that a poison may be liberated in the
process of blood coagulation does not seem to us to be beyond
the range of possibility. Blood coagulation is a fermentative
process and that there is no cleavage in the protein molecule
in this process has not been shown.
As Doerr points out, so long ago as 1877 Kohler showed
that fresh defibrinated blood, whether homologous or heter-
ologous, is an active poison. This has been confirmed by
others, and recently it has been reinvestigated by Moldovan,
who has shown that blood freshly defibrinated by shaking
with glass beads causes acute death when injected intra-
venously into guinea-pigs and rabbits. In the former animals
the typical anaphylactic lungpicture after death is seen. When
the dose is slightly sublethal there is marked fall in -tempera-
ture with subsequent fever. When the doses are smaller
there is marked fever. On standing for fifteen to forty-five
minutes defibrinated blood looses in toxicity. Serum obtained
by rapid centrifugation of defibrinated blood is poisonous.
The same is true of the deposited and once-washed cor-
puscles. When coagulation is delayed by the presence of
sodium citrate, neither the supernatant fluid nor the cor-
puscles are poisonous, but both become so when coagulation
has been induced by shaking with porcelain beads. Later,
Doerr has shown that blood received in paraffined vessels
becomes poisonous; but when coagulation is complete the
toxicity disappears. When coagulation is made to proceed
slowly by the addition of hirudin solution or a 0.7 per cent.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 335
solution of colloidal silicic acid, it retains its toxicity for
several hours. The source of the poison in coagulation
blood has been discussed and variously explained. Kohler
thought the fibrin ferment is the poison, but this was shown
not to be true by Boggs. Studzinski suggested that the
poison might come from the mechanical disruption of the
red corpuscles, but Moldovan and Doerr have shown that
the plasma or serum may be absolutely free from both
hemoglobin or cells and still be poisonous. Freund thought
that the poison might come from the disrupted blood plate-
lets, but in rapid centrifuges even the platelets may be re-
moved and even then the plasma or serum may be poisonous.
However, the role that the platelets play in coagulation still
suggests that they may furnish the matrix for the poison.
The toxicity of extracts of normal tissue, especially of
the lungs and lymph nodes, is most interesting in this
connection. If the normal lungs of a rabbit or guinea-pig
be digested for two hours in physiological salt solution, the
solution kills promptly on intravenous injection. Homolo-
gous organ extracts are more active than heterologous.
Rabbits are the most susceptible animals used so far. The
poison seems to be destroyed when heated to 70°; it does not
pass through a Berkefeld filter, and is absorbed by kaolin
as is anaphylatoxin. The addition of serum either homolo-
gous or heterologous seems to destroy or neutralize the
poison after contact of one hour or more. Doerr states that
the fresh vaccine growth just scraped from a calf with a
curette furnishes an extract which kills rabbits instantly
on intravenous injection. These extracts seem to owe their
effects to a coagulating ferment.
Section of a dying animal shows the left heart, and the
pulmonary arteries and veins filled with coagula. Doerr
admits that poisoning with organ extracts from normal
animals is quite unlike death from anaphylactic shock.
In the former the blood is coagulated; in the latter it is
fluid. But Doerr holds that the formation of thrombi during
life is not the sole cause of death after the administration
of extracts from normal tissues or after the intravenous
336 PROTEIN POISONS
injection of fresh defibrinated blood. He states that Bianchi
failed to find intravascular thrombi after sublethal doses of
the extracts, and that Moldovan met with a like observation
after poisoning with defibrinated blood. To us the difference
between poisoning with normal tissue extracts and the
effects of anaphylactic shock seem quite clear. The organ
extracts do not contain a chemical poison, but a ferment.
This ferment coagulates the blood and leads to the forma-
tion of thrombi. This is a process of protein digestion, and
whether a protein poison is set free in it remains for future
research to determine. In anaphylactic poisoning the
ferment is in the body cell and splits up the protein introduced
with the liberation of a protein poison.
In this connection the work of Blaizot1 is of interest.
When dog's serum is treated with an extract from the intes-
tinal mucosa of the dog, or rabbit's serum with an extract
of the intestinal mucosa of the rabbit, after a few minutes
of contact, if either preparation be injected intravenously
into a guinea-pig, acute death results. Extensive thrombi
are found in the heart and large vessels. The serum of the
guinea-pig is not rendered poisonous by homologous extracts,
but is made poisonous by heterologous extracts.
We must, however, admit with Doerr, that the matrix
of Friedberger's anaphylatoxin remains undetermined, with
much probability in favor of the possibility of its being in
the so-called complement, or the serum of the guinea-pig,
the one constant factor in its production. This does not
mean that it is not a protein poison. It must be borne in
mind that anaphylatoxin is recognized only by its effect,
and it has never been even partially isolated from the serum.
Our protein poison comes certainly from the protein molecule.
It cannot be a ferment as we understand ferments at present.
It is thermostabile, and it elaborates no antibody and yet
it may be identical with anaphylatoxin, for whether the
latter comes from bacterial cells or from the serum it is of
protein origin.
1 Compt. rend. Soc. biol., 1910-11-12.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 337
Friedberger, Mita, and Kumagi1 have prepared anaphyl-
atoxin by the action of the normal serum of the guinea-pig
on the crude toxins of tetanus and diphtheria, and the
venom of the cobra. They assume that the poison comes
from the cleavage of the toxins, an assumption which seems
to us wholly without warrant. Some years ago we precipi-
tated the crude toxins of tetanus and diphtheria with alcohol
and split up the precipitate with chemical agents by our
method, and obtained the protein poison, but we never
felt justified in even supposing that the poison came from
the toxins. Crude toxins and venoms are complex protein
substances, and because, the protein poison can be obtained
by the cleavage of these is far from proof that the poison
comes from the toxin constituent of such a mixture. Indeed,
one does not know that the active principle in these mixtures
is a protein. Until the toxins have been obtained in some-
thing like a pure state it is useless to speculate concerning
their split products.
Bordet2 has shown that anaphylatoxin can be obtained
by the action of the normal serum of the guinea-pig on agar,
and this has been confirmed by Nathan.3 One-half gram of
agar is added to 100 c.c. of 0.85 per cent, salt solution and
sterilized by boiling. From 0.5 to 1 c.c. of this agar solution
is incubated with 5 c.c. of normal serum from the guinea-
pig at 37° for from one to twenty-four hours and then centri-
fuged. Many tubes may be employed and the supernatant
fluid from these mixed. From 3 to 5 c.c. of this fluid injected
intravenously into a guinea-pig of about 250 grams kills with
typical anaphylactic symptoms in from three to five minutes.
Even 0.1 c.c. of the agar solution furnishes enough poison
to kill. When the amount of agar solution employed is
greater than 1 c.c. or less than 0.1 the amount of poison
formed is, as a rule, not sufficient to kill, but may induce
anaphylactic symptoms of varying intensity. When the
1 Zeitsch. f. Immunitatsforschung, 1913, xvii, 506.
2 Compt. rend. Soc. biol., 1913, Ixxiv, No. 5.
3 Zeitsch. f. Immunitatsforschung, 1913, xvii, 478.
22
338 PROTEIN POISONS
serum is inactivated by being heated for one-half hour at
55° no poison is formed. Experiments of this kind, as well
as those with kaolin already referred to, have led to various
suggestions. They have caused many to suspect that the
poison comes from the serum. It has been suggested: (a)
That the agar or kaolin or bacteria absorbs the complement
from the serum and that this renders the serum poisonous.
(6) That the poison is preformed in the serum but that its
action is neutralized by some other constituent of the serum
which is absorbed by the agar or kaolin, (c) That the absorp-
tion of some constituent of the serum by the agar, kaolin, or
bacteria leads to a disturbance of the equilibrium of the
protein constituents of the serum which as a consequence
break up with the liberation of the poison. These suggestions
assume that the poison comes from the serum, and this
may be true. Further experimentation will determine this,
but it must be borne in mind that agar contains small
amounts of protein, and this has a large surface exposure
and is in a physical state most favorable to the action of a
proteolytic ferment in the serum.
Zinsser1 has shown that the action of complement "upon
typhoid bacilli strongly sensitized or not at all sensitized
may be carried on, at body temperature, for considerably
longer than twelve hours, without leading to a destruction
of the poisons, and that this is true when the quantities of
the bacteria used vary within the wide range of from one
to twelve agar slants. It has been found, in fact, that in
the case of this microorganism prolonged exposure at the
higher temperature of considerable quantities of bacteria
constitute an unfailing method of regularly obtaining power-
ful poisons. The results obtained by the use of smaller
quantities and the less vigorous action at low temperatures
are far less regular or satisfactory." He thinks that his
results throw more weight on the assumption that anaphyl-
atoxins are responsible to a large extent for the toxemic
manifestations of typhoid fever. He also states: "If we
1 Jour. Exp. Med., 1913, xvii, 117
PROTEIN SENSITIZATION OR ANAPHYLAXIS 339
leave out of consideration bacteria which, like the diphtheria
bacillus, produce true secretory poisons, it would be the
ability to gain a foothold in the body, the degree of invasive
power, the predilection in the choice of a path of entrance,
and the specific local accumulation, upon which the speed and
quantity of toxin 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, the point of view suggested by this work again
brings out the great importance of the work of Vaughan
and Vaughan and Wheeler on the non-specific poisonous
fraction obtained by hydrolysis of bacterial and other pro-
teins, and makes it desirable that the particular conditions
of anaphylatoxin and endotoxin production in the case of
individual pathogenic bacteria should be carefully studied."
We regard the work of Jobling and Bull1 as confirmatory
of our studies in every particular. These investigators
have studied the action of the cellular substance of the
typhoid bacillus and its split products, produced by the
action of a proteolytic ferment obtained from leukocytes,
and state their findings as follows: "Freshly washed,
unheated typhoid bacilli intravenously injected into dogs
cause the development of definite symptoms as early as
twenty minutes after the injection. Boiling for ten minutes
does not destroy the toxic effects of a freshly washed bac-
terial emulsion. Complete solution of the bacteria (in dilute
alkali) of a fresh emulsion does not prevent the removal of
the toxic substance with the coagulable proteins. The action
of leukoprotease splits the toxic substance to a non-coagulable
state, the digested mixtures being toxic after removing the
coagulable portion. The mere presence of the leukocytic
ferment is not responsible for the toxicity of the filtrate
from the digested mixture, and continued digestion destroys
the toxicity of a previous toxic mixture. From these observa-
tions it is concluded that the toxic properties of freshly
1 Jour. Exp. Med., 1913, xvii, 453.
340 PROTEIN POISONS
washed typhoid bacteria are not entirely due to preformed
secretory toxic bodies that are stored in the bacterial bodies,
but that these properties are due largely to products formed
by hydration of the bacterial proteins through the agency
of ferments present in the circulation of the animal previous
to the injection, or which become mobile subsequent to
the entrance of the foreign bodies into the blood-stream.
Since leukocytic ferments can attack the bacterial proteins
in vitro, it is possible that the leukocytes are a source of the
ferments which are active in experimental and natural cases
of intoxication with the whole bacteria."
Nolf1 has stated a theory of anaphylaxis which has come
to be known as "the physical theory." It supposes that
the active constituent of proteins is a thromboplastic sub-
stance which disturbs the colloidal equilibrium of the blood
and leads to the deposition on the surfaces of the leukocytes
and the endothelial cells of capillaries, of a delicate film of
fibrin. Thus stimulated, these cells pour out an unusual
amount of antithrombin. On account of the consumption
of a part of the fibrinogen and the increased formation of
antithrombin the blood fails to coagulate after anaphylactic
shock or peptone poisoning. On account of the coagulation
deposits on the endothelial cells the viscosity is increased
and the leukocytes adhere to the vessel walls, thus accounting
for the leukopenia observed after protein injections. The
endothelial cells are injured and the walls of the capillaries
become more permeable, thus accounting for the local edema
often seen in anaphylaxis. The fine capillaries of a given area
may be occluded by thrombi, and this explains the necrosis
characteristic of the Arthus phenomenon. The irritation
of the endothelial cells extends to the smooth muscle, and
this leads to vasoparalysis, and the characteristic fall in
blood-pressure. The affinity of the endothelial cells for
the protein is stimulated by the first injection, and acts in
a fulminating way on reinjection, and thus the suddenness
of anaphylactic shock is explained.
1 Archiv. intern, de physiol., 1910.
PROTEIN SENSITIZATION OR ANAPHYLAXIS 341
This is a plausible and attractive statement, and we are
inclined to believe that there is truth in it, but we fail to
see any reason for designating it as a "physical theory."
It starts out with the assumption that proteins contain a
poison and the theory is an explanation of the modus operandi
of the chemical poison. The endothelial cells are sensitized
and pour out a ferment, antithrombin, in increased quantity.
That the endothelial cells are involved in sensitization we
held as long ago as 1907. That the permeability of the walls
of the capillaries is increased under the action of protein
poison has been frequently demonstrated by the general
diapedesis. We found this true even when rabbits died as
the result of a single large dose of egg-white.
There is, however, one very important point in which
the theory of Nolf differs from ours. In his theory the cause
of death on reinjection is not due to the cleavage of the
protein introduced, but is due to the action of the antithrom-
bin on the blood. He holds that the fact that intravenous
reinjections are so much more effective both in dose and in
time than intraperitoneal and subcutaneous administrations
is in favor of his theory, and we are inclined to agree with
him on this point. However, anaphylactic shock cannot be
due wholly to rendering the blood non-coagulable, because
this may be done by injections of hirudin without shock.
Doerr objects to our theory on the ground that anaphylactic
shock follows reinjection so quickly that there is not time
for a ferment to split up the injected protein and liberate
a poison; but Nolf's theory also depends upon ferment
action. The sensitized endothelial cells must be awakened
by the reinjection, must pour out their abnormally accumu-
lated ferment, and this must act, not on the injected protein,
as we suppose, but on the blood. According to either theory,
certain cells are sensitized and store up zymogen, which is
activated on reinjection. This ferment acts either upon the
injected protein or on the proteins of the blood. It seems
to us that the time objection made by Doer to our theory
is quite as applicable to that of Nolf. The latter is, after
all, only a modification of the former.
CHAPTER XII
THE PARENTERAL INTRODUCTION
OF PROTEINS1
FOR a long time it was thought that the proteins of our
food undergo but slight modification before absorption
through the walls of the alimentary canal. The studies
of Beaumont laid the foundation of the scientific investi-
gation of proteolytic digestion, and soon it was shown that
the digestive juices convert proteins into peptones.
After experiments had demonstrated that peptone is
formed in alimentary digestion and had shown the com-
paratively ready diffusibility of the digestive products,
several questions arose. Among these may be mentioned
the following: (1) Is all the protein converted into peptone
in the alimentary canal, or is part of it absorbed in unaltered
form? (2) What is the fate of peptone after absorption?
Briicke,2 whose studies on pepsin and its action made him
one of authority in this matter, held that only a part of
the protein is converted into peptone in alimentary diges-
tion, and that much of the soluble protein of the food is
absorbed unchanged. Furthermore, he taught that the
fate of the two after absorption is different. The peptone,
he taught, is rapidly oxidized and serves as a source of
energy, but is not utilizable in the building of tissue, the
latter function devolving solely on the protein absorbed in
unaltered form. Briicke's arguments in support of this
theory may be briefly stated as follows:
1 This is largely taken from an article by Vaughan, Gumming, and
McGlumphy (Zeitsch. f. Immunitatsforschung, 1911, ix, 16).
2 Sitzungsber. d. k. Akad. d. Wissensch. zu Wien, 1859, Band xxxvii,
ibid., Band lix.
THE PARENTERAL INTRODUCTION OF PROTEINS 343
(a) At best the gastric juice forms peptone slowly, and
the time during which the food is detained in the stomach
does not permit of its complete peptonization. It will be
understood that the action of the pancreatic juice was
not known, nor had erepsin been discovered when Briicke
wrote. (6) In an animal killed while in digestion, Briicke
found forty-eight hours after death coagulable protein not
only in the chyle vessels in the intestinal walls, but in the
intestinal villi, and he concluded that this could come only
from the absorption of unaltered protein, (c) Briicke argued
that the absorption of unbroken protein is quite as possible
as that of fat, since the molecule of the former could not
be larger than that of the latter. This argument assumed
that the absorption of both proteins and fats is simply a
process of filtration.
Diakonow1 supported the theory of Briicke because
peptone cannot be found in large amount in the blood.
Voit and Bauer2 and Eichhorst3 concluded that unaltered
protein is absorbed because they found that the introduction
of protein in the large intestine is followed by increased
elimination of urea. This certainly is proof that the protein
is absorbed, but not proof that it is absorbed unaltered.
Eichhorst showed that a glycerin extract of the mucous
membrane of the large intestine had no digestive action,
but he did not show that the large intestine did not contain
any pancreatic juice and this might have digested the
proteins. Fick4 took an aqueous solution of peptone and
precipitated it with alcohol, then dissolved the precipitate
in water and injected it into nephrotomized dogs. He found
that the blood after this treatment yielded a larger amount
of nitrogenous material soluble in alcohol and precipitable
with mercuric nitrate, and he concluded that peptone
introduced into the blood is speedily converted into urea
without being employed in tissue building, while unaltered
1 Hoppe-Seyler's med. chem. Untersuchungen, 1867.
2 Zeitsch. f. Biol., 1869, v
3 Pfliiger's Arch., 1871, iv,570; 4 Ibid., 1872, v, 40.
344 PROTEIN POISONS
protein is used in tissue metabolism. Fick's conclusions
support Briicke's hypothesis. Maly1 pointed out a possible
error in Fick's work, showing that while peptone may be
precipitated from aqueous solution by alcohol, it is not
wholly insoluble in this menstruum, and the increase of
alcohol soluble nitrogen in the blood might be due to peptone
and not to the conversion of this into urea.
Evidently if Briicke's theory were true, the animal body
could not maintain its health and vigor if fed exclusively
on peptone, which according to the theory is not utilizable
by the animal in the repair of tissue. Plosz2 fed animals
exclusively, so far as their nitrogenous food is concerned,
on peptones and found that they did not loose weight or
suffer in any detectable way. Maly3 confirmed this finding
which has been repeated many times and under divers
conditions, so that now Briicke's contention that the
absorption of unaltered protein is essential to health has
no support.
The second question, What is the fate of the absorbed
peptone, became for a time one of much importance. Plosz
and Gyergyai4 injected from 200 to 300 c.c. of a 10 per
cent, solution of peptone into the stomachs of fasting dogs
and after periods of from one to four hours searched for
the peptone in the blood and tissues of various organs. The
method of recognizing peptone consisted in the applica-
tion of the biuret and Millon tests to the filtrate after the
removal of all the coagulable protein by acid and heat.
They found the largest amount of peptone in the mesenteric
veins and in extracts of the mesentery, much less in the
liver, and only traces in hepatic and carotid blood. Next,
they injected dogs and cats intravenously with a 10 per
cent, solution of peptone, employing from 100 to 200 c.c.,
and introducing it at the rate of from 2 to 3 c.c. per minute.
A dog received 200 c.c. during one and one-half hours, and
after three hours the carotid blood showed only a small
1 Pfluger's Arch., ix, 605. 2 Ibid., 325.
3 Loc. cit. 4 Pfluger's Arch., 1875, x, 536.
THE PARENTERAL INTRODUCTION OF PROTEINS 345
amount of peptone which had wholly disappeared after
four hours. When larger amounts were used a small quantity
appeared in the urine, but the proportion eliminated in this
way was only a small part of that injected. They also
transfused certain organs and tissues with blood to which
peptone has been added, and found that the peptone soon
disappeared from the blood. These investigators concluded
that peptone is soon so changed in the organism that it
can no longer be detected by the method which they
employed. Whether it is changed directly into albumin
or is so altered by cell activity by combining with other
substances, they could not decide. They were of the
opinion that the capability of effecting this change is not
confined to any one organ or tissue, but that it may occur
in the liver, muscle, or other tissue. They were quite con-
vinced, however, that the conversion is not essentially one
of oxidation, since the amount of oxygen in the blood did
not affect it.
Schmidt-Muhlheim1 injected from 5 to 10 grams of
peptone into the jugular vein of dogs and found that the
peptone disappeared from the blood within sixteen minutes
after completing the injection. He also concluded that the
injected peptone undergoes a rapid conversion into albumin
and globulin.
According to Hofmeister,2 peptone, when injected into
the blood, does quickly disappear from that fluid, but is
not converted into albumin or globulin. It quickly diffuses
through all the tissues undergoing a dilution which is
determined by the total fluid in the body, and which is
so great that its detection by chemical tests is impossible.
Diffusion into the brain results in certain characteristic
symptoms, the most marked of which are muscular weak-
ness and somnolence. These symptoms may be observed
in a 10 kg. dog after the subcutaneous injection of from
0.2 to 0.4 gram of peptone, but the fatal dose is large; as
1 Du Bois Raymond's Arch. f. Phys., 1880.
2 Zeitsch. f. phys. Chemie, 1881, v, 127.
346 PROTEIN POISONS
high as 1 gram per kilo. Injections of peptone lead to
lowered blood pressure and much of the peptone, according
to Hofmeister's finding, may be deposited in the tissue
where it may be detected at a time when there is none in
the blood. He found one-seventh of the peptone injected
into the blood in the kidneys, which organs were equiva-
lent to only yir °f the body weight, and concludes that the
peptone injected into the circulation has a special predilec-
tion for renal tissue. When the amount of peptone injected
is large, Hofmeister recovered as much as 84 per cent, of it
from the urine.
Neumeister1 reviewed the literature of this subject up
to that time, and made some additional contributions.
He stated that some proteins are absorbed unchanged,
that others need only to be dissolved, and that still others
must be digested. He stated that the compound proteins,
as casein and hemoglobin, when injected into the blood,
act like foreign bodies, and are eliminated in the urine, while
the simple and denatured proteins, when injected into the
blood, do not cause albuminuria. Stockvis did not observe
albuminuria after injecting dog, rabbit, or frog serum into
dogs or rabbits, but did when he used egg albumen. Leh-
mann invariably induced albuminuria by injecting egg
albumen intravenously in dogs, but failed to do so when
he employed sodium albuminate, or syntonin, prepared
from frogs' muscle, myosin, or fibrin. Ponfick found that
dogs bear astounding amounts of lamb serum, free from
corpuscles, when the injections are made slowly, and under
gentle pressure. "The amount of urine was not appreciably
increased, although the color became darker, owing to the
greater concentration, while not a trace of albumin could
be detected." There is no statement concerning the effect
upon the elimination of nitrogen; Forster injected large
amounts of horse serum into dogs, while the urine remained
free from albumin. Neumeister injected into the jugular
veins of dogs without inducing albuminuria, the following
1 Zeitsch. f. Biol., 1891, xxvii, 309
proteins in large amounts: Syntonin and albuminate from
egg albumen, syntonin from ox flesh, crystalline phyto-
vitellin from pumpkin seed, and pure serum albumin from
the ox.
According to Neumeister, Salviolo was the first to show
that peptone is transformed by the living intestinal wall,
but this investigator only demonstrated that peptone
disappears when placed in an intestinal loop and cannot
be found in either the blood from the part, or within the
loop. The nature of the transformation was not determined.
The fact that peptone is synthesized into albumin seems
to have been first suggested by two women, students of
Kronecker, Nadine Popoff and Julia Brinck. It was thought
by these investigators that this synthesis is accomplished
partly by the epithelial cells of the intestinal wall, and
partly by a microorganism, to which Julia Brinck gave
the name micrococcus restituens.1 Hofmeister2 suggested
that the leukocytes in the intestinal wall might combine
with peptone much as hemoglobin does with oxygen in the
lungs, and Heidenhain3 thought that the leukocytes might
play a part in the absorption of peptones, but that it could
not be as suggested by Hofmeister,4 otherwise the leuko-
cytes in the circulating blood would combine with peptone
injected intravenously.
As early as 1874 Tschiriew,5 working under Ludwig's
direction, found that dog serum transfused into another
dog increased the elimination of urea much more slowly
than when given to the dog by mouth, but Forster6 found
that horse serum affects the urea output in dogs equally,
both in amount and time, whether given intravenously or
by mouth.
Zunz and von Meering7 injected solutions of peptone
1 Zeitsch. f. Biol., 1889, vii, 427, 453.
2 Zeitsch. f. phys. Chemie, 1881, v, 151.
3 Pfltiger's Arch., 1888, liii. 4 Loc. cit.
6 Arb. a. d. phys. lust, zu Leipzig.
6 Zeitsch. f. Biol., 1895, ii, 496.
7 Pfluger's Archiv, 1883, xxxii, 173.
348 PROTEIN POISONS
and egg-white intravenously in rabbits. The injections
were single and too large to be followed by marked rise,
and too small to result in marked depression of temperature.
However, that they noticed the ill effects of these injec-
tions is shown by the following quotation : " The poisonous
action of peptone was unknown to us at the time when the
experiments were made, but some of our rabbits died soon
after rather large injections of peptone. Moreover, un-
changed egg-white is not an indifferent substance, and it
has long been known that its direct introduction into the
blood may cause albuminuria and deep-seated changes
in the kidneys."
Giirber and Hallauer1 employed casein for the reason
that it may be distinguished from other proteins by the
action of rennin. Solutions of this protein were injected
intravenously into rabbits and the casein was detected in
the bile. Evidently the foreign protein was on its way to
the intestines where it might be properly digested. These
authors quite properly point out that because a protein
injected into the blood does not appear, or appears only
in part, in the urine, is no proof that it has been assimilated
in unchanged form by the tissues, because it may have
been carried to the intestine and there properly digested.
They also brought out another point, confirmed later
by our own work, that when the foreign protein diffuses
from the blood it carries with it some of the blood proteins.
This is true whether it is poured out into the intestine or
eliminated through the kidneys.
The results obtained by Giirber and Hallauer have been
confirmed by Burckardt,2 who injected hemielastin intra-
venously and found it in the wall of the small intestine.
He concluded that it had been brought to this locality
preparatory to its being properly digested and fitted for
assimilation.
Friedemann and Isaac,3 after extensive experimentation,
1 Zeitsch. f. Biol , 1904, xlv, 372.
2 Zeitsch. f. physiol. Chem., 1907, li, 506.
3 Zeitsch. f. exp. Path. u. Ther., 1907, iv, 830.
concluded that both homologous and heterologous proteins
when parenterally administered are digested and probably
assimilated. They say: In fasting animals the parenteral
administration of protein leads to increased protein metab-
olism. The increased nitrogen elimination is the same
whether the injected protein be the serum of the same or
of a different species, or egg albumen. In dogs in nitrogen
equilibrium protein parenterally administered behaves
the same as that given by mouth. Carbohydrates hinder
increased nitrogen elimination while on a carbohydrate
free diet there is increased nitrogen elimination. In her-
bivorous animals (goats and sheep) there seems to be a
tendency to retain some of the nitrogen given parenterally,
but the results obtained were not constant. We cannot
speak of a toxic protein metabolism unless symptoms imme-
diately follow the administration. If heterologous proteins
be poisonous, when administered parenterally, we have
not been able to demonstrate it in fasting animals. If
this does not exclude a toxic metabolism it renders its
assumption wholly hypothetical. In our experiments
increased nitrogen elimination corresponds to increased
administration, and a toxic protein metabolism is char-
acterized by the fact that nitrogen ingestion and elimination
bear no relation to each other. The parenteral adminis-
tration of protein has an advantage over the enteral, because
in the former we know just how much protein enters the
blood. When blood serum is introduced there can be no
increased concentration of proteins in the blood, but in
fasting animals the introduction of serum, either homolo-
gous or heterologous, does lead to increased nitrogen
elimination. The transfusion of blood even in large amount
from one dog to another is not followed by any marked
nitrogen elimination by the recipient. It seems, therefore,
that even homologous serum behaves like a foreign protein,
possibly on account of the changes that have taken place
in it during coagulation. In sensitized animals the parenteral
introduction of the homologous protein leads to explosive-
like increase in nitrogen elimination.
350 PROTEIN POISONS
In the light of later research, some criticism of the above
given conclusions reached by Friedemann and Isaac may
be made. Foreign proteins when introduced parenterally
are poisonous. In some, the poisonous action is due to
ferments, but all are poisonous, even when the ferment
action has been destroyed by heat. It is only a question
of dosage. Even egg-white or horse serum injected intra-
venously in sufficient doses will kill dogs and rabbits.
Saturat on of the body cells with any foreign protein inter-
rupts their function. The protein foods of the body cells
are carefully prepared physiologically. The foreign pro-
teins eaten by the animal are broken into non-protein
bodies, and these pieces are put together again after a
model which is peculiar to that species of animal to which
the feeder belongs. In this way the specific albumins and
globulins of the blood of each species are constructed,
and these supply the normal protein foods for the body
cells. Foreign proteins parenterally introduced are digested,
if the amount be not too large. This digestion is carried
out in part in the intestines, and other body cells acquire
the function of digesting a limited amount of the protein
introduced. Certainly the digestive products formed in
the intestine are fit for assimilation, and it may be that
those formed in other parts are also utilizable, but the
capability of the body of taking care of proteins parenterally
introduced is limited, and in large doses all foreign proteins
thus administered are poisonous. Whether this is also
true of homologous sera we do not have sufficient data to
determine.
Bankowski and Szymanowski1 find that normal human
blood, when injected intravenously into guinea-pigs, kills
with the symptoms of anaphylactic shock in doses of 0.5
per cent, of the body weight of the animal. In typhoid
fever the minimum fatal dose falls to 0.25 and in scarlet
fever and measles to 0.13 per cent, of the body weight of
the animal. Fetal human blood is well-nigh atoxic; when
1 Zeitsch. f. Immunitatsforschung, 1913, xvi, 330
THE PARENTERAL INTRODUCTION OF PROTEINS 351
the dose is as much as 2.5 per cent, of the body weight it
kills slowly. The mother's blood kills in doses of 0.5 per
cent., and fetal blood in doses of 2.5 per cent. This is not
due to biological differences between the blood of the
mother and that of the fetus, because one sensitizes to
the other as well and in as small doses as to itself, but is
due to the relative freedom of the fetal blood from ferments.
When human blood is injected into the blood of a guinea-
pig the former carries the ferment and the latter supplies
the substrate. In the infectious diseases the proteolytic
ferment in the blood is increased and consequently the
minimum fatal dose is decreased.
Dehne and Hamburger1 state that white mice do not
produce a precipitin when treated with horse serum, and
Celler and Hamburger2 find that white rats fail to respond
to ox serum by elaborating a precipitin.
Uhlenhuth3 and Michaelis and Oppenheimer4 found
that when rabbits are repeatedly fed through a tube with
egg-white or serum, they develop precipitins. Celler and
Hamburger say that this may be due: (1) To injury of
the esophagus or stomach by the tube, and the introduction
of the protein in this way. (2) To the failure of the secre-
tion on account of the unnatural method of feeding. (3) To
the direct introduction of the protein into the intestine
where, according to Oppenheimer and Rosenberg,5 serum
proteins resist tryptic digestion.
Celler and Hamburger found that in forced or tube
feeding the protein may be absorbed unchanged, on account
of the lack of digestive juice.
Chiray6 has studied the effects of the administration of
heterologous proteins. The intravenous injection of a very
small amount of egg-white in rabbits causes, after from ten to
thirty minutes, a transitory albuminuria with increase in the
1 Wien. klin. Woch., 1904, No. 29. 2 Ibid., 1905, xviii, 271.
3 Deutsch. med. Woch., 1900, p. 734.
4 Arch. f. Phys., phys. Abt., Suppl.-Band, p. 336.
6 Hofmeister's Beitrage, 1903, v, 412.
6 These de Paris, 1906; Jahresber. d. Tierchemie, 1907, xxxvi, 805.
352 PROTEIN POISONS
volume of urine, but without glycosuria, hemoglobinuria, or
hematuria. The intravenous, subcutaneous, or intraperito-
neal injection in increasing doses, even with long intervals,
causes in rabbits a gradual decrease in weight. When the
egg-white is injected into the portal vein the albuminuria
appears much later than when the injection is made into
the general circulation. Subcutaneous injections induce
an albuminuria which appears later, is less marked, and
persists longer than when the injection is made intravenously.
The intramuscular injection of 2 c.c. of egg-white in man
was without effect, but when tried upon one who had renal
deficiency, albumin appeared in the urine from fourteen
to twenty-four hours later. In cases of marked albuminuria
injections of egg-white did not materially affect the excre-
tion of albumin. In rabbits, dogs, and men the introduction
of large amounts of egg-white into the stomach was fol-
lowed by albuminuria, though this did not invariably
occur in the men. The injection of egg-white into the
rectum of rabbits was followed by an albuminuria which
appeared later and was more persistent than when it was
given by the stomach. Egg-white administered by rectum
to man, especially to convalescents from infectious diseases,
was followed by an albuminuria, but this did not occur
when the albumin was mixed with an active trypsin before
injection. In some, not in all, the administration of peptone
by rectum was followed by albuminuria as well as peptonuria.
Injections of egg-white in rabbits decrease the proteins of
the blood, as shown by the refractometer. It appears that
not only a part of the foreign protein, but also a part of
the blood proteins passes into the muscles. All of the
injected protein, probably not the greater part, 'does not
pass through the kidney, at least not within the time of
an observation, but much is withdrawn from the blood
and held in the tissues. Repeated injections of egg-white
lead to marked structural changes in the kidneys. Alimen-
tary albuminuria is not due to poisons resulting from the
splitting of the protein, but the absorption and elimination
of the unbroken protein which is a foreign and poisonous
THE PARENTERAL INTRODUCTION OF PROTEINS 353
body. When a milk diet is presented, casein may in
some instances be found in the urine. The prohibition of
egg diet in albuminuria is justified.
The statement that the injection of a foreign protein
leads to the exudation of the normal proteins of the blood,
as made by Chiray, is interesting, and if confirmed it may
be found to be of marked importance. It has been prac-
tically confirmed by Wolf,1 who found that the proteins
of the plasma were diminished by the injection of Witte's
peptone in 11 out of 14 tests.
Oppenheimer2 estimated that as much as 49 per cent,
of egg-white injected intravenously or intra-abdominally
in rabbits is eliminated in the urine. However, he does not
claim any great exactness for this work, because he is aware
of the fact that all the protein in the urine does not come
from that injected, and that a part of it is serum albumin.
It is probable that the latter makes up the larger part.
Castaigne and Chiray3 hold that heterologous proteins
injected subcutaneously are absorbed and eliminated in
the urine unchanged. They act as poisons, causing destruc-
tion of the proteins of the blood and increased elimination
of nitrogen, urea, and sulphur. The decrease in the normal
proteins of the blood may be as high as from 1 to 3 per
cent. This is not due to hydremia as shown by determina-
tion of total solids. Repeated injections of heterologous
proteins, either subcutaneously or intravenously, lead to
cachexia.
Nobecourt4 introduced egg-white into the alimentary
canal of rabbits. He used 46 animals, 31 adults, weighing
from 1650 to 2270 grams and 15 young, weighing from
320 to 4030 grams. These received by the stomach or
rectum from 5 to 13 c.c. of egg-white, at each injection
at intervals of one, three, seven, ten, and fifteen days.
The mortality was as follows:
1 Arch. int. Phys., iii, 343.
2 Hofmeister's Beitrage, iv, 263.
3 Compt. rend. Soc. biol., 1906, Ix, 218.
* Ibid., 1909, xlvi, 850.
23
354 PROTEIN POISONS
Per cent, of mortality:
In adults: In young:
Intervals. Stomach. Rectum. Stomach. Rectum.
Every day ... 0 0 100 100
Every three days . 100 0 100 100
Every seven days . 50 50 33 33
Every ten days .50 0 33 33
Every fifteen days 0 50 33 33
From the above test, 24 animals, 20 adults and 4 young
survived. After a rest of from twenty to forty-four days
each of these received rectal injections, every seven days
of from 3.3 to 9.6 c.c. of egg-white.
Per cent, of mortality.
Interval in first series. Method of administration in first series.
Stomach. Rectum.
Every day 66 0
Every three days ... 66 adults 0
Every seven days . . . f50 adults 75 adults
\50 young 100 young
Every ten days ... 0 50
Every fifteen days . . 0 0
From the second ordeal, 13 animals, 12 adults and 1
young, survived. After a rest of from seventeen to forty
days these received every seven days rectal injections of
from 5.2 to 8.9 c.c. of egg-white, with the following results:
Per cent, of mortality.
Intervals in first series. Method of administration in first series.
Stomach Rectum.
Every day 0 50
Every three days ... 0 100
Every seven days / 50 adults 100
\100 young
Every ten days ... 100 100
Every fifteen days . . 100 0
From this test, 4 animals, all adults, survived. After
a rest of from twenty-four to forty-five days these received
every seven days from 4.6 to 7.5 c.c. of egg-white. Three
died.
It appears from the experiments of Uhlenhuth and Nobe-
court that egg-white is absorbed, at least in some instances,
unchanged from the stomach and intestines of rabbits.
THE PARENTERAL INTRODUCTION OF PROTEINS 355
When tubes are used for the introduction of the egg-white
it is possible that a small amount of the material may be
introduced through some slight wound or abrasion in the
mucous membrane. Sensitization might be induced in
this way, but it is hardly conceivable that subsequently
enough would be introduced in this way to kill the animal.
It seems, therefore, that we must conclude that in forced
feeding, at least, unbroken egg-white may be absorbed from
the alimentary canal of the rabbit. It must be understood,
however, that apart from any injury to the mucous mem-
brane, the conditions of forced feeding are not exactly the
same as those in natural feeding. Celler and Hamburger
have called attention to this point. By continued tube-
feeding of rabbits with the serum and blood of the ox, in
only one instance did they obtain a precipitin for the serum,
and an hemolysin for the corpuscles, while they found
that rabbits after fasting took the serum and blood willingly
when mixed with milk, and in none of these was there any
evidence of absorption without digestion. They admit the
possibility of wounding the mucous membrane with the
tube, or of carrying the material through the tube into
the intestine, but they are inclined to the opinion that in
the unnatural tube feeding the digestive secretions are not
poured out so freely or are less effective than in natural
feeding. This is in accord with the findings of Pawlow,
who holds that desire for food is an important factor in
securing thorough digestion.
With this brief and imperfect review of the literature of
the subject, we turn to our own experimental work. Our
method is to inject egg-white into the animals and test for
its presence in the blood and extracts of tissue by sensitizing
guinea-pigs, having first demonstrated that the blood of
the rabbit and extracts from its tissue do not sensitize
guinea-pigs to egg-white. The details of the method will
be developed in the report of the experiments. Our findings
are as follows:
1. Egg-white injected into the stomach of a rabbit may
be in part absorbed unchanged.
356 PROTEIN POISONS
December 7, 1909, at 9.30 A.M., 50 c.c. of egg-white was
introduced through a tube into the stomach of a rabbit that
had been kept without food for two days. Neither at the
time nor subsequently did this have any recognizable
effect upon the rabbit. 3 c.c. of blood was drawn from the
heart of this rabbit at 10.30 and 11.30 A.M., and at 12.30
2.30, and 4.30 P.M., and each of these portions of blood was
injected intraperitoneally in a fresh guinea-pig. January
3, 1910, each of these pigs received intra-abdominally
5 c.c. of a dilution of egg-white with an equal volume of
physiological salt solution.
Only one of these pigs developed symptoms of sensitization
and this one received blood drawn from the rabbit's heart
three hours after the introduction of the egg-white into the
stomach. Neither the blood drawn earlier nor that drawn
later sensitized guinea-pigs.
January 8, 1910, at 8 A.M., 50 c.c. of a dilution of egg-
white with physiological salt solution (1 to 1) was intro-
duced through a tube into the stomach of a rabbit which
had not been kept without food.
Hourly 2.5 c.c. of blood was drawn from the heart of this
animal, and injected intra-abdominally into guinea-pigs.
January 22, 1910, these pigs were treated each with 5
c.c. of the egg-white dilution intra-abdominally. The first,
second, and third hour pigs showed no sensitization; the
fourth and fifth hour ones were sensitized, while the sixth,
seventh, and eighth were not.
That absorption from the stomach of the fed animal
should have been more tardy than from the fasting one is
easily understood.
2. Egg-white injected into the rectum of a rabbit may
be, in part at least, absorbed unchanged.
January 8, 1910, at 8 A.M., 50 c.c. of egg-white diluted
with physiological salt solution (1 to 1) was introduced
through a tube into the rectum of a rabbit. Hourly, 2.5
c.c. of blood was drawn from the heart and injected intra-
abdominally into guinea-pigs.
January 22, 1910, these pigs were tested and all from the
first to the seventh hour were found to be sensitized to
egg-white.
It appears from this that egg-white may be absorbed
from the rectum of a rabbit without being so far altered
as to destroy its specific sensitizing properties and that
absorption into the blood begins within the first hour and
continues for at least seven hours.
3. Egg-white injected into the peritoneal cavity of a
rabbit may be absorbed unchanged.
December 7, 1907, at 9.30 A.M., a rabbit received intra-
peritoneally 50 c.c. of a dilution of egg-white with an equal
volume of physiological salt solution. Hourly, 2.5 c.c.
of blood was drawn from the heart of this animal and
injected intra-abdominally into guinea-pigs.
January 3, 1910, these pigs were treated with the egg-
white dilution given intraperitoneally.
All, from the first to the fourth hour, died, the first two
in fifteen and the latter 'in twenty minutes. The fifth hour
one was not sensitized. It should be stated that in all these
experiments guinea-pigs found not to be sensitized to egg-
white were subsequently tested and found to be sensitized
to the blood serum of the rabbit.
4. Egg-white injected intravenously in rabbits quickly
disappears from the circulating blood.
January 3, 1910, a rabbit received intravenously 50 c.c.
of a dilution of egg-white with physiological salt solution
(1 to 1). Every half hour blood was drawn from the heart
of this animal and injected intra-abdominally into guinea-
pigs.
January 12, 1910, these pigs were tested with the egg-
white dilution.
The first two were found to be sensitized to egg-white
while the others were not. The pig that received blood
drawn at the end of the first hour died in a typical way
within thirty minutes, while the blood drawn at the expira-
tion of one and one-half hour failed to sensitize.
December 6, 1909, a rabbit received intravenously 50
c.c. of the egg-white dilution (1 to 1). Hourly blood was
358 PROTEIN POISONS
drawn from the heart and injected into guinea-pigs. The
first two were found to be sensitized while the others were
not.
5. Egg-white injected intravenously in rabbits may be
detected in the peritoneal cavity after it has disappeared
from the circulating blood.
January 3, 1910, a rabbit received intravenously 50 c.c.
of the egg-white dilution. Two and one-half hours later
and after the egg-white had disappeared from the heart's
blood, as was shown by a subsequent test, some physio-
logical salt solution was injected into the peritoneal cavity,
withdrawn and injected into a guinea-pig, which later was
found to be sensitized to egg-white.
6. Egg-white injected intravenously into rabbits may
be detected in the bile.
November 15, 1909, a rabbit received intravenously 50
c.c. of the egg-white dilution, one and one-half hours later
the abdominal cavity was opened, the animal being under
ether, and small amounts of bile and washings from the
small intestine were injected into guinea-pigs, all of which
later were found to be sensitized to egg-white. Of four
pigs thus treated all but one died, and this one developed
marked symptoms when treated with the egg-white dilution.
7. Egg-white when injected intravenously into a rabbit
may be detected by the sensitization test in certain organs
after it has disappeared from the circulating blood.
December 6, 1909, at 8.45 A.M., a rabbit received intra-
venously 50 c.c. of a dilution of egg-white with an equal
volume of physiological salt solution. 5 c.c. of blood was
drawn from the heart of this animal at 9.45, 10.45, and
11.45 A.M., and at 1.45 P.M. Each of these portions was
injected into the abdomen of a guinea-pig and one hour
after the last blood was taken the rabbit was killed with
ether, and extracts of the brain, liver, kidney, and spleen,
with physiological salt solution were made and injected
into other fresh pigs. December 17, 1909, each of the pigs
that had been treated with the blood of the rabbit had
5 c.c. of egg-white dilution (1 to 1) intra-abdominally.
THE PARENTERAL INTRODUCTION OF PROTEINS 359
The only pig that gave any evidence of sensitization was
the one that had received the first blood, drawn one hour
after the injection of the egg-white. The symptoms in
this animal were slight and transitory. The other pigs
showed no indications of having been sensitized. It seems
from this that after one hour there was no egg-white in the
circulating blood of the rabbit. All of these pigs had been
sensitized to the proteins of the rabbit's blood as was shown
by treating them with rabbit serum.
January 4, 1910, the guinea-pigs that had received the
extracts of the organs were treated with the egg-white
dilution. All were affected within a few minutes.
The one that had received the kidney extract was most
seriously disturbed and passed to the convulsive stage, but
ultimately recovered. The one that had the spleen extract
came next in the severity of the symptoms developed. The
first and second stages were well-marked in this animal,
somewhat less so in the pig that had received the extract
from the brain. Much to our surprise, the pig that had
received the extract from the liver was least affected.
However, failure to sensitize with the extract from an
organ does not necessarily mean that the tissue of the organ
contained none of the foreign protein. It may combine
with certain tissues so firmly that it is not removed by a
simple solvent, like physiological salt solution. The fact
that from certain organs the extracts did sensitize the pigs
shows that these tissues had absorbed the egg-white, but
failure to sensitize or to sensitize so fully with other extracts
does not conclusively show that such tissues have not
absorbed the protein.
8. Egg-white carried into the tissue after intravenous
injection may be washed back into the blood current by
transfusion with salt solution.
A rabbit received intravenously 50 c.c. of the egg-white
dilution. After one hour egg-white had disappeared
from the circulating blood. Two and one-half hours after
the injection of the egg-white, the animal was transfused
with physiological salt solution. During the transfusion,
360 PROTEIN POISONS
2 c.c. portions of the fluid were drawn from the heart and
injected into guinea-pigs. The last of these portions was
drawn after one liter of the salt solution had passed through.
All of these portions sensitized the guinea-pigs. After the
transfusion, the brain, liver, spleen, kidney, and the deltoid
muscle were removed, and rubbed up with physiological
salt solution. 2 c.c. of each of these extracts, after filtra-
tion, was injected into guinea-pigs. The one having the
brain extract was not affected by the subsequent injection
of egg-white. The one having the liver extract was in
convulsions within five minutes after receiving the egg-
white. The ones having the extracts from the spleen and
muscle developed first and second stages of sensitization,
but recovered, while the one that received the kidney
extract was not affected.
We conclude from this that the brain and kidney were
washed free of the egg-white by the transfusion, while the
muscle, liver, and spleen held the egg-white more tena-
ciously.
9. The injection of egg-white intravenously in rabbits
decreases after a few hours the total protein in the blood.
In reviewing the literature we have referred to the finding
of Chiray that the intravenous injection of foreign proteins
decreases the total proteins of the blood. His experiments
were made with a refractometer. We deemed this of
sufficient importance to justify further study. On one
day blood was drawn from a rabbit and the serum obtained.
On the next day this animal received 50 c.c. of the filtered
egg-white dilution intravenously. Each cubic centimeter
of this dilution of egg-white contained 26 mg. of protein,
as calculated from a nitrogen determination. In other
words, with the dilution there was introduced into the blood
of the rabbit 1.3 grams of foreign protein. On the day after
the injection more blood was drawn and the serum from
this secured. The total nitrogen in these sera was deter-
mined and the protein content calculated with the following
results:
THE PARENTERAL INTRODUCTION OF PROTEINS 361
Per cent, of protein in the blood serum before the
injection of egg-white 10 . 50
Per cent, of protein in the blood serum after the
injection of egg-white 8.18
Loss 2.32
This experiment was repeated on a second animal with
the following results:
Per cent, of protein in the blood serum before the
injection of egg-white 9 . 33
Per cent, of protein in the blood serum after the
injection of egg-white 7.36
Loss 1.97
<
In a third animal the following results were obtained:
Per cent, of protein in the blood serum before the
injection of egg-white 7.90
Per cent, of protein in the blood serum after the
injection of egg-white 6.30
Loss 1.40
It appears from these figures that the injection of egg-
white intravenously in rabbits is followed by the disap-
pearance of an appreciable amount of the normal proteins
from the circulating blood. This confirms the finding of
Chiray.
10. The injection of a large amount of egg-white intra-
venously in rabbits proves fatal.
No. 1. 35 c.c. of undiluted egg-white filtered through
cotton was slowly injected into the ear vein. The respira-
tion was immediately embarrassed, and with a slight con-
vulsive movement the animal died before it could be removed
from the table. On opening the thorax the heart was found
to be still beating and irregularly distended. The right
side was dilated and filled with dark fluid blood. Markedly
anemic areas were plainly seen in the lungs.
No. 2. 32 c.c. of the same was injected more slowly and
through a finer needle. The result was practically the same.
362 PROTEIN POISONS
No. 3. 40 c.c. was injected. The respiration became
difficult and the animal quite limp. The right side was
found to be paralyzed, but the animal lived for two hours,
when it died with failure of respiration, and without a
movement. The heart was dilated and contained dark,
fluid blood. Anemic areas were seen in the lungs and the
muscles also were anemic.
Van Alstyne and Grant1 injected dilute egg-white intra-
venously into a dog and sensitized guinea-pigs with blood
drawn from one-quarter to seventy-two hours. Pearce2
injected foreign proteins intravenously into rabbits, and
sensitized guinea-pigs with organ extracts. His conclusions
are stated as follows: "Extracts of the kidneys of normal
rabbits prepared one, two, three, and four days after the
intravenous injection of egg albumen and horse serum have
the power to sensitize guinea-pigs to a second injection of
these proteins. The sensitization by first- and second-day
extracts was constant and intense, that by the third-day
extracts was less marked and sometimes was not evident,
and that by the fourth-day extracts was only occasional,
and when present was always weak. Comparative studies
of the power of the blood, liver, and kidney to sensitize
indicate that this sensitization depends upon the content
of the foreign protein in the circulatory blood and not upon
its accumulation or fixation in the tissues of an organ. This
opinion is supported by other experiments in which the
sensitizing power of the blood and of the extracts of unwashed
kidneys was compared with the sensitizing power of washed
kidneys. The weak sensitizing power of washed kidney
extract is taken as evidence that foreign proteins of the
kinds used are not held in the tissues of the kidney, and if
these results may be applied to nephrotoxic proteins, it
follows that nephritis is not due to selection and persisting
fixation of a protein by the renal cells, but is due to the
action of such proteins merely during the process of elimina-
1 Jour. Med. Research, 1911, xxv, 399.
2 Jour. Exp. Med., 1912, xvi, 349.
THE PARENTERAL INTRODUCTION OF PROTEINS 363
tion. In experimental acute nephritis of the type due to
uranium nitrate, the power of sensitization to egg albumen
is prolonged for twenty-four hours, and in the chromate
type for forty-eight hours, thus indicating that in nephritis
of the acute type at least, the elimination of a foreign protein
is delayed."
In our opinion the possibility of harm coming to the
kidney or any other organ from the deposition of a foreign
protein in it is not due to any directly poisonous effect of
the foreign protein but to the liberation of the poisonous
group when the body cells become sensitized and split up
the foreign protein.
Abderhalden1 has shown by both dialysis and by the
polariscope that foreign proteins injected into animals are
digested by ferments. However, he does not find evidence
that specific proteolytic ferments are formed. Indeed, it
still remains a question whether the sensitizer leads to the
development of an entirely new ferment or causes the
common non-specific proteolytic ferment of the blood to
develop specific properties. We regard this question as
only of academic value. In either case the proteolytic
ferment becomes specific, whether formed by an altered
rearrangement in the molecules of the cells or by alteration
in the molecular structure of a non-specific proteolytic
ferment. Abderhalden believes that the ferment is always
present in the blood, and that it is a secretion of the leuko-
cytes. We agree with him insofar as the non-specific
proteolytic ferment of the blood is concerned. The blood
is a digestive fluid, but we believe that specific ferments
are developed in various fixed cells under the influence of
foreign proteins or sensitizers. Abderhalden holds that
the ferment is always present in the blood, and that the
ferment and the sensitizer may both be present as they
are on first injection, but that for the production of ana-
phylactic shock a third and unknown factor is necessary.
He seems to be influenced in this belief largely by the
1 Schutzfermente, 1912.
364 PROTEIN POISONS
phenomena of so-called antianaphylaxis, but he admits
that rapid digestion in this state may be prevented by the
accumulated products of digestion. He says: "From recent
studies we know that the ferment forms a compound with
the substrate before the equilibrium of the latter is destroyed.
After the cleavage the ferment is again free, so far as it is
not bound by the cleavage products." It seems to us that
this is all that is necessary to explain the known facts in
so-called antianaphylaxis.
In his work on the digestive action of blood serum Abder-
halden has largely employed polypeptids and purified
peptones. Of course, he does not expect these denatured
proteins to act as sensitizers and lead to the development
of specific ferments, but they are especially suited for diges-
tive experiments, because the split products as soon as
formed are easily recognized by their effect on the rotation
of light. In this way he has shown that peptones and poly-
peptids are quickly split into their constituent amino-acids
by the proteolytic ferment normally present in serum. In
other instances he has employed native proteins as sensi-
tizers. In one case he divided a lot of guinea-pigs sensi-
tized to egg-white into three .groups. The members of
the first group while in the sensitized state were bled, and
the serum thus obtained was digested with egg-white, and
it was demonstrated both by dialysis and the optical method
that the egg-white was digested by the serum. Now, had
this been done with the serum of normal guinea-pigs there
would have been no recognizable digestion. It must follow,
therefore, so far as we can see, that the blood-serum of the
sensitized guinea-pig contains a ferment which is not
present in the blood-serum of the fresh guinea-pig. Further-
more, had it been tried, it would have been found that the
blood-serum of the guinea-pig sensitized to egg-white would
either have no digestive action, or but slight effect, on other
proteins. It follows, therefore, that the sensitized animal
differs from the unsensitized in the fact that its body cells
elaborate a specific ferment which digests the protein by
which it was called into existence, and no other. It will,
THE PARENTERAL INTRODUCTION OF PROTEINS 365
of course, be understood that the non-specific proteolytic
ferments are capable of digesting a more or less extended
group of proteins, but with this ferment the digestive process
proceeds slowly, and it is not supposable that all proteins
would be digested by it.
From the second set of sensitized guinea-pigs Abder-
halden took the serum and dialysed it by itself, the purpose
being to see if, while in the sensitized state, the blood-serum
contains any diffusible biuret body. The serum from
six sensitized guinea-pigs was tested in this way, and only
in one instance did the dialysate respond to the biuret
test. On the eighteenth day after sensitization the remaining
guinea-pigs (6) were reinjected and blood was taken five,
fifteen, thirty, forty-five, sixty, and ninety minutes after
the reinjection. The serum obtained from the samples
of blood was dialyzed and the dialysate subjected to the
biuret test. The samples taken five and fifteen minutes
after reinjection failed to yield dialysates which responded
to the biuret test, while the remaining four did. This
experiment demonstrates that at the very moment when
anaphylactic shock is being developed, peptone-like bodies
are being formed in the blood. It is probable that the
specific digestion of the protein begins at the very moment
that the reinjection is made, but that further time is neces-
sary for the digestive product to accumulate in a few cubic
centimeters of blood serum in recognizable amount. The
biuret test is not a highly delicate means of recognizing
proteins.
H. Pfeiffer and Jarisch1 have repeated diffusion experi-
ments. The method of procedure may be briefly described
as follows: Guinea-pigs are sensitized with horse serum,
and after varying intervals they are bled and serum obtained.
The serum of the sensitized animals is mixed with varying
amounts of horse serum, and the mixture incubated in
small dialyzers. If digestion takes place, peptone-like
products are formed, diffuse through the membrane, and
1 Zeitsch. f. Immunitatsforschung, 1912, xvi, 38.
366 PROTEIN POISONS
are detected in the dialysate by the biuret test. First, it
was shown that the serum of normal guinea-pigs with or
without mixture with horse serum does not supply biuret
bodies. After guinea-pigs had been sensitized to horse
serum for six days, then for the first time serum obtained
from them and mixed with horse serum and the mixture
dialysed, did the biuret test, when applied to the dialysate,
prove positive. This does not mean, as we take it, that
the formation of the specific ferment begins on the sixth
day after the injection of the sensitizer. It means that
with the amount of the sensitizer employed, the specific
ferment had accumulated sufficiently and was efficient
enough when brought into contact with horse serum in vitro
to digest it enough to show its action by the biuret test when
applied to the dialysate. The serum of sensitized animals
continues to digest the homologous protein (that to which
the sensitization is due) up to about the thirtieth day.
This does not mean that the formation of the ferment
ceases after this time. We know that this is not the case,
because Rosenau and Anderson have shown that guinea-
pigs sensitized to horse serum remain in this condition for
two years at least, and probably throughout life. This is
an important point and one which H. Pfeiffer nowhere
discusses, so far as we can find, although it has been brought
out by his work more prominently than by anyone else.
In order that it may be understood, we will try to state it
plainly. Guinea-pigs sensitized to horse serum furnish,
from about the sixth to about the thirtieth day, serum
which in vitro digests horse serum, as is shown by the forma-
tion of diffusible biuret bodies. After the thirtieth day or
thereabouts, the serum of the sensitized animal no longer
has this digestive action on horse serum in vitro; at least
such action is not demonstrable. And yet the guinea-pig
remains sensitive to horse serum. This, in our opinion, is
due to the fact that certain fixed cells in the animal body
remain sensitive and responsive to reinjection long after
the leukocytes lose their sensitization. Pfeiffer and Jarisch
found that in the so-called antianaphylatic state the blood-
THE PARENTERAL INTRODUCTION OF PROTEINS 367
serum of the guinea-pig does not digest the horse serum,
at least not to the extent of supplying the dialysate enough
of the digestive product to be detectable by the biuret
test, and that no response to this reaction can be obtained
until the third or fourth day after the reinjection. This is
not due to the absence of the specific ferment in the blood-
serum, but is due to the accumulation of the digestive
products, leading to an increase in the antitryptic titer
of the serum. It has been known for some time that the
addition of normal blood-serum to a mixture of casein and
trypsin prevents or arrests the digestive action of the latter
on the former. This phenomenon has been investigated
by Rosenthal,1 who concluded that the antitryptic action
of blood-serum is not due to the presence of antiferment.
His reasons for this conclusion may be stated as follows:
(1) It takes at least twenty-four hours to produce an anti-
ferment, and this effect of blood-serum on tryptic digestion
is immediate. (2) The antitryptic action of blood-serum
is not increased by ligature of the pancreatic duct, and it
should be if it were due to increased formation of anti-
trypsin. (3) The antitryptic constituent of blood-serum
is thermostabile and non-specific, and it would be thermo-
labile and specific were it an antiferment. (4) The anti-
tryptic action of blood-serum is increased in full digestion,
and in those diseases in which there is excessive protein
metabolism, and is decreased in hunger. Rosenthal con-
cluded that the antitryptic action of blood serum is due to
the presence of digestive products. When these are abun-
dant the digestive effect of blood serum is decreased or its
antitryptic property is increased. When the blood is
relatively poor in the products of protein metabolism the
digestive property of this fluid is increased or its antitryptic
property is decreased. This is a striking illustration of at
least one of the ways in which the parenteral digestion of
proteins is regulated, and it seems to us quite sufficient
to explain the phenomena of so-called antianaphylaxis.
1 Folia Serologica, 1910, vi, 285.
368 PROTEIN POISONS
Rusznjak1 has shown that when a sensitized animal recovers
from a reinjection or is in the so-called antianaphy lactic
state, its blood is laden with digestive products. He has
demonstrated this by showing that as early as thirty minutes
after the reinjection the antitryptic titer of the blood-serum
is greatly increased. In this way the animal body strives
to protect itself from the effects of unusual parenteral
digestion. Parenteral digestion is a normal process. It is
continuous and the protein poison in small amount is being
formed constantly in the body, and in part converted into
a harmless substance by further digestion, and in part
eliminated as such in the urine. After anaphylactic shock
it is found in the urine in unusual quantity. The regulation
of the formation and disposition of this poison is dependent
upon the fine adjustment between cell metabolism and the
digestive action of the blood. Parenteral digestion, as a
physiological process, is carried on by the non-specific
proteolytic ferments of the blood, and tissues. When a
substance easily acted upon by this non-specific proteo-
lytic ferment is suddenly thrown into the blood, life may be
endangered. This is apparently the case when a hemo-
lytic ferment is injected into the body in the form of a
foreign active serum, or a venom. The hemolysis, thus
caused, results in the liberation of a large amount of protein
substance which is readily split up by the normal, non-
specific proteolytic ferment, and the poison thus formed
may destroy life. However, if the dose be not overwhelm-
ingly great the digestive products retard the action of
the ferment and tend to conserve life. H. Pfeiffer and
Jarish attempt to distinguish between primary and second-
ary protein toxicoses. When the protein poison, preformed,
as in peptone, urinary residue or /3-i (ergamin) is injected
into an animal, the antitryptic titer of the serum is decreased.
On the other hand, when the protein is broken up in the
animal body, as in anaphylactic shock or hemolytic poison-
ing, the products of the cleavage increase the antitryptic
1 Deutsch. med. Woch., 1912, No. 4.
THE PARENTERAL INTRODUCTION OF PROTEINS 369
titer of the serum, or, in other words, lessen the digestive
action of the blood, and they propose to distinguish between
primary and secondary protein toxicoses by determining
the antitryptic titer of the blood serum.1 They state that
the curve of the antitryptic serum titer in retention uremia
after double nephrectomy in rabbits is similar to that of
anaphylactic shock and hemolytic poisoning. Before the
development of symptoms of poisoning the antitryptic
titer rises above the normal (owing to the accumulation
of digestion products).
It must be evident that the presence of a specific proteo-
lytic ferment is not necessary in all cases to split up proteins
with the liberation of the protein poison. As has been
stated, Friedberger and many others have found that the
protein poison is liberated from bacterial proteins by diges-
tion with the normal serum of guinea-pigs. In this there
can be no question of a specific ferment. The animals
supplying this serum have not been sensitized with bacterial
or any other proteins. They are normal, untreated animals;
besides, there is nothing specific in this reaction, since the
same poison is obtained from diverse bacterial proteins.
As we have held for years, every protein molecule contains
a poisonous group, and whenever and by whatever agent
the protein molecule is disrupted, the poisonous group
may be set free. The disrupting agent may be a chemical
substance, a specific or a non-specific ferment. Failure
to grasp this point has, in our opinion, led more than one
investigator into error. At one time Friedberger stated
that our poison cannot be the true anaphylactic poison
because its formation is not specific. If this be true of our
poison it is also true of his so-called anaphylatoxin. When
a protein is digested or split up there is one stage in the
process when the poisonous group is liberated. This may
not always be evident because when the cleavage is carried
one step farther the poison itself is destroyed. This is true
1 The details of this procedure are given in their paper, Zeitsch. f.
Immunitatsforschung, 1912, xvi, 38.
24
370 . PROTEIN POISONS
of our poison and of Friedberger's anaphylatoxin. As we
have stated more than once, parenteral digestion is a
normal, physiological process, and in this process the
protein poison is liberated. There are many proteins, not
all, which are digested by the normal, non-specific proteo-
lytic ferment of the blood and tissues. In this, in our
opinion, lies the explanation of the results obtained by
Szymanowski,1 who has found that the intravenous injection
of varied protein-precipitating agents, such as copper
nitrate, copper sulphate, mercuric chloride, lead acetate,
phosphomolybdic acid, tannin, and picric acid in small
doses, may cause all the symptoms of acute anaphylactic
shock and death. In our opinion, the most probable explan-
ation of this is that the precipitates formed in the blood by
these substances act like foreign proteins and are digested
by the non-specific proteolytic ferment of the blood with
the liberation of the protein poison.
There has been some difference of opinion as to the
source of the protein poison in anaphylaxis. At one time
HU Bfeiffieri thought that it must come from the proteins of
t&e-^body. kHe was led to this conclusion by the smallness
exfihthB)od®be of the anaphylactogen necessary to induce
shock on reinjection, but in his latest paper
the poison comes from the anaphylactogen
(antigafi)).fi ^or like reason Friedemann2 was inclined to
fchuliepiniointhat the poison is furnished by the serum of
the srarfdtiHed animal, but he seems now to think that it
the anaphylactogen (antigen). Wassermann
thought that the source of the poison is in
the fiermertt9(amboceptor). They shook horse serum with
kadlfn aondxthen separated the kaolin from the horse serum
arfJanceigsJfeuge and digested the kaolin with guinea-pig
iygmaniEiiirfid .Obtained a poison. They explained this by
feapposmg 9$E#t the kaolin absorbed the amboceptor from
fcheJ koEseCEeriim, and when this was acted upon by the
.1 i.2eitffi&. ,fo( Jhpmunitatsforschung, 1912, xvi, 1. * Ibid., ii, 591.
3 Folia Serologica, 1911, vii.
THE PARENTERAL INTRODUCTION OF PROTEINS 371
complement of the guinea-pig serum the poison was
formed. But Friedberger showed that kaolin does not
absorb amboceptor, but does absorb the protein of horse
serum, and this, when acted upon by the ferment in the
guinea-pig serum, furnishes the poison. Bauer1 thinks
that the digesting serum becomes poisonous from the loss
of its complement; thus, when bacteria are digested with
the normal serum of the guinea-pig the bacteria absorb
the complement from the serum and on account of this
loss the serum becomes poisonous. Bauer thinks that de-
complemented serum acts as an anaphylactic poison. This
claim is deserving of further study. The ease with which
Friedberger and his students prepare anaphylatoxin from
all kinds of bacteria by digesting them with normal guinea-
pig serum has caused some to suspect some flaw in the
experiment. Besredka and Strobel2 claimed that the
poisonous effect obtained is due to traces of peptone trans-
ferred from the medium on which the bacteria have been
grown, and that when bacteria grown on peptone-free
media were employed the results were negative. They also
found that when peptone-agar was digested with normal
guinea-pig serum the latter became poisonous. In answer
to this communication, Lura3 stated that bacteria grown
on peptone-free agar and those grown on potato furnish
the poison, when digested with serum, just as abundantly
as those grown on peptone-agar. The work of Lura has
been confirmed by others, and if there be a flaw in the
preparation of anaphylatoxin by digesting bacteria with
normal guinea-pig serum it has not been detected up to
the present time.
Pearce and Eisenbrey4 showed by the following experi-
ment that the specific ferment formed in sensitization is,
under certain conditions at least, a product of the fixed
cells: "Our procedure has been to exsanguinate, under
1 Berl. klin. Woch., 1912, 344
8 Compt. rend, de la Soc. biol., 1911, Ixxi.
3 Zeitsch. f. Immunitatsforschung, 1912, xii, 701.
4 Journal of Infectious Diseases, 1910, vii, 565.
372 PROTEIN POISONS
ether anesthesia, a small normal dog (A), and to transfuse
this animal by Crile's method, with the blood of a larger
sensitized dog (B), until the blood pressure reached approxi-
mately its original level. After sufficient blood has been
obtained from B to raise the pressure of A, the sensitized
dog is then bled to exsanguination and transfused from a
third normal dog (C) until its pressure reaches its previous
normal level. At the proper moment, the normal dog
containing the blood of the sensitized dog and the latter
containing the blood of the normal dog, each receives
intravenously the toxic dose of horse serum. In the former,
a fall in pressure does not occur, and in the latter it does,
thus proving that the phenomenon of anaphylaxis is due
to a reaction in the fixed cells, and not either primarily or
secondarily in the blood." That the blood does under
certain conditions at least contain the specific ferment is
shown by the production of passive anaphylaxis.
CHAPTER XIII
PROTEIN FEVER1
IT is interesting and instructive to read the older litera-
ture on fever in the light of the knowledge which has been
gained in the study of sensitization. It has long been
known that the parenteral introduction of proteins in small
amounts, and especially repeated introduction, leads to
fever. The older literature on this subject as well as an
account of his own work was given in 1883 by Roques.2
In 1888, Gamaleia3 showed quite clearly that fever accom-
panies and results from the parenteral digestion of bacterial
proteins, and a year later Charrin and Ruffer4 confirmed
this work and extended it to non-bacterial proteins. In
1890 Buchner5 produced the characteristic phenomena of
inflammation — calor, rubor, tumor, and dolor — by the
subcutaneous injection of diverse bacterial proteins. In
1895, Krehl and Matthes6 induced fever by the parenteral
introduction of albumoses and peptones, but they did not
obtain constant results, which we now know are secured
only by regulation of the size and frequency of the dosage.
In 1909, Vaughan, Wheeler, and Gidley7 demonstrated that
any desired form of fever (acute fatal, continued, inter-
mittent, or remittent) can be induced in animals by regu-
lating the size and frequency of the doses of foreign protein
administered parenterally, and in 1911, Vaughan, Gumming,
and Wright extended the details of this work.
1 This chapter is taken in part from an article by Vaughan, Gumming,
and Wright, Zeitsch. f. Immunitatsforschung, 1911, ix, 458.
2 Substances Thermogenes, Paris, 1883.
3 Ann. de 1'Institut Pasteur, xii, 229.
4 Compt. rend. Soc. de biol., 1889, 63.
o Berl. klin. Woch., 1890, 216.
6 Arch. f. exp. Path. u. Pharm., 1895, xxxv, 232.
7 Jour. Amer. Med. Assoc., August 21, 1909.
374 PROTEIN POISONS
When a man drinks water containing typhoid bacilli and
proves susceptible, he does not immediately manifest symp-
toms of this disease. There is a period of incubation which
in typhoid fever is about ten days. During this time the
bacilli are multiplying in the man's body in great numbers,
and are converting his proteins into bacterial proteins. The
period of incubation stops and that of the active disease
begins when the cells of the man's body become sensitized,
elaborate a specific proteolytic ferment, and with this begin
to split up the foreign protein.
The Production of Continued Fever in Rabbits by Repeated
Subcutaneous Injections of Dilutions of Egg-white. — If the
above be true it should be possible to cause a continued
fever in animals by repeated injections of a foreign protein
in small doses. This was first tried on rabbits with egg-
white, and the results are shown in Fig. 12.
This animal was kept under observation and its tempera-
ture taken for six days before the injections were begun.
The temperature of this fore period varied from 101.8° to
102.5° F. The injections consisted of egg-white with an
equal volume of 0.5 per cent, phenol solution, and were
freshly prepared each day. The injections were made
under the skin over the back and repeated at intervals of
two hours from 7 A.M. to 9 P.M. The urine was collected,
measured, its specific gravity taken with a picnometer,
and its nitrogen content determined by the Kjeldahl method.
The animal was weighed once a day. The first injection
of 2 c.c. of the egg-white dilution was made at 1 P.M., May
27, 1909. This dose was continued at the intervals stated
until 3 P.M., June 1, when it was doubled, and again doubled
at 7 A.M., June 4. The animal received 40 doses of 2 c.c.
each, 20 doses of 4 c.c. each, and 82 doses of 8 c.c. each of
the egg-white dilution; in all 816 c.c. Albumin appeared
in the urine when the dose was increased to 4 c.c. The last
dose was given at 9 A.M., June 15, after which the albumin
in the urine gradually diminished and wholly disappeared
June 26. The day before the first dose was given the animal
weighed 2525 grams, and on the day of the last dose it
375
weighed 2180 grams. After discontinuing the treatment,
June 15, the weight continued to decrease until June 21,
when it reached its lowest, 1850 grams. After this the
weight gradually increased until June 26, after which it
was not taken.
The following figures give the weight of the animal, the
amount of urine, the specific gravity, percentage of ash and
percentage of N in the urine :
Date,
Weight,
Urine,
Specific
Ash,
N.
1909.
gm.
c.c.
gravity.
per cent.
per cent.
' May 26
. 2525
May 27 .
. 2450
260
1.0106
1.61
0.16
May 28 .
. 2440
125
1.0124
2.12
0.15
May 29 .
. 2380
54
0.16
May 30 .
. 2330
275
1.0110
1.73
0.19
May 31 .
. 2270
266
1.0130
2.41
0.23
June 1 .
. 2355
104
1.0294
3.75
0.53
June 2 .
. 2335
370
1.0114
1.89
0.16
June 3 .
. 2320
290
1 . 0093
1.26
0.30
June 4 .
. 2290
155
1.0101
1.55
0.28
June 5 .
. 2290
225
1.0129
1.20
0.42
June 6 .
. 2250
95
1.0148
1.71
0.56
June 7 .
. 2320
100
1.0165
2.00
0.66
June 8 .
. 2350
52
1.0267
2.02
2.53
June 9 .
. 2355
115
1.0134
1.56
1.00
June 10 .
. 2345
155
1.0106
1.66
0.56
June 11 .
. 2230
165
1.0100
1.64
0.60
June 12 .
. 2330
95
1.0170
.1.56
0.92
June 13 .
. 2280
190
1.0140
1.81
0.54
June 14 .
. 2250
115
1.0148
1.15
0.98
June 15 .
. 2180
140
1.0179
2.60
0.65
June 16 .
. 2130
155
1.0194
2.99
0.66
June 17 .
. 2080
170
1.0139
2.76
0.50
June 18 .
. 2040
205
1.0184
2.10
0.43
June 19 .
. 1960
150
1.0198
2.80
0.64
June 20 .
. 1880
22
. .
2.90
June 21 .
. 1850
10
June 22 .
. 1885
55
June 23 .
. 1900
June 24 .
. 1980
June 25 .
. 2070
June 26 .
. 2120
It must be admitted that this chart bears a striking
resemblance to one of typhoid fever. Indeed, we have
here a condition practically identical with an infectious
376
PROTEIN POISONS
fever, and yet without infection. In the infectious diseases
the invading, multiplying cell supplies the foreign protein;
in this experiment the supply of foreign protein has been
kept up by the frequency of the injections. The amount of
urine is variable, but averages less than normal as it does
in the continued fever of man. The elimination of nitrogen
is increased. During the twenty-four hours immediately
following the first dose the temperature kept below the
normal average, and each time the dose was doubled a
marked rise followed in twenty-four hours. There is the
5-2O-0
1 £
-21
fr
•It
s
i:;
5
x
24
x
0-
.!.-,
ft.
21!
5
27
4
5
a
2M
.">
2'J
j
:.:>
5-
x
:n
(
-1
x
<;-
2
i
C,-
S
G-
4
<;-:•
< Q-
/^
.
I
\
/
,
r
S
^ J
\
__T
\
^r
1
"
\
r
v-
\_
^
^
E F
Eh
>ID
r
t
^
-\
^
^
-_/
V
-J
+FIR3T INJECTION OF SERIES
i-.-ld r.-l i G-12 6-13 6-14 6-15 6-1G 6-17 C-1S
C-20 C-21 6-22
=
a{
7
>
j
^
i
>
^
^
<
^
^
?
T
•s
I
•
*
J
^
^
S.
•
x
a
x.
X
x
*.
a.
x.
a a
r^
V
V
\
r^
•^
'VV,
A
-T
\
__ f
y
1 \
^
f"
^
A
^
'
~J-
' "
^A
^
N
yv
^/
^
-
k^-
101
*
+ LA3T INJECTION OF SERIES
FIG. 12. — The production of continued fever in a rabbit by repeated
subcutaneous injections of egg-white.
gradual rise in the morning temperature, so frequently seen
in typhoid fever and the fact that for the most part the
highest temperature for the day falls in the afternoon is
also interesting. When the injections were discontinued
the temperature gradually fell and remained somewhat
below the normal, as is often observed in convalesence
from typhoid fever.
We wished to ascertain the source of the albumin in the
urine; did it consist wholly of egg albumen or serum albumin,
or did it contain both? In order to determine this we
injected 2 c.c. of the filtered urine on June 18, intra-abdomi-
PROTEIN FEVER 377
nally, into each of 5 guinea-pigs for the purpose of sensitizing
them to whatever proteins the urine might contain. Twelve
days later 2 of these animals received intra-abdominally,
each 5 c.c. of egg-white dilution (with an equal volume of
water); 2 others had, each 5 c.c. of fresh serum from a
rabbit, and the fifth had a mixture of 2.5 c.c. of each of
these fluids. All were found to be sensitized, thus showing
the presence of both egg-white and serum protein in the
urine of the febrile rabbit.
It is worthy of note that while these animals developed
the three stages characteristic of protein sensitization, the
second and third stages were unusually prolonged and less
acute than those generally observed in sensitized guinea-
pigs. Of the two treated with egg-white, one died at the
end of two hours and the other fifteen minutes later. Of
the two treated with rabbit serum, one died at the end of
one hour, while the other lived for three hours. The one
that had the mixture of proteins developed the symptoms
more promptly than any of the others, but did not die.
A continued fever was maintained in another rabbit by
injections of the same strength of egg-white solution from
April 30 to May 18, 1909. In this instance the size of the
dose was not altered. The animal received four doses daily
from April 30 to May 11, after which five were given until
May 15, and then for three days we returned to four doses
daily. The fever continued, after the injections were
discontinued, until the evening of May 20, when it fell by
crisis below the normal, slowly returning to the normal.
The urine was collected and nitrogen determined as in the
other instance, but the charts are so similar that we do
not consider it necessary to present the second one.
The Production of Continued Fever in Rabbits by Repeated
Subcutaneous Injections of the Poisonous Group of the Typhoid
Protein. — Fig. 13 shows the effects of repeated subcutaneous
injections of sublethal doses of the poisonous group split off
from the cellular substance of the typhoid bacillus with a 2
per cent, solution of sodium hydroxide in absolute alcohol.
The material used was the crude soluble poison containing
378
PROTEIN POISONS
about 10 per ceixt. of the poison in the purest form in which
we have been able to obtain it. This crude soluble poison
was administered every two hours from 7 A.M. to 9 P.M
from May 3 to May 18, 1909. Each dose consisted of
200 mg. of the crude poison, 300 mg. being a fatal quantity
for rabbits of the size used.
When a fatal dose of the protein poison is administered
the temperature rapidly falls, but with smaller repeated
doses a continued fever results. The more nearly the dose
approaches the fatal amount the more speedily will the
5-3.09 .'.-I 5-5 .-,-;! 5-7 5-8 .,.'.) 5-19 5-11 5-12 5-13 5-14
X
z
Z
z
s
2
T
i
z
z
£
2
z
2
z
z
z
?
z
z
2
I
2 Z
A
A
,A
a
u
i
~l ,
.
A
z
p
\J
V
V
/V
N /
-s/
•Vs
\;
k/^
\
j
31—
t
v
V,
\J
1(1".
I
*
V
+ FIRST INJECTION OF SERIES.
5-15 5-10 :.-17 5-18 5-19 5-20 5-21 5-22 5-23 5-24 5-25 5-20
X Z
m
^
5
a
^
a
•
2
Z
•
2
£
2
2
£
Z
2
3:
2
2
X X
\l*
s/
/\
A
^
7 L
SA
N
\r
^-v
VL
_— •*-'
^
^
,>
<\
-^
+ LAST INJECTION OF 8FRIE8.
FIG. 13. — The production of continued fever in a rabbit by repeated
subcutaneous injections of the poisonous group from the typhoid bacillus.
animal succumb. Death may be sudden and under a high
temperature, or it may be slow and preceded by a fall of
several degrees below the normal. From sublethal doses
of the poison, animals recover quickly and apparently,
completely. This seems to indicate that the poisonous
effects are quickly neutralized in the animal body, but we
are of the opinion that this neutralization is secured by
more or less chemical disintegration in the protein molecule
of certain cells in the body. The protein poison is acid to
litmus, and we have secured continued fever in rabbits by
repeated injections of either the acid solution or the same
PROTEIN FEVER 379
after neutralization with sodium bicarbonate. Fig. 13
needs no further explanation.
The Effects of Intra-abdominal Injections of Egg-white. —
Large single or repeated doses of egg-white injected intra-
abdominally in non-sensitized rabbits have but little effect
on the temperature. Generally the temperature runs slightly
subnormal after such injections.
August 22, 1909, we injected the whites of three eggs
into the abdominal cavity of a rabbit. The highest tem-
perature of the fore period was 100.9°. After the injection
the temperature was taken every two hours from 8 A.M.
to 6 P.M. up to September 6. The animal was weighed
each day and its urine measured and tested for albumin.
There was no fever; indeed, the morning temperature
fell some days to 97° and one day to 96.6°. The animal lost
in weight, slightly more than one-fifth of its original weight.
The volume of urine averaged normal, and at no time did
it contain albumin.
On the other hand, 0.05 c.c. of egg-white, filtered through
cotton, injected intra-abdominally every half hour from
8 A.M. until 4 P.M. produced the following results:
Time. Dose in c.c. Temperature.
8.00 A.M. 0.05 102.8°
8.30 0.05 101.0°
9.00 0.05 102.6°
.9.30 0.05 102.9°
10.00 0.05 102.6°
10.30 0.05 103.2°
11.00 0.05 103.4°
11.30 0.05 103.5°
12.00 0.05 103.4°
12.30 P.M. 0.05 104.4°
1.00 0.05 104.8°
1.30 0.05 105.1°
2.00 0.05 105.3°
2.30 0.05 105.3°
3.00 0.05 105.8°
3.30 0.05 105.8°
4.00 0.05 105.8°
4.30 0 106.6°
5.00 0 106.6°
5.30 0 106.1°
8.00 0 104.1°
8.00 A.M. 0 103.0°
380 PROTEIN POISONS
The Production of Fever in Rabbits by Repeated Intra-
venous Injections of Egg-white. — The injection of a large
amount of egg-white intravenously in a single or in repeated
doses in non-sensitized rabbits does not cause any marked
elevation of temperature.
After keeping a rabbit under observation for three days
and finding that its temperature at no time reached 102°,
we injected into its ear vein every two hours from 8 A.M. to
6 P.M. 4 c.c. of a dilution of egg-white with an equal volume
of physiological salt solution, which dilution had been
passed through a Berkefeld filter, and each cubic centimeter
of which contained 26 mg. of protein as ascertained by a
Kjeldahl determination. This dosage was continued for
six days. During the greater part of this time the tempera-
ture, which was taken before each injection, remained
normal, sometimes subnormal, and only once did it reach
102°. Then the dose was increased to 10 c.c. and con-
tinued for four days. Twenty-four hours after the increase
in dose there was an irregular, but not marked elevation
of temperature, the highest point reached being 104.4°.
During the whole of the time the animal seemed quite well.
Its greatest weight was observed during the time when the
largest injections were being given, and at the same time
the daily elimination of urine greatly increased, from 114
c.c., the average of the fore period, to as much as 650 c.c.
at the time of the largest injections. The urine was tested
daily for albumin with negative results. This experiment was
continued from August 6 to 19, 1909. In March, 1910,
a single injection of 5 c.c. of the dilution of egg-white was
followed by a gradual rise in temperature to 105.8° within
a few hours.
In order to induce fever in rabbits by the intravenous
injection of dilutions of egg-white the doses must be small
and the most striking results are obtained when the size
of the dose is gradually increased. We have made many
experiments along this line and some of them will be
detailed.
PROTEIN FEVER
381
Group /.—The initial dose in this group was 1 c.c. of
egg-white diluted with an equal volume of physiological
salt solution and passed through a Berkefeld filter. Each
cubic centimeter of this dilution contained 26 nag. of protein.
The doses were increased by 1 c.c. at each repetition, which
was hourly.
In rabbit No. 2 the highest temperature of the fore period
was 101.4°. The following table shows the results:
Time.
8.00
9.00
10.00
11.00
12.00
3.00
5.00
Dose in c.c.
1
2
3
4
5
6
7
8
9
0
0
Temperature.
100.8°
103.2°
104.3°
104.6°
105.4°
105.6°
106.1°
105.2°
102.2°
100.5°
Death
This animal died with a sudden convulsive movement.
The urine of the day and that in the distended bladder after
death was tested for albumin with negative results. Fig. 14
106
2 F
M.
4 P.M.
6 P.M.
1 CC.
9 A.M.
t CC.
10 A.M.
3 CC.
11 A.M.
4 CC.
I CC.
1 P.M.
6CC.
t P.M.
7 CC.
3 :>.M.
8 CC.
4 P.M.
9 CC.
5 P
M.
^J^-
\
105
10-1
1C3
102
101
11)0
/
.
\
t
/
f
/
\
/
\
/
\
/
/
\
/
\
S
~—~
X
/
\
"^
\
h-
XFIRST INJECTION CF SERIES
-{-DEATH
FIG 14. — Acute fatal fever produced in a rabbit by repeated
intravenous injections of egg-white.
382 PROTEIN POISONS
shows the temperature curve of this animal, including the
long fore period.
In No. 5 the highest temperature of the fore period was
101.1°.
Time. Dose in c.c. Temperature.
8.00 A.M. 0 100.1°
9.00 1 100.4°
10.00 2 101.3°
11.00 3 102.2°
12.00 4 102.7°
1.00 PM. 5 102.8°
2.00 6 103.5°
3.00 7 104.7°
4.00 8 105.1°
5.00 9 105.6°
6.00 10 104.0°
7.00 11 103.8°
8.00 12 104.0°
8.30 0 Death
This animal showed no symptoms until the final con-
vulsive movement.
The following is the record of No. 34.
Time.
Dose in c.c.
Temperature.
9.00 A.M.
1
102.8°
10.00
2
104.0°
11.00
3
105.4°
12.00
4
105.2°
1.00 PM.
5
105.4°
2.00
6
105.4°
3.00
7
106.0°
4.00
8
105.4°
5.00
9
105.3°
6.00
10
105.8°
By
10 P.M. the temperature had fallen to 102.2°.
The
record of No.
35 is shown by
the following:
Time
Dose in c c.
Temperature.
9.00 A.M.
1
102.8°
10.00
2
103.6°
11.00
3
105.4°
12.00
4
105.6°
1.00 P.M
5
106.2°
2.00
6
106.7°
3.00
7
106.6°
4.00
8
106.8°
5.00
9
106.6°
PROTEIN FEVER
383
By 10 P.M. the temperature had fallen to 103.4° and the
next morning the animal was apparently normal. Twenty-
six days later this animal was treated in the same way,
with fatal results. Fig. 15 shows the curve for both
treatments.
1
107
100
105
104
103
102
O-20-O9 10-27-09
10-28-O9
I f
M.
1 CO.
2
C.
3 C
c.
4 CC.
5 CC.
6 CC.
7 1
C.
1
C.
91
C.
to
M.
,c.
/
, — -
~~-
-— _
--— -
. 1
/
/
\
^.
X
\
'
/
/
/
r
ORE
PER
OC
/
^
/
x
\
/
N
/
FIG. 15. — Acute fever produced in a rabbit by intravenous injections
of egg-white. The continuous line represents the temperature in the
non-sensitized animal. The broken line represents the temperature in
the same animal, sensitized.
Group II. — The dilution used contained 13 mg. of protein
in each cubic centimeter.
384 PROTEIN POISONS
The following is the record of No. 15:
Time. Dose in c.c. Temperature.
9.00 A.M. 1 102.6°
10.00 2 104.0°
11.00 3 104.9°
12.00 4 105.8°
1.00 P.M. 5 105.1°
2.00 6 105.6°
3.00 7 105.6°
4.00 8 106.2°
5.00 9 105.6°
6.00 10 104.2°
The temperature gradually fell, and the next morning
at 10 it was 101.8°.
The following is the record of No. 16:
Temperature.
103.5°
104.1°
103.4°
104.2°
104.5°
104.5°
103.8°
103.4°
103.2°
104.2°
The next morning the temperature was 101.8°.
Group III. — Each cubic centimeter of the dilution
contained 6.5 mg. of protein.
The following is the record of No. 18:
Time. Dose in c.c. Temperature.
8.00 A.M. 1 103.2°
9.00 2 102.0°
10.00 3 102.5°
11.00 4 103.8°
12.00 5 104.0°
1.00 P.M. 6 104.6°
2.00 7 104.0°
3.00 8 103.6°
4.00 9 104.0°
5.00 10 103.8°
6.00 11 103.8°
Time.
Dose in c.c.
9.00 A.M.
1
10.00
2
11.00
3
12.00
4
1.00 P.M.
5
2.00
6
3.00
7
4.00
8
5.00
9
6.00
10
PROTEIN FEVER
385
The temperature was normal the next morning.
After an interval of 174 days this animal was treated in
the same way, with a fatal ending. Fig. 16 shows the curve
for both treatments.
10-11-O9 10-12-09
M
M.
-,.
Cr-
-*
-
-«
-^
10-1
X
^
x
\
^
---J
X
,oc
/
^--
•— «,
^-
<^
^
/
-^,
.-~~
\
/
.in-.;
\
^
^
+FIRST INJECTION
N—
FIG. 16. — Acute fever produced in a rabbit by the intravenous injection
of egg-white. The continuous line represents the temperature in the non-
sensitized animal. The broken line represents the temperature in the
same animal, sensitized.
Group IV. — Each cubic centimeter of the dilution con-
tained 3.25 mg. of protein.
The following is the record of No. 20:
Time.
9.00
10.00
11.00
12.00
1.00
2.00
3.00
4.00
5.00
Dose in c.c.
1
2
3
4
5
6
7
8
9
Temperature.
103.0°
103.0°
103.4°
103.8°
104.0°
103.8°
104.2°
104.4°
104.4°
25
386 PROTEIN POISONS
The following is the record of No. 21 :
Time. Dose in c.c. Temperature.
9.00 A.M. 1 103.0°
10.00 2 101.8°
11.00 3 102.4°
12.00 4 102.8°
1.00 P.M. 5 103.2°
2.00 6 104.0°
3.00 7 103.6°
4.00 8 103.8°
5.00 9 104.0°
After an interval of 140 days this animal was again
treated with the following record:
Time. Dose in c.c. Temperature.
9.00 A.M. 1 102.0°
10.00 2 103.4°
11.00 3 105.2°
12.00 4 105.8°
1.00 P.M. 5 106.6°
2.00 6 106.4°
3.00 7 105.6°
4.00 8 106.2°
5.00 9 105.6°
6.00 10 106.4°
7.00 0 106.2°
8.00 0 105.6°
9.00 0 106.4°
10.00 0 103.4°
After the second injection Cheyne-Stokes respiration
appeared and continued through the day, but the injections
were continued until 6 P.M., and the animal recovered.
Group V, — In this group but one injection was made
each day. The dilution contained 26 mg. of protein in
each cubic centimeter. The beginning dose was 3 c.c.
and each day it was increased by 2 c.c. In No. 6 the dose
was given each day at 10 A.M., and was repeated for seven
consecutive days. The beginning dose was 3 c.c., two of
the larger doses being repeated. In this way a well-marked
intermittent fever of mild type was established. After
PROTEIN FEVER 387
each injection the temperature arose within from two to
four hours, and returned to normal during the evening,
and continued so until the next injection. With increase
in the size of the dose the tendency was to take the remittent
type, so that at no time of the day did the fall quite reach
the normal limit.
Group VI. — We have found some rabbits which do not
respond to the pyrogenic effect of intravenous injections
of egg-white until the size of the dose is reduced. On
receiving a new consignment of rabbits, we were surprised
to find that these animals did not develop fever on receiving
repeated doses of a dilution of egg-white (1 to 1) with salt
solution. The following illustrates our experience:
No. 101. This animal, weighing about 2600 grams, was
treated with increasing doses of the dilution (1 to 1):
Time. Dose in c.c. Temperature.
9.00 A.M. 1 102.0°
10.00 2 100.4°
11.00 3 100.5°
12.00 4 100.2°
1.00 P.M. 5 100.0°
2.00 6 96.4°
3 . 00 0 Death
The abdominal cavity was filled with bloody fluid. The
urine found in the bladder contained a small amount of
albumin.
No. 102. Thinking that No. 101 was an individual
exception, No. 102 was treated with the same dilution:
Time. Dose in c.c. Temperature.
10.00 A.M. 1 101.2°
11.00 2 101.4°
12.00 3 101.5°
1.00 P.M. 4 101.5°
2.00 5 101.8°
3.00 6 101.3°
4.00 7 101.5°
4.45 0 Death
388 PROTEIN POISONS
No. 103. In this experiment the egg-white dilution was
reduced (1 to 2) :
Time. Dose in c.c. Temperature.
10.00 A.M. 1 101.4°
11.00 2 101.7°
12.00 3 102.0°
1.00 P.M. 4 102.5°
2.00 5 102.8°
3.00 6 102.4°
4.00 7 101.8°
5.00 0 97.6°
6.00 0 97.4°
6 . 30 0 Death
No. 104. In this experiment the dilution was further
reduced (1 to 3).
Time. Dose in c.c. Temperature.
7.30 A.M. 1 101.6°
8.30 2 101.2°
9.30 3 101.7°
10.30 4 101.8°
11.30 5 102.6°
12.30 P.M. 6 103.4°
1.30 7 103.8°
2.30 8 104.1°
3.30 9 104.4°
4.30 0 104.5°
5.30 0 105.2°
These differences in response to the injections of egg-
white seem to be characteristic of certain groups or batches
of animals and not of individuals within the group. Our
laboratory buys its rabbits in lots of from twenty-five to
one hundred. If one out of a given lot responds to a certain
dose of egg-white we have found that others of the same
lot respond in much the same way. Whether this is due
to differences in food or in breed we have not determined,
but are inclined to believe that breed has much to do with
it. Possibly, age is an individual factor.
A Brief Statement of the Autopsy Findings in Acute Poison-
ing of Rabbits with Egg-white. — This work has been turned
over to our colleagues of the department of pathology, and
PROTEIN FEVER 389
Morse has been kind enough to make a few autopsies for
us. We make a short abstract of his report:
Rabbit A received intravenously four doses of 10 c.c.
each of a dilution of egg-white with an equal volume of
physiological salt solution. The doses were administered
at intervals of one hour. During the administration of the
fourth dose the animal died in convulsions. A gray rabbit
of average size and well-nourished; external orifices are
normal; mucous membranes cyanotic; body, cold; rigor
mortis, moderate. The peritoneal cavity contains a small
amount of blood-tinged, serous fluid. Superficial inspection
reveals nothing else of note. There is no displacement of
organs, no peritonitis, no area of hemorrhage in the serosa.
The pericardial sac is normal. There is no increase of
pericardial fluid and the pleural cavities are dry. In the
anterior mediastinum there- is a moderate amount of pale
fat with a few petechial hemorrhages. The thymus is
large, swollen, and edematous. It spreads over the anterior
mediastinum covering the great vessels and it contains
hundreds of miliary hemorrhages. There are no large
areas of blood in the tissue. The heart is moderately
dilated and filled with red clot. There is no imbibition
of hemoglobin in the intima of the great vessels. The
heart valves are normal, the myocardium is darker than
normal, and drips blood too freely. The lungs are reddish
pink, though slightly darker than normal and rather moist
on section. There is no pneumonia and no solid areas are
seen in the lung tissue. The spleen is slightly congested
and darker than normal. The kidneys are dark, congested,
and drip blood on section. The adrenals are apparently
normal. The stomach and intestine show no abnormality.
The bladder is empty and normal in appearance. The liver
is large, dark, and bleeds freely on section. In the retro-
peritoneum there is a moderately large suffusion of blood
through the cellular tissue and partially involving the head
of the pancreas. The brain appears normal, but section
shows the tissue somewhat congested and moist.
The chief microscopic findings may be stated as follows:
390 PROTEIN POISONS
The myocardium shows slight increase in hemofuscin,
and there seems to be an excessive fragmentation of the
muscle bundles (myocardite segmentaire of Renault).
However, this may be due to the fixing fluid. The most
striking thing is the occurrence of numerous miliary hemor-
rhages into the muscle substance, forcing the fibers apart
in places. There is a diffuse distribution of blood throughout
the heart muscle, blood cells being found here and there
outside the capillaries in the muscles, forcing the fibers
apart as though there had been a general diapedesis. The
ventricular cavity shows a homogeneous red clot. The
lungs show marked acute, passive congestion and localized
areas of moderate congestion. There are a few small
hemorrhages near the veins along the bronchi and also
beneath the pleura. The liver shows extreme passive
congestion, all the capillaries being gorged with blood.
The whole liver substance appears as though soaked in
blood, which lies everywhere, between the liver cells and
in the bile capillaries. Some of the smaller ducts contain
blood. The liver cells have a cloudy appearance and the
nuclei are farther apart than normal, due to the engorgement
with blood. The kidneys show some cloudy swelling and
marked acute passive congestion. A condition similar to
that seen in the heart and liver is found throughout the
kidney. The renal tissue is full of miliary hemorrhages,
and appears to be soaked in blood. Everywhere between
the tubules and scattered throughout the parenchyma are
red blood cells. Many of the glomeruli have red blood
cells lying free within Bowman's capsules, and the capillaries
of the tufts are extremely dilated. The epithelium of the
proximal convoluted portion of the tubules is markedly
desquamated. Many of the collecting tubules contain
pale blood cells which have lost their hemoglobin. The
pelvic fat is in part displaced by large suggillations of blood,
and there are a few areas of coagulated blood around the
kidney capsule. The pancreas is passively congested and a
portion of this organ has been included in a large clot in the
retroperitoneal region and is wholly necrotic. There is also
marked fat necrosis and infiltration of the tissue with blood.
PROTEIN FEVER 391
In sensitized rabbits killed by injections repeated after
six months, Morse has found the microscopic lesions of the
same character, but much less marked than those described
above as resulting from acute poisoning. In fresh rabbits
hemolysis and hemorrhage seem sufficient to account for
death, but this does not appear to be the case in sensitized
animals dying suddenly from relatively small doses.
The Effects of Intravenous Injections of Laked Human Red
Corpuscles on the Temperature of Rabbits. — The blood was
drawn into a solution of sodium citrate, and the corpuscles
thrown down in a centrifuge. The corpuscles were repeatedly
washed with physiological salt solution, and then dissolved in
distilled water and diluted to the volume of the original blood.
One dose of 5 c.c. of this solution was injected into the
ear vein of rabbit No. 7. The highest temperature of the
fore period was 101.8°. The effects of this injection are
shown by the following figures:
Time. Dose in c.c. Temperature.
8.00 A.M. 0 101.8°
9.00 5 101.5°
11.00 0 104.2°
12.00 0 105.2°
1.00 P.M. 0 106.0°
2.00 0 105.6°
3.00 0 105.6°
4.00 0 104.7°
5.00 0 104.2°
6.00 0 104.0°
8.00 A.M. 0 101.8°
In No. 10, 2 c.c. of the same solution had the effect
showrn in the following:
Time. Dose in c.c. Temperature.
8.00 A.M. 0 103.0°
10.00 2. 103.1°
11.00 0 104.4°
12.00 0 105.6°
1.00 P.M. 0 107.0°
2.00 0 106.4°
3.00 0 106.2°
4.00 0 105.6°
6.00 0 104.6°
8.00 A.M. 0 102.8°
12.00 0 102.6°
392
PROTEIN POISONS
It will be observed that this animal had a temperature
of 103.1° before the injection was made. Fig. 17 gives this
record.
In No. 11, 1 c.c. of the same solution had the following
effect :
Time. Dose in c.c. Temperature.
8.00 A.M. 0 102.3°
9.00 1 102.0°
10.00 0 104.2°
11.00 0 104.4°
12.00 0 105. 4~
1.00 P.M. 0 104.0°
2.00 0 103.0°
3.00 0 102.2°
4.00 0 102.6°
9-294)9 !i.:;ii.iiit 10-1-09 1(1-24)9 10-3-0
4PM.
(.A
M
l!
M
4t
M.
M
M.
M
M
10*
11A
.V
"
w.
7
^k
tf
M.
3t
M.
M.
V
M.
SA
.M.
'•
"'
4C
H.
CA
M.
:i
M. 4P
t
'
" — .
' .
/
^
\
/
\
/
s
/
\
/
s
?
— _
\
X
— — '
_±
\
1 ^
X
1
^^
X
^^~~
+DOSE 2 C.C.
FIG. 17. — Acute fever produced in a rabbit by an intravenous injection
of washed human blood cells hemolyzed.
Fig. 18 gives the curve in this case.
100
lU'.
H4
m
100
101
HO
9-33-
*. i
S ?
99 9-24 9-25 9-26 9-27 9-28 9-29 9-80 . 10-t 10-3
1
\
/
\
^
\
1
\
1
\
/
s
\
X
\
s
/
\
/
\
/
\
/
x
\x
/
/
\
/
s
s
^
/•
/
/
\
/
/
\
/
\
/
\
/
\
+ DOSE 1 C.C.
FIG. 18. — Acute fever produced in a rabbit by an intravenous
injection of washed human blood cells, hemolyzed.
PROTEIN FEVER
393
In No. 12, 0.5 c.c. of the same solution produced the
following effects:
Time.
8.00 A.
10.00
11.00
12.00
1.00 P.
2.00
3.00
4.00
Dose in c.c.
0
0.5
0
0
0
0
0
0
Temperature.
102.4°
102.0°
103.0°
104.0°
105.2°
104.4°
103.4°
102.4°
The highest . temperature in the fore period covering
seven days was 102.6°.
In No. 17, 10 c.c. of the same solution caused a precipitate
fall in temperature, and death in seven hours.
The laked blood corpuscles of either man or rabbit after
filtration cause an elevation of temperature when injected
into rabbits either intra-abdominally or intravenously.
In rabbit No. 55, rabbits' corpuscles prepared as already
stated and filtered were injected intra-abdominally as
shown by the following figures:
Time.
8.00
9.00
10.00
11.00
12.00
.00
.00
3.00
4.00
5.00
Dose in c.c.
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0
Temperature.
102.8°
102.2°
102.8°
102.8°
104.3°
105.2°
105.1°
104.8°
104.4°
104.6°
394
PROTEIN POISONS
In No. 56 the unfiltered laked corpuscles were injected
intra-abdominally :
Time.
8.00
A.M.
8.30
9.00
9.30
10.00
10.30
11.00
11.30
12.00
12.30
P.M.
1.00
1.30
2.00
2.30
3.00
3.30
4.00
4.30
5.00
5.30
6.00
6.30
7.00
7.30
8.00
8.30
9.00
10.00
10.00
A.M.
2.00
P.M.
4.00
6.00
8.00
A.M.
9.00
Dose in c.c.
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
0.55
0.60
0.65
0.70
0.75
0.80
0.85
0.90
0.95
.00
.05
.10
.15
.20
.25
.30
.35
0
0
0
0
0
0
0
Temperature.
102.0°
101.8°
101.6°
101.8°
101.8°
101.8°
101.8°
102,4°
102.4°
103.0°
103.2°
M)3.2°
103.8°
104.6°
105.0°
105.2°
105.4°
105.4°
105.4°
105.6°
105.4°
105.7°
104.0°
105.0°
105.2°
105.2°
105.0°
105.6°
105.2°
105.0°
103.6°
104.6°
103.4°
103.6°
The Destination of Egg-white Introduced into the Circulating
Blood of the Rabbit. — We have shown (p. 357) that egg-white
injected into the ear vein of a rabbit soon disappears from
the circulating blood and diffuses through the various
tissues, from which it may be extracted with physiological
salt solution and its presence demonstrated by sensitizing
guinea-pigs. Since reporting on this we have found that
the distribution of egg-white, injected into the blood,
through the tissues extends not only to the organs mentioned
PROTEIN FEVER 395
in the article referred to, but also to the skin and walls of
the alimentary canal.
We have also attempted to determine how long after
injection egg-white can be detected in the tissues.
Three rabbits received intravenously 25 c.c. of egg-white
dilution (1 to 1). One was killed twenty-four hours later,
and sections of its skin, kidney, brain, liver, spleen, and
intestinal and stomach walls rubbed up with salt solution.
After standing in the cold room overnight these emulsions
were filtered and the filtrates injected intra-abdominally
into guinea-pigs. The second rabbit was killed after forty-
eight hours, and the third after seventy-two hours, and their
tissues treated in the same way. All the guinea-pigs that
received extracts from the first and second rabbits were
found to be sensitized, though none died. Choking symp-
toms were very marked, most pronounced in those that
received extracts from the spleen and kidney. The symptoms
were quite as marked in those that received the extracts
from the second rabbit as in those treated with the extracts
from the first. The pigs that received the extracts from the
third rabbit showed absolutely no symptoms.
From these experiments we conclude that egg-white
diffused through the tissues after injection into the blood
becomes, sometime between two and three days, either
so far changed as to loose its identity or so fixed in the
tissue that it cannot be washed out with salt solution. This
time interval probably varies with the kind and amount
of foreign protein introduced, and in different species of
animals.
The Digestive Action of the Blood Serum of Rabbits in Which
Fever Has Been Induced with Egg-white. — The following
illustrates some of our experiments on the digestive action
of the blood serum: The temperature of two rabbits was
raised to 106° by hourly intravenous doses of a dilution
of egg-white (1 to 1). One hour after the last injection
both of these animals were bled to death from the jugular
vein and the serum obtained.
Two cubic centimeters of this fever serum without any
396 PROTEIN POISONS
addition, after standing for twenty-four hours in the incu-
bator, was diluted to 10 c.c. with normal salt solution and
deprived of normal proteins by acetic acid and heat. The
filtrate gave a slight biuret test but no Millon. At 11.15,
2.5 c.c. of the filtrate was injected intracardiacally into a
guinea-pig. Temperature before the injection was 98.8°.
At 11.30, 97.9° and at 11.40, 98.4°. The animal was not
visibly disturbed, with the exception of slight tremor.
A second sample of 2 c.c. of this serum which had been
mixed with 2 c.c. of milk and kept in the incubator for
twenty-four hours was treated in the same way. The biuret
was slight and the Millon negative. The guinea-pig was
not disturbed nor the temperature lowered.
A third portion of the serum mixed with an equal volume
of a 2 per cent, solution of Witte's peptone was tested in
the same way. The filtrate gave a beautiful biuret, but
no Millon. The temperature of the pig fell 2.6° in ten
minutes, but otherwise the animal was not affected.
A fourth portion of the serum mixed with an equal
volume of a dilution of egg-white was tested in the same
way. The filtrate gave a splendid biuret and also a good
Millon. The pig received only 1.25 c.c. of the filtrate, half
the quantity given to the others, but it immediately
developed the symptoms characteristic of the protein
poison and died within five minutes. Postmortem examina-
tion showed no injury and the heart-apex still beating.
We took a mixture of 2 c.c. of the fever serum and 10 c.c.
of the egg-white dilution (1 to 1), which had stood in the
incubator for five days. This was diluted to 20 c.c. and
heated, after being made distinctly acid with acetic acid.
After the removal of the normal blood proteins the filtrate
gave both the biuret and the Millon tests very distinctly.
Five cubic centimeters of the filtrate was evaporated on
the water-bath and the yellowish residue extracted with
20 c.c. of absolute alcohol. The portion insoluble in alcohol
was extracted with 5 c.c. of salt solution. The part soluble
in salt solution responded feebly to both the biuret and the
Millon tests. The part insoluble in both alcohol and salt
PROTEIN FEVER 397
solution, when suspended in water, did not give the biuret,
but did give an intense Millon reaction, while the part
soluble in alcohol gave neither. A duplication of this
experiment gave identical results.
We are not ready to conclude from our work, which has
been more extensive than detailed here, that the ferment
of the fever serum is strictly specific. An exhaustive meas-
urement of its specificity will take much time and close
work. The digestive products probably vary much, in
amount at least, with conditions, and a close study of
parenteral digestion offers a promising field for research.
We are inclined to think that the rabbit is a good animal
in which to study parenteral digestion, and we suspect
that this form of digestion is not altogether abnormal in
this animal. We have already shown (p. 355) that egg-white
introduced into the stomach or rectum of a rabbit is, in
part at least, absorbed unchanged into the blood. Besides,
we have observed that our laboratory rabbits when abun-
dantly supplied with food often show a rectal temperature
of 103° or over, and when the food supply is limited the
temperature is lower and more constant.
The Production of Acute Fever, Followed by Immunity, by
Repeated Intra-abdominal Injections of Bacterial Suspensions.
In this group of experiments guinea-pigs have been used.
The usual method has been to take a standard loop from
an agar slant four days old, suspend this in 10 c.c. of normal
salt solution, and with a beginning dose of 0.1 c.c. of this
suspension, the dose is increased by 0.1 c.c. each time and is
repeated every half hour. As a basis for these experiments
two guinea-pigs were treated, in the manner described, with
the salt solution alone. The result in one of these is shown
in Fig. 19. The total range in this case covers 2.7°, and it
is possible that these injections stimulate parenteral diges-
tion slightly. The general agreement of the curve in kind
with those to be presented later is quite as interesting,
though not so striking as its difference from them in quantity.
It is. worthy of note that during the continuance of these
injections the temperature did not fall below the initial.
398
PROTEIN POISONS
Figs. 20, 21, 22, and 23 are illustrations of the results
obtained by this mode of treating guinea-pigs with sus-
TIME
s .
FIG. 19. — Showing effect of repeated injections of physiological salt
solution.
"1:00
2:00
3:00
4:00/
^5:00
6:00
7:00
8:00
9:00
10:00
/
^
\
/
\
/
\
/
\
/
\
/
— \
\
.
\
/
^
\
/
V
\
\
/
\
/
\
\
/
\
/
FIQ. 20. — Showing effect of repeated injections of bacillus subtilis
FIG. 21. — Showing effect of repeated injections of bacillus prodigiosus.
PROTEIN FEVER
399
pensions of living bacteria. All the animals whose tem-
peratures are shown in these curves recovered. Additional
information concerning these experiments is given in tables
XLII to XLV. The first one or two animals of each
group were treated with a beef-tea culture of the bacillus
a 10:00
11:00
12:00
Jl:00
2:00
3:00
4:00
5:oo
6:00
7:00
/
\
/
\
/
\
"\
/
\
\
/
\
x
t
^
/
\
.
^
\
/
\
1
\
/
\
/
N
1
\
/
FIG. 22. — Showing effect of repeated injections of bacillus cholerse.
".11:00
12:00
».i :oo
2:00^
AJ:oo
4:oo
6:00
6:00
7:00
8:00
9:00
10:00
11:00
/
\
/^
ss
1
s
j
\
7
\
\
/
\
/
/\
/\
\
/
\,
,/
\
7
\
/
V^
/
\
/
/
\
/
\
s /
FIG. 23. — Showing effect of repeated injections of bacillus typhosus.
twenty-four hours old. After this, we changed to the
suspension in salt solution and the exact dilution used in
each animal is shown in the tables. The average bacterial
content of each loop is also given in the tables, and the
first dose contained one one-hundredth of this number,
in case the dilution was 10 c.c.
400
PROTEIN POISONS
i
-Q T3 T3
?' u
3 33
Second injection.
•
g 2 §
.2 .2.S
o o a
.a M >>
O OH
Q QQ
i-i 33
S SS
<N CO CO
«> $ •
*i
<N 00 00
Result cf series.
Died after last injection
Recovered
Died 12 hrs. after series
Recovered
"o.S
. O
O 0)
fc'3
•»J<t-Tj<<Ol-Hl-ll-HO
<Ni-H<N<Ni-Hi-i<N<N
•a
Rd
OOCOOINOOlNt^
8"3
«2~
OOCO'*'-<O5Tj<CT>
1
COCDOSOOOiOO
2
rt^H<N'<l<'-iO<NO
•g-
OOt^^TfiOINOO
.S a
(JC,S-1
OOO^HOOOO
"3ft
OOtOOO-<l<00-<lilN
•3 a
tS-s
O3O5O5--IOOOO
O5O3C35OOOOO
cjd------
•ength of
pension.
5 6-
•S22S2222
o
•°.s
-*j —
CO g
5 §
T(t ~"
C^i— 1>— li— (i-Hi— 1 i-l i-l
6
K
iOOOO^H(N-*<C
^H ^ (N CO C<5 C«5 CO
I-H i-H i-H i-H
1
K
•O 73
II
TO
3
S
•43
0) 33
O *i
^ CO
a"o
0
|
HO
T5
a
o
qq
(N <N
Days
interval
oo oo
)0 bacteru
BTILIS
1
B
a
o
•43
o
:l
0 P
uf »
1
"8
1
o ®
» 5
CO j
1
•a h •&
• JK w
1 1
O T3 O " " " "
u 5 o
o ,_;
o B
O, *"H.
O I-]
2 M
O &
• "o
O V
fc'fl
1C t^ CD "^ CO r* i-l
IN rt (N <N iM (N (N
a H
TJ
CD O CO CD t- i-i O5
§ 5?
£ H
§•2
CO
CO CD <N IN ^H Tj< Tj<
CO
i
•«}< t* 00 O t^ 00 CD
H
a
CM ^H i-H (N O rH ^H
2=3
00 O t~ •* 0 "5 00
s-
O CN O O O O O
•3 ft
a> •* oo •* CD CM •*
11
00 O5 O O 00 CM I-H
O5 O5 O O O5 O O
I-H ^H 1-H I-l
o3 c5 - - -
"o ej
J3.2
^ 0 "
a §
£§•
asg
j. a: : : :
-= o
N CM i-H i-H i-H i-l i-l
6
%
•* CD 01 10 a> >o co
i-H i-H CO CO
i— I i-H
PROTEIN FEVER
401
nj
H
if
«
« a
31
e <u
g a
T3 T3-OT3T3
<D 0>Q^OO
> >>>>
3 3333
Q Q Q Q Q Q Q Q
3 3333
T3
"o
A
a,
>>
C
g T3
"
W
B
-OOT3O
m O m O
i, O •—
• SO) .i, O •— i)
Q tf Q pj Q tf
26
402
PROTEIN POISONS
H
1
i
o
i
w
5
x
ia
J
cq
1*
§2
2s
p pp P PP
CO CN CN CO •* »O
pp
CO t» CO CO »H I-H l>
So
-
ser
2 2:
OS OmO SuSo
•SI : B >a S a; •-. o -S oj -S o
Q tfP tfPtf PtfPtf
•iT -g ' jj
CO CN <N [S £
CO i-H i-H £ . 03
T3 -O O " -o
flJ ^3 ^ w fl)
(3 3" « P
Oi co co co cf ^ Tt* ^ ^^ ^^ *~^ f^ i^ ^^
T3
rt
X CO 00 IO O
CD fH TjH T-H 1C CO
00 T}< 10 SN ^H O
rH iH CNCOCOCNCO ^
3
o o ^ co co
O O O 1-1 i-H CN
00 t> CN CO 00 -
OCOOO5'-iO3XOOOOOX^HO5O O5 O O OS -H O
Ok 9 Q Oft O Ofr Cv Ov C& O O) O O) O ~ OOO5OO
•^ -^ 1
d
•3
*3
p]
OJ C----, ------ O O. . . ,
o o o o o o o
.3 2 .2 -< .9
Q<;;;:::::;::'-|D<mO>::::
o m o o o
o £ o -o o
—^ Q ^H ^-< j-
C^Cqr-li— I^H.— li— I^H^H^HrHr-H^H.— I'D— • — • — •
i-Hi-iCNCNCN-*-*-*-*-*iO
CO >-H CM CO Tfl IO
H
PROTEIN FEVER 403
It will be seen that in the large majority of the animals
the first effect is a slight fall in temperature. We designate
this as the primary or short fall. It should be understood
that we estimate the falls and rises from the initial tempera-
ture. The short fall is followed by the primary rise. In a
few instances the primary fall does not occur, or what is
more probable, has not been detected. In one animal
(No. 136, Table XLII) neither primary fall nor primary rise
was detected. The rise is followed by the secondary or
long drop. We call attention to some of the great drops;
for instance, Nos. 131 and 134, Table XLII; No. 140, Table
XLIV; and Nos. 46, 141, and 142, Table XLV. We did not
suspect that this treatment would give immunity, and for
this reason in our earlier work the recovered animals were
not tested, nor have we as yet determined the limits of the
immunity secured by these treatments. As is shown by the
tables, the immunity is not, qualitatively at least, specific.
Animals treated with subtilis or prodigiosus bear, at least
from two to three M. L. D.'s of cholera or typhoid bacilli,
and the immunity secured by the latter is interchangeable.
When treated animals are inoculated with living cultures
they become sick some hours before the controls, and we
are reminded of the immunity induced some years ago
in this laboratory with the haptophor, or non-poisonous
groups of the colon and typhoid bacilli, and reported by
V. C. Vaughan, Jr. (p. 144), by Vaughan and Wheeler
(p. 157), and by Vaughan.1
The Production of Fever by Repeated Injections of Vegetable
Proteins. — As was shown in this laboratory some years
ago, the vegetable proteins contain the same poisonous
group found in bacterial and animal proteins; conse-
quently there seemed no reason why these should not
induce fever, and such we have found to be true. We will
give here only one illustration. We extracted 10 grams of
oat-meal with 100 c.c. of normal salt solution, and with a
beginning dose of 0.1 c.c. of the slightly opalescent fluid
1 Zeitsch. f. Immunitatsforschung, 1909, i, 263.
404 PROTEIN POISONS
thus obtained we have induced fevers similar to those
already described in this chapter.
We have made many experiments on the production of
fever with non-protein bodies, giving special attention to
the amino-acids, the xanthin group, inorganic salts of
ammonia, and certain carbohydrates, but a report upon
these findings must be postponed.
General Conclusions. — Protein fever, and this includes
the great majority of clinical fevers, results from the paren-
teral digestion of proteins. Bouillaud1 was practically
right when he said: "La fievre est une maladie, dont la
nature est toujours la meme." Proteins, living and dead,
occasionally find their way into the body. They may come
from without or from within. Crushed bone, muscle, or
other tissue, on being deprived of its vitality or detached
from its normal surroundings, becomes foreign material
and must be broken up preparatory to its elimination.
Under certain conditions proteins taken into the alimentary
canal escape enteral digestion and are in part absorbed
unbroken. When this happens, they are disposed of by
parenteral digestion. In a finely divided form, as in the
pollen of plants, proteins are absorbed from the respiratory
tract and give rise to the condition designated as hay or
rose fever. But in the great majority of instances proteins
gain entrance to the body in unbroken form, as living
proteins, bacteria, or protozoa. The parenteral proteo-
lytic ferments are of two kinds, non-specific and specific.
The former are normally present in the blood and tissues,
especially in the former, of all animals. They differ in
kind in different species and in amount and efficiency in
individuals. Their purpose is to break up foreign proteins
that find their way into the blood and tissues. They are,
within limits, general proteolytic ferments, as are those
of the alimentary canal; though the variety of proteins
upon which they can act is more limited. They constitute
the most important factor in racial and individual immunity.
1 Trait6 clinique et experimentale des Fifevries, 1826.
PROTEIN FEVER 405
Man is immune to most bacteria, not because they do not
elaborate poisons, for every protein molecule contains its
poisonous group, but because they are destroyed by the
general proteolytic enzymes as soon as they enter the
tissue and consequently are not permitted to multiply
in man's body. These non-specific, parenteral proteolytic
enzymes are probably secretions of certain specialized
cells. Under natural conditions these enzymes are capable
of digesting those proteins upon which they do act only in
small amounts, but the cells which elaborate them may
be stimulated to increased activity by proper treatment,
and the method detailed in this paper seems to accomplish
this purpose. Whether or not these enzymes become
qualitatively specific under such treatment as we have
detailed can be determined only by further study. The
immunity secured by these enzymes is limited in extent
and transitory in duration.
The specific, parenteral proteolytic ferments are not
normal products of the body cells, but are brought into
existence under the stimulation of those proteins, intro-
duced into the blood and tissues, which on account of their
nature or amount escape the action of the non-specific
ferments. It is to the development of these ferments that
the phenomena of sensitization (wrongly called anaphyl-
axis) are due. A protein introduced into the blood and
not promptly and fully digested by the non-specific enzymes
is discharged from the blood current and deposited in some
tissue, the cells of which after a time develop a specific
ferment which splits up this protein and is not capable of
acting upon any other. For certain proteins there are
certain predilection organs and tissues in which they are
stored, either exclusively or most abundantly: the pneu-
mococcus in the lungs; the typhoid bacillus in the mesen-
teric and other glands; the viruses of the exanthematous
diseases in the skin, etc. For the development of the specific
proteolytic ferments time is required, and this varies with
the protein and probably with the tissue in which it is
deposited. The development of these ferments necessitates
406 PROTEIN POISONS
changes in the chemical constitution of the protein mole-
cules of the cell, and by this means the cell acquires a new
function, which subsequently is brought into operation
only by contact with that protein to which its existence is
due. As a result of this rearrangement in molecular struc-
ture, the .cell stores up a specific zymogen which is activated
by contact with its specific protein. This explanation of
the phenomena of sensitization originated in this laboratory,1
and was not simply a fortunate guess, as has been assumed
by some. The same is true of the statement made at the
same time, that protein sensitization and bactericidal
immunity are identical, and not antipodal, as they may
appear to the superficial observer. A close study of the
split products of bacterial, vegetable, and animal proteins,
and especially of the poisonous group found in all proteins,
had already been made in this laboratory. A study of the
symptoms induced by the protein poison and of those
following a second administration to the sensitized animal
was certainly good and sufficient ground for concluding, as
then stated, that the man that dies from the administra-
tion of morphine and the one that dies from opium both
owe their death to the same poison. The most valuable
experiments of Friedberger and his assistants, and of
Pfeiffer and Mita, have, in our opinion, fairly established
the validity of this explanation.2
Whether the products of digestion with the non-specific
ferments and those elaborated by the specific enzymes are
identical or not remains to be ascertained. The presence
of a poisonous group in the protein molecule is disclosed in
both enteral and parenteral digestion, as well as by our
process of splitting up the protein with dilute alkali in
absolute alcohol. In the first case it appears in the peptone
1 Jour, of Infect. Dis., June, 1907.
* In 1910 Friedberger (Berl. klin. Woch., Nos. 32 and 42) made very
plain the relation between sensitization and the infectious diseases, and
in his address at the meeting of the German naturalists at Konigsberg in
September, 1910 (Munch, med. Woch., 1910, Nos. 50 and 51), he dwelt
most instructively upon the wide application of the facts learned in his
studies of sensitization.
PROTEIN FEVER 407
molecule, which is large and complex. By the chemical
process it is obtained as a less complex, more diffusible, and
consequently more active body. Between the two there
are probably several intermediate substances. The prompt-
ness in action manifested by all proteolytic ferments is
determined, in part at least, by the proportion between the
surface of the substrate and the mass. We observed some
years ago1 that the more finely divided the cellular sub-
stance of bacteria is, the smaller the dose which proves
fatal. This is due to the greater surface exposure, and the
same apparently holds good for colloids in solution. When
soluble proteins are expelled from the blood and diffused
throughout the animal body, the conditions for their rapid
cleavage are most favorable, and consequently the fulmina-
ting phenomena observed after the second injection into
a sensitized animal.
When a protein deposited in mass is rapidly acted upon
by the parenteral enzymes, more or less marked inflamma-
tion results. This may be demonstrated by injecting sus-
pended, dead, bacterial cellular substance into the peritoneal
cavity of a guinea-pig when a diffuse peritonitis results,
and we wish to suggest that the exanthems are due to the
rapid digestion of proteins deposited in the skin. We admit
that this is largely theoretical, but we have found, as already
stated, that egg-white is in part deposited in the skin of
rabbits after intravenous injection. This may be an
explanation of the Arthus phenomenon.
The fact that every protein molecule contains a poisonous
group does not mean that the products of protein digestion
must contain a poison, for the poison itself may be split up
and rendered inert, as happens when the proteins in the
alimentary canal are broken up into amino-acids. It may
therefore happen that in certain forms or stages of parenteral
digestion no poison is formed.
The low temperature seen in some of our charts
undoubtedly indicates the liberation of the poisonous
1 Trans. Assoc. Amer. Phys., 1902.
408 PROTEIN POISONS
group, and consequently the subnormal as well as the high
temperature is a result of parenteral digestion, and it is
in this stage that the greater danger to the life of the animal
lies, as is plainly shown in our results. However, there is
danger to life in the high temperature in and of itself. A
rabbit is not likely to survive a temperature above 107°,
and this was reached in at least one of our experiments, and
closely approached in many others.
Fever must be regarded as a conservative process, although
like many of nature's processes it often leads to disaster.
But its purpose is the disposal of foreign and dangerous
material, and therefore must be regarded as beneficent.
In parenteral digestion the following sources of heat
production must be evident: (1) The unaccustomed
stimulation and consequent increased activity of the cells
which supply the enzymes must be the source of no incon-
siderable increase in heat production. (2) The cleavage of
the foreign protein means the liberation of heat. (3) The
reaction between the products of the digestion and the
tissues, especially when an active and irritant poison is
liberated, must lead to increased heat production. We
regard the first and last of these as the more important
sources of the overproduction of heat in the febrile state.
Special Conclusions. — 1. Large doses of unbroken protein
administered intra-abdominally, subcutaneously, or intra-
venously have no effect upon the temperature; at least,
do not cause fever.
2. Small doses, especially when repeated, cause fever,
the forms of which may be varied at will by changing the
size and the interval of dosage.
3. The effect of protein injections on the temperature is
more prompt and marked in sensitized than in fresh animals.
4. The intravenous injection of laked blood corpuscles
from either man or the rabbit causes in the latter even in
very small quantity, either in single or repeated doses,
prompt and marked elevation of temperature.
5. Laked corpuscles after removal of the stroma by
filtration have a like effect.
PROTEIN FEVER 409
6. Protein fever can be continued for weeks by repeated
injections, giving a curve which cannot be distinguished
from that of typhoid fever.
7. Protein fever is accompanied by increased nitrogen
elimination and gradual wasting.
8. Protein fever covers practically all cases of clinical
fever.
9. Animals killed by experimentally induced fever may
die at the height of the fever, but, as a rule, the temperature
rapidly falls before death.
10. Fever induced by repeated injections of bacterial
proteins and ending in recovery is followed by immunity.
11. The serum of animals in which protein fever has
been induced digests the homologous protein in vitro.
12. Fever results from the parenteral digestion of proteins.
13. There are two kinds of parenteral proteolytic enzymes,
one specific and the other non-specific.
14. The production of the non-specific ferment is easily
and quickly stimulated.
15. The development of the specific ferment requires a
longer time.
16. Sensitization and lytic immunity are different mani-
festations of the same process.
17. Foreign proteins, living or dead, formed or in solu-
tion, when introduced into the blood soon diffuse through
the tissues and sensitize the cells. Different proteins have
predilection places in which they are deposited and where they
are, in large part at least, digested, thus giving rise to the
characteristic symptoms and lesions of the different diseases.
18. The subnormal temperature which may occur in
the course of a fever or at its termination is due to the
rapid liberation of the protein poison, which in small doses
causes an elevation, and in larger doses a depression of
temperature.
19. Fever per se must be regarded as a beneficient phe-
nomenon, inasmuch as it results from a process inaugurated
by the body cells for the purpose of ridding the body of
foreign substances.
410 PROTEIN POISONS
20. The evident sources of excessive heat production in
fever are the following: (a) That arising from the unusual
activity of the cells supplying the enzyme; (6) that arising
from the cleavage of the foreign protein; (c) that arising
from the destructive reaction between the split products,
from the foreign protein and the proteins of the body.
In 1910 Friedberger1 studied the effects of graduated
doses of foreign proteins on the temperature of both normal
and sensitized animals. With lambs' serum intravenously
administered to normal guinea-pigs he obtained the following
results :
5.0 c.c. equal fatal dose.
0.5 c.c. equals limit for fall in temperature.
0.01 c.c. equals upper constant.
0.005 c.c. equals fever plane.
0.001 c.c. equals lower constant.
In sensitized guinea-pigs the above figures were changed
to the following:
0.005 c.c. equals fatal dose.
0.0005 c.c. equals limit for fall.
0.00001 c.c. equals upper constant.
0.000005 c.c. equals limit for fever.
0.000001 c.c. equals lower constant.
In 1911 Schittenhelm, Weichardt, and Hartmann2
experimented upon the effect of the parenteral administra-
tion of diverse proteins on animal temperature and came
to the following conclusion, which in our opinion is well
stated : " In severe experimental anaphylaxis there is a
fall in temperature; in the lighter manifestations there is
fever." We regard this as a confirmation of our conclusion
reached some years earlier. "Small, especially repeated,
doses of the protein poison cause fever, while large doses
depress the temperature."
Some years ago Friedmann and Isaak3 showed that after
the parenteral introduction of foreign proteins the increase
in nitrogen elimination is greater than can be accounted
1 Berl. klin. Woch., 1910, No. 42.
2 Zeitsch. f. exp. Path. u. Ther., 1911. 3 Ibid., 1905, 1906, and 1908.
PROTEIN FEVER 411
for by the protein injected. This has been confirmed by
the work of Schittenhelm and Weichardt,1 and as has been
stated, we found the same in protein fever. Our explanation
for the marked increase in nitrogen elimination has been given.
In intermittent and remittent fevers and in relapses in
all infectious diseases the phenomena of protein sensitiza-
tion are fully demonstrated. In the different forms of
malaria, chill and fever correspond to the discharge of
foreign protein into the blood, just as promptly as anaphyl-
actic symptoms follow the injection of the homologous
protein in a sensitized animal. The moment the blood
cells rupture and the protozoal protein is disseminated
the sensitized cells discharge the lytic ferment by which
the foreign protein is disrupted and destroyed, but in this
process the poison is liberated.
Local sensitization is frequently established in the mucous
membrane of the air passages and of the alimentary canal,
also in the skin for two reasons. In the first place, foreign
proteins are frequently brought into direct contact with
these tissues, and in the second place, foreign proteins intro-
duced into the blood are frequently deposited in the skin
and in the walls of the alimentary canal. These local sensi-
tizations characterize many of the infectious diseases. The
work of Dunbar and Weichardt on hay fever is a good
illustration. These investigators injected each other sub-
cutaneously with minute quantities of pollen suspension.
Immediately Dunbar, being a hay-fever subject, became
dizzy, and within a few minutes began to sneeze, then a
whooping-like cough began. The eyes were congested, and an
abundant secretion flowed from the nose. The face became
swollen and cyanotic, and soon the body was covered
with an urticarial rash. After twenty-four hours these
symptoms subsided. Weichardt, not being a hay-fever
subject, was not affected. That this and kindred affections
are not benefited by antisera was abundantly and positively
demonstrated by the failure of the so-called hay-fever
1 Zentralbl. f. d. ges Physiol. u. Pathol. d. Stoffwechsel, 1910.
412 PROTEIN POISONS
serum, which was found in no instance to be of special
value, and in some it greatly intensified the symptoms.
Our common colds are instances of local sensitization.
Schittenhelm and Weichardt tell of a man who was so
deeply sensitized by the inhalation of Witte's peptone that
he could tell on entering the laboratory whether the peptone
flask was open or closed, and some moist peptone painted
on the skin caused the area covered to become red. The
high degree of susceptibility to odors from the horse shown
by some people has already been referred to. It seems in
some instances that this susceptibility is transmitted from
mother to child.
A volume might be filled with citations of cases of food
and medicine idiosyncrasies. That these are, in large part
at least, instances of protein sensitization has been demon-
strated by rendering animals susceptible to the same food
or medicine by injecting them with the serum of the sus-
ceptible individual. In other words, passive anaphylaxis
has been established in the animal. In this way Briick has
sensitized animals to iodoform and antipyrin with the sera
of persons especially susceptible to these agents.
Thiele and Embleton1 have made an extensive study
of temperature variations. They employed well-fed guinea-
pigs. First, they tried non-protein bodies, and with these
reached the following conclusions:
1. Sodium chloride varies in its effects according to
its degree of concentration when injected into feeding
animals :
As normal saline it produces only a slight rise.
As 2 to 2.5 per cent, saline it produces a marked rise.
As 3 per cent, saline it produces a fall.
As 5 per cent, saline it produces a rise.
2. Calcium salts intraperitoneally produce a fall.
3. Ringer's fluid has no effect on temperature.
4. Alkalies, very dilute, produce a rise, and stronger,
produce a fall.
1 Zeitsch. f. Immunitatsforschung, 1913, xvi, 178.
PROTEIN FEVER 413
5. Acid, same effects.
6. Lecithin injected intraperitoneally as a water emulsion
has no effect on the temperature.
7. Charcoal (animal) very fine powder; in suspension
intraperitoneally in large doses produces a marked fall, and
in small doses a slight rise.
They notice, confirmatory of the findings of Krehl and
Matthes, that fasting animals do not easily get febrile
reactions with these substances. This agrees with Hirsch
and Roily, who state that glycogen-free animals cannot
become febrile with sodium chloride injections. Kulz
showed that injections of sodium chloride cause glycosuria.
This was confirmed by Fischer, who believed the glycosuria
to be due to irritation of the central nervous system, and
showed that it is abolished by cutting the splanchnics, and
that marked glycosuria results when direct injection into
the cerebral vessels is made. Calcium salts inhibit both the
glycosuria and the fever induced by sodium chloride. Freund
showed that adrenalin causes glycosuria, and in large doses
a fall, and in small doses, a rise in temperature. Again,
calcium salts inhibit both the glycosuria and pyrexia due
to adrenalin. The conclusion is reached that the fever
caused by these non-protein substances is due to sympathetic
irritation and consequent increased glycogen metabolism.
Turning now to the work done by Thiele and Embleton
with protein bodies, we find a complete confirmation of the
results obtained by previous investigators. They say:
"With regard to endotoxin proper, by which we mean a
toxic substance, which is specific to the bacteria or protein,
there is but little evidence. The virulent and non-virulent
bacteria, as well as simple egg albumen, appear to have
the same temperature effects when inoculated into healthy
animals in the way mentioned above, and also performed
by Vaughan and Wheeler. The so-called endotoxin has
been liberated from protein, from bacteria, etc., by Vaughan
and Wheeler and others, also by Schittenhelm and Weichardt.
In all these experiments the protein, whether simple or
bacterial, has been treated for some time with caustic
414 PROTEIN POISONS
alkali either in watery solution or in alcohol, so that a
certain amount of protein degradation has occurred which
is apparently just sufficient to form the toxic bodies which
cause the acute toxic symptoms or temperature variations
under discussion, according to the size of the dose. The
work of Friedberger, Abderhalden, and others with ana-
phylatoxin production in vitro, and the demonstration of
the presence of proteolytic degradation bodies during the
process of the formation of toxic substance, is in favor of
the view that the toxin (poison) is purely a degradation
product of the protein simple or bacterial, as again there
is no specificity in the toxins (poisons) produced from the
various antigens (sensitizers) used. The reason why bacteria
have a much more potent action in such small quantities
would appear to be in the chemical composition of their
bodies, and in the presence of normal specific ferments to
them. Thus, according to Schittenhelm and Weichardt,
and our own observations from the protamine injections,
etc., it appears that some proteins of the normal animal
body are much more toxic than others, and the toxic ones
are those which have a high percentage of the diamino
bases. In the experiments brought forward here the toxicity
of protamine as regards producing temperature variations
is almost as great as that of the tubercle bacillus. This
is important in view of Ruppel's work showing that the
tubercle bacillus has a large amount of diamino bases.
Further, the formation of toxic substances in the case of
bacteria suspended in normal saline would appear to be
due to the formation of cleavage bodies from autolytic
changes, just as occurs when tissues undergo autolysis. A
final argument that cleavage products of protein and bac-
teria are the causation of temperature reactions, etc., is
the observation of Matthes, who showed that in a digesting
tuberculous animal, albumose injection gave rise to hyperemia
of the small intestine and around the tuberculous foci just
as tuberculin injection does, as we have noted in our
present experiments. Here we have a protein cleavage
product giving rise to the same effects as the specific antigen.
Hence it would appear that the cleavage products of the
PROTEIN FEVER 415
antigen are the cause of the reaction, cleavage going on
continually locally at the tuberculous foci, and the addition
of a little more cleavage body producing a cumulative
effect, and in the intestine where cleavage is also going
on in the cells during digestion, the same occurs whether
the reacting dose is from simple or bacterial protein."
These investigators agree with us that it is the same
substance which in larger doses causes a fall in temperature,
and in small doses a rise. They say: "(1) In sensitized
animals, owing to the presence of a specific enzyme, the
homologous antigen undergoes more rapid degradation
than in non-sensitized ones, and consequently certain
degradation products are liberated in sufficient quantities
from relatively small amounts of the antigen to cause
temperature depression in these animals, and from still
further amounts to cause fever. (2) The pyrogenic bodies
are not a further stage in the degradation of the antigen, but
the same degradation body or bodies cause depression or ele-
vation according to the amounts present at any given time."
The relative effects of egg-white and tubercle protein on
fresh and sensitized animals are shown by Thiele and
Embleton as follows:
EGG-WHITE
Normal animal, Sensitized,
Limits of grams. grams.
Temperature fall .... 0.05 0.005
Constant temperature . . 0 . 02 0 . 0002 to 0 . 0001
Temperature rise . . . . 0.01 to 0.001 0.0001 to 0.000002
TUBERCLE EMULSION
Normal animal, Sensitized,
Limits of grams. grams.
Temperature fall . . . . 0.005 to 0.002 0.0005
Constant temperature . . 0.002 to 0.001 0.0001
Temperature rise .... 0.001 to 0.00001 0.00001 to 0.000001
In another interesting way these investigators confirm
some of our earliest work (Chapter III) when they say:
"The more finely divided the bacterial protoplasm is, the
more rapid are the temperature effects and the more toxic
is the bacillary substance."
CHAPTER XIV
SPECIFIC FERMENTS OF THE CANCER CELL1
THE chief distinction between a living protein and any
more or less complex chemical group or dead cell is that
the living protein contains within itself different ferments
or enzymes. Many of these ferments are common to several
different forms of living protein or at least bear a close
resemblance to each other, but it is also probable that each
and every form of living cell contains within itself a specific
enzyme which is distinctive for that given form of protein,
and is present in no other.
While as yet ferments have not been isolated in a chemic-
ally pure state, and we have not an exact knowledge of
their chemical nature, yet they are commonly regarded as
albuminous, and their functions are specific and fairly
well known. This at least is true of the soluble ferments
such as the amylases, proteoses, and lipases. Again, there
are examples of ferments of similar or nearly like consti-
tution which are formed by more than one variety of cell
(thus ptyalin formed by the salivary glands and diastase
formed by the pancreas are both starch-splitting ferments).
The above examples are all of ferments which are water
soluble, and which are active outside of the cell that pro-
duces them. They are ferments which are excreted by the
cell and are used by the more highly developed forms of
animal life for the purpose of converting complex chemical
bodies into more simple forms, so that they may be of
service to the whole, through their reaction with other
ferments contained in other body cells.
1 This chapter is contributed by J. Walter Vaughan and the work has
been done by the aid of the Chase Cancer Fund in the research laboratory
of Harper Hospital, Detroit, Michigan
SPECIFIC FERMENTS OF THE CANCER CELL 417
It is well known, however, that all unicellular forms of life
contain ferments which it is safe to divide into two classes.
These we would designate as soluble and insoluble, or extra-
and intracellular ferments; a soluble ferment being one
that is excreted by the cell and which performs its ferment
action without the cell, and an insoluble ferment being an
enzyme that acts only within the cell structure. Obviously,
with our rather limited comprehension of the nature of
ferments, our knowledge is confined chiefly to the soluble
variety. In this class of ferments formed by unicellular
organisms might be mentioned zymase formed by the
yeast cell, and the putrefactive ferments formed by many
bacteria. V. C. Vaughan would also classify such sub-
stances as diphtheria toxin and tetanus toxin in this group,
since it is his belief that these substances are not poisons
in themselves, but liberate poisons through their ferment
action.
While the known functions of most enzymes are those of
decomposition, either through such processes as hydrolyza-
tion or oxidation, yet we also know that ferment action
may be one of construction; and both processes may be
carried on by one and the same ferment. Thus we know that
ethyl butyrate may be converted by the action of lipase
into alcohol and butyric acid, while with a change in the
acting masses, alcohol and butyric acid plus lipase will
form the ester.
Inasmuch as we do know of the presence of constructive
ferments belonging to the soluble variety, in which class
we would naturally suppose the greater number to be
destructive, since their function is to reduce complex
proteins into simpler forms for the easier assimilation by
the cell, it seems reasonable to suppose that most intra-
cellular or insoluble ferments are of the constructive variety,
and that it is through their aid that simple chemical sub-
stances are converted into the complex distinctive proteins
of the cell itself.
While much of the above is theoretical, yet there is
sufficient knowledge concerning some of the facts to make
27
418 PROTEIN POISONS
the theoretical portion logical, and we have previously
advanced the following theory as to the etiology of malig-
nant disease with the foregoing in view.
Every living cell has within itself a constructive ferment
whose specific action is to construct proteins of the same
specific composition as the cell itself. Through the aid of
this ferment a sufficient amount of cell protein is formed,
so that one cell may form two daughter cells and these in
turn may again do likewise. As an example of this may be
cited the growing of the typhoid bacillus upon an agar slant.
In the original few organisms, which are spread upon the
media, is contained a ferment which is capable of constructing
the specific typhoid protein as long as sufficient required
chemical substances are present in the agar medium. In
the case of a unicellular organism like this the property of
cell division and multiplication must be inherent in the
cell, and the process of cellular reproduction progresses as
long as suitable media are furnished. When, however, we
come to consider more complex organisms, where the
functions and relations of one cell are dependant upon
outside cells, we have a somewhat different and more com-
plex problem to consider. Such cells cannot continue to
multiply without limit, else the cell that reproduced the
fastest would soon predominate and outstrip all others.
Here the growth of cells must be governed by outside
circumstances to some extent, although it would appear
reasonable to suppose that the property of reproduction is
inherent in the cell as well as in the case of the simple
unicellular organisms. This property, however, must lie
dormant at times, only to be aroused by outside stimuli.
A more lucid and concise way of expressing this idea would
be to state that the reproductive ferment was normally
stored up within the cell as a zymogen or inactive ferment,
which becomes active when called upon by outside stimuli.
Even in the unicellular organisms it is a well-known fact
that cell division can be hastened or retarded by outside
conditions, such as heat and cold or various chemical or
electrical stimuli. With regard to the bearing of the fore-
SPECIFIC FERMENTS OF THE CANCER CELL 419
going upon the cancer cell, we would state that it seems
possible to conceive that a cancer cell is one which has lost
its power of forcing its reproductive ferment back into an
inactive stage. It is a cell whose chemical nature has become
so altered that the reproductive ferment is uppermost and
can no longer be influenced by outside stimuli. Just like
the typhoid bacillus in the test-tube, its sole purpose is now
one of active reproduction, and the reproduction will
persist until the cancer cell has used Up all protein that it
is capable of transferring into its own specific protein, or
until it itself is destroyed through the formation by the
body cells of an antiferment in sufficient amount to destroy
all cancer cells. How often the body itself does this it is
difficult to surmise, but we do know that once such cells
have multiplied so as to form a palpable tumor, the body
is very seldom able to cope with it.
With the foregoing in mind the following experiments
will throw some light upon the subject, especially with
regard to the formation of an antiferment for the cancer cell.
As has been brought out in a previous article,1 certain
definite and characteristic blood changes are brought about
by the injection of small amounts of dead cancer protein
into a living host. Two forms of vaccine have been used
in this work, cancer residue and a vaccine made of the
cancer cell in its entirety.
Cancer residue is prepared by dissecting as freely as
possible the cancer material from all surrounding tissues.
The cancer material is next ground in a meat grinder, then
is washed with water, diluted salt solution, alcohol, and
lastly ether. This process removes salts, fats, wax, several
protein bodies, and traces of carbohydrates. The remaining
substance is then heated in a flask with a reflux condenser,
with from fifteen to twenty times its weight of a 2 per cent,
solution of sodium hydroxide in absolute alcohol, and by
this means it is split into a toxic and a non-toxic group.
The toxic portion is soluble in the alcohol, the non-toxic
1 Jour. Amer. Med. Assoc., November 16, 1912.
420 PROTEIN POISONS
is insoluble in alcohol, but soluble in water. This portion
is dissolved in normal saline solution in sufficient quantities
to form a 1 per cent, solution, and is used in the same
manner as a vaccine.
Cancer-celled vaccine is prepared by grinding cancer tissue
finely in a meat grinder, after which it is rubbed up as a
suspension in alcohol in a sterile mortar. Next, it is rubbed
through a very fine-meshed sieve, the alcohol filtered off,
and the collected cell substance air-dried. This is weighed
and then placed in normal saline solution, making a 2 per
cent, cell suspension. To this 0.5 per cent, phenol is added
for the purpose of rendering it less likely to become
contaminated. In the preparation of both residues and
vaccines it has been found that satisfactory blood changes
can be obtained only when the tumor is of firm consistency
and without necrotic or infected areas. The average injec-
tion of a 1 per cent, residue is from 5 to 20 minims; that of
a 2 per cent, cancer-cell emulsion is from 5 to 10 minims.
Sheep and rabbits have been injected intravenously,
intra-abdominally, and subcutaneously with both cancer
residue and cancer vaccine, and frequent blood-counts
made. In over 600 animals the blood changes have been
practically uniform except in about 10 animals in which
the vaccine used had been allowed to stand too long. The
accompaning charts show that while the percentage of
polymorphonuclear and small mononuclear leukocytes are
not affected with any degree of regularity, the proportion
of large mononuclear cells is invariably increased from 100
to 400 per cent, within from twenty-four to forty-eight
hours. This increase, however, is of short duration and
recedes with rapidity after reaching its height.
Fig. 24 illustrates the average blood change obtained
through the injection of an active residue. The preparation
used was a 1 per cent, sarcoma residue which had been
made from a small round-celled sarcoma of the mediastinum.
Three subcutaneous injections of 5 minims each were given
at hourly intervals, the first blood count being made before
the injections were commenced. The second count, made
SPECIFIC FERMENTS OF THE CANCER CELL 421
seven hours later showed an increase in polymorphonuclear
cells from 21 to 37 per cent., and a corresponding decrease
in small mononuclear leukocytes. The third count made,
twenty-five hours after the first, showed the characteristic
increase of large mononuclear cells which were here regis-
8-14-12 2-15-12 2-10-12
MAST (
FIG. 24. — Rabbit: 5 minims of 1 per cent, sarcoma residue injected
subcutaneously at 8, 9, and 10 A.M. The line designated as P indicates,
in this and the following charts, count of polynuclears; S indicates small,
and L large mononuclears.
tered at 27 per cent., with a normal at the first injection of
10 per cent. Two and one-half hours later the percentage
of this form of cell had returned to normal, which was
again followed by a slight increase. It is probable that the
highest registration of large mononuclear cells occurred
between the second and third counts.
422
PROTEIN POISONS
Fig. 25 represents four rabbits which were given 0.5 c.c.
each of cancer-cell vaccine by different methods. The first
was given a simple subcutaneous injection of rectal adeno-
carcinoma. The fourth represents the same amount of
the same tumor given intraperitoneally, as does also the
third. The second shows the injection of vaccine made
from a rapidly growing round-celled sarcoma of the neck.
B
11-1-1111-2-11 11-3-11
ll-2-llll-3-llll-4.il
1C
S
P
L-
C
-10-11 10-17-1110-18.il
/\
^
\
\
*
\
A
/
V
/-
/
-X
x
D
10.10-1110-17-1110.18-11
FIG. 25. — A, rabbit given subcutaneous injection of 0.5 c.c. rectal
adenocarcinoma (Morris) ; B, rabbit injected with 0.5 c.c. small round-celled
sarcoma; C, rabbit given intraperitoneally 0.5 c.c. rectal adenocarcinoma
(Morris) ; D, rabbit given intraperitoneally 0.5 c.c. rectal adenocarcinoma
(Morris).
From these charts it can be seen that the intraperitoneal
method of injection gives a more rapid reaction than the
subcutaneous.
Fig. 26 illustrates the blood changes occurring in a sheep
following the injection of 0.5 c.c. of sarcoma residue.
Figs. 27 and 28 show more frequent differential counts
following repeated subcutaneous injections.
SPECIFIC FERMENTS OF THE CANCER CELL 423
Fig. 29 illustrates daily counts following a single intra-
peritoneal injection of cells from a breast adenocarcinoma.
This is rather an extreme reaction, inasmuch as the pro-
portion of large mononuclear cells is increased from 6 per
cent, to 35, about 500 per cent.
Fig. 30 shows a very slight reaction following the. injec-
tion of a breast-cancer residue which has almost lost its
activity.
Day
1 2 3
POLY
\
SMALL\
\
\>
s
£
^1
/
/
!GE /
FIG. 26. — Sheep injected with 0.5 c.c. sarcoma residue.
Figs. 31, 33, and 34 illustrate rabbits giving the average
reactions. Fig. 32 represents the effect of giving an
intraperitoneal injection of 10 c.c. of breast-carcinoma
cell-emulsion into a rabbit sensitized to sarcoma vaccine,
the injection being given when the percentage of large
mononuclear cells was at its highest point. Death resulted
in from eight to ten hours.
In order to ascertain just what bearing this change in
percentage of large mononuclear leukocytes had to the
formation of a specific ferment, several rabbits were sensi-
424
PROTEIN POISONS
tized to the cancer-cell and then at varying percentages
10 c.c. of cancer-cell emulsion was injected intravenously
into the animals. In unsensitized rabbits there was no
noticeable effect. Rabbits with a percentage of above
FIG. 27
<--l-9-12--X 1-10-12 **-l-ll-12-->
FIG. 28
-1-3-12--X 1-4-12- -
.19 4 8 10.15 a
.M. P.M. A.M. A.M. P.M.
8-
\
7
V
FIGS. 27 and 28. — A, rabbit injected subcutaneously with 5 minims of
1 per cent. s. r. sarcoma at 12 M., 2 P.M. and 3 P.M. ; B, rabbit injected
subcutaneously with 0.1 c.c. of 2 per cent, mixed residue at 12.15 P.M.,
2 P.M., and 4 P.M.
30 of large mononuclear cells usually died within one to
three hours, and rabbits with a lower percentage, but with
a marked increase, were made sick, but recovered.
Sickness is immediate after injection; the animal at once
falls on its side and begins violent scratching. The respira-
SPECIFIC FERMENTS OF THE CANCER CELL 425
tion is labored and, when death ensues, always stops before
the heart-beat. Rabbits which die from a dose, after being
sensitized, usually have a stage of apparent rest from one-
half to three hours after the first stage of excitability, which
lasts for from three to five minutes. A possible explanation
DAY
1 2 3
DAY
GO
55
50
45
40
35
30
25
20
15
•m
/
/*
/
/
SMALL
-POI
Y
\
\
\
\
/
/*
<£
GE
FiG. 29. — Rabbit given a single
intraperitoneal injection of cells
from a breast adenocarcinoma.
FIG. 30. — Rabbit injected with
a breast-cancer residue which has
almost lost its activity.
of this is that the first stage of excitement is due to the
destruction of cancer cells and consequent liberation of
their toxic radical by the specific ferment present in the
blood serum, while the fatal result which follows later is
due to the reaction between the cancer cell and the large
mononuclear leukocytes.
426
PROTEIN POISONS
The increased percentage of large mononuclear leukocytes
is but a transitory affair, however, which usually lasts from
four to ten hours, and it is impossible to produce fatal
results in rabbits by intravenous injections of cancer-cell
FIG. 31
1-SO-18 1-81-12 2-1-18
P.M. 8 A.M.
FIG. 32
FIGS. 31 and 32. — A, rabbit 167 W., given 5 minims of 1 per cent, sar-
coma at 8.30, 9.30, and 10.30 A.M.; B, rabbit 54 T., given 5 minims of 1
per cent, sarcoma at 1, 2, and 3 P.M.; at 12.15 P.M., 10 c.c. of breast car-
cinoma-cell emulsion (Mel.) was injected intraperitoneally ; rabbit died
in eight to ten hours.
emulsion after this stage is passed. Consequently we
have applied the term "transitory sensitization" to this
phenomenon.
If we bleed a rabbit at the height of this transitory sensi-
tization and obtain the serum, this, when mixed with cancer-
SPECIFIC FERMENTS OF THE CANCER CELL 427
cell emulsion and incubated for one hour will produce
marked symptoms of poisoning when injected intravenously
into a normal rabbit. The severity of symptoms depends
upon both the amount of cancer-cell emulsion and serum.
FIG. 33
A
1-23-12 1-24-12 1-25-12
2.49 4.30 8.30
4 P.M. 11 A.M.
FIG. 34
B
1-17-13 1-18-12 1-19-12
1 P.M. 4 P.M. 8 P.M. 4 P.M. 8 A.M. 12 IK
p
s.
7\
N
/
A
FIGS. 33 and 34. — A, rabbit 54 Br., given 5 minims of 1 per cent, sarcoma
at 1, 2, and 3 P.M.; B, rabbit 54 Bl., given 5 minims of 1 per cent, sarcoma
at 1, 2, and 3 P.M.
Given 5 c.c. of cancer-cell emulsion and 10 c.c. of the serum,
a rabbit of from 500 to 1000 grams will be sick but will
invariably recover.
The next step was to ascertain whether the specific
ferment could be removed from the large mononuclear leuko-
cytes. For this purpose rabbits were sensitized to the
428 PROTEIN POISONS
cancer cell and when the percentage of large mononuclear
leukocytes was at its highest point, the blood was collected
under sterile conditions, in 0.333 per cent, acetic acid.
This was next centrifugated until the leukocytes were
thrown down and the supernatant fluid decanted. The
leukocytes were then placed in a sterile mortar and mixed
with a sufficient quantity of sterile quartz sand to cover.
This was then ground up with vigor for fifteen minutes,
with the frequent addition of sterile normal salt solution,
until five times the volume of the leukocytes obtained had
been added. Next, this normal saline extract was separated
by passing through a Berkefeld filter and tests made for
the presence of the specific ferment by adding varying
amounts of the leukocyte extract to cancer-cell emulsion,
incubating for one hour and then injecting intravenously
into rabbits. Through repeated experiments it was ascer-
tained that 5 c.c. of 2 per cent, cancer-cell emulsion, plus
10 c.c. of leukocyte extract, which was obtained when the
percentage of large mononuclear leukocytes was 25 or
above, would, when injected intravenously into a rabbit
of from 500 to 1000 grams, kill within from one to five
minutes. This is well illustrated in the accompanying
table.
TABLE XL VI. — SHOWING RESULTS OF INJECTING INTRAVENOUSLY INTO
RABBITS CANCER-CELL EMULSION PLUS LEUKOCYTE EXTRACT
INCUBATED ONE HOUR.
Weight, Rabbit leukocyte
Animal. gm. extract. + Vaccine. Result.
343 G. 646.5 10 c.c care. 100 5.0 c.c. 100 Died 1 min.
167 G. 750.0 10 c.c. sarc. 51 res. 5.0 c.c. 100 Died 1 min.
157 W. 1500.0 10 c.c. sarc. 100 5.0 c.c. 100 Very sick £
hr. ; rec.
53 Bl. 1050.0 7 c.c. sarc. 51 res. 5.0 c.c. 100 Died 5 min.
96 G. 1317.0 7 c.c. sarc. 51 res. 2.5 c.c. 103 Died 4J hr.
251 W. 645.0 10 c.c. sarc. 51 res. Not sick.
251 Br. 842.0 15 c.c. care. 100 Not sick.
167 W. 700.0 10 c.c. normal sal. 5.0 c.c. 100 Not sick.
The entire mass of leukocytes, when removed from the
outside of the Berkefeld filter and mixed with 5 c.c. of
SPECIFIC FERMENTS OF THE CANCER CELL 429
cancer-cell emulsion and incubated one hour, failed to have
any effect on a rabbit when injected intravenously, thus
showing that the specific ferment was soluble.
Again, 10 c.c. of leukocyte extract, prepared in the same
manner from a normal rabbit which had not been sensitized,
added to 5 c.c. of cancer-cell emulsion and incubated one
hour, produced no effect when injected intravenously into
a normal rabbit. This would show that the sensitized
animal possesses some specific chemical substance which
reacts with cancer tissue and which the normal rabbit does
not possess. That this substance is specific for malignant
cells is further shown by the fact that 10 c.c. of leukocyte
extract from a sensitized animal, plus 5 c.c. of a 2 per
cent, normal skin vaccine, incubated one hour, produces
no effect when injected intravenously into a rabbit of
450 grams.
The same table shows also that sarcoma residue or
vaccine sensitizes to carcinoma as well as sarcoma, and
vice versa, so that the probable conclusion is that the chemical
change within the cell is the same for both sarcoma and
carcinoma.
Organ extracts from kidney, liver, brain, spleen, and
heart, made by grinding these organs in normal saline on
successive days, after sensitization to cancer protein, have
no effect when mixed with cancer-cell emulsion, incubated
one hour; and injected intravenously into normal rabbits.
When as small an amount as 1 c.c. or more of leukocyte
extract is injected directly into the tumor of a cancer
patient it may cause sudden and severe symptoms. In four
cases, when this procedure was adopted, the patient has
complained within from one to five minutes of difficulty
of respiration. Next, he would lose consciousness, which
would be accompanied by rather violent muscular twitchings
and lowered pulse-rate. This stage would last from five
to ten minutes, and would be followed by a period of rest,
from fifteen minutes to one hour in duration, which in turn
would be followed by a violent chill, and temperature
ranging from 103° to 106° F. This would last from one to
430 PROTEIN POISONS
six hours and would be followed by from twenty-four to
forty-eight hours of extreme exhaustion. At no time has
such a reaction been obtained, even with doses of 10 c.c.,
when given subcutaneously or intravenously away from the
tumor; although a chill from one to three hours after injec-
tion has been observed. We wish now, however, to call
attention only to the animal experiments, the above being
mentioned simply because it is additional proof of the
presence of a specific ferment.
It is of interest to note here also that a vaccine prepared
from human carcinoma gives a much higher percentage of
large mononuclear leukocytes when injected into a rabbit
or sheep than when injected into a human being. This
has been observed regardless of whether the human being
had malignant disease or not. I have injected cancer
vaccine into myself, and the highest resulting percentage
of large mononuclear leukocytes was 15. This fact is of
interest when we consider that with experimental cancer
the animal injected must always be of the same family as
the one that furnishes the tumor in order to obtain a
"take."
It should not be understood that we consider the ferment
causing reproduction in the cancer cell to be of exactly the
same chemical nature as the active ferment of a reproducing
normal cell, but rather that we are dealing with a chemically
altered constructive ferment, a fact that we will demon-
strate later.
While attempting to use leukocyte extract from animals
sensitized to the cancer protein in a therapeutic way, it
was found that extract prepared by filtering through a hard
filter paper would upon repeated usage in the same case
cause symptoms of sensitization; while the use of extract
prepared by Berkefeld filtration would not be followed by
these characteristic phenomena. At the same time, as
previously mentioned, all of the specific ferment passes
with ease through the filter.
In order to ascertain what constituents were removed
from the leukocyte extract by passage through the^Berkefeld
SPECIFIC FERMENTS OF THE CANCER CELL 431
filter, definite amounts were analyzed by the Scherer and
Hammersten methods for total albumins and globulins.
Determination of Albumin and Globulin. — 25 c.c. of leuko-
cyte extract which had been centrifugated to throw down
all corpuscles and sand used in its preparation was passed
through a filter paper. This was rendered acid with acetic
acid, 4 grams of NaCl added, and then heated for one-half
hour on a water-bath. After coagulation had occurred this
was filtered through a previously dried and weighed filter
and washed with hot water until the filtrate ceased to give
a reaction for chlorides. Next, the residue was washed
with absolute alcohol and then ether, after which it was
dried at 130° to a constant weight.
Paper and albumin plus globulin . . . 1 . 66658 grams
Weight of paper 1.61150 grams
Weight of albumin and globulin . . . . 0.05508 gram
The same method was applied to 25 c.c. of the same lot
of leukocyte extract after passing through a Berkefeld
filter.
Weight of filter plus albumin and globulin . 1 . 55000 grams
Weight of filter 1.53125 grams
Weight of albumin and globulin .... 0.01875 gram
In order to make a separate determination of globulins
and albumins the following method of Hammersten was
adopted: 25 c.c. of leukocyte extract before filtration
through a Berkefeld was added to 30 grams of pulverized
magnesium sulphate. This was warmed to 30° with fre-
quent stirring, and then placed in the cold for twenty-four
hours. This was then filtered through a weighed filter,
previously dried at 110°, and washed with magnesium
sulphate until the filtrate ceased to give a reaction for
albumin when heated with acetic acid. Next, the filter
was dried for four hours at 110° to coagulate the globulin,
after which the magnesium sulphate was washed out with
432 PROTEIN POISONS
hot water. It was then washed with alcohol and ether,
and dried to a constant weight at 110°.
Weight of globulin and filter ....'. 1.5864 grams
Weight of filter 1.5766 grams
Weight of globulin 0 . 0098 gram
25 c.c. of the same leukocyte extract after Berkefeld
filtration was treated in the same manner.
Weight of globulin and filter 1.4913 grams
Weight of filter 1.4766 grams
Weight of globulin 0.0147 gram
By subtracting these globulin weights of before and
after filtration from the above-given albumen plus globulin
weights we arrive at the albumin weights:
Albumin and globulin before filtration . . 0.05508 gram
Globulin before filtration 0.00980 gram
Albumin before filtration 0.04528 gram
Albumin and globulin after nitration . . 0.01875 gram
Globulin after filtration 0.01470 gram
Albumin after filtration 0 . 00405 gram
From the above it can be seen that we have removed a
large percentage of the albumin through the passing of the
extract through the Berkefeld filter. This is represented
by the difference between 0.04528 gram and 0.00405 gram,
which is 0.0412 gram.
The difference in the globulin weight gives an apparent
increase of 5 mg. in the after-globulins, an amount so small
that it can be neglected when compared with the albumin
difference.
From this it can be seen that the sensitizing portion of
the extract is in all probability contained within the albumin,
and that the specific enzyme is not of albuminous nature.
SPECIFIC FERMENTS OF THE CANCER CELL 433
While the globulins are filterable this does not prove that
the ferment is a globulin, but only that it filters through
with the globulin. The enzyme itself may be of much
simpler construction.
In conclusion I may state that the most valuable deduc-
tions to be drawn from the work as so far conducted are
as follows:
1. Transitory sensitization. The fact that an animal
may be sensitized to certain proteins, and that such sensi-
tization is active for only a few hours is of extreme interest
and importance.
2. The active transitory ferment formed by the intro-
duction of cancer protein into an animal is not an albumin.
It is either a globulin or of simpler chemical structure.
3. Sensitization, in the use of leukocyte extract at least,
is probably caused by the albumin content of the solution.
Other Methods Used. — In order that a more definite
knowledge of the chemistry of this specific ferment might
be obtained, the following experiment was made: 120 c.c.
of sensitized leukocyte extract, after filtration through a
Berkefeld filter, was mixed with an equal volume of satu-
rated ammonium sulphate solution. This was allowed to
stand overnight and the precipitate was then filtered off.
The precipitated globulins, while still slightly moist, were
removed from the filter paper and dissolved in a saturated
solution of sodium chloride. This was next rendered slightly
acid with 0.25 per cent, acetic acid, which again precipitated
the globulins. When the precipitation was complete and
had settled to the bottom, for which twelve to twenty-four
hours should be allowed, the globulin was collected upon
a hard filter paper. From this it was removed, while still
slightly moist, to a sterile watch-glass, where it was allowed
to dry. The globulin was next weighed and dissolved in
normal saline solution in the proportion of 1 mg. of globulin
to 1 c.c. of normal saline.
The above method is given after using many different
modifications of the same. In many experiments I have
precipitated the globulins with semisaturated ammonium
28
434 PROTEIN POISONS
sulphate solution, then dissolved in water, equal in amount
to the original leukocyte extract, and then precipitated again
with ammonium sulphate before dissolving in saturated
sodium chloride solution, but such a procedure is unneces-
sary when the extract has been previously filtered through a
Berkefeld, inasmuch, as previously shown, four-fifths of the
albumin is removed at this time.
Again, it was my custom to filter after dissolving in
saturated sodium chloride solution, but this has been
abandoned because of the fact that if insufficient saturated
sodium chloride is used, much globulin may be retained
by the filter, and, inasmuch as the excess of ammonium
sulphate is all that is desired to be rid of, both time and
material can be saved by not filtering. The saturated
sodium chloride solution is always filtered before using.
Again, it was my custom to dialyze through parchment
paper after collecting the precipitate thrown down by 0.25
per cent, acetic acid, but repeated experiment has shown
that the percentage of acetate when the globulins are
dissolved in the proportion of 1 mg. to 1 c.c. of normal
saline is so small as to be of no moment, and it is much
preferable to have a solution of normal saline than one of
sterile water for injection in the patient.
It should be added that precipitation is aided by allowing
hot water to pass over the outside of the flask until the
temperature of the contained fluid is about 30° C.
The following experiment has been repeated many times,
so that the possibility of error is slight.
(a) 10 c.c. of soluble globulins from sensitized leukocyte
extract plus 5 c.c. normal saline were incubated one hour,
and 10 c.c. of this injected intravenously into a small rabbit.
There was no change in the animal.
(6) 10 c.c. of soluble globulins from sensitized leukocyte
extract plus 5 c.c. of 2 per cent, cancer-cell emulsion were
incubated one hour and then injected intravenously into
a moderate-sized rabbit. The rabbit died in one-half
minute.
From the above it can be seen that the specific ferment
is in all probability a part of the globulin content, and its
activity is much increased by obtaining it in the more
purified state.
It is now my practice to test the strength of each batch
prepared by the above experiment before using it in any
given case. Sensitization phenomena have been entirely
lacking whenever used therapeutically, which was not
true of any former preparations. The filtered leukocyte
extract when used in small amount subcutaneously rarely
produced sensitization phenomena, but when injected
intravenously in increased dosage this symptom complex
was frequently observed. The unfiltered leukocyte extract
frequently produced sensitization even with small amounts
given subcutaneously, so the conclusion previously arrived
at; that the albumin content was responsible for these
untoward symptoms is apparently confirmed. It must not
be understood, however, that the globulins do not sensitize.
The work so far simply shows that it is possible to remove
the albumin which contains no specific ferment, and inas-
much as four-fifths of the protein has been removed, much
larger doses must be given before the phenomena of sensiti-
zation can be observed.
The above substance, because of its specific ferment
action and its apparent chemical nature, I would designate
as anticancer globulin.
CHAPTER XV
THE PHENOMENA OF INFECTION
IT may be of interest to go somewhat into detail con-
cerning our ideas of the phenomena of infection. In all
infections there are two principal factors — one the infecting
virus and the other the body cell. In addition to these
there is the environment in which the struggle for supremacy
between the virus and the body cell takes place. This con-
sists of the unorganized fluids of the body, and is of great
weight in determining the result of the contest. In the first
place, wrhat do we know of the infecting virus? As we
have seen, bacteria are particulate, specific proteins. Since
they are particulate, we speak of them as bacterial cells.
It is not, however, essential that an infecting virus be
particulate in the sense that it be possessed of substance
and form recognizable to our limited sense of sight even
when aided by the most perfect microscope. There are'
many filterable viruses. Some pass through our finest
porcelain filters and cannot be deposited from the fluids in
which they exist even when kept for hours in the most
efficient centrifuge manufactured. Theoretically, there
is no reason why a virus may not exist in any degree of
lability of structure. The bacteria are particulate and
solid, which means that their structure is so radically
different physically from the medium in which they exist
that they can be recognized by our sight, aided by proper
magnifying lenses, but viruses may be semi- or wholly
fluid. In such instances their structure is not sufficiently
differentiated from the medium that we can recognize
them. According to our conception, a living protein does
THE PHENOMENA OF INFECTION 437
not necessarily possess a form recognizable to our limited
sense even when aided by the most perfect lenses.
One of the most important results of our work, in our
opinion, is the demonstration that bacteria are chemically
not simple, but quite complicated in structure. Morpho-
logically, they show but little or no differentiation in struc-
ture, but chemically they are quite as complicated and
complex as many of the cells of the higher animals. They
contain carbohydrates, nuclein bodies, and polymers of the
mono- and diamino-acids. They are glyconucleoproteins.
We interpret this as signifying that functionally they are
highly developed.
While an infecting virus may be solid, semisolid, gela-
tinous, or liquid, we will, in the further consideration of
the phenomena of infection, take the particulate type, the
bacterium, as an example of an infecting agent.
What are some of the capabilities of a bacterial cell? In
the first place it possesses that attribute which distinguishes
and characterizes all living matter — the capability of growth
and reproduction. In order to grow and multiply its
molecular structure must be labile — in a sate of constant
change. Some bacteria under certain conditions may pass
into a resting state characterized by the formation of spores,
but these are awakened into life when the environment
becomes fit, and the spore develops into the active form
when it infects. In all instances the active, infecting agent
is a living protein, capable of growth and multiplication.
In order to do this it must carry on a constant exchange
in matter with the medium in which it exists. It must
assimilate and eliminate. It must absorb groups from the
molecules about it, and cast out those which it has already
used. Stop this process and the continuation of life is
impossible. Every living cell, be it bacterial, vegetable,
or animal, must feed or cease to exist. Besides, a cell is
limited in its food supply by that which lies within its
reach. There must, therefore, be a certain supporting
relation between the bacterial cell and the medium. The
groups derived from the medium must fit into the molecular
438 PROTEIN POISONS
structure of the cell, otherwise they would be of no service
to it. This necessitates the cleavage of the molecules of
the medium along definite lines. Many kinds of cells may
live in the same or like media, but for each kind of cell the
cleavage of the medium must be specific. From this it
follows that the agent by which the cleavage products are
secured must be supplied by the cell itself, and must be
peculiar to that kind of cell. These cleavage agents which
prepare foods for the cell from the medium are known as
ferments, and each kind of cell has its own characteristic
and specific ferments. As to the real nature of ferments,
we know little or nothing, but that every kind of cell has
its specific ferment or ferments, we do know. The same
ferment may not be able to break up all proteins. In this
respect there are great variations in the proteolytic fer-
ments. Some digest a wide variety of proteins while others
are capable of acting only on one specific protein. There
must be a relation between the ferment and its substrate.
As Fischer once said, the former must fit the latter as a
key fits into the lock, and as there are master keys that
open many doors, so there are general proteolytic ferments,
and as there are special keys that fit only one lock, so there
are specific proteolytic ferments. It will be observed that
we have used the word " specific" in two senses in speaking
of proteolytic ferments. Each kind of cell has its specific
ferment, and each protein may have its specific ferment.
This double use of the term "specific" should be borne in
mind, since there seems to be no way to avoid it.
It follows from what has been said that a bacterium
placed in a medium in which its ferment is ineffective cannot
grow and multiply. A bacterium which cannot grow and
multiply in the animal body cannot cause an infection. Its
inability to grow and multiply in the animal body may be
due to the fact that its ferment or ferments cannot digest
or properly break up the proteins of the animal body. This
is one of the reasons why the great majority of bacteria
are non-pathogenic or are harmless. These organisms
when grown on suitable media produce just as much poison
THE PHENOMENA OF INFECTION 439
as the pathogenic bacteria, but not being able to feed upon
the proteins of the body they die. This, however, is not
the sole, and probably not the most important, cause
of the failure of so many varieties of bacteria to do harm
to the higher animals. What has been said about the pro-
duction of ferments by the bacterial cell is equally true of the
body cell. In fact, it is true of every living cell. The body
cell has its specific ferments, and the bacterial cell being
protein substance is liable to be digested by the 'ferments
elaborated by the body cells.
In the inability of the bacterial cell to grow in the animal
body either because it cannot feed upon the proteins of
the body, or because it is itself destroyed by the ferments
elaborated by the body cells lies the fundamental expla-
nation of all forms of bacterial immunity either natural
or acquired. Toxin immunity needs further explanation.
Certain bacteria, of which the diphtheria bacillus may be
taken as a type, elaborate soluble, extracellular substances
known as toxins. These are probably ferments or closely
allied bodies. They resemble ferments in the following
particulars: (1) They are destroyed by heat. (2) They
act in very dilute solution. (3) When repeatedly injected
into animals in non-fatal doses they cause the body cells
to elaborate antibodies which neutralize the toxin both
in vivo and in vitro. (4) In the development of their effects
a period of incubation is required. (5) It has been shown
by Abderhalden, by optical methods, that they have a
cleavage effect upon proteins. They split complex proteins
into simpler bodies. In other words, they have a proteo-
lytic action. (6) They are specific in two senses, (a) They
are specific according to the cell which produces them.
Diphtheria toxin is elaborated by the diphtheria bacillus
and by no other organism. The toxin of snake venom is a
specific product of the poisonous gland of the snake, and
this is further specific inasmuch as that produced by the
glands of one species is different from that elaborated in
another species. (6) They are specific in the antibody
elaborated in the animal body after repeated injections of
440 PROTEIN POISONS
non-fatal doses. Diphtheria antitoxin protects only against
diphtheria toxin, and not against that of the tetanus or
dysentery bacillus, or that of snake venom.
The side-chain theory evolved by the genius of Ehrlich
best explains the action of toxins and the production of
antitoxins. Without subscribing to all the details of this
theory, we believe that it is a biological law that when a
living cell is attacked by a destructive ferment or toxin it
tends to elaborate an antiferment or antibody. This is
one of the ways in which the living cell may protect itself.
The formation of such antibodies in multicellular animals
is one of the factors in the fine adjustment essential to
harmony of action between different tissues and organs.
It best explains the fact that the digestive organs do not
harm themselves, and the antitryptic action of blood-serum
is one of the most interesting and important phases of
parenteral digestion.
The number of pathogenic bacteria which produce
toxins, at least in appreciable quantity, is small, and the
action of toxins and antitoxins in infections due to those
organisms which do not produce such bodies is of minor
importance. Since all bacteria, and in fact all living cells
produce ferments, and since every ferment, so far as we
know, may lead cells acted upon by them to produce anti-
ferments, there may be some toxin and antitoxin action
in all infections, but in most bacterial infections such
action is overshadowed by processes much more powerful
in their effects.
In our opinion the action of the diphtheria bacillus may
be stated as follows: The organism finds lodgement and
the conditions for growth favorable in the upper air pas-
sages. Here it grows in mass and may kill by mechanical
obstruction. It produces its soluble, diffusible toxin,
which has the properties of a ferment and splits up the
proteins of the body, setting free the protein poison. In
case of recovery or in the production of antitoxin in animals,
the body cells elaborate an antiferment or antitoxin which
neutralizes the toxin and prevents its cleavage action.
THE PHENOMENA OF INFECTION 441
The bacilli in the throat are not destroyed by natural
recovery or by cure with antitoxin, but the action of the
toxin is prevented by the antibody. It is not, in our opinion,
the toxin itself which kills, but a cleavage product which
results from the action of the toxin on the proteins of the
body.
All ferments are of cellular origin. This does not mean
that ultramicroscopic forms of life or non-particulate living
organisms, if there be such, do not produce ferments. It
would probably be better to say that all ferments are the
products of living organisms and that there can be no living
organism which does not produce its specific ferment.
We cannot conceive of life without ferment action, because
all living things must feed and food assimilation without
ferment action is inconceivable. Food must be fitted for
assimilation, and this is dependent upon ferment action.
Ferments are intra- and extracellular. All are formed
within the cell, but some diffuse into the medium while
others do not. In some instances at least cell permeation
by the pabulum is essential to the feeding of the cell. In
other instances it is highly probable that the ferment is
accumulated on the cell surface and there acts upon the
pabulum. In still other instances the ferment diffuses
into the medium more or less widely from the cell which
elaborates it. Many cells produce both intra- and extra-
cellular ferments, and these are not necessarily the same.
In some instances, probably in most cells, the intracellular
ferment cannot be extracted from the cell or obtained in
soluble form without destruction of the cell. This does
not mean that it must exist in the soluble form before it
can manifest its cleavage action. The pabulum may per-
meate the cell and in this location be split up by the intra-
cellular ferment. We have insisted upon this as an explana-
tion of the well-established fact that soluble proteins
sensitize much more readily and completely than insoluble
ones.
It will be well to illustrate what we have said about
cellular ferments by a condensed sketch of the work that
442 PROTEIN POISONS
has been done on the germicidal properties of the blood.
As early as 1872 Lewis and D. Cunningham1 demonstrated
that non-pathogenic bacteria injected into the circulation
soon disappear. In the blood of 12 animals thus treated
bacteria could be found after six hours in only 7. Of 30
animals, bacteria were found in the blood of only 14 after
twenty-four hours, and of 17 animals bacteria were found
in the blood of only 2 when the examination was made
from one to seven days after the injection. In 1874,
Traube and Geschiedlen2 found that arterial blood, taken
under aseptic precautions from a rabbit into the jugular
vein of which 1.5 c.c. of a fluid rich in putrefactive bacteria
had been injected forty-eight hours previously, failed to
undergo decomposition when kept for months. These
investigators attributed the germicidal properties of blood
to its ozonized oxygen. Like results were obtained by
Fodor3 and Wysokowicz.4 The latter accounted for the
disappearance of the bacteria, not through the germicidal
action of the blood, but by supposing that they found
lodgement in the capillaries. The first experiments made
with extravascular blood were conducted by Grohmann5
under the direction of A. Schmidt in his researches on the
coagulation of the blood. It was found that the virulence
of anthrax bacilli, as demonstrated by their effect on rabbits,
was diminished by being kept in blood. He supposed that
the bacilli were altered in some way by the process of
coagulation. In 1887 Fodor6 made a second contribution
on this subject, in which he combated the retention theory
• of Wysokowicz. One minute after the injection of 1 c.c.
of an anthrax culture into the jugular vein, in eight samples
of blood, Fodor found only one colony of the bacillus. He
1 Eighth Annual Report of the Sanitary Commission of the Government
of India.
2 Schlesische Gesellschaft f. Vaterland. Cultur.
8 Archiv f. Hygiene, iv, 1886. 4 Zeitsch. f. Hygiene, i, 1886.
8 Ueber die Einwirkung d. Zellenfrien Blut-plasma auf einige pflanzliche
Mikroorganismen, Dorpat, 1884.
6 Deutsch. med. Woch.
THE PHENOMENA OF INFECTION 443
also took blood from the heart with a sterilized pipette,
and added anthrax bacilli to it. This was kept at 38°, and
plates made from time to time showed rapid diminution
in the number of bacteria, until after a time, when the blood
having lost its germicidal properties, the number rapidly
increased. In 1888 Nuttall,1 working under the direction
of Fliigge, used defibrinated blood taken from various
species of animals, rabbits, mice, pigeons, and sheep, found
that the blood destroyed the bacillus anthracis, b. subtilis,
b. megatherium, and staphylococcus aureus. He also con-
firmed the finding of Fodor that after a while the blood
loses its germicidal properties and becomes a suitable
culture medium. Continuing this work, Nissen2 reached
the following conclusions: (1) The addition of small
quantities of salt solution or bouillon to the blood does not
destroy its germicidal properties. (2) The bacilli of cholera
and typhoid fever are easily destroyed by fresh blood. (3)
For a given volume of blood there is a maximum number
of bacilli that can be destroyed. (4) Blood whose coagu-
lability has been destroyed by peptone injection is still
germicidal. (5) Blood in which coagulation is prevented
by the addition of 25 per cent, of magnesium sulphate
has its germicidal properties decreased. (6) Filtered blood
plasma from the horse is germicidal. Behring3 attributed
the germicidal action of the blood of the white rat on the
anthrax bacillus to its great alkalinity. In 1890, Buchner
and his students4 published their first contribution on the
germicidal properties of blood serum. At first Buchner
believed that the germicidal constituent of serum is the
serum albumin and the conclusions were stated as follows:
(1) The germicidal action of blood is not due to the phago-
cytes, because it remains after destruction of the leuko-
cytes by alternating freezing and thawing. (2) The germi-
cidal properties of the cell-free serum must be due to its
1 Zeitsch. f. Hygiene, iv, 353.
2 Ibid., 1889, vi, 487. 3 Ibid., 1889, vi, 1.7.
4 Archiv f. Hygiene, 1890, x, 84, 101, 121, 149.
444 PROTEIN POISONS
soluble constituents. (3) Neither neutralization of the
serum, nor the addition of pepsin, nor the removal of
carbon dioxide gas, nor treatment with oxygen has any
effect upon the germicidal properties of the blood. (4)
Dialysis of the serum against water destroys its activity,
while dialysis against 0.75 per cent, salt solution does not.
In the diffusate there is no germicidal substance. The
loss by dialysis with w^ater must be due to the withdrawal
of the inorganic salts of the serum. (5) The same is shown
to be the case when the serum is diluted with water, and
when it is diluted with the salt solution. In the former
instance the germicidal action is destroyed, while in the
latter it is not. (6) The inorganic salts have in and of them-
selves no germicidal action. They are active only insofar
as they affect the normal properties of the albuminates of
the serum. The germicidal properties of the serum reside
in the albuminous constituents. The difference in the
effects of the active serum and that which has been heated
to 55° is due to the altered condition of the albuminate.
The difference may possibly be a chemical one (due to
cleavage within the molecule) or it may be due to changes
in mycelial structure. The albuminous bodies act upon
the bacteria only when the former are in an active state.
Vaughan and McClintock1 called attention to a contra-
diction between Buchner's work and his conclusions, in
the following language: "We wish at this point to call
attention to an inconsistency between the results obtained
by Buchner and the conclusions that he draws: In experi-
ment 45 he renders the serum slightly acid, and adds 0.1
gram of pepsin to each 5 c.c. of serum (showing by a side
experiment that this pepsin actively digests coagulated
egg albumen in neutral solution) and finds that the digestive
action of the pepsin does not lessen the germicidal properties
of the serum. In fact, he states this in his conclusion, but
his ultimate opinion and the one held by him in his latest
contribution, is that the germicidal constituent of the
1 Med. News, December 23, 1893.
THE PHENOMENA OF INFECTION 445
blood is the serum albumin. How much serum albumin
remains in blood serum after it has been thoroughly digested
with pepsin ? He could scarcely have chosen a more positive
method of demonstrating that the germicidal constituent
is not serum albumin. Either his pepsin was not active
and on this supposition his experiment was without value,
or the active constituent of the blood-serum is a substance
that is not destroyed or materially altered by peptic diges-
tion. We know that the peptones not only have no
germicidal properties, but that they belong to that class
of proteins that is most favorable to the growth of bacteria.
We recognize this fact when we add peptones to the various
artificial media on which we cultivate bacteria." We will
return to this point after proceeding farther with the
chronological order in which this research has developed.
Prudden1 found that ascitic and hydrocele fluids restrain
the development of certain bacteria. Rovighi2 reported
that the germicidal action of the blood is increased in
febrile conditions. Pekelharing3 enclosed anthrax spores in
bits of parchment and introduced these under the skin
of rabbits. Thus treated the spores soon lost their viru-
lence and finally their capability of growth. The destruc-
tion of these spores could not have been due to phagocytes
which did not penetrate the parchment, but must have
been caused by soluble substances. Behring and Nissen4
found that the serum of the white rat, the dog, and the
rabbit destroy anthrax bacilli, while serum obtained from
the mouse, sheep, guinea-pig, chicken, pigeon, and frog
has no such action. It will be observed that in this there
is no constant relation between the germicidal action of
the blood of animals of different species and their suscep-
tibility to the infection. Thus the rabbit is highly suscep-
tible to anthrax notwithstanding the fact that its blood
destroys large numbers of this organism. On the other
hand the chicken is immune to anthrax from the moment
1 Medical Record, 1890.
2 Atti della Accad. Med. di Roma, 1890.
3 Ziegler's Beitrage, viii. 4 Zeitsch. f. Hygiene, 1890, viii, 412.
446 PROTEIN POISONS
when it comes from the shell, and yet the bacillus grows
luxuriantly in the extravascular blood of the chick. Hankin1
was one of the first to show that other cells, besides the
leukocytes, contain germicidal substances. He made several
contributions to the study of so-called "defensive proteins,"
which he believed to be globulins. This is interesting in
view of the fact that ferments are often carried down with
globulins on precipitation with neutral salts. Bitter2 was
unable to confirm Hankin's work, but it is needless to go
into this because we now know that many cells elaborate
germicidal substances. Christmas3 prepared a germicidal
substance from the spleen and other organs by the following
method: The animal was killed with ether, opened under
aseptic precautions, the organ removed, cut into fine pieces,
covered with 50 c.c. of glycerin, and allowed to stand for
twenty-four hours and then filtered. The filtrate is treated
with five times its volume of alcohol and the precipitate
is immediately collected and washed with absolute alcohol.
Traces of alcohol are removed, so far as possible, by pressure,
and the precipitate is dissolved in 25 c.c of distilled water,
and air is blown through the solution to destroy last traces
of alcohol; then the fluid is filtered and its germicidal
action tested. Bitter4 strove hard to find fault with this
agent and its method of preparation. He found it a power-
ful germicide, but he could not reconcile the fact that the
preparation of Christmas still proved a powerful germicide
after it had been heated to 65°, while blood-serum loses its
germicidal effect when heated to 55°. Buchner5 had the
following to say on this point : " A method given by Christ-
mas for the preparation of germicidal solutions from the
organs of normal rabbits has also been tested by Bitter.
Germicidal solutions were indeed obtained, which, however,
differed materially from active serum, for in three experi-
ments, notwithstanding heating to 65°, the germicidal
1 Centralbl. f. Bakt., 1891, ix, 336.
2 Zeitsch. f. Hygiene, 1892, xii, 328.
3 Annales de 1'Institut Pasteur, 1891, v, 487.
4 Loc. cit. 6 Archiv f. Hygiene, 1893, xvii, 112.
THE PHENOMENA OF INFECTION 447
action remained." We have gone into this detail concerning
the preparation of Christmas for the following reasons:
(1) It remains to day a good method of preparing a germi-
cidal agent from the spleen or other tissue. (2) Its method
of preparation indicates that it is a ferment. (3) It is an
illustration of the fact that the degree of heat borne by a
ferment, without being inactivated, is dependent in part at
least on the character of the solvent in which the ferment
is found.
Emmerich and his students1 made the following experi-
ments: A serum was dialyzed against distilled water until
its globulin was precipitated. The globulin-free serum
was precipitated with alcohol, and the serum albumin
thus thrown down was dissolved in 0.05 per cent, solution
of potassium hydroxide. This solution was found to be
markedly germicidal, and the conclusion reached was that
the germicidal constitutent of blood serum was an alkaline
albuminate.
Vaughan and his students2 published their first paper
upon the germicidal properties of nuclein. In their first
contribution they showed that nucleins prepared from
testes, thyroid gland, and yeast cells are markedly germi-
cidal to both pathogenic and non-pathogenic bacteria.
In 1894 Kossel3 quite independently announced the dis-
covery of the germicidal action of nuclein and nucleic
acid. Vaughan not only demonstrated the germicidal
action of nuclein in vitro, but also showed (1) that it
protected rabbits against subsequent inoculation with
the pneumococcus; (2) it also protected a considerable
percentage of rabbits against inoculation with the bacillus
tuberculosis; (3) that it had a curative effect on rabbits
already inoculated with tuberculosis; and (4) that it
apparently benefited initial tuberculosis in man.
It now turns out that the germicidal action attributed
by Vaughan and Kossel to nuclein was probably not due to
1 Centralbl. f. Bakt., 1892, xii, 364.
2 Medical News, May 20, 1893.
3 Archiv f. Anat. u. Physiol., Physiolog. Abtheilung, 1894, 194.
448 PROTEIN POISONS
this agent, but to ferments which came out with the nuclein
from the cells used in its preparation. This should have
been known at the time the work was done, because both
of these investigators were aware of the fact that a tempera-
ture short of boiling destroyed the germicidal properties
of their solutions, but we did not know so much about
cell ferments then as we do now. This, of course, does not
mean that all the results obtained with preparations of
nuclein, such as an increase in the number of leukocytes,
were due to the ferment contained in the preparations,
but it is more than probable that the germicidal action was
due to the ferment. The whole matter demands reinves-
tigation.
It should be stated that Vaughan and McClintock1
demonstrated the presence of nuclein in blood-serum. This
was done by precipitating a large amount of serum, obtained
under aseptic conditions, wyith alcohol and digesting the
precipitate with artificial gastric juice so long as digestion
proceeded, the completion of digestion being indicated by
failure to respond to the biuret test. The small amount
of protein material which wholly resisted gastric digestion,
and which could be only nuclein, was dissolved in 0.12
per cent, potassium hydroxide, and its germicidal action
demonstrated on the bacillus of cholera, anthrax bacillus,
typhoid and colon bacilli, and on various cocci. At the
same time it was shown that a 0.5 per cent, solution of the
alkali was without effect upon these organisms.
It then seemed that the whole question of the germicidal
action of the blood was practically settled. The leukocytes
contain large quantities of nuclein. The blood serum
contains small quantities of the same substance. That of
the serum comes from the leukocytes either in the form of
a secretion or as a result of the breaking-down of the cells,
and nuclein is a powerful germicide. The phagocytes
destroy bacteria either by engulfing and then digesting
them, or through the action of the nuclein dissolved in the
1 Loc. cit.
THE PHENOMENA OF INFECTION 449
blood. We say all this seemed clear and probably it is all
right, except now it seems probable that although nuclein
is abundant in the leukocytes and present in small amount
in plasma and serum, it is not the germicidal agent in
either. The germicidal agent in the cell and that dissolved
in the plasma or serum are both most likely ferments, the
one intra- and the other extracellular, and the two are not
identical. Metschnikoff's phagocytic theory and Buchner's
alexin theory are both in a way right, but whether the
germicidal substance in the serum is a secretion of the
phagocyte or a disintegration product of the cell is an
interesting question.
The germicidal constituent of blood serum, studied by
Buchner and named alexin by him, is inactivated by
heating the serum to 55°, while the germicidal substance
obtained by Kossel and Vaughan from cells, and believed
by them at the time to be nucleic acid, required a tempera-
ture of 85° to render it inert. Evidently these must be two
quite different bodies, or if the same substance, their behavior
under the influence of temperature must be markedly
affected by the conditions under which they were tested.
The researches of Schattenfroth1 showed further differences
between the intra- and extracellular germicidal constituents
of the blood. The former has no hemolytic action on the
red corpuscles of other species, while the latter may have.
The intracellular germicide is not affected by the salt
content of the medium, retaining its activity in a salt-free
menstruum, while the extracellular substance is inactivated
by the removal of salt from the serum by dialysis. Daubler2
came to the conclusion that the germicidal constituents
of the serum and of the leukocytes are not identical, the
latter remaining active after being heated to 60°. He also
found that the germicidal substances obtained from the
leukocytes of different species differ in measurable degree
as tested upon the same bacteria. Many other investi-
1 Archiv f. Hygiene, 1897, xxxi, 1; ibid., 1899, xxxv, 135.
2 Centralbl. f. Bakt., 1899, xxv, 129.
29
450 PROTEIN POISONS
gators produced evidence of the fact that the intra- and
extracellular germicidal constituents of the blood are not
identical, but since the literature of this subject has been
collected by Kling,3 we will not go into detail but will
content ourselves with the reproduction of the summary as
given by this author. It should be stated that Petterssen
designates the intracellular germicidal constituent of leuko-
cytes and other cells as "endolysins." Kling's conclusions
from the .work of others and himself are stated substantially
as follows: (1) The germicidal substances (endolysins)
of the polymorphonuclear leukocytes may be obtained
from the protoplasm by the following methods: (a) By
digesting the cells for half an hour at 50° in bouillon. (6)
by extracting the cells with weak acid or alkali, or (c) by
alternating freezing and thawing of the cells. They cannot
be obtained by digesting with bouillon at 37°, nor with
physiological salt solution, nor with 5 per cent, "inacti-
vated" serum. (2) As tested on bacillus subtilis, the
endolysin bears a temperature of 65° without recognizable
effect on its germicidal action, and it is not until the tem-
perature is increased to 75° that any such effect is noticed.
The endolysins can, in daylight at room temperature, and
in the dark at 37°, be evaporated to dryness, and in this
state they may be heated for half an hour at 100° without
being destroyed. The serum alexins may be obtained in
the dry state in the same manner, but when heated to this
temperature they are inactivated. The endolysin as tested
on the subtilis does not pass through a Pukall filter, while
the serum alexin does. The endolysins as tested on the
subtilis, the anthrax, and the typhoid bacillus are destroyed
by the Rontgen ray, while the serum alexins are not. The
endolysins cannot be extracted with ether, but are not
injured by ether, while the serum alexins are destroyed by
ether. (3) The activity of an inactivated extract of the
leukocytes of the rabbit, as tested on the subtilis, may
be restored by the addition of a small quantity of the same
3 Zeitsch, f, Immunitatsforschung, 1910, vii, 1.
THE PHENOMENA OF INFECTION 451
extract in a fresh state. Likewise, an inactivated normal
serum of the rabbit or the inactive serum of the guinea-pig
may be complemented by the addition of a small amount
of the leukocytic extract from the rabbit or guinea-pig,
respectively. Furthermore, an inactivated leukocytic
extract from a guinea-pig can be activated by the addition
of a small amount of the normal serum of a rabbit. (4)
Extracts from the polymorphonuclear leukocytes of rabbits,
guinea-pigs, and cats destroy in vitro the timothy bacillus,
the grass bacillus II, Korn's acid-fast bacillus I, and Rubner's
butter bacillus. The extract from rabbits' leukocytes has
a bactericidal action on the bacillus tuberculosis of man.
Extracts of rabbit, guinea-pig, and cat macrophages do not
destroy these acid-fast bacilli in vitro. The same is true of
the extracts from the thymus gland of the rabbit. Living
polymorphonuclear leukocytes injected into guinea-pigs
decrease the virulence of the human tuberculosis bacillus.
The leukocytes of the guinea-pig do not have this effect.
These experiments do not permit us to draw positive con-
clusions concerning the action of living macrophages and
lymphocytes on tubercle bacilli, but it appears that rabbit
macrophages may have a protective action against these
organisms. (5) Extracts from rabbit, guinea-pig, and cat
macrophages have no hemolytic effect upon the erythrocytes
of chickens, goats, rabbits, or guinea-pigs.1
1 This is interesting in view of the statement made by Vaughan (Med.
News, December 15 and 22, 1894), from which the following quotation is
taken: "On March 19, 1894, I inoculated rabbits 1, 2, 3, 4, 5, 6, a and
6 with a virulent culture of the (tubercle) bacillus. Animals from 1 to 6
inclusive had had previous treatments with a 1 per cent, solution of nucleinic
acid as follows:
March 9 10 13 14 15 16 17 19
Amount of solution in
cubic centimeter . . 0.3 0.5 0.6 0.7 1111
a and b had had no nuclein. All of the animals were half-g own, and weighed
respectively: No. 1, 714 grams; No. 2, 724 grams; No. 3, 740 grams; No.
4, 729 grams; No. 5, 647 grams; No. 6, 614 grams; a, 709 grams; b, 705
grams. On July 6, 1894, I killed No. 6, a and b. No. 6 weighed at this
time 1557 grams. I found a nodule the size of a pea at the point of inocu-
lation. In all other respects this animal was normal, I could find no
452 PROTEIN POISONS
The above-mentioned facts, ascertained by experimental
study, have been cited to show that the existence of both
intra- and extracellular germicidal substances in the blood
has been demonstrated. These substances have been
called alexins, antibodies, and by other names. It seems
to us that at present they should be classed as ferments.
As has been said, we do not know much about ferments,
but it is evident that these bodies have a lytic action. They
break up complex molecules into simpler bodies. Their
primary function seems to be to supply the cells which
elaborate them with food. In doing this they also protect
the cells, to which they belong, by the destruction of harmful
bodies both particulate and formless, both animate and
inanimate. The digestive ferments of our alimentary
canals serve the same double purpose. Any unbroken
foreign protein having found its way into the blood is a
poison, but in the alimentary canal it is broken up and
prepared as a food for the body cells. Every living cell
has such a ferment or such ferments. Their presence and
bacilli in the nodule, which was rubbed up with beef-tea and injected into
the abdominal cavity of guinea-pig No. 186, weighing 385 grams. On
October 10, 1894, I killed this pig, and found a nodule the size of a pea at
the point of inoculation. Three small tubercles were found in the peri-
toneum; the omentum and liver were filled with tuberculous nodules. One
testicle was tuberculous. This is an interesting case, showing that the
germ, which had not spread in the rabbit, had, when transferred to the
more susceptible guinea-pig, induced a widespread tuberculosis
"Rabbit a weighed 1030 and b 1100 grams. In both, nodules as large
as filberts were found at the point of inoculation, and smaller nodules in
the omentum. On October 10, I killed No. 1, weight, 2134 grams. This
animal was found to be wholly free from tuberculosis. On October 4, I
killed No. 2, weight, 2150 grams, which was found perfectly normal. No. 3
was found dead October 2. Postmortem examination showed a pear-shaped
tumor in the omentum. This was three inches long and one and one-half
inches in diameter at the base. It consisted of three cysts, which contained
very fetid pus, in which were found a short bacillus and a large micrococcus.
There was no evidence of tuberculosis. No. 4 was killed October 10,
weight 1990 grams. I found a small nodule at the point of inoculation.
This was not attached to the abdominal wall, but was in the connective
tissue, between the skin and the muscle. I could find no germ. In all other
respects this rabbit was normal. No. 5 was killed October 10, weight 2000
grams, and found perfectly normal
"These experiments indicate that rabbits may be rendered immune to
tuberculosis by previous treatment with yeast nucleinic acid "
THE PHENOMENA OF INFECTION 453
activity distinguish living from non-living matter. We
have taken the leukocyte as an illustration, but the bac-
terium is also supplied with its ferments, some of which are
intra- while others are extracellular. We do not know that
all cells elaborate both kinds of ferments, but all have at
least one kind.
Before proceeding further it may be well to call special
attention to some of the properties of these ferments. The
extracellular ferments are diffusible. They not only pass
out of the cells in which they are prepared, but they diffuse
more or less widely through the medium which surrounds
the cell. This suggests that in molecular structure they
are relatively simple. At least some of them may pass
through membranes and collodion sacs, as is shown by the
fact that bacteria and other proteins enclosed in such
receptacles and left in a body cavity are destroyed.
The extracellular ferments are, in part at least, filterable,
passing with more or less readiness through porcelain.
In their activities they are easily affected by modification
in the medium through which they diffuse. The alexin
of the blood serum is highly sensitive to the salt content
of the serum, and by variations in this the activity of the
ferment may be hastened, lowered, or wholly arrested.
The same is true of bacterial ferments. In one species of
animal a given bacterium multiplies with great rapidity;
in another it grows slowly, while in a third it cannot grow
at all. There are like variations in individuals of the same
species. The extracellular ferments, at least some of them,
are susceptible to slight changes in temperature. It is
believed that every ferment has its optimum temperature,
but the range in which continued activity is possible is
narrow with some and relatively wide with others.
The intracellular ferments are non-diffusible, or at least
less diffusible than the extracellular. They remain in the
cells in which they are elaborated. They cannot be extracted
from the cell by indifferent solvents. As a rule, they can
be obtained from the cell only after partial or complete
destruction of the cell. Some, probably most, are best
454
PROTEIN POISONS
extracted from the cell with dilute alkali, while others are
best obtained by dilute acid. In either case the reagent
must not be strong enough to destroy the ferment itself.
They are non-filterable, or pass through filters slowly and
imperfectly. We suspect that their molecular structure
is relatively complex, or that they are more colloidal than
the extracellular ferments. Under natural conditions the
intracellular ferments act only on those bodies which are
taken into the cell. The inclusion of bacteria by phago-
cytes is essential to the digestion of the former by the
intracellular ferment of the latter. This is a phenomenon
which may be seen, but cell permeation by foreign bodies
is certainly necessary before such bodies can be acted upon
by the intracellular ferments, and occurs with soluble
proteins as well as with participate ones. The intracellular
ferment bears a wider variation in temperature, and is not
so easily and delicately influenced by variations in the
composition of the medium in which the cell exists. So
far as we know the intracellular ferments do not diffuse
from living cells. They are, however, recognizable in the
fluids of abscess cavities as the leukocytes disintegrate.
We are of the opinion that they are essential constituents
of the chemical structure of cells. The reason for this
belief will be developed later. The extracellular ferments
may be regarded as secretions of cells. Much has been
written about cellular and humoral theories. In our opinion
every living thing has a chemical structure, which we may
designate as a cell if we wish, understanding that a cell is
not necessarily something that can be seen, and that it
may possess widely different degrees of lability, but we are
quite certain that there is no ferment which is not the
product of life processes. We have been somewhat sur-
prised to find it stated that our own theory of protein
sensitization or anaphylaxis is a humoralistic doctrine.
All ferments are products of life processes, and all life
processes are more or less responsive to outside influences,
to change in environment. In our opinion the most valuable
fact that we have learned in the study of protein sensi-
THE PHENOMENA OF INFECTION 455
tization is that life processes manifested through ferment
action are modified and may be modified at will by changes
in environment. The blood-serum and organ extracts of
normal guinea-pigs do not digest egg-white, but these
fluids from an animal sensitized to this protein do have
this action. The virus of smallpox is pathogenic to the
man who has never had smallpox, and has not been vacci-
nated, but to the man who has had the disease or been
properly vaccinated the virus of smallpox is non-pathogenic.
We explain this, and in our opinion, the experiments of
Pirquet have so demonstrated, that this is due to the fact
that the ferments of the man's body cells have been so
influenced by the disease or by vaccination that they have
acquired a new function — that of digesting and thus
destroying the virus of the disease. If this explanation be
true, it opens up a wide field for the possible extension of
the beneficial effects of preventive treatment.
There is another point of difference between intracellular
and extracellular ferments, which is of the greatest impor-
tance in a study of the phenomena of infection. The extra-
cellular ferments are comparable to those of the digestive
juices of the alimentary tract in the higher animals. They
roughly prepare foods for the cells. Their function is
solely a lytic one. They break up complex proteins into
simpler bodies, but the products thus formed are not,
without further treatment, ready to be built into the struc-
ture of the cell. Proteins in the medium are rendered
soluble by the extracellular ferments. They are so altered
that they may be taken into the cell, but they are not so
patterned that they are ready to be built into the structure.
They are fitted for absorption, but are not ready for assimi-
lation. The extracellular ferments are in a sense destructive
agents. They break down complex molecules into simpler
structures. The intracellular ferments are constructive.
They are cell builders. They shape the material brought
them and fit it into place. They build up specific proteins.
They convert the raw material brought them into specific
proteins, bacterial, vegetable, or animal. This does not
456 PROTEIN POISONS
mean that the intracellular ferments have no cleavage
action. They chip the rough stone so that it fits in at the
right place. It is by virtue of their activity or through their
agency that cells grow and multiply. In case of an infec-
tious disease the intracellular ferment of the infecting
organism during the period of incubation converts man's
proteins into bacterial proteins, and continues to do this
with more or less success during the course of the disease.
This seems to be accomplished in pome diseases, at least,
like typhoid fever, without any marked disruption of the
cells of the man's body. The bacteria multiply rapidly
during the period of incubation, and at this time the man is
unconscious of the fact that his body is serving as a culture
flask. We must conclude from this that the conversion of
human proteins into typhoid proteins in the growth of the
infecting agent is not accompanied by the liberation of the
poisonous group in the protein molecule. This group,
probably attached to other groups, or as a constituent of
a more complex group, is used in the construction process.
The poisonous group is common to all proteins. The syn-
thesis of specific proteins from other specific proteins is
accomplished without the -liberation of the poisonous
portion. It is one of the building stones, and changes in
specificity do not occur in this, but in the secondary or
characteristic groups. This is, in our opinion, the explana-
tion of the fact why incubation — a period of rapid repro-
duction in the infecting agent — proceeds without any
recognizable disturbance in the health of the host. The
typhoid bacillus therefore does not feed upon the cells of
the man's body, but upon the formless, soluble proteins.
Cell building is accompanied by the absorption of the
poisonous group in the proteins serving as food. However,
when the body cells become sensitized and elaborate a
ferment which breaks down the bacterial cells, the poisonous
group in the proteins of the latter is set free, and it is the
effect of this poison that develops the symptom complex
of the disease. The symptoms of one infectious disease
differ from those of another largely according to the organ
THE PHENOMENA OF INFECTION 457
or tissues in which the infecting agent is located. In acute
miliary tuberculosis and in typhoid fever, both conditions
arising from a bacteremia caused by different organisms,
the symptoms are only too frequently identical, and it is
only by bacteriological methods, a suggestive history, or
the finding of a preexisting tuberculous focus in some part
of the body that a differential diagnosis may be reached.
A cholecystitis is the same, not only in symptomatology,
but frequently in gross pathology as well, whether the
infecting organism be the pneumococcus, the streptococcus,
the colon, or the typhoid bacillus. The most skilful diag-
nostician cannot tell from the symptoms alone the specific
bacterial cause of a meningitis.
During the period of incubation of an infectious disease,
the infecting organism supplies the ferment, the body pro-
teins constitute the substrate, the process is essentially
constructive, no poison is set free, and there are no recog-
nizable clinical symptoms. During the active progress of an
infectious disease, the body cells supply the ferment, the
infecting organism constitutes the substrate, the process is
essentially destructive, the protein poison is set free, the
symptoms of disease appear- and life is placed in jeopardy.
Our work seems to show that the body cells, when over-
whelmed with a foreign protein of the blandest kind — such
as egg-white — may fail to function and death may result.
There is no reason for suspecting that in these cases there
is any cleavage of the foreign protein or the liberation of
any poison. The body cells are simply clogged with the
foreign protein and fail to function. We are not sure that
this phenomenon has any parallel in the infectious diseases.
There is, however, something closely related to it in cholera
infantum, cholera nostras, and Asiatic cholera.
We have already referred to the fact that ferments may
be modified in their activities. These modifications may
be so radical that it is generally believed that cells may be
trained, as it were, to develop new ferments. There can
be no doubt that change in environment does alter activity
as manifested through the ferments. As we have stated,
458 PROTEIN POISONS
it seems to be a biological law that when a living cell is
brought in contact with or permeated by a foreign protein,
it tends to furnish a ferment which will digest and destroy
the foreign body. The ferments of the cells of man's body
may be modified or new ones developed by (a) disease,
(6) vaccination, and (c) sensitization. Many of the infec-
tious diseases give immunity to subsequent exposure. In
some of the chronic infectious diseases the altered behavior
of the body cells to the infecting agent is evident even
while the disease continues.
That the tuberculous animal behaves differently from
the non-tuberculous on receiving injections of the tuber-
culin protein, whether it be in the form of the living bacillus,
in dead cells, or in solution, has been abundantly demon-
strated. Before Koch gave us tuberculin, Arloing and
Courmont had come to the conclusion that the tubercle
bacillus produces soluble substances which reduce the
natural resistance of the body and render it more susceptible
to reinfection. This corresponds closely with the first
impression made by observation of the phenomena of ana-
phylaxis; the impression that led Richet to select this
term. In 1891, Koch described a perfect example of protein
sensitization as we understand it today. He stated that
when a healthy guinea-pig is inoculated with the living
tubercle bacillus there is no change at the site of inoculation
until from ten to fourteen days later, when a hard lump
forms, finally opens and ulcerates, and continues until the
animal dies. On the other hand, when a tuberculous guinea-
pig is Inoculated with the living bacillus, on the second or
third day a lump forms, soon becomes necrotic, falls out,
ulcerates for a time, and finally heals without any infection
of the neighboring lymph glands. In 1897 Trudeau observed
that when healthy rabbits receive injections of virulent
cultures in the eye, there is little to be seen for about four-
teen days, when with increasing vascularity tubercles form
in the iris, after which inflammation extends and the eye
is practically destroyed within from six to eight weeks.
Like treatment of tuberculous rabbits develops an iritis
THE PHENOMENA OF INFECTION 459
within from two to five days, but at the end of the second
or third week, at a time when the controls begin to develop
destructive changes, the inflammation begins to subside.
Later studies have confirmed and amplified these, and it
has been found that death may be induced within twenty-
four hours by injecting a large amount of a living culture
into a tuberculous animal.
The same difference between healthy and tuberculous
animals has been observed in their response to injections
of dead cultures of the tubercle bacillus. The first observa-
tion along this line, so far as we know, was made by Strauss
and Gamaleia, who found that when large numbers of dead
tubercle bacilli are injected into tuberculous animals death
results, while similar amounts are without immediate effect
upon healthy animals.
When we come to tuberculin, every phase of its action
or its failure to act is explainable on the ground that the
tuberculous animal is a sensitized one. Koch found that
0.5 gram of his preparation killed tuberculous guinea-pigs,
and induced no symptoms in healthy ones. A fraction of
1 mg. may cause marked symptoms in a tuberculous man,
while many times this amount is borne easily by a healthy
man. The inflammatory reaction about local tuberculous
lesions caused by injections of tuberculin is explained by
the fact of the high degree of sensitization in their localities,
and the cleavage of the bacilli. The ophthalmic, cutaneous,
subcutaneous, and intravenous tests with tuberculin are
all typical sensitization reactions. Even in the failure to
respond to tuberculin seen in advanced tuberculosis we have
the condition known as anti-anaphylaxis, which simply
means that the anaphylactic ferment is partially exhausted
by the large amount of material supplied by the bacilli
in the body.
There is a second factor in the failures of advanced cases
of tuberculosis to respond to the tuberculin test, which has
been generally overlooked, but to which we have already
referred in our discussion of so-called anti-anaphylaxis.
This is the fact that in such cases the body is saturated with
460 PROTEIN POISONS
the products of the digestion of tuberculoproteins. It is a
well-established fact that the accumulation of fermentation
products retards and finally arrests the fermentative process.
The instance in point is a perfect illustration of this law of
fermentation. It must be evident from this how unscientific
it is to treat advanced tuberculosis with tuberculin. It
has been argued that the tuberculin reaction is not an
example of sensitization, because as the treatment proceeds,
larger and larger doses of tuberculin are necessary in order
to induce the reaction, as shown by the development of
fever. To anyone who has followed the evidence that we
have given so far, the explanation must be plain. It lies
in two facts, of each of which we think that we have given
abundant proof. In the first place we have shown that a
certain degree of tolerance for the protein poison is easily
and quickly established. In the second place, the accumu-
lation of fermentation products retards fermentation.
Tuberculosis, in most instances at least, begins as a strictly
local infection. This is true even when the first recognition
of it has been in the acute miliary form. There has been
a previous focal infection. Only those body cells in the
immediate vicinity of the infection are sensitized, and only
these supply a ferment capable of digesting the tuberculo-
protein. It may well be that in this stage benefit may be
secured by the proper use of tuberculin, which may act as
a sensitizer, and develop more of the ferment to split up and
destroy the tubercle bacilli. It should be always borne in
mind that tuberculin contains a poison and should be used
with caution.
There is another line of evidence that in tuberculosis there
is a condition of specific protein sensitization. This is to be
found in the fact that this disease is much more deadly in
lands and among people who have recently come under its
influence than it is where it has prevailed for many genera-
tions. In other words, the widespread and long-continued
existence of the disease, slowly, and at the cost of much
sickness, and many deaths, brings a certain degree of
immunity. The readiness with which the North American
THE PHENOMENA OF INFECTION 461
Indian has succumbed to this disease is a striking illustration,
and Calmette has recently collected additional evidence
on this point. He states that tuberculosis is being widely
disseminated among peoples who have until recently been
free from it. The world-wide wanderings of the white man
are carrying the disease to every people, from the Laplander
and Esquimaux of the Arctics to the negroes and Malays
of the tropics. Iceland, the Faroe Islands, and the steppes
of Russia are being infected, and in these new regions
tuberculosis exists in its most speedily fatal forms. The
same author points out that recently discovered methods
for the recognition of this disease, even in latent states,
shows that among Europeans not more than 7 or 8 per cent,
reach more than twenty years of age without receiving the
infection. Those who survive the first infection become
more or less immune, and after that develop, when they do
acquire the disease, the more chronic forms.
Romer1 concludes that the less widely tuberculosis is
distributed among a people the greater is the case mortality,
and the wider the distribution the smaller is the case
mortality.
Still another fact of importance is that the most speedily
fatal forms of tuberculosis, such as the miliary and menin-
geal, are more frequent among children than among adults.
There is another matter of much importance in this
connection which we must discuss. We have found the
tubercle bacillus highly resistant to lytic agents, and it
appears that its long experience as a parasite has led it to
protect itself with deposits of wax and fat, but proteolytic
enzymes digest the most firm proteins. Friedberger has
found that at least some strains of this bacillus are digested
by the serum of healthy guinea-pigs, and the researches of
Markl, Bail, and Kraus and his students have shown that
tubercle bacilli placed in the peritoneal cavity of tuberculous
animals respond to Pfeiffer's reaction. Some strains are
dissolved in the peritoneum of healthy guinea-pigs, but
1 Beitrage z. klinik d. Tuberk., 1912, xxii, 301.
462 PROTEIN POISONS
dissolution occurs more promptly and more completely in
the peritoneum of a tuberculous animal. The healthy animal
may have to depend upon its phagocytes to combat the
invading bacillus, but the tuberculous animal supplies a
specific proteolytic enzyme, and to this the fresh invader
succumbs.
Nature is slowly immunizing the white man to tuber-
culosis, and the question arises whether or not the process
employed by Nature -can be aided in any way. There is
before the medical profession at this time no greater question
than this: Is it possible to aid in eradicating tuberculosis
by vaccination? As Romer says, the problem of securing
immunity to tuberculosis with a non-infective virus is of
great practical importance, and recent work brings the
possibility of doing this more and more to the front. What
we need is a vaccine. Various methods of modifying the
tubercle bacillus so that it could be used as a vaccine have
been tried. The bovo vaccine of Von Behring was tried,
but the increased resistance given by it was found to be of
short duration. Attempts to reduce its virulence by age,
heat, chemicals, and by submitting it to ultraviolet and
other rays and emanations have been made. What we
need is a tubercle protein sensitizer. It should be soluble,
and it should be free from the poisonous group in the protein
molecule. In our opinion the nearest approach to this
desired substance is the non-poisonous portion of the tubercle
protein. So far we have not been able to secure a uniform
product. Some preparations seem to fill every requirement.
They sensitize animals to the unbroken bacillus, dead or
alive, and in surface tuberculous lesions they cause inflam-
mation about the tuberculous area, and we have seen
the tuberculous tissue slough off and complete recovery
result; but other preparations made from the same cellular
substance by the same method seem inert. We have had
similar difficulties with the sensitizing groups from other
proteins. Some preparations from egg-white sensitize to
unbroken egg-white, while others seem wholly without
effect, and still all are prepared from the same material and
THE PHENOMENA OF INFECTION 463
in the same way. Evidently the sensitizing group in the
protein molecule is a highly labile body and susceptible
to influences which so far we have not been able to recog-
nize. We have no difficulty in obtaining the poisonous
group uniformly, but it is otherwise with the sensitizing body.
Further work along this line is needed, and if an efficient
and uniformly reliable sensitizer for the tuberculous protein,
free from the poisonous group, can be secured, all children
should be vaccinated for tuberculosis; then with protection
against natural infection the restriction of tuberculosis
will be as completely under man's control as is that of
smallpox. It should be clearly understood that the pro-
tection afforded by vaccination is relative and not absolute.
The studies inaugurated by Wright have demonstrated
that vaccination is of service not only in prevention, but
also in cure. Bacteria and protozoa are particulate, and in
many diseases they are confined to limited localities. As
we have seen, sensitization may also be local. No body
cell is sensitized against a foreign protein until the latter
comes in contact with the former and penetration of the
body cell is probably essential to the most efficient sensi-
tization. The microorganisms of acne are located in the
cutaneous tissue, and being particulate and not in solution,
the area sensitized by them is small, if there be any sensiti-
zation at all. By vaccine therapy the area of sensitization is
greatly extended and the amount of lytic agent formed and
made available is greatly increased. This being in solution
and diffusible, digests and destroys the bacteria located in
the skin. The same is true of the treatment of localized
tuberculosis, or of any other localized infectious disease.
In vaccine therapy, as in vaccination, the great need is for
soluble sensitizers free from poisonous content. When these
are secured, and not until then, we may develop a vaccine
therapy along scientific lines, and expect to secure important
results with it.1
1 The following pages are taken with but little change from an article by
Vaughan, Jr., in "International Clinics."
464 PROTEIN POISONS
The importance of sensitization as a factor in the case of
tuberculosis is evident in the widespread use of tuberculin
as a diagnostic measure. The various reactions of the
body to tuberculin, whether they occur as the general
reaction following subcutaneous injections, or as the more
local reaction following the vaccination of the skin with
tuberculin, the application of a tuberculin containing
ointment to the skin, or the instillation of tuberculin into
the conjunctival sac, are all evidences of the sensitization
of the body of the tuberculous individual to tuberculin.
Thus, when a small amount of tuberculin is injected into
the fluids and tissues of a normal individual, no effects are
noticeable, since the enzyme which causes proteolysis of
tuberculin is not present in the body. When, however, the
same amount of tuberculin is injected into the tuberculous
individual, it practically corresponds to a second injection
of this foreign protein. The enzyme present in the body of
the tuberculous individual attacks the tuberculin, liberating
the poisonous cleavage products, which in turn give rise
to the well-known symptom-complex designated as the
tuberculin reaction. In addition to the general symptoms,
such as fever, which accompany the presence of protein
poisoning within the body, poisonous proteins have a
decidedly irritant local effect upon the tissues with which
they are brought directly in contact. This is seen in the
hyperemia and inflammation of the peritoneum in cases of
infection within the abdominal cavity, and is also evidenced
by the local reaction of inflammatory type following the
application of tuberculin to the mucous membrane or the
abraded skin of the tuberculous individual.
Sensitization to tuberculin may be either local or general
in type, as is quite evident to anyone who has employed
the conjunctival test as a means of diagnosis in tuberculous
disease. This test consists in the application of a 1 per cent,
solution of specially prepared tuberculin to the conjunctival
sac. The reaction following this method of applying the
tuberculin test may be divided into two distinct types,
the first of which we may call the reaction of general sensi-
THE PHENOMENA OF INFECTION 465
tization, or the tuberculous reaction, in contradistinction
to the second, or the reaction of local sensitization.
When a solution of tuberculin is applied to the conjunc-
tival sac of a tuberculous individual, no changes are usually
noticed for an interval varying from six to forty-eight
hours. At the end of this time there is a slight smarting
or gritty sensation complained of, the patient often referring
to it as a sensation of sand in the eye. The examination of
the conjunctiva at this time reveals a reddening and swelling
of the mucous membrane of the lower lid and the caruncle.
This inflammatory reaction gradually increases in intensity
until from ten to fifteen hours have elapsed, at which time
it has usually reached a maximum, and after which a gradual
recession occurs, until at the end of from two to four days,
occasionally after a longer interval, the conjunctiva has
again regained its normal appearance. At the height of
the reaction, and on awakening in the morning, it is not
uncommon to observe a slight fibrinous or fibrinopurulent
exudate accompanying the inflammatory reaction.
When a solution of tuberculin is applied to the eye of a
normal individual, no reaction is obtained. If, however, a
second instillation is made in the same eye after an interval
of seven days, a reaction will be observed in a large propor-
tion of cases. This reaction is quite distinct from that
previously described as occurring in the eye of the tuber-
culous individual. The reaction is rapid in appearance,
explosive in type, and subsides with great rapidity. Thus,
it is not rare to find, as a result of a second instillation, within
from three to four hours after the application, a highly
inflamed conjunctiva associated with considerable chemosis
of the lids and a profuse purulent discharge. The symp-
toms, however, in spite of their severity, rapidly subside.
These differences in type of reaction find a satisfactory
explanation if we consider the fact that in the tuberculous
reaction we are dealing with what may be termed a phenom-
enon of general sensitization. In this case the cleavage
of the tuberculin introduced within the conjunctival sac
is brought about through the action of the proteolytic
30
466 PROTEIN POISONS
enzyme which has been developed in the body of the tuber-
culous individual as a result of his disease. During this
cleavage certain poisons are liberated which act as irritants
to the conjunctival mucous membrane, and the degree of
irritation present will be directly proportionate to the amount
of toxic cleavage product present at a given time. However,
the amount of cleavage product present at a given time
will be determined by the rate of proteolysis, which depends
in turn upon the quantity of proteolytic enzyme directly
available. Since this enzyme is available only in such
proportions as may be present in the circulating fluids of the
conjunctiva, it necessarily follows that only a small amount
can be operative at a given time. The result is that we
have a foreign protein slowly broken up, with the liberation
of a small quantity of irritant poison over a considerable
interval of time. For this reason the reaction of general
sensitization is slow in its development, maintained at its
maximum for a considerable period, and subsides gradually.
When tuberculin is instilled into the eye of a normal
individual, no apparent result is obtained, since no ferment
is present in the body capable of splitting up tuberculin.
However, as a result of the instillation itself, certain cells
of the mucous membrane are stimulated to produce a
specific ferment which will be stored up as a zymogen for
future use. If subsequently a solution of tuberculin is
brought in contact with these sensitized cells, the zymogen
is activated, liberated in a concentrated form, and splits
up at once all of the tuberculin introduced. The result is
that we obtain the reaction characteristic of local sensitiza-
tion, which is rapid in onset, comparatively severe in type,
and disappears with great rapidity. Owing to the high grade
of inflammatory reaction obtained in connection with the
second instillation, it is well to use a more dilute solution
of tuberculin and to ask the patient to present himself for
examination within from two to four hours following the
instillation. At this time, if any noticeable redness is
present, the eye should be thoroughly washed out with a
saturated solution of boracic acid in order to remove any
THE PHENOMENA OF INFECTION 467
excess of tuberculin which may be present. The information
obtained through the employment of the second instillation
in an individual who has previously failed to react is of
value in that it indicates that the body cells of the patient
are capable of producing a ferment which will split up
tuberculin, and consequently we should have obtained a
primary reaction provided the individual was actively tuber-
culous. Failure of the ophthalmo-reaction occurs under
the following conditions: (1) In early cases in individuals
who are incapable of producing the specific ferment. Such
individuals will fail to react to the second instillation.
(2) Normal individuals who are not afflicted with tuber-
culous disease will fail to react to the primary instillation.
They may, or may not, react to the second instillation,
depending on whether or not their body cells are capable
of producing the specific ferment. (3) Patients suffering
from acute tuberculous disease or advanced cases fail to
react to either the first or the second instillation. In these
cases the failure of the reaction is due to the exhaustion of
any specific ferment which may have been present through
the overwhelming of the system with tuberculin, or to the
accumulation of split products, as has been stated.
While the importance of sensitization in connection with
the infectious diseases is not as yet thoroughly appreciated,
later investigations have been conducted largely along
these lines. Thus, sensitization is undoubtedly an important
factor in the treatment of bacterial diseases through the
employment of vaccine therapy. This is true whether the
vaccine employed consists of the whole bacterial cell or
the split products, such as those obtained after our method.
The injection of foreign proteins as such into the body
always represents an abnormal condition. The symptoms
of sensitization following the administration of horse serum
in man may be divided into two classes, according to the
interval of time elapsing between the administration of the
serum and. the development of symptoms. In general, it
may be stated that symptoms of sensitization, provided they
occur at all, show themselves either very shortly after the
468 PROTEIN POISONS
administration of the serum, or, if not at this time, on the
seventh to the tenth day following the injection. In instances
in which effects are not noticeable until from seven to ten
days following the injection, the symptoms are largely
confined to those of peripheral irritation, as evidenced by
urticarial lesions accompanied by intense itching. On the
other hand, in cases in which a reaction follows within
twenty-four hours, the symptoms of poisoning are more
pronounced, and where occasionally a fatal result follows
it occurs usually within an hour after the injection. In
these cases, which are fortunately rare, we find that the
symptoms are very similar to those obtained through the
injection of the poison obtained by Vaughan through
protein cleavage. Thus, Gillette1 reports the case of
an asthmatic fifty-two years old, to whom he gave 2000
units of antitoxin globulin, administered under the left
scapula. While dressing, following the injection, the
patient complained of a prickling sensation in the chest and
back of the neck. He at once sat down in the chair and
complained of inability to breathe. The physician felt his
pulse and found it full and regular. Immediately thereafter
the patient was seized with a tonic spasm, during which
death ensued, the whole interval elapsing between the
injection and the fatal outcome not exceeding five minutes
in duration. In spite of the rapidity with which death
occurred in this case, we can still recognize evidences of
the three stages characteristic of fatal protein poisoning:
the stage of peripheral irritation indicated by itching
sensations in the skin, the stage of partial paralysis or
weakening of the lower extremities, and the convulsive
stage, during which death occurred.
In cases of sudden death following within a few minutes
after the injection of horse serum, it is not infrequent that
one of the stages is absent or ill-defined. Thus, in the
instance cited above, the loss of ability to move the lower
limbs was not specifically mentioned, although in other
1 Jour. Amer. Med. Assoc., January 4, 1908, p. 40.
THE PHENOMENA OF INFECTION 469
reported cases patients have before death remarked on their
inability to walk.
It is quite evident from a study of the untoward results
following the administration of horse serum, that the
apparent differences existing between immediate mani-
festations and those occurring after an incubation period
of from seven to ten days are of degrees of intensity rather
than of character of the poisoning. Thus, in the instances,
fortunately rare, in which death occurred within thirty
minutes following the injection, the symptoms are due to
the liberation of a fatal amount of poisonous substance at
once, and in instances in which alarming but not fatal
symptoms arise shortly after injection, recovery from the
intoxication is usually prompt and complete. On the other
hand, where symptoms appear only after an interval of
from seven to ten days, and are confined to those of peripheral
irritation, as evidenced by the development of urticaria,
we find that complete recovery is slow and tedious.
That such differences should exist appears but natural
when we consider the mechanism involved in sensitization,
and the fact that immediate effects are due to the injection
of the serum into a sensitized individual, whereas the remote
effects are to be looked upon as a manifestation of the
sensitization of the patient as the result of the injection
itself. In the first instance the individual has stored up in
his body cells a ferment which, liberated by the injection
of the serum, splits up the foreign protein introduced at
once, and sets free all of the poison contained therein imme-
diately. The symptoms resulting therefrom are necessarily
acute in character, sudden in development, and transitory
in nature, since the effects of the poison rapidly disappear.
In the individual developing symptoms after an incubation
period of from seven to ten days, conditions are decidedly
different. In this case no special ferment capable of pro-
ducing proteolysis of the foreign proteins contained in
the serum is present within the body at the time of injec-
tion, and as a result the foreign proteins continue to
exist as such within the body for a certain length of time.
470 PROTEIN POISONS
However, under their influence certain body cells are stimu-
lated to produce a ferment which will split up the foreign
substances into simpler non-specific bodies. The fact that
animals injected with serum do not become hypersensitive
to a second injection until after the lapse of from seven to
ten days indicates clearly the length of time necessary for
the new ferment to be formed in appreciable amount.
Symptoms developing after an incubation period are,
therefore, to be explained by the fact that the foreign
proteins still existing in the tissue are acted upon by the
enzyme called forth by their presence. Under these con-
ditions, however, no large amount of ferment will be active
at any given time, and consequently the amount of poison
liberated through protein cleavage at any one period will
be small in amount, although the cleavage itself will con-
tinue over a comparatively long interval and the resulting
poisoning will be more chronic in type. This affords a
plausible explanation of the fact that late manifestations
of serum sickness are milder in character, being confined
for the most part to those of peripheral irritation, and also
of longer duration.
Provided the theory advanced above is correct, one would
expect to find a difference in the time interval elapsing
between the injection of the foreign serum and the subse-
quent appearance of symptoms of poisoning, depending on
whether the individual had previously been treated with
serum or not. While it is true that, in many instances in
which alarming symptoms develop immediately, a history
of previous treatment is unobtainable, the results quoted
by Pirquet are interesting as bearing on this point. Thus
it has been found that of 214 individuals who developed
symptoms after the first injection of serum for therapeutic
purposes, 111, or 51.8 per cent., manifested symptoms of
poisoning on from the seventh to the tenth day inclusive;
while in 172 patients who received a second injection, 89,
or 51.7 per cent., showed signs of poisoning within the first
forty-eight hours.
As has been previously mentioned, alarming symptoms
THE PHENOMENA OF INFECTION 471
following the use of diphtheria antitoxin and other thera-
peutic sera are fortunately so rare that they should not be
considered too seriously when indications for the use of
such sera arise. However, there are certain precautions
which can and should be employed, and which may aid
materially in avoiding the untoward effects following the
administration of these remedies. Much has been accom-
plished by the efforts of various pharmaceutical houses to
prepare an antitoxin from which a large proportion of the
foreign albumin contained in horse serum has been removed,
and such products should be used exclusively whenever
possible. It wrould, furthermore, appear to be well to make
a preliminary test with regard to the sensitiveness of any
given individual to the serum employed. This may be
done by the injection of a very small test dose (0.05 c.c.) of the
serum and watching for rapid evidences of toxic action in
the patient. Alarming signs, if they occur, develop usually
within an hour after treatment; and if no sign of poisoning
occurs within this time, it may be safely assumed that the
individual does not contain within his body the special
ferment required to split up the material injected, and a
second injection may be made with impunity, provided
the interval of time elapsing between each is not sufficiently
long to admit of the development of sensitization. In a
disease such as diphtheria this is, of course, a matter which
does not enter into consideration in the treatment of any
given attack. The preliminary injection of atropine has
been advised by Auer, who found that 18 out of 25 sensi-
tized guinea-pigs which had been given atropine sulphate
recovered from the second injection, while of 24 untreated
controls, only 6 survived.
When symptoms of sensitization appear immediately or
soon after the injection, the use of ether by inhalation is
to be recommended, as Besredka found that in experi-
mental sensitization animals narcotized with ether did not
succumb to the second injection. When it is ascertained
that a given individual is sensitive, and nevertheless the
use of therapeutic sera is imperative, it may be given by
472 PROTEIN POISONS
fractioning the total amount of serum, and, instead of
using a single dose, give several doses at frequent intervals
over a considerable period of time. In this way it may be
possible to exhaust the ferment present in the body for
the time being, after which further amounts may be given
with impunity. A temporary exhaustion of the special
ferment would explain cases of diphtheria such as are
reported, in which, although alarming symptoms followed
the first injection, on account of the condition of the patient,
a second injection seemed advisable, and was given within
a few hours without any untoward effects whatever. Experi-
ences such as this find their analogy in experimental work
in the fact that sensitized animals which have recovered
from the poisoning following a second injection are not
again susceptible to that particular protein until after the
lapse of several days.
The idiosyncrasies which certain individuals possess
with regard to certain protein articles of diet would appear
to be explainable on the ground that, through some abnor-
mal condition of the intestinal mucosa, certain protein
substances are allowed to enter the body in an unchanged
state. The symptoms which develop are certainly strikingly
suggestive of those described as appearing in connection
with sensitization. The symptom most constant in appear-
ance is an urticaria, more or less generalized in extent and
accompanied by intense itching. Thus, Bruick reported
the case of a man who reacted with urticaria every time
after eating pork. Smith reports the case of an individual
who developed a severe urticaria within a short time after
partaking of any article of diet which contained buckwheat.
Numerous cases of severe urticaria accompanied by dyspnea,
and occasionally incoordination of the lower extremities,
have been reported as occurring in susceptible individuals
after partaking of any food containing egg albumen, and
one of us has recently had opportunity to observe an indi-
vidual who developed a generalized urticaria accompanied
by marked edema and intense itching within half an hour
after partaking of peas as an article of diet.
THE PHENOMENA OF INFECTION 473
The most striking peculiarity mentioned in connection
with the above idiosyncrasies is the rapidity with which
the symptoms of poisoning develop after the introduction
of the attending cause within the alimentary tract. As
has been mentioned, these individual peculiarities are
possibly best explained by the supposition that the individual
has become sensitized to certain specific proteins, the sensi-
tization arising from the fact that the particular protein
has gained entrance into the body through the intestinal
mucosa in an unchanged state. When this occurs a foreign
protein is present in the tissues and fluids of the body, and
to counteract the abnormal condition thus produced certain
body cells are called on to develop a proteolytic ferment
which will have for its function the cleavage of the particular
protein present in any given case. This ferment, once
formed, is stored up in certain cells as a zymogen for future
use. The same protein cleavage then occurs as normally
takes place within the intestinal canal, with the important
difference that the toxic substances formed are liberated
within the body itself and consequently are capable of
exerting their harmful action. That these ferments are
not present in the body in inexhaustible amount is shown
experimentally by the fact that animals which have recovered
from the effects of sensitization following a second injection
of egg albumen are not subsequently sensitive to this
protein until after the lapse of several days. This is un-
doubtedly due to the fact that the ferment has been largely
used up in bringing5 about the cleavage of this particular
protein and time must be allowed for the body cells to
produce an additional amount. In other words, it is possible
in a susceptible individual to destroy their susceptibility
with regard to any particular protein through an exhaustion
of the special ferment present in their body. This is well
illustrated clinically by the following example: A woman
who was fond of strawberries, but developed an intensely
disagreeable urticaria after each indulgence, was accustomed
to eat this fruit two or three times during the season.
Finally, being firmly convinced that the rash was^simply a
474 PROTEIN POISONS
nervous manifestation, she determined to eat them con-
tinuously, in order, as she said, "to break herself of the
nervous habit." After the first week no unpleasant symp-
toms whatever were noted following the daily use of this
article of diet. This appeared to the patient to be an
entirely satisfactory proof of the effectiveness of Christian
Science, and yet the phenomenon is explainable on a
rational basis. The daily use of strawberries had led
to an exhaustion of the special ferment in her body, and
subsequent indulgence was consequently not followed by
untoward symptoms. Whether or not the experience was
repeated during the succeeding summer we have not been
able to ascertain.
In conclusion, we may state that sensitization primarily
represents an important phenomenon of lytic immunity.
Sensitization occurs whenever a foreign protein as such
gains entrance into the fluids and tissue of the body, and
results from the development within the body of a special
ferment which will attack the particular protein introduced.
When individuals become sensitized through the introduc-
tion of dead protein substances, such as egg albumen or
horse serum, the results obtained prove unfavorable to the
individual. In these cases our attempt should be to bring
about a desensitization of the individual through the
exhaustion of the special ferment. On the other hand,
sensitization occurring as a result of the entrance of bacterial
cells into the body represents a beneficial process, and
plays an important part in the development of active
immunity to the specific infections. Since, under ordinary
circumstances, pathogenic bacteria represent the only
proteins which gain entrance into the body in an unchanged
state, we may conclude that sensitization arises as an
attempt of nature to protect the individual against bacterial
disease.
INDEX
AMINO acids, 74, 78
Anaphylactic state, 242
Anaphylatoxin, 296
Anaphylaxis, 214
in vitro, 274
mechanism of, 247
passive, 254
Animals, action on, 119
Anthrax protein, 189
Arthus phenomenon, 262
Anti-anaphylaxis, 258
BACILLUS, action of, 119
colon, 37
diphtheria, 38
participate proteins, 17
pathogenicity of, 21
reducing action of, 62
Bacterial cellular substance, 37
/3-iminazolyle thylamin, 291
CANCER cell, specific ferments of, 416
Cellular protein, hydrolysis of, 81
substance, 37
action of, 121
carbohydrates in, 57, 66
chemistry of, 52
diamino-acids of, 74
fats of, 60
immunization with, 138
mono-amino acids of, 78
nucleins in, 73
of pneumococcus, 205
proteins of, 52
Cleavage of proteins, 95
Cultures, massive, 29
DEFIBRINATED blood, the poisonous
action of, 334
Diamino-acids in cellular substance,
74
Diseases, infectious, 23
E
EGG-WHITE, cleavage of, 98
the disposition of, when intro-
duced parenterally, 355
Ergamin, 291
FEVER, acute fatal, 381
continued, 374
digestive action of blood in, 395
followed by immunity, 397
intermittent, 386
protein, 373
remittent, 387
GASTRIC juice, action of, 42
Group, poisonous, 19
sensitizing, 234
HAPTOPHOR, properties of, 112
Histamin, 291
476
INDEX
IMMUNITY, natural, 24
toxin, 27
with split products, 157
Immunization with poisonous por-
tion, 138
with residue, 144
Infection, the phenomena of, 436
KYRINS, the, 295
M
MONO-AMINO-ACIDS in cellular sub-
stance, 78
N
NUCLEUS, chemical, 20
PARENTERAL digestion, 342
Peptone, the fats of, 342
Pneumococcus, cellular substance of,
205
Poison, action of, 125
crude soluble, 101
immunization with, 138
physiological action of, 315
protein, 17, 284
Protein fever, 373
poisons, nitrogen in, 111
sensitization, 25, 214
Proteins, cleavage of, 95
particulate, 17
physiological action of, 320
RED corpuscles, production of fever
with, 391
Reinjection, the, 243
Residue, immunization with, 144
SENSITIZATION, cellular, 321
period of incubation, 241
protein, 214
symptoms of, 245
Sensitizers, 219
volatile, 231
TANKS for massive cultures, 30
Theories, 323
Theory of Friedberger, 324
of Nolf, 340
of Vaughan and Wheeler, 327
Toxic sera, 264
Toxicity of extracts from normal
tissue, 325
Toxogens, 266
Trypsin, action of, 43
Tubercle bacillus, cellular substance
of, 164
cleavage of, 165
poison in, 165
residue of, 165
split products of, 164
toxophor of, 178
Tuberculosis and sensitization, 181
VACCINATION, 24
Vaccines, 26
Vegetable proteins,
fever with, 403
production of
*w
•a
3
-P
o
PH
University of Toronto
Library
DO NOT
REMOVE
THE
CARD
FROM
THIS
POCKET
Acme Library Card Pocket
Under Pat. "Ref. Index File"
Made by LIBRARY BUREAU