1 OT ee ee ers pS oo 2 ‘ > - ~“ E i Ae. * meek 2 s 4. Spree se er e eee freee enorene ee Pei paren gecesi Pacer nak rte emt Spee : : e rd abt ~ > oh Raveena Yate RATORIES SCIENGE & MEDICINE DEPT, THE JOURNAL OF IMMUNOLOGY Libravv pul of ‘sic Rutverstiy af Gore VOLUME VII Library Srtonvi of wyaicre Rutmerstey vf Tartmio BALTIMORE, MD. 1922 CONTENTS NuMBER 1, JANUARY, 1922 A Study of the Virulence of Meningococci for Man and of Human Suscepti- bility to Meningococcie Infection. George D. Heist, Solomon Solis- ohen anay Wiver SOMs-COnen etc s. \< co c5- cries, quae memes see eee The Action of Various Metallic Salts on Hemolysis. Helen A. Purdy and Tae ol PUG dl va 0 echoed cee en es te ey ca aR aA An Allergic Reaction of the Tuberculous Uterine Horn. G.H.Smith...... Relationship of Various Antiorgan Sera. Moyer S. Fleisher.............. Bacillus Diphtheriae. Immunological Types; Toxin-Antitoxin Relation- Ships We. iH, Paxson and Baward Redowrtz. 00°"! or eh ie. oes. eae NuMBER 2, Marcu, 1922 Bronchial Asthma and Allied Conditions. Clinical and Immunological Observations. Nils P. Larsen, Royce Paddock and H. L. Alexander.. Studies in Specific Hypersensitiveness. I. The Diagnostic Cutaneous Reaction in Allergy. Comparison of the Intradermal Method (Cooke) and the Scratch Method (Schloss). Aaron Brown..................-. Studies in Specific Hypersensitiveness. II. A Comparison of Various Pollen Extracts with Reference to the Question of their Therapeutic Value im Hay Meyer. Albert Vander Veer, dis.) co... fae see ee Studies in Specific Hypersensitiveness. III. On Constitutional Reactions: The Dangers of the Diagnostic Cutaneous Test and Therapeutic Injection giAilergens= (Robert Ad Cooke: : 20h a sae en wi eel si a erent Studies in Specific Hypersensitiveness. IV. New Etiologic Factors in Bron- GhimleActhimease io bert) Any! Cooker ta Sah a: tude moan) deve epandt apetecuio cate Studies in Specific Hypersensitiveness. V. The Preparation of Fluid Ex- ‘tracts and Solutions for Use in the Diagnosis and Treatment of the Allergies, with Notes on the Collection of Pollens. Arthur F. Coca.... Studies in Specific Hypersensitiveness. VI. Dermatitis Venenata. W. C. SS LDS ane ORS he 5k exes Gs avaigeatane eo ee Ream un dicein ooo inal Rk Studies in Specific Hypersensitiveness. VII. The Age Incidence of Serum Disease and of Dermatitis Venenata as compared with that of the Natural Mileroves | Ad OUr pe MOO atin eck a's ctaryaile Wo clea kia; ala okie da abet nserieiegane Studies in Specific Hypersensitiveness. VIII. On the Relative Suscepti- bility of the American Indian Race and the White Race to the Allergies and to Serum Disease. Arthur F. Coca, Olin Deibert and Edward F. OM ISLCR TSP SPOR I RSEREE ES 2.0.7 Crea RSME Coe EIRP nee ne tee IE NO Pe y 81 97 113 119 163 179 193 lv CONTENTS Studies in Specific Hypersensitiveness. IX. On the Phenomenon of Hypo- sensitization (the Clinically Lessened Sensitiveness of Allergy). Robert NuMBER 3, May, 1922 Immunological Studies on Types of Diphtheria Bacilli. I. Agglutination Characteristics. II. Protective Value of the Standard Monovalent Antitoxin. William H. Park, Anna W. Williams and Alice G. Mann.... 243 The Relationship of Lipoids and Proteins to Serum Reactions in Tuber- culosis. W. Ray-Hodge and M. F. MacLennan...................... 253 _ The Toxicity of Acids for Leucocytes, as Indicated by the Tropin Reaction. Alsee°Cs Bivens - 2 cic .vecos oe eck cus dates Geet ere ale eae eee 271 Numser 4, Jury, 1922 A Serological Study of the Gonococcus Group. John C. Torrey and George EPP AcOll. 5 a9 Gi aioe wien oS lS s+ ain bracllor che eee) Seep ok tee ee aoe ee 305 Studies on Acute Respiratory Infections. XI. A Serological Study of Alpha Streptococci from the Upper Respiratory Tract. AgnesGoldman. 361 NuMBER 5, SEPTEMBER, 1922 On the Photolability of Serum Complement. E.G. Lundberg..... Ba eg it 389 Prophylactic Treatment for Rabies by Means of Standardized Glycerinated Virus. dames wiclivame. Phillips 3 yeaa eee ee 409 A Study of the Precipitin and Complement Fixation Reactions with Tuber- culous Exudates with Special Reference to Tuberculous Pleuritis. MSS OPAWA. 2. .nh0ie «02 ere inls 3:2 < wnyahg + cate eee ieee 423 On the Origin and Nature of Alexin (Complement) in Guinea-Pig Blood. L. Ie MA OYEISO Mts. css oo aloe acs otn.s Roce Oe Ie Spee 435 NuMBER 6, NOVEMBER, 1922 A Study of the Hemolytic Antibody-Antigen Combination. H.W.Cromwell. 461 Studies on the Toxicity of Human Blood Piasma for Guinea-Pigs. I. Rela- tive Toxicity of Fetal and Maternal Plasma. S.A. Levinson.......... 497 Studies on the Toxicity of Human Blood. Plasma for Guinea-Pigs. II. Coagulation Toxicity.) ~S.- A... Levinson’.(0022"* 222s. == i.e 511 A STUDY OF THE VIRULENCE OF MENINGOCOCCI FOR MAN AND OF HUMAN SUSCEPTIBILITY TO MENINGOCOCCIC INFECTION! GEORGE D. HEIST, SOLOMON SOLIS-COHEN, anp MYER SOLIS-COHEN From the Jules E. Mastbaum Research Laboratory of the Jewish Hospital, Philadelphia, Pennsylvania Received for publication October 28, 1921 Epidemic cerebrospinal meningitis presents a number of immunological and epidemiological problems. Chief among these are the réle of the meningococcus carrier in the dissemination of the infection and the factors that make him a distributor rather than a victim. Why, in an individual who develops meningitis, do the meningo- cocci invade the deep tissues, and in the carrier confine themselves to the surface of the mucous membrane? Morphologically and culturally the cocci from the two sources are identical; both series are capable of setting up infection in susceptible men and animals; both are agglutinated by polyvalent, antimeningo- coccic serum. In searching for an explanation, two phenomena must be con- sidered. One is the virulence of the meningococci for human beings; and the other is the susceptibility of human beings to the attacks of meningococci. It is important to recognize that these are two separate and distinct phenomena—that the dis- tinction is not one of phrasing, only. We have not to do, as in 1 Read before the joint meeting of the American Association of Pathologists and Bacteriologists and the American Association of Immunologists, at Cleve- land, Ohio, March 25, 1921. The preliminary work required by this study was made possible through the kindness of Mr. Samuel S. Fels of Philadelphia. The experiments with meningo- cocci were chiefly carried out by the late Dr. George D. Heist at the Base Hospital, Camp Zachary Taylor, Kentucky, in 1919. To Major Herbert Fox, U.S. A., who commanded the Base Hospital laboratory, that author desired to express his gratitude for kindly advice and support. 1 THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 1 2 G. D. HEIST, S. SOLIS-COHEN AND M. SOLIS-COHEN the case of resistance and susceptibility, with two sides of the one shield—with a single problem positively or negatively stated—but with two different problems. The one concerns a variable x, relative to a constant y; the other concerns a variable y, relative to a constant x. In final result, increasing either of these factors may, indeed, be equivalent to decreasing the other; but the process—the mechanism—is different. And it is the latter we have to study. The virulence for lower animals of the microdérganisms in question is known to vary within wide limits, and there is no a priori reason why their virulence for man should not vary similarly. That individual human beings (and animals) differ greatly in their susceptibility to all bacterial infections is a matter of common knowledge; and the meningococcic infection does not differ essentially from any other. The object of the present study is to throw light upon each and both of the two phenomena mentioned: (a) meningococcic viru- lence and (b) the varying susceptibility of human beings to meningococcic infection. It will be convenient to consider the last-mentioned, first. THE WHOLE-BLOOD TEST FOR IMMUNITY MAY BE APPLIED TO MEASURE BOTH BACTERIAL VIRULENCE AND ANIMAL SUSCEPTIBILITY 1. Susceptibility.2 Several years ago the writers described a test that has been applied successfully by others as well as them- 2 We use the term susceptibility in the common acceptation of lack of resistance, without reference to any special factor or factors; high susceptibility being low resistance and low susceptibility being high resistance, and soon. Representing complete susceptibility by S and complete resistance by R, the mathematical formula would be R + S = 0; hence R = —S and S = —R. So too, if virulence be represented by V and non-virulence by A, then V + A =0,andso on. But to get rid of the minus signs we may take z as a variable indicating resistance, which at the point of mazimum susceptibility becomes equal to 0; and y as a variable indicating virulence which at the point of avirulence becomes equal to 0. zx If then 7 represent immunity we have the formally = 7 and the value of 7 will vary inversely as y with a constant z, and directly as x with a constant y; while the effect of multiplying y by n will be the same as that of dividing x by the same factor, and vice versa. VIRULENCE OF MENINGOCOCCI FOR MAN 3 selves, to the demonstration, 77 vitro, of the relative susceptibility of animals to various infections—particularly pneumococcic, meningococcic and diphtheric infections. In this test, whole, coagulable blood as it comes from the vessel, is brought in con- tact with small numbers of bacteria adhering to the inner wall of capillary glass tubes. The bacteria and blood are sealed in the tubes and incubated for twenty-four hours. At the end of that time the contents are blown out and examined micro- scopically to see whether or not the bacteria have developed. In a general way, all the writers who have worked with this capillary tube test agree that if the bacteria fail to multiply in twenty-four hours, sufficiently to be seen on microscopical examination, it may be definitely inferred that the animal or man from whom the blood was taken, possesses some degree of immunity against that particular bacterium. In estimating what degree, of course, careful attention must be paid to the essential conditions of the test; particularly as regards the numbers of bacteria seeded in the tubes. Heist, Solomon Solis-Cohen and Myer Solis-Cohen (1) culti- vated pneumococci in the blood of rabbits, mice and pigeons. To quote from the conclusions reached through this study: ‘If small numbers of pneumococci are seeded, by a suitable method, in pigeon blood before it coagulates, the pneumococci fail to multiply. On the contrary, if pneumococci are seeded in mouse or rabbit blood before it coagulates, the pneumococci grow with great vigor.”’ It will be remembered that the pigeon is immune toward pneumococcic infection, while mouse and rabbit are not. Bull and Bartual (2) carried the work further. They observed that if incubation of the tubes containing pigeon blood is pro- longed sufficiently, growth always occurs, but not until after a period of delay or lag. Lag was absent when mouse or rabbit blood was employed. ‘They found no proof of a true bactericidal action, but considered the phenomena observed to indicate merely a delay or inhibition of growth by the blood of the immune animal. In a subsequent personal communication, however, Bull states that he has been convinced that the whole fresh blood 4 G. D. HEIST, S. SOLIS-COHEN AND M. SOLIS-COHEN of resistant animals may kill a certain number of pneumococci. He gained this impression from the fact that capillary tubes charged with high dilutions of culture and then filled with blood failed to show growth after any length of incubation period. He still feels that it is not possible to say definitely whether this result is due to actual killing of the organisms or to an inhibition of multiplication with spontaneous death. The real point, however, is this—and Bull now concedes it—that given a certain relation between the nwmber of pneumococci (or other organisms introduced) and the volume of blood, the organisms perish in the blood of the immune animal. Matsunami and Kolmer (3) incubated meningococci in capillary tubes of the whole blood of rabbits and mice. Rabbits are immune from meningococcic infection and mice are fairly sus- ceptible. In rabbit blood little or no growth of meningococci had taken place after twenty-four hours incubation, but in the blood of the mouse, growth was vigorous. Matsunami, in a later communication (4) states that, since the publication of the article referred to, he has found evidence that the action of the immune blood on meningococci is truly bactericidal. Black, Fowler and Pierce (5), who seeded with undiluted virulent cultures of the typhoid bacillus and of Bacillus dysen- teriae the blood of rabbits immunized against these organisms and incubated the tubes for different periods of time, found that the organisms were destroyed at the end of five minutes, no bacilli or recognizable fragments being seen in the leucocytes. By means of our method they were able to demonstrate in rabbits the progress of the development of an artificially induced immunity against these organisms; and they conclude, as the result of comparative studies with various methods, that the bactericidal power of blood, thus determined, “‘is the most dependable criterion of the actual immunity of the animal.”’ More recently (6) we have studied the behavior of diphtheria bacilli in the whole coagulable blood of certain animals, in- cluding the guinea-pig, rat, mouse and rabbit—in each instance using a number of animals of different ages and sizes—and in the defibrinated blood of these species and of the steer. The whole, VIRULENCE OF MENINGOCOCCI FOR MAN 5 coagulable blood and the defibrinated blood of all the animals tested, with the exception of the rat, failed to inhibit the growth of diphtheria bacilli. The whole, coagulable blood of the rat proved bactericidal; its defibrinated blood did not. 2. Virulence. In a second series of experiments, instead of starting with a single virulent strain of pneumococci as they did in the first series, Heist and Solomon Solis-Cohen (7) re- versed their angle of approach. They made a normal rabbit their fixed point, and cultivated strains of pneumococci of differ- ent degrees of virulence in normal rabbit blood. They found that in such blood virulent pneumococci grew much more readily than did non-virulent ones; from which fact, taken with their measurements and inoculation experiments, they concluded—and this is the conclusion that interests us most at present—that the ‘“mathematical expression”’ (in terms of the dilution of culture) “of the ability of a strain of pneumococci to grow in the blood of normal rabbits in vitro is an expression of the virulence of that strain for rabbits. ”’ Myer Solis-Cohen, Heist and Borow (8) studied the virulence of 88 strains of diphtheria-like bacilli by growing a 1:10 dilution of a suspension of a twenty-four hours serum culture of each strain in guinea-pig’s whole, coagulable blood, and noting the shortest time in which growth of the bacteria could be de- tected. Each strain also was injected into an unprotected guinea-pig and into a control guinea-pig that had received 500 units of antitoxin one hour before the injection of the culture. Fourteen of the cultures failed to grow up within eight hours in guinea-pig’s whole coagulable blood, and had no effect when injected in either the unprotected or the control guinea-pigs. These were regarded as Bacilli hoffmanni. Seventy-four of the 88 cultures grew up decidedly within eight hours in guinea-pig’s whole coagulable blood. Sixty-five of these, which killed only the unprotected guinea-pigs, the controls being saved by the protecting antitoxin, were regarded as virulent Klebs-Loffler bacilli. Three others of the seventy-four cultures that grew up within 8 hours produced in the unprotected guinea-pigs, a transient illness, from which the control pigs 6 G. D. HEIST, 8. SOLIS-COHEN AND M. SOLIS-COHEN were slightly protected. ‘These were regarded as Klebs-Léffler bacilli that were ‘‘slight”’ or “weak” toxin producers. The re- maining 6 of the eight hour cultures killed both the unpro- tected and the protected guinea-pigs and were regarded as probably identical with the diphtheria bacilli that are usually found at autopsy in the blood and internal organs. In the case of the pneumococcus, meningococcus and diphtheria bacillus, then, animal susceptibility (both as to species and as to individuals of a given species) and bacterial virulence (both as to species and individual strains of a given species) may be clearly demonstrated and, to some extent, measured in vitro by culti- vating the bacteria in whole, coagulable blood.’ THE MENINGOCOCCUS PROBLEM It was thought that similarly, if meningococci from the spinal fluid of cases of meningitis and those from the throats of carriers were grown under like conditions in whole, coagulable human blood, some light might be thrown on their comparative virulence; and that if a strain of meningococci from a severe case of men- ingitis were grown in the blood of different individuals, some light might be thrown on the comparative susceptibility of men to meningococcic infection. Before concluding, however, that what is true of animals is likewise true of man, we experimented with bacteria that are known to be practically harmless to men under ordinary condi- tions, and bacteria known to be commonly harmful to men, to see whether or not the harmful ones would grow more readily in human blood than those that are harmless. It would have been ideal if throughout this study the different bacteria examined could have been tested against the blood of the same individuals. But this was not practicable. We were obliged to grow each strain in the blood of a different group of persons. The groups were made as large as possible in order that an average might be reached. The tests were made on 3 We are not here concerned with the explanation of the phenomenon; or with the part played by the factors or process of coagulation; merely with the test and its results. VIRULENCE OF MENINGOCOCCI FOR MAN 7 young men in good health. In addition to theblood test, cultures from the throats of the majority of the groups were studied for meningococci. In two instances carriers were detected, and note of this is made in the tables. OBSERVATIONS ON COLON BACILLI The colon bacillus was selected as a type of the ordinarily non- virulent group of organisms. Colon bacilli are sometimes found in diseased human organs, and probably at times initiate morbid processes when introduced into regions other than their normal habitat; but there is no definite colon bacillus malady in the sense that there is a pneumococcus malady or a meningo- coccus malady. In this and subsequent experiments our technic as previously described (1) was followed exactly. A strain of colon bacilli was isolated from feces. The forty- eight hours growth on an agar plate was washed off with sufficient broth to make a dense white emulsion, containing 4 to 6 bil- lion bacilli per cubic centimeter. Four dilutions of thissuspension were made in broth, 1:10, 1:100, 1:1000, 1:10,000. The five tubes of a Wright’s many-stemmed pipette were filled to a mark; one from the undiluted suspension and one from each dilution. The tips of the tubes were then touched to moist, sterile gauze, and the fluid was permitted to run out, leaving on the inner wall of each tube a film containing bacteria. Blood from a finger prick was immediately filled in up to the mark, and the tubes were sealed and incubated for twenty-four hours. At the end of that time the contents were expelled on to glass slides, stained with dilute carbol-thionin and examined. If great numbers of bacteria were seen in every field a double plus was recorded; if only a few were seen in some fields and in some none at all, a single plus; if only two or three bacteria could be found in the entire preparation, a plus-minus was put down; and if no bacteria at all were seen, a minus. As a further control to each experiment, one set of tubes was loaded with defibrinated blood instead of whole, coagulable blood. When growth took place in the defibrinated blood and 8 G. D. HEIST, S. SOLIS-COHEN AND M. SOLIS-COHEN not in the whole coagulable blood, we were assured that there was active in the whole, coagulable blood some anti-bacterial factor (or factors) lacking in the defibrinated blood.! TABLE 1 Cultures of Bacillus coli in whole, coagulable human blood DILUTIONS OF A HEAVY SUSPENSION OF B. COLI IN BROTH Undilu- 1:10 1:100 1:1000 1: 10,000 MST TAD CHARS Re AN SOR — _ _ _ — IVIGIEN OLY Baa otk oe tent Senses - a -- ~ — IMATE OSRIOE ee ee cia hsjapcicles eevee ae -- _ _ _ — Man nO. 622 sce cated Se ees _ _ _ _ — WEED TONG Es Eee em Ou a nin on Mis _ - - _ - IVEAIEN ORO eee ere chad Tees Oe ee _ - = _ — Manan OMROEE ee a, Ge Cee BART _ _ _ — —- IV Teamnera oil Oat tees ey niche Medea hte see eps _ + — _ — IV USINTO Meet ee oso! Aer ye ep ae gt _ — IVETE TO Uo ett © hn heirs oe ore oe te — = IMIANENO A Gta Reo eaeotae an _ — NTN OLAS aay. eect hls ky die iae — = | Controls: + - + Dextrose. brathes.c6sice ckceehioeek ++ | ++ S535 | lata Se | As will be seen in table 1 almost no growth of colon bacilli had occurred in the whole blood of normal men after twenty- four hours incubation. OBSERVATIONS ON HEMOLYTIC STREPTOCOCCI Few microérganisms are responsible for more infections of the human race than are the streptococci. Streptococci are able to invade any and all tissues, setting up severe and ofttimes fatal 4 Were it not for our previous observations of this phenomenon, it would be astonishing; for a priori one would expect defibrinated blood to be the more bactericidal, owing to the release, by the disintegration of leucocytes, of bacteri- cidal endolysins allied to the microcytase of Metchnikoff. ee N, Hie * Soya. eee ae % “So 7) as a = ——s VIRULENCE OF MENINGOCOCCI FOR MAN 9 inflammations and intoxications. This may not, indeed, be true of all the Gram-positive, chain-forming microérganisms we are accustomed to call by that name. Some of them may be as harmless as the colon bacillus in its normal domain. But it is certainly true of many of the hemolytic variety. ‘Therefore, if our hypothesis concerning the relation of whole coagulable TABLE 2 Cultures of Streptococcus hemolyticus from a leg ulcer, in whole, coagulable human blood DILUTIONS OF TWENTY-FOUR HOURS GROWTH ON BLOOD AGAR TUBE SUSPENDED IN 2 CC. DEXTROSE BROTH Wadi: ith ted 1:100 | 1:1000 | 1:10,000 I NUESTOVS cg Bae IRE) Sk a a he rE ++ +--+ Sas Sic apse erat oO ean Oe ee ie ee eee -- -- — -- _ IVAN EINO smote re. eee eas ee eee a _ — _ + UIC Taio Cat he Oa ei eae deed ++ ++ + — + TAIT ON eos ys se Ha Sane ees Was ++ ++ +--+ — = NUTT aes: MOS See Acie ve ne aR Pe ne ck Se ae = = Wistar cerns cies cesses Nerves 55 + + _ ae = Vien O MES mere rere Aer see So ++ ++ ++ _ = Mina nernrOre ek aueeietet, eu ne mb! os +++ ++ — — INIT aVo\a TI) ee ere Sea eee ee pe _ _ — — a 1s) [erro hy GG) es OS Ree ee a oe ++ ++ +--+ a5 = YS W Og OX Osc] Ce iat: Sere i ars ane --+ ++ —-+ Sar = INET MYO) Sal layin. sea OO ee ++ + ++ = ++ = NLT LOWa)ea C2 BS as eg Ge eV ore + - Soar — = IVIMeT npn es eee lk cis. cate nea hay ake ++ ++ ++ _ — Controls: RBar Re 2s Ses Cyr aS erg ge hey +--+ + +--+ + -t- +b Defibrinated blood.............. ++ “E+ +--+ ++ ++ Bloodsbrouneycet a neccaee ce + <5 == =F a blood to immunity be correct, a hemolytic streptococcus, fresh from some human infection, should grow in whole coagulable human blood much more readily than the colon bacillus does. To test this a hemolytic streptococcus was isolated from a severe leg ulcer, following a shell wound, and cultured in the blood of aseries of men. The result is shown in table 2. 10 G. D. HEIST, S. SOLIS-COHEN AND M. SOLIS-COHEN In evaluating this result the difference between the number of streptococci in the fluid from which tube 1 was seeded, and the number of colon bacilli in the fluid from which tube 1 was seeded in the preceding experiment, must likewise be taken into consideration. The undiluted suspension of streptococci con- tained about one billion bacteria per cubic centimeter, while the undiluted suspension of colon bacilli contained four to six times that number. Yet there were sufficient streptococci present to produce vigorous growth in the blood of all but two of the fifteen men tested, even in the 1:100 dilution; while colon bacilli grew in only one of the entire lot of capillary tubes. This seems to give considerable emphasis to the trustworthiness of the criterion. But we were not entirely satisfied with this experiment, despite the growth of the streptococci. They did, indeed, grow better than did colon bacilli; yet the whole, coagulable blood of normal men was able to offer considerable resistance to them, as compared with the lack of resistance shown by defibrinated blood. When a virulent pneumococcus is cultivated in the blood of rabbits, it grows just as well in whole blood as in defibrinated blood. Experiments were made with many bacteria, but in no in- stance could a parallel be found in man to the pneumococcus phenomenon in the rabbit. The failure is not surprising. The observations were confined to microérganisms such as pneumo- cocci, Bacillus influenzae (Pfeiffer), staphylococci, Bacillus diphtheriae and others, which ordinarily affect man, and excluded what may be called the artificial or extraordinary infections of man, such as glanders and anthrax. There are not so many infections to which man is universally susceptible, of which the agent is known, and, at the same time, suitable for study by the capillary tube method. Bubonic plague is one, but we were not able to obtain a freshly isolated culture of Bacvllus pestis. However, the two experiments recorded are sufficient to prove that what Matsunami and Kolmer, and Bull and Bartual, as well as ourselves, have found to be true of animals,is true of human beings also; namely, that there is a certain correlation VIRULENCE OF MENINGOCOCCI FOR MAN il between the ability of bacteria to grow in the whole, coagulable blood of man and their virulence for man. We felt therefore that we could look with favor on the method, as a means of testing the relative virulence of strains of meningococci for human beings. EXPERIMENTS WITH MENINGOCOCCI The medium on which the meningococci were grown was beef infusion agar plus 7 per cent horse serum and 1 per cent dextrose. The sterile serum and dextrose solution were added after cooling the agar to 55°C., and the reaction was then ad- justed, sterilely, to pH 7.4. The atmosphere in which the meningococci were grown was kept saturated with moisture by closing all the ventilating holes and placing large, shallow pans of water on the floor of the incubator. The method adopted by the United States Army for the detec- tion of meningococcus carriers was followed, excepting that serum-dextrose-agar replaced blood agar. It should be stated, specifically, that in this study no coccus with the cultural and morphological characteristics of meningo- cocci was accepted for the work unless it was also completely agglutinated by polyvalent anti-meningococcus serum, 1:100, with clearing of the supernatant fluid after 18 hours at 55°C. Moreover, any trace in the controls of agglutination with normal horse serum, 1:50, or with salt solution, ruled the culture out. Normal and para serums were used in typing. Distinct agglu- tination by one serum alone was accepted as sufficient evidence that the strain belonged to that type. Complete sedimentation with clearing was not required. All strains which were not agglutinated by either normal or para serum were called irregular. After isolation, the meningococci were transplanted at least twice daily and no transplant over twelve hours old was used for the whole, coagulable blood tests. The use of a young culture of meningococci for tests with whole blood is a most vital point. Usually but a few transplants intervened between the isolation of the meningococci and their culture in whole, coagulable blood. 12 G. D. HEIST, S. SOLIS-COHEN AND M. SOLIS-COHEN The suspensions for the tests were made by washing off the growth of a large agar slant with 2 cc. of warm serum-dextrose broth. All strains grew about equally well on agar so that each suspension contained approximately the same number of cocci. The suspensions were of such density that a wire held between them and the light was barely visible. This corresponded to from four billion to six billion cocei per cubic centimeter. The dilutions were made in warm serum-dextrose broth. All of the tables covering the meningococcus work have been placed together, with a few clinical notes concerning each case. Case 1. H. QO. White. Aged twenty-one. The attack began with headache and vomiting followed by unconsciousness. The patient was admitted to the hospital March 17% with a temperature of 101°. On March 18 his spinal fluid contained over 5000 leucocytes per cubic millimeter, 95 per cent of them being polynuclears. Gram- negative, intracellular diplococci were seen. They grew well on serum- dextrose agar and proved to be normal meningococci. They were cultured in whole blood, as shown in table 3. The patient was given serum, intraspinally and intravenously. After the third dose there was a marked improvement in his condition. His recovery was not interrupted by any complication. By April 1 convalescence was well established. During April, cultures of his naso-pharynx were taken fifteen times. Meningococci were found on five occasions. The positive culture of April 24 was tested with whole, coagulable blood, as shown in table 6. Case 2. E. B. White. Aged twenty-three. The attack began two days before admission to the hospital, with severe headache and pain in the calves of the legs. A maculo-papular eruption appeared over the entire body. The patient was admitted to the hospital on February 17, with headache and vomiting. His spinal fluid was cloudy and contained over 5000: leucocytes per cubic centimeter, ninety-five per cent of them being polynuclears. No bacteria were seen, but in a similar specimen, drawn the following day, Gram-nega- tive intracellular diplococci were found. They failed to Err on serum-dextrose agar. From February 17 to April 2 the patient was, alternately, desperately ill and slightly improved. His temperature ranged from 98° to 104°. Serum was given without apparent effect. In all 450 cc. were given VIRULENCE OF MENINGOCOCCI FOR MAN 13 intraspinally and 63 cc. intravenously. At no time could improvement be ascribed to the use of the serum. Meningitis was complicated at different times by pneumonia of one lobe of the right lung, endo- carditis, and orchitis. At times the spinal fluid was heavily clouded and at other times slightly opalescent. Repeated cultures were made from it, but it was not until March 22 that a growth occurred. The organisms proved to be meningococci. They were completely agglutinated by polyvalent serum, but not by either normal or para type sera. They were tested with whole, coagulable blood (table 4). The patient’s temperature reached normal by April 2, and two weeks later he was definitely convalescent. Frequent cultures were made from his naso-pharynx. A strain of meningococci was isolated on March 1. It was tested with whole,. coagulable blood (table 7). It needs but a glance at these tables to see the marked differ- ences, as to ability to grow in whole, coagulable blood, between meningococci from the spinal fluid of the two cases of meningitis and those from the throats of the carriers. The strain from the spinal fluid of case 1 (H. O.) (table 3) grew well in whole blood, and the strain from the spinal fluid of case 2 (E. B.) (table 4) grew very well. Among the carrier strains there was some variation. Strains from the throat of carrier M. W. (table 16) and from the throat of carrier G. R. 5. (table 18), grew to some extent, but the others scarcely grew at all. From the reports of others who have used the capillary tube method of testing bacterial virulence, and from our own pre- liminary experiments with both virulent and _ non-virulent bacteria, we are justified in concluding that this disparity of growth in whole, coagulable human blood connotes a like disparity in virulence; and that while the meningococci from spinal fluids are in general decidedly virulent for man, most of those from the throats of carriers are but slightly so. It is true that some strains from carriers are more virulent than others, and that some strains from spinal fluids are less virulent than others; but the carrier throat-strains are always relatively weak, the spinal-fluid strains always relatively powerful. 14 G. D. HEIST, S. SOLIS-COHEN AND M. SOLIS-COHEN TABLE 3 Cultures of normal meningococci, isolated March 18 from the spinal fluid of case | (H. O.), in whole, coagulable human blood parup | 2:10 | 1:100 | 1:1000 |1:10,000 | March 19—Dilutions of 8 hours culture: Wepre ee Seo eset ee sec ee eee Spee) Se Seem sees |) === 1M [EM es a0 oI Ee ei PR Pa a a ++ + +++ =_ a Wiantnomosen ceeds Cee ne eee Ree ++ ),4+4+)+4++ + ++ IIE Tc ha: Sp ee ep Raney ty ee ee + + _ _ Vis en Ove eis areheerers + _ -- _ —_ INEANSN Os, Sos och i Pe ike eect ~ _ Si = Manreno. [420 5. Sect contrat ee cee cee eee +++ _ _ — = ian hn Oe) Oe nck ae OAS ee eee ee ee — _ — — = WMianenose .62% soos 62 sas ca ce tee eet Moree ee — -- - — — iN EST Wa Oe ay ee Oe Ree oes Stents hiro Se - - = _ — IVE TIT Oo Sc tocine a es ord ane bee eee oie ee oe + _ — = WMiaay TOO satis coos ttre, Walelettoe oes Sea eie one creer -- — _ — — WaT OAL Oo sa cocvc sass ro Seore ctetee stein eee oe es er _ _ _ _ _ Controls: QUlett EE a oe ees sea e ae eer Seo estes fo siesta teeta ie Defibrinated blood: ...4.s.6 a2 ose +4 )/++/]/4+4+)++4+)]+4+4+ TABLE 7 Cultures of un-typed meningococci, isolated March 1 from the throat of case 2 (E. B.), in whole, coagulable human blood UNPI- | 14:10 | 1:100 | 1:1000 |1:10,000 LUTED March 12—Dilutions of 11 hours culture: Vins Wace ee ne ears nen oie ee eee a - _ = — a sO eee oe ee ake Rist ciate eee — a —~ _ _ Controls: 1,0 2% cpl & Cea nee Pr permed PM me TENG go? + -— -- o- + iDetibrinated blood... 45-4-00eee eee -- — - + + TABLE 8 Cultures of normal meningococci, isolated February 18 from the throat of carrier (A. F.W.), in whole, coagulable human blood. (Contact of case of normal menin- gitis. Quarantined twenty-three days. Throat cultures 10. Positives 2) UNDI- | 1:10 | 1:100 | 1:1000 |1:10,000 LUTED February 21—Dilutions of 9 hours culture: 167 E28 003 60 ae ne ge ye I Ce hace ee — — — = = Mla GG tee cee ys ec Mierke ce iecatis On cleraeetes — _- = = = 14 BS Wart © (0 as pea espe Ce I ng a 5 _ = = = = Controls: VAs EN heck arog aeons A Tore — _ _ = = Wekbrinated blood=-.-.-+-5-.26 es: eee “= - + + _ TABLE 9 Cultures of un-typed meningococci, isolated February 20 from the throat of carrier (T. R. B.), in whole, coagulable human blood. (Contact of case of normal menin- gitis. Quarantined twenty-nine days. Throat cultures 13. Positives 3) vepap | 1:10 | 1:100 | 1:1000 |1:10,000 February 21—Dilutions of 9 hours culture: | IVE ear TNO Desa rc. ask ai svsischin oe els ahs eal siolauptelolet aera — = = - _ INEATION Ose soe a aiokereketn tierels ins since cise eter _ _ - _ _ IYIENTIBE KOR N35 be Grab ieee Gao ain cx Ree a ae ane _ — _ _ — Controls: Merri bie eee ety ot. reek efi aeclcie sof chareiees iat + — + — — Detibrinated blood: oc. cc... isccjas<'eersisie + + + + | + TABLE 10 Cultures of normal meningococci, isolated March 14 from the throat of carrier(Y.), in whole, coagulable human blood. (Contact of case of normal meningitis. Quarantined eighty-three days. Throat cultures 30. Positives 18) vorgp | 1:10 | 1:100 | 1:1000 |1:10,000 March 17—Dilutions of 12 hours culture: | Walrer anos: 1h es eee eis Seen eis Area Eee een Ne con ae a -- _ — ~ WALA in a8 els ee sae a a aa A A ++ )+4-4 ~- _— | — Tilalnstaty Supe ice 0 ite ie Ae Ce aC aan STINE val IIE Hf ATL Pol fu | Z Controls: | SV Tins eee ett b tek Stn ote a le 7g sear |) Sear | seae arse |) Sear Dehibrinatedsploods. cece. ose eee = + + a - TABLE 11 Cultures of normal meningococct, isolated March 18 from the throat of carrier (H.T.), in whole, coagulable human blood. (Contact of case of normal meningitis. Quarantined thirty-seven days. Throat cultures 18. Positives 6) NDT At TO met 00 1:1000 |1: 10,000 March 21—Dilutions of 8 hours culture: Vieira One lee see eae tye epg soe cls, _- _ _ — _ IMAI GT Ose na er Te ye caaicoicias ace ++ )+4+4+ -— _ — MORIN OS Os te ee eC een Cite a sos +4 — -- _ - Mitanien 8 24 '.t62 rent baie PNT ore lara cay see + -- — _ -- Van MO Ole estes Reverie Ce yean soe ae ee _ _ — _ — VIPAT ORG si3) 5:5: ers cha oe RE re dedi oe ecine ~ -- — _ _ IVESINETY Osa s jc ines Siac ie eterno oe oe oe eee 4-4 —~ _ _ — IVE AINET OES ori c a.s Sova, oS aia eee eee — — _ ++ — VEN NORA) efor core c MER NG Pe crcte rae Ei _ —_ _ _ — Controls: LNCS GOR a i ek aa ie ERE sea ara ara dleaeae se Webrinated: bloodes.4.0 eee ++)++/]/++/])4+4+;)4++4+ TABLE 12 Cultures of normal meningococet, isolated April 5 from the throat of carrier (J. J.), in whole, coagulable human blood. (Contact of case of normal meningitis. Quarantined twenty-five days. Throat cultures 9. Positives 5) UNPI- | 4:10 | 1:100 | 1:1000 |1:10,000 April 8—Dilutions of 11 hours culture: MMT Ee Te O's CAL 55. oop Sales aot erknereS Soca cae —— _ -- — - IVI SRRIVOs 1c iste. SaaS wate caper Meera a _ — = = Wisin tniG Boece eee nos Bho er ate aie - - — _ — 1) CATV LG Ce Pe APS Sree kt 4 —~ _ _— — — Miaminios 5: oo tiscs ces cette io ae ee ee - — = — — IVEATTINIO AIO eer eaa nts sare Oe end oe ee a _ — _ — Wien 0 hse oo alis sda fe erate ee _ ah ee - _ IM IGH OV TAY Lp te eR eRe RE ea en cn GMs & _ - — _ -- IMamin O54 (9 oic 2 coe bed aed eee eee - _ ~ ~ _ Misrsnos 1 Oi. Aoscics chek en Ce Oe eee -- — + + _ IVEATIO PUL nic siete create sieaee ¢ hve tenet eee ~ — — _ - ABSA CO ee ae pS EF Ts 5 — -- _— _ ~ IVAN ETO’ eles os Sst = tacts cee ciclo Eee eae — + BE = — Miananon gla a. ssnrss chek sie teens oleae _ + _ _— —_ Wine Owls access Se eck, eee eee -- + os - _ Controls: 13 © Rage coe a een Meee)! =. + == =f = Denbrinated bloods...4.544.6 2266 eee: - a -- a+ TABLE 13 Cultures of normal meningococci, isolated April 5 from the throat of carrier (J.S.C.), in whole, coagulable human blood. (Contact of case of normal meningitis. Quarantined thirty-four days. Throat cultures 12. Positives 5) UNDI- ‘ | 9 : pNPT| 1:10 | 1:100 | 1:1000 |1:10,000 April 8—Dilutions of 11 hours culture: VANE TOS melee oie: <1c sets cet aie eee af = = = ee VERITON 2 eee tee Lie cies ee tee ee eee + = = = = WENO S| so then co ecsticoe-e (ae Aa ee =F =F = = = IMM BTTIO) FAG eee os yon oreinis Om Aaa ee BS =F = = = MamCnOL§ Oi.0) isc Oso... ae ee ae = a= = = aa Mammo: (658.2272. 300 Foe eee ee — — = = = VERTIGO Sic s cts ok eee ae ys PE =F = = a = WP TIO NE Bice ft Oe ose aa ee + + = = = MUCUETI ODOM rs Rui te BA tos a — “= = = = MEE OMI ons coh he tee, as 8 ee ee + + = = = RAO Yoana, Siiccos MA een oe ee Ie + _ = = = Controls: ~ IIE 18 Pa Ee nee NUR Tuer Dewey ee — t AF = Defibrinated blood. 225: ..<...-6.5..52 - + =e ae = TABLE 14 Cultures of normal meningococci, isolated April & from the throat of carrier (P van P.), in whole, coagulable human blood. (Casual* from Camp Diz. Quarantined forty-one days. Throat cultures 14. Positives 6) UNPI- | 1:10 | 1:100 | 1:1000 |1:10,000 April 11—Dilutions of 8 hours culture: MVIUTMBTVO ag deets,- crave shes o's: a) chasa i aboterete eect ehcc _ —: — _- Ts [ere tra GSR ee ey a) dA edn in on — _ _ _ _ VEAP O Mises: revere arorade iors toners ake lel od eee cul soker ara ++ {++ + — — Mane OiMae Sy sete Aer. estate eR Dene ++),++] — _ ~- WVEATUTI OL. MDGS Serle) Bok 5 ORS are niece ees —_ — _ - _ MV lenriern ones Gisk Sek at ete erevatn tarecaclae st eetenese rena _ ar _ = _ INTs AOR Teese cca tha pel sc hier secre te easier a's ake +. _ + _ _ VISTO MER OC ree Eee PSTN (0 tothe ss sch atermiorm -- + _ _ - AV ISrmeTi Ont O ecttesiccsinchierearioraiciom ove mianstas cereals - + + + + Ie TO EL OSes trees Sete coc eleva mala chere eee _ — _- — -- Tear er O MLW rvs ta a eee ease accion > MEISE we SiR eie eee + + _ - - IMAI S AY Tao) WAN ee ae 2 Ses re Sei ee eer Params ene + ++ _ — — TA Lew a SCV ONS, 1B 15 CRN ety > RS Pen ere Ree A A — = — = = IV Urea EN TN CORMAN et tere oe Me 5) PA SINID ENS. GRE T te vce renal — —_ _ — _ UN Terrains Oren per tees Bo Ware te wiehstiahote dicho tesele — _ — - a IMipirin@s tly Sess ee See oedema tocee ar _ _ _ — Controls: JN Eran 21 RG Pe aa On er a RT eens OE ate - + _ - Misnapiteeant. orev aeeees 4 ha ie aes + at ciel bea eua don as Defibrinatedtbloodsnss.. 4.2 --eeee ee ++t}++}]++;]4++)]4++4+ * Cultures were taken from the throats of all men transferred to Camp from any camp reporting cases of meningitis, before admission. These men were ‘termed ‘‘casuals.’’ TABLE 15 Cultures of irregular meningococci, isolated A pril 3 from the throat of carrier (J. P.), in whole, coagulable human blood. (Casual from Camp Dix. Quarantined forty-one days. Throat cultures 14. Positives 5) UNDI~ | 14:10 | 1: 100 | 1:1000 |1:10,000 April 11—Dilutions of 11 hours culture: IIHT TEVO)GY UY OA dcr Ne eh i ge - — = = 5 WH [ena veTeyO)s: 974.5. 08 eee Met Aa Ai oe a a = = = = = VILA ITO San oor erie rete) er tis es Sic os cureteeiete ++ — = = = IWIATRCINO RT eevee cesecbvoms oie siehe o Hee arisen — _ = = = VEDI GTT ON One ae ee Ak A a cee Se tall ane - — _ = = JME G0 HN OE Sacto SRR RICA ERA ara Rea _ — — = = AV URATN TNO et ind Seer cee oe ened coetas eee ae ict one - _ = = = IVEAINSTLOMG (Oise acicis cea cvente in oe ee eens _ _ _ = BS Terme 9)", center cio tetas Se eee ate — - = == = VETTE ONAL Os 2 c.\ 5 ste re oe eesteccco eee —- - — = = Controls: 114 Ufa ed 3 BRS ee aera ESA A RN ee ei +--+ | ++ = a = Detibrinated bloods i)... Me. vise els ++} ++ })]++)++)]+4++ 22 G. D. HEIST, S. SOLIS-COHEN AND M. SOLIS-COHEN TABLE 16 Cultures of normal meningococci, isolated April 10 from the throat of carrier(M.W.), in whole, coagulable human blood. (Casual from Camp Diz. Quarantined seventy-seven days, and cultures still positive. Throat cultures 24. Positives 12) UNDPI- | 14:10 | 1:100 | 1:1000 |1:10,000 April 14—Dilutions of 12 hours culture: INRA ATO self ccc.c reco os cks oe aus ee eee eee a a -- _ = Manon Os 2 a6 ste ads cts v8 ae eee t+})+4++ 4- + 4 IM aNANO koe. . Sootns «pinches eos eee ter a os + — = INGAAS Os BA mari cye oe bok a. " ae “we casei - 7 = ' *j vt Mey al lee & ~ ot ’ ; \ f rk oe © : Tee Stn Ly th : eet ae y ri Ay . i Sanity pine a } * Lone Pee i is at - —s yy ; ’ a ie, x i) ‘ i Es 7 , ¥ d i , wits? “ Cee erricyt a - a ow i re ~ = Ort ee Br tet a mee . eer ee i. ‘if i a V0, cg Wace a c 4 fur re cit OY aaa seen ae | Mt. - tea ; “ yay ay “ ‘ THE ACTION OF VARIOUS METALLIC SALTS ON HEMOLYSIS HELEN A. PURDY anp L. E. WALBUM From the Statens Seruminstitut in Copenhagen Received for publication November 7, 1921 During a series of investigations concerning the formation conditions for various bacterial toxins, one of us (W.) made the observation, that the presence of quite negligible quantities of various metallic salts in the substratum often had a rather con- spicuous effect on the amount of toxin formed in the cultures. These experiments, which were performed with staphylolysin and diphtheria toxins (1), showed that this action, to a great extent, depended on the concentration of the metallic salt; a certain concentration exerted a maximal effect, the action dimin- ishing with increasing as well as with decreasing concentrations. This observation, which is of a certain theoretical interest and which, moreover, will probably become of practical value in connection with the preparation of diphtheria toxin, induced us to undertake an investigation—necessary for the estimation of the results—of the sensitivity of the blood-corpuscles employed in the staphylolysin experiments towards small quantities of certain metallic salts. During the first of these investigations the metallic salts were observed in some cases to augment and in other cases to reduce the resistance of the blood-corpuscles to the staphylolysin, and this often in a very considerable degree. Therefore, we thought it would be of interest to enter into a more thorough investigation of these relations, and we extended the experiments to comprise besides staphylolysin-goats’ blood-corpuscles, also saponin-horses’ blood-corpuscles, and complement-amboceptor-sheep blood- corpuscles. 35 36 HELEN A. PURDY AND L. E. WALBUM Such investigations should of course be performed in such a way that the action of the individual salts be examined in many concentrations from just the minimal dose that produces a dem- onstrable reaction up to the maximal dose applicable at all; as most metallic salts produce either hemolysis, agglutination or discoloration of the suspension of blood-corpuscles, there will always be a limit to the quantity of such substances that may be employed for experiments of this kind (see tablel). We have, TABLE 1 broom | su0op- | conreacues Pron Gene SALTS eet CORPUSCLES CORPUACLES 1 PERCENT | 1PERCENT 8 cc. 1.25 cc. TaiGhogt sng fies fh RE 4 molecular | 0.2 0.2 0.2 0.03 CaCl rah yg Meee molecular | 0.2 0.2 0.2 0.03 DLCNO 3) ocr. hnscae tio ee molecular | 0.2 0.2 0.2 0.03 BaON@Os)eccc-cc. daeee eee, Ss amoleculan: 170.2 022 0.2 0.03 Pb(C2H;O.)o.............| 4% molecular | 0.0003 | 0.0003 | 0.0003 | 0.00005 Be Glo: «ta. tosh nate 4+ molecular | 0.0002 | 0.0003 | 0.0003 | 0.00005 MgSO iS2 facads cree ene molecular | 0.2 0.2 0.2 0.03 TATOO RSE ae molecular | 0.0001 | 0.0003 | 0.00025 | 0.00004 CdCl eet Me A Ee molecular | 0.005 0.005 0.003 0.0005 CaiSO eo. seis ee molecular | 0.00001 | 0.0005 | 0.00006 | 0.00001 INTO eae tale ert yn molecular | 0.0035 | 0.015 0.003 0.0005 POUNOg)s' 5h sci owen molecular | 0.005 0.005 0.003 0.0005 EG see ccc ne Oe ots molecular | 0.00001 | 0.00004 | 0.00012 | 0.00002 Mn SO ys 5-135 ae tee molecular | 0.2 0.2 Ost 0.015 (On Cl SAA Rses be peers alee molecular | 0.00003 | 0.00004 | 0.0003 | 0.00005 ATH (SO asia ieee eee zs molecular | 0.0001 0.001 0.0012 | 0.0002 ONO Nath ccratid ok oe molecular | 0.0005 | 0.0005 | 0.00025 | 0.00004 13 [4 @) Pes Geet Rh ameh bre Fe ae molecular | 0.0001 | 0.0001 | 0.00006 | 0.00001 EAM lesen ces Sea molecular | 0.00005 | 0.00004 | 0.00012 | 0.00002 7A S167 G27 ©] DE eee ee molecular | 0.025 0.015 0.025 0.004 however, been obliged to confine our investigations of these relations to the two extremes. It is of course possible that several optimal or minimal concentrations may be found within these limits. In those cases, however, in which we have watched the progress through many concentrations, the action gradually diminished with decreasing doses. The actions on complement of Hg, Mg and Ag show, however, an inciting effect at one con- centration and an inhibitive effect at another (see table 4). ACTION OF METALLIC SALTS ON HEMOLYSIS 37 The technic applied is described in connection with the indi- vidual experiments. The method employed for the hemolytic tests is the one usually employed at this institute, described by Madsen (2). In a series of experiments we first determined the maximal dose of the various solutions of metallic salts which, without causing perceptible alterations (hemolysis, agglutination or dis- coloration), might be added to the volume of blood-corpuscles employed for the experiments. The results of these determinations are shown in table 1. In the ease of sheep’s blood-corpuscles the permissible quantity is stated for 8 ec. of blood for the sake of comparison and also for 1.25 cc. which is the quantity employed in the following experi- ments. The blood-corpuscles were washed as stated in the preliminary accounts of the respective experiments. No larger doses than 0.2 ec. of the molecular solutions have been employed. As appears from table 1, the sensitivity of the blood-corpuscles towards the various metallic salts is extremely variable, nor does there seem to be any homogeneity within the various groups of metals. The figures stated are of course not absolute, but may vary somewhat in the experiments performed at different times, presumably chiefly owing to a varying content of serum in the suspension. ATTEMPTS AT DETERMINING THE MINIMAL CONCENTRATION OF THE RESPECTIVE METALLIC SALTS, AT WHICH THEIR ACTION IS MEASURABLE These experiments are performed in the following manner: To a series of test-tubes containing decreasing doses of the metal- lic salt, hemolysin is added in such a quantity as to produce only a slight hemolysis (about 20 to 30 per cent); after filling up with a physiological solution of sodium chloride to the volume stated, the blood-corpuscle suspension is added, this mixture being then, after shaking, placed in the thermostat and afterwards in the ice-box overnight. The next day a determination is made of how much hemoglobin (in per cent) has been dissolved in each tube. 38 HELEN A. PURDY AND L. E. WALBUM EXPERIMENTS WITH SAPONIN For the experiments with saponin, horses’ blood-corpuscles have been employed; the defibrinated blood was centrifugalized and washed twice in a 0.9 per cent solution of sodium chloride, and from these washed corpuscles was prepared a 1 per cent sus- pension, to which was added 2 per cent of a 2 per cent solution TABLE 2 Saponin; horses’ blood-corpuscles RECIPRO- cane, | See eee eee, | armeeoe | ea ATION es. eek cies 0.0035 | 0.0031 | 0.0035 | 0.0034 | 294.0 HgCl. 0.007 0.0065 | 0.0065 | 0.0067 | 149.0 AgNOs. 0.04 0.037 | 0.04 0.039 25.6 Pb(C2H30s)e..... 0.1 0.08 0.1 0.09 Het Favoring CoaG(NO sec... 24 0.5 62755 0.5 0.52 t92 hemolysis Mi Oana 5 cca tee 2.5 1.8 Did 0.476 BaUNOs) zs. <. .. er 70 10.0 8.0 8.0 02125 MegSOii50. 8s. sacaiccs =| UO hemolysis || MgSO,.......... 0.905 WeENOy 2.2! 0.01 iG es so toes 0.5 MgSO.,...... 16.0 SLCNOs)essaes ote 1.0 Ba(NOs)e.....-. 0.8 Od @lessoasec eset 0.08 AgNO; sfefotefetaintelar 0.03 BeCl, pisterelcialetnrauane 0. 032 ons CrCl. Siateletat wis ietetw re 0.03 ae HeClas skeen ee: 0.02 Pb(C2H302)2....| 0.015 Als(SO,4)3.......- 0.01 He Oiyes frie ee 0.0075 CuSO icc. aoe 0.0055 YMSO ieee 0.005 MnSO, woencereces 0.005 d 217 rl ie Mar eet 0.0005 INK O) RAN ea eee 0.0004 EXPEBI- EXPERI- MENT 2 MENT 3 0.005 0.0055 0.0053 | 0.0052 0.0052 | 0.0055 0.0055 | 0.0055 0.008 0.008 0.45 0.45 6.0 19.0 lath 1.6 122 1.5 0.75 0.78 0.085 0.08 0.033 0.035 0.033 0.032 0.032 0.03 0.018 0.023 0.013 0.015 0.009 0.01 0.0067 | 0.0083 0.0052 | 0.0055 0.0048 | 0.0055 0.0045 | 0.0045 0.00052 | 0.00055 0.00042 | 0.0004 The figures are to be understood as in table 2. METALS SAPONIN btbtttoottt+++4+4+ +] ie TABLE 5 STAPHYLOLYSIN b+ttttt+4+ fe of AVERAGE a eee k 0.0052 | 192.0 0.0052 | 192.0 0.0052 | 192.0 0.0053 | 189.0 0.0087 | 115.0 0.47 2.13 7.0 0.558 1.6 0.625 1.2 0.833 0.78 1.28 0.082 12.2 0.033 30.3 0.032 31.3 0.031 32.3 0.02 50.0 0.014 71.4 0.0097 | 103.0 0.0075 | 133.0 0.0054 | 185.0 0.0051 | 196.0 0.0047 | 213.0 0.00052 |1920.0 0.00041 |2440.0 COMPLEMENT- AMBOCEPTOR -+ Means favoring, + inhibitive of, and 0 no action on hemolysis. 41 42 HELEN A. PURDY AND L. E. WALBUM Mg, which all acted favorably on hemolysis in the saponin and staphylolysin tests, assumed two forms in the complement test, the salts of these metals acting favorably at a certain concentra- tion and inhibitively at another concentration. The effect of the other metallic salts tested is summarized in table 5. Whether this action of the salts may be due to a direct influence on the hemolytic agent (for instance a destructive action on the complement) or to the blood-corpuscles, will rest with future investigations to ascertain. EXPERIMENTS WITH THE MAXIMAL CONCENTRATIONS OF THE VARIOUS METALLIC SALTS THAT CAN BE APPLIED WITHOUT PRODUCING HEMOLYSIS, AGGLUTINATION OR DISCOLORATION OF THE BLOOD CORPUSCLES These experiments are so performed that the respective hemo- lytic agents are employed in such quantities as to produce only a partial hemolysis; the doses of the various metallic salts being as those stated in table 1. For each experiment the hemolysin was pipetted off in 3 to 5 different doses (all giving partial he- molysis but, of course, in a different degree), to this was added the solution of the metallic salt and a solution of physiological sodium chloride up to a definite volume and, finally, the blood-corpuscle suspension; otherwise the technic was quite identical to what has been previously described. By means of a control series of hemolysin and blood-corpuscles alone, the percentage augmentation was calculated for each mix- ture that allowed of such calculation, that is, each mixture in which total hemolysis did not occur. From the 3 to 5 figures thus obtained for each experiment and for each metallic salt, the average augmentation percentage was calculated. Of such experiments, four withsaponin-horses’ blood-corpuscles, five with staphylolysin-goats’ blood-corpuscles, and five with complement-amboceptor sheep blood-corpuscles were performed at different times. The results of all these investigations are stated in tables 6, 7 and 8, the figures thus stating the average —S ee eee TABLE 6 Saponin; horses’ blood-corpuscles AVERAGE SALTS EXPERIMENT I | EXPERIMENT II |EXPERIMENT III| EXPERIMENT IV AUGMENTATION per cent Li. 20 40 57 25 36 CR ee 2233 1900 1900 1300 1833 SEA eee: 617 233 900 140 473 13 hot id tas BEE 75 33 213 108 107 12a oar ae eee 13 90 19 0.7 30.7 1B CG eae erties ae 13 57 +4 13 20 Mg. 617 167 150 140 260 Pinitor hy eyes Fs 53 74 132 46 76 ONG aaa crea eae 31 47 13 30 (Car EARS eam Sane 57 23 +4 25 INSEE ence peer ees: 20 47 0 +7 15 COM ie oct 27 Pf 13 128 49 Te oone aeey ea 74 +7 18 28 Mn. 484 567 650 500 550 ores Oe a an 57 56 18 44 Al 0 74 21 29 31 occ Neate a aa 13 33 29 43 30 Hg. 104 200 67 219 148 TAN ts a eee ee 57 7 35 33 121s She Re ee 42 44 38 47 43 The figures state average augmentation in per cent. TABLE 7 Staphylolysin; goats’ blood-corpuscles AVERAGE SALTS EXPERIMENT} EXPERIMENT] EXPERIMENT EXPERIMENT! EXPERIMENT AUGMENTA- I II III IV Vv TION IN PER CENT Nes ee hs +6 48 83 +10 5 24 a PONE AE ascot Total abolition of the action of the lysin +100 Bate +98 +66 +100 +100 Op Pb: +71 4 +79 +40 +53 +48 B@aethtys ee). A2% +82 0D = 97 02 +88 +75 (NEA ay ReaNOS 1629 40 37 569 Zn... 1107 543 80 577 (iy) Se BES 450 44 335 20 100 190 Cu.. +76 +66 +87 caAG +63 +74 a The augmentation cannot be calculated fromthe >92000 coe... experiments, but is sure to be over 2000 per cent Be, 4! +46 176 +58 30 +67 ii LCT 1629 40 54 574 Cid ie a ae +87 +88 +100 +50 +95 +84 aN | Si a ee a +83 +81 +92 +96 +91 +89 Je hte 8 li aR 98 50 1 ify +30 +64 34 15 [es 184 71 +13 =-30 +38 35 13 ae ea +29 71 +41 =-20 +59 =-16 Pere weep cescices-. 117 106 460 60 64 161 44 HELEN A. PURDY AND L. E. WALBUM augmentation in percentage for the individual experiments and, in the last column of the tables, the average augmentation for all the experiments taken together. While the previous experiments, in which the purpose was to determine the minimal active dose of the metallic salts, showed very close congruity in spite of their being performed at TABLE 8 Complement-amboceptor ; sheep’s blood-cor puscles AVERAGE EXPERIMENT) EXPERIMENT| EXPERIMENT! EXPERIMENT| EXPERIMENT| AUGMENTA- Vi SELES TION IN PER CENT 1 Ps 55 32 50 CR sciatic +100 +100 +98 Sy Rha eer cee +100 +100 +100 Mot: tnd cee es ee +86 +79 +71 IBD es ee ect ace +32 +3 6 +9 Bee ee eee +17 0 0 +6 Mowe ese% +100 +98 +49 Zn... +71 +52 +75 +68 LOfa le sete ig +91 +73 +88 +81 CUS as ee +18 +16 +10 +17 ING seic ree one 69 74 29 59 Co +2 0 +25 +11 1 +2 +6 0 +6 Minter eso +90 +74 +74 +83 Cress Ae 14 0 16 7 PAN ors Soiree it oh 18 0 13 11 ped Pe Sa +10 0 15 z Hg +10 +10 0 +11 Au 2 3 5 2 Pit pee hess ee +91 +32 +69 +62 different times, there is a very conspicuous variation in these latter experiments. As the only varying factor in the experi- ments is a difference of the concentration at which the metallic salts were employed, the cause of the incongruity should presum- ably be sought therein; it seems as if blood-corpuscles, to which is bound as large a quantity of a metallic salt as can be bound without acting in a perceptibly destructive manner acquire strongly varying stability towards various hemolytic agents. ACTION OF METALLIC SALTS ON HEMOLYSIS 45 The considerable disagreement, which is found almost through- out, between these latter experiments and those previously described in this article, should probably be assigned to relations of this kind. We have made several experiments in order to determine whether the action of these metallic salts may be due to the cation or to the anion. These experiments were performed with chlorides, sulphates and nitrates of Mg, Mn, Zn and Ni; it was ascertained in all cases that the anion was without signifi- cance in this respect. SUMMARY What has been examined in this study is the significance of the presence of small quantities of metallic salts on the hemolytic action of saponin on horses’ blood-corpuscles, of staphylolysin on goats’ blood-corpuscles, and of complement-amboceptor on sheep blood-corpuscles. By determining the minimal dose of the individual salts (in molecular solution) at which their action is demonstrable, it is possible to obtain a comparison between the action of the different salts. While some salts exert an inciting action on hemolysis, others exert an inhibitive one (positive and negative catalysis?) ; some show an inciting effect at one concentration and an inhibiting one at another (table 5). On examining blood-corpuscles to which is bound as large a quantity of a metallic salt as can be bound thereto without acting perceptibly destructively, the stability of such blood- corpuscles towards the action of hemolytic substances will be found to vary greatly, to which fact the considerable experi- mental inconstancy in the experiments described should probably be assigned. The anion in the salts seems to be without significance as regards their action in either favoring or inhibiting hemolysis. REFERENCES (1) Watsum, L. E.: Compt. rendus des séances de la Soc. de Biologie, Paris, 1921, 85. (2) Mapsen, Tu.: Kraus and Levaditi; Handbuch . . . d. Immun., Jena, 1908, 1, 35-144. (3) Arkin, E. E.: Zeitschr. f. Immunitatsforschung, Orig., 1909, 1, 387. (4) Warzoum, L. E.: Zeitschr. f. Immunitatsforschung, Orig., 1909, 3, 70. ha Fil ' . he nae iy: eli WDA = yet gt Capa eR i ed ule we praninn r vt a add: Fan Lad atsiitmi| sae HAG i ey an sa of wre” ree lt ae pn “ids in eine A: Tae A Fagyotdoy oiled a SaeihaD 95 iis nb cnrgiaet/a awl bain sine ae Pe Ce he ae oe ee ae mail othe wena Ws, epi diy -olbhy Dire ey a Le he AiR ey ch eres 7 Gatick gaan a Pe RA 2 wa. to = De eee eae res HE Hae emia RAR Gees a nant Pe Bas: | ONE DAY TWO DAYS |THREE DAYS; FOUR DAYS NUMBER poses 1n- | ANTITOXIN JECTED 10 1 iE; rE 5 8 110-A 40 1 ip L i if | 80 1 L iF L az 10 1 L if i L 110-B 2 50 1 i i; i i 100 1 L L D 50 1 L i if if 110-C 1 100 1 e D 200 1 D 10 1 : L Li 111 i 50 1 ie D 100 1 | D 10 1 i if 5 f 111-A 2 50 1 i 7 L i, 100 ae (ee iy D | 10 1 L L if iD, 111-B 2 50 1 re | IF iP ij \| 100 1 ele ee D 19 1 i iz L i 111-C 1 L i i I i DA) $e py 4 L = lived; D = dead. The amount of protection in one immunity unit of standard antitoxin against the toxins of either group, was found to range between 50 and 100 minimal lethal doses with the exception of two lots of group I toxin (lots 111 and 112), in which the pro- tection fell below 50 minimal lethal doses. Lot 113-A shows BACILLUS DIPHTHERIAE 15 proof of protection against 100 minimal lethal doses of group II toxin. This result is, we believe, due to the estimation of too high a potency for lot 113-A, so that the doses given in this TABLE 4 Protection tests TOXIN NUMBER SESUE NUMBER OF MINIMAL LETHAL DOSES IN- | JECTED Wa Uae ONE Day | TWO DAYS |THREE DAYS| FOUR DAYS ANTITOXIN 112 1 112-A bo | | | | 112-C 1 113 1 | | mis | lols alll roles mroukeeaites aku lll sallert Sclelile leet ee sl aI cola alee clr) lel call er] [cal eal eee) lea el leis colo leah [ss sail ea salle It ls! EAGES — eat Sil eat gal ecole ral eal eileals e cloalelen! [eg = clerl eects fealaenl eal Weicalseplteal= © ra “ ’ . eae » SOOT ophses eta AR AR. shneeomd ; ee Rilhe res? noone nee Ak) RR a eee, ona cpeleanifisaaes | htnlbaatt Wwodihs Iw! toawerrehanl= hib hiiae arene are Sleetebbehdie ol deactoeclaa pedpiber es ie | Rakucuaaeniaal ea sink sehdht WONG rthaingeaih eile Seenres ag ‘habe tol tersh Bape emi) Kia 5 $i) ... ..-:.s..e seen omes aeons 393 Cases with constitutional reaction to pollen extract on test............. 3 Cases of pollen allergy therapeutically injected.....................005- 310 Number of pollen extract injections given..:.............2.:--2050ss00s 4,215 Cases of pollen allergy with immediate constitutional reaction on injection 28 Number of immediate constitutional reactions to pollen extract on injec- PROT cep Set oe elas «nn gee es Bae ck Se ee ee eae 38 Cases of pollen allergy with delayed constitutional reaction on injection 15 Number of delayed constitutional reactions to pollen extract on injection 38 Cases of pollen allergy with constitutional reaction on injection, onset LTE TUAW IN ss ap 02's rece 9 epee eles nie EU eo ease he ate eee 5 Number of constitutional reactions on injection with pollen extract, onset PHUG WOKIOWED. ©. ono cisces 208 Sete pete eee oad ee ae ea en ae 6 No previous attempt has been made seriously to consider this subject in its important relation to, and as a consequence of, the diagnostic study and the treatment of hypersensitiveness in STUDIES IN SPECIFIC HYPERSENSITIVENESS. III 127 the human subject. It is the object of this paper to present and to correlate all the facts relating to such general reactions, as this type of work is being, and will continue to be, much more generally used and especially because of the dangerous and sometimes fatal results which may occur. In all of these cases the diagnostic tests were made by the intradermal method and a definite diagnosis of allergy was based upon a marked cutaneous reaction: that is, a wheal with obvious pseudopod formation surrounded by a definite zone of hyperemia. It was further required that this marked positive reaction be repeatedly confirmed. The preparations used were those pre- pared by the writer standardized according to the nitrogen con- tent. There are no data with which to compare the relative frequency of general reactions occurring with the hypodermic as opposed to the scratch method of cutaneous diagnostic test. It can only be pointed out that reactions do occur with the latter method, as, for example, the reaction produced by Rufus Cole . with buckwheat, reported by Smith (1), using the scratch test. Some years ago the writer witnessed a constitutional reaction to timothy pollen where the pollen was applied to an abrasion on the forearm and dissolved in 0.8 per cent salt solution. Within five minutes there was a local reaction with a very difinite lym- phangitis extending to the axilla, with intense itching in the cubital fossa and axilla. This was almost immediately followed by the general symptoms of erythema, urticaria, general pruritus, asthma and coryza, characteristic of the general allergic reac- tion. This lymphangitis is very characteristic of all marked positive reactions obtained on the anterior aspect of the forearm, even in the absence of general symptoms, and gives a clue as to the part the lymphatics generally must take in all local or general reactions. It may be that constitutional reactions occur more frequently after the intradermal than after the scratch test. Since the general reaction depends solely upon the amount of the active allergen absorbed, it is possible that if such a difference does exist it will be found to be due to the use, in the intradermal test, of a more potent extract than is generally used in the scratch 122 ROBERT A. COOKE ; PROTOCOL ; = CLINICAL SYMPTOMS INTRADERMAL TEST a a Bb we | art Constitu- r Pollens giving Other substances giving nual Outlot Beis am A ee reactions symptom Mise S | Asthma Coryza Asthma Coryza frm 2166 Orris Orris j 2176 Timothy 2177 Dust, cat 2180 Horse Horse 2185 Ragweed | LePage LePage LePage Immediate | Asthma, coryza, caria 2193 Timothy 902 Ragweed Orris, horse and dog epi thelium 2197 Timothy 2199 Timothy 2213 Timothy 2388 Timothy 2393 Timothy Orris Ragweed | Ragweed 2415 Timothy | Ragweed | Ragweed | 2416 Ragweed | 2428 | Ragweed | Ragweed | 2440 Ragweed | 2441 Ragweed 2447 Horse 2473 | Ragweed | Ragweed Orris, horse 2479 | Ragweed | Ragweed | Orris | | LePage Immediate | Urticaria : Timothy 2492 | Ragweed | Ragweed | 2520 | Ragweed | | 2526 Ipecac | Ipecac Linseed | Immediate | Asthma, coryza, Linseed | Linseed caria. Angio-ni Buckwheat Buckwheat tic edema of 1 ' Feathers | Feathers Jasting 2 days STUDIES IN SPECIFIC HYPERSENSITIVENESS. III PROTOCOL ON SUBCUTANEOUS INJECTION ASSOCIATED REACTION : ROBABLE CAUSE AT SITE OF stitutional reactions] Injection Onset 2 pattie cick io nonod from reacting of symptons 2 Pelee SENS OR TEST is 4th 8 hours 2-3 days | Asthma x4 Marked 5th 8 hours 10 hours | Asthma ? Marked nothy 10th 1 hour 3 hours | Cough and _ urti- x8 None caria t epithelium 5th 8 hours ? Asthma ? ? ree epithelium Ist Immediate 6 hours | Asthma, coryza, | Patient very sensitive | Marked } urticaria, also f nausea and head- ache lasted one week Page Ist Immediate 24 hours | Asthma, coryza, | Patient very sensitive | Marked urticaria nothy 8th 8 hours 12 hours | Pruritus of arms | Change of extract Marked and neck eed 8th 12 hours 8 hours | Coryza, cervical ? Slight glands 9th 12 hours 8 hours | Coryza, cervical ? Slight glands nothy 10th Immediate 10 days | Asthma, coryza, | Concentrated extract ? ; ‘ cervical glands nothy 8th Immediate ? Urticaria Concentrated extract ? nothy 5th Immediate 4 hours | Coryza, erythema,| Dose increased too} Moderate is headache and] _ rapidly ¢ pruritus . fithy 7th 1 hour lday | Asthma, coryza 2? Moderate 8, 10th Immediate lhour | Asthma, coryza 2 ? Breed 17th 6 hours 2 days ? Overdose ? gZweed 16th ? lday | Coryza ? Marked gZweed 12th 24 hours lday | Coryza and edema | Change of extract None of lip gZweed 14th 1 hour ? Asthma, coryza, | Change of extract ? a urticaria gweed 16th 24 hours 6 days | Coryza, headache ? ? zweed 18th ? ? Coryza, edema of ? ? yt eyes se epithelium 9th and ? ? Asthma ? ? . 10th Tse epithelium and 10th 4 hours 2days | Coryza ? ? 11th 4 hours 2 days | Coryza 1S Patient very sensitive | Marked 10th 1 hour 3 hours | Asthma Change of extract Slight 17th 4 hour ? Asthma Change of extract ? 7th Immediate ? Asthma ? Moderate 10th Immediate ? Asthma ? ? Patient very sensitive | Marked 124 CASE NUMBER 2274 2276 2216 2227 2233 2643 ROBERT A. COOKE CLINICAL SYMPTOMS Pollens giving Asthma Timothy Timothy Ragweed Daisy Dandelion Ragweed Ragweed Ragweed Timothy Ragweed Coryza Timothy Ragweed Ragweed Timothy Timothy Ragweed Ragweed Daisy Dandelion Ragweed Ragweed Ragweed Ragweed Ragweed Ragweed Ragweed Timothy Timothy Timothy Ragweed Ragweed Timothy Timothy Ragweed Timothy Timothy PROTOCOL Other substances giving Asthma Peach Celery Raspberry Horse, cat Feathers, dust Horse Horse Feathers Coryza Peach Celery Raspberry Horse, cat Feathers Orris Horse Orris; dust INTRADERMAL TEST | Constitu- tional Onset of reactions symptoms from Ee ek : ee ™ ; " STUDIES IN SPECIFIC HYPERSENSITIVENESS. III PROTOCOL ON SUBCUTANEOUS INJECTION ASSOCIATED 2 PROBABLE CAUSE ae ante ee nstitutional reactions] Injection Onset eas Ritacet a INJECTION from reacting | of symptoms |°' 7 SLD OR TEST mothy 5th Immediate 3 hours | Coryza Change of extract ? azweed 21st ? ? Asthma, coryza, | Change of extract ? urticaria ngweed 7th Immediate ? Coryza Directions not fol- ? lowed. Intervening doses omitted mothy 14th Immediate ? Asthma Concentrated extract ? azweed 18th Immediate ? Asthma, coryza, | Change of extract ? ; urticaria aisy, dandelion 2nd Immediate lhour | Asthma, coryza Patient very sensitive | Marked izweed 8th Immediate 2 hours |} Urticaria in all Patient very sensitive ? 9th Immediate 4 hours Patient very sensitive ? izweed 9th Immediate 2-3 hours | Coryza 4 ? 13th Immediate 2days | Asthma ? ? 15th Immediate ze Coryza, asthma ? ? wzweed 10th to § hours 36 hours | Urticaria ? 2 | 20th Immediate ? Coryza Change of extract ? agweed 15th Immediate 3 hours | Coryza, asthma, | Change of extract 2? urticaria izweed 19th ? 2 Asthma ? None izweed 16th 4 days 4 days | Urticaria Moderate izweed 18th 2 hours 2 Urticaria Marked weed 12th 4 hours 8 hours | Asthma, coryza Change of extract Marked | 17th 3 hours 3 days | Urticaria, edema| Dose increased too ? - of eyes, coryza| rapidly mothy llth Immediate 12 hours | Asthma, .coryza,| Dose increased too ? urticaria rapidly mothy 14th 8 hours 4 hours | Urticaria ? ? : 15th 8 hours 4 hours | Urticaria ? 19 mothy Ist Immediate ? Coryza, erythema | Patient very sensitive | Marked weed 17th 6 hours 12 hours | Coryza, edema of | Change of extract Marked eyes iZweed 6th Immediate 4hours} Asthma, coryza, | Concentrated extract ? urticaria 20th 3 days Several | Urticaria Change of extract ? mothy 15th Immediate 2 hours} Asthma, coryza, ? 2 urticaria, edema of lips 21st 1 hour 2 hours | Urticaria Change of extract ? mothy 8th Immediate ? Asthma, coryza, | Concentrated extract 2 urticaria eed 1st Immediate ? Severe abdominal | Patient very sensitive | Marked pains othy 8th» + hour 2days | Asthma, coryza, ? None urticaria, edema | face : 13th 1 hour 8 hours | Asthma, coryza, | Change of extract ? | urticaria, edema face ris 15th 24 hours 3 days | Asthma ? Slight 125 126 ROBERT A. COOKE PROTOCOL : 3 CLINICAL SYMPTOMS INTRADERMAL TEST \ [oa] 8 p Pollens giving Other substances giving | Constitu- a tional Onset of s @ SS SEE OGEL TUCO symptons ymptoms Ps Asthma Coryza Asthma Coryza from 2651 Timothy | Horse, dog Ragweed Immediate | Coryza, asthma | Ragweed | Buckwheat Headache, erythema. | 1058 Timothy Timothy Immediate | Coryza 4 2673 Egg, milk LePage, horse } Rabbit, feathers | LePage Immediate | General edema. Asthma, coryza, deat) : from asphyxia 2681 Orris, wheat Orris Immediate | Asthma, urticaria Dust 2684 Dust, feathers 2258 | Ragweed | Ragweed 2260 Timothy Ragweed 2273 | Timothy | Timothy Ragweed | Ragweed 2547 | Ragweed | Ragweed 2581 Timothy Ragweed 2627 Ragweed | Orris Ragweed Immediate | Coryza, asthma, urti- : caria 2707 Feathers, dog, cat Dog saliva | Immediate | Coryza, asthma, urti- caria 2708 Timothy | Chicken, rabbit Flaxseed Immediate | Asthma Ragweed | Flaxseed, mustard Mustard Immediate | Asthma 896 Timothy Cat 2541 | Ragweed | Ragweed 2532 | Ragweed | Ragweed test. This surmise is supported by the experiences of Brown (2) and of Larsen, Paddock, and Alexander (3), who found only 50 per cent of diagnostic efficiency using the dry preparation generally employed for the scratch test as compared with the fluid preparations used in the intradermal test. HISTORICAL A survey of medical literature discloses the fact that the year 1894, when diphtheria antitoxin was first introduced for general STUDIES IN SPECIFIC HYPERSENSITIVENESS. _ III 12¢ PROTOCOL ON SUBCUTANEOUS INJECTION ASSOCIATED REACTION ; PROBABLE CAUSE AT SITE OF mstitutional reactions; Injection Onset enon Sentra INJECTION from reacting | of symptoms ronda ymp OR TEST orse epithelium ust extract imothy agweed agweed i ) Concentrated extract Marked 4th 12 hours 2 days | Nausea ? Marked Concentrated extract Marked Patient very sensitive | Marked Concentrated extract Marked 5th 1 day 5 days | Urticaria if ? 7th Immediate ? Asthma ? None 10th Immediate ? Asthma ? Marked 10th 5 days 1 week | Urticaria Dose increased too ? rapidly 16th Immediate ? Asthma, urticaria ? % 2nd Immediate ? Coryza, asthma ? Moderate Ist 1 hour 2 hours | Urticaria, edema} Patient very sensitive ? of uvula Concentrated extract Marked ? ? Concentrated extract Marked Concentrated extract Marked 5th Immediate 3 hours | Urticaria ? ? 7th Immediate 2 hours | Coryza, edema of | Change of extract ? eyes 12th Immediate 3 days | Coryza, edema of ? ? eyes, headache, urticaria Ist Immediate ? Coryza, erythema | Patient very sensitive | Marked 10th 24 hours 1 day Erythema ? ? use, marks the beginning of a period in which there are rather frequent reports of the violent reactions and sudden death follow- ing the injection of the antitoxic horse serum. The nature of the reaction was not understood. Gottstein (4) reported several such cases in 1896. In 1906, Rosenau and Anderson (5) state, ‘“We have collected from the literature 19 cases of such unfor- tunate results and know personally of several more which have not been reported.” Gillette (6) collected 30 cases Park (7), speaking of the frequency of such general reactions with diph- 128 ROBERT A. COOKE theria antitoxin, says that since January 1, 1895, using either whole serum or the globulin fraction, thirty thousand cases had been injected, with collapse in two cases; that over a period of five years, when an immunizing dose was used in all scarlet fever cases, sixteen thousand cases were injected, with no collapse; and that in New York City, inspectors have given injections to one hundred and five thousand cases, with two deaths, but no record was kept of cases with reactions which were not fatal. Every death which did occur followed a primary injection. The literature since 1913 has not been searched for the antitoxin reactions because the further collection of cases does not add materially to our knowledge of the reaction and because compara- tively few of the reactions that have occurred have been reported. We are not interested here in the relative frequency of the allergic reaction to diphtheria antitoxin. Gillette (6) appears to have been the first to remark the similarity of the clinical reaction to the experimental anaphylactic reaction in the guinea-pig and he further noted the important fact that in man most of the severe reactions with dyspnea, edema, urticaria and pruritus, occurred immediately after the first injection, in this way differing from the experimental reaction. Since this time the appreciation of the condition now known as human hypersensitiveness or allergy has greatly increased and, as we now look back over the literature, we find occasional instances reported of peculiar reactions to many different sub- stances which chemically bear no relation to antitoxic serum. Indeed, a great many reactions, such as those of infants and chil- dren to milk, egg, and animal danders, were well known to the laity but occupied no place in medical literature. Peculiar drug re- actions, known as idiosyncrasies, were well recognized, but they were not identified as allergies until 1916, when the writer noticed the relatively large number of such cases occurring among hyper- sensitive individuals and remarked the similarity of symptoms of drug reactions with those of foreign proteins in specifically hypersensitive persons. From our present vantage point we can easily appreciate the reactions obtained by Blakeley (8) with pollens of grasses, by Dunbar (9) with ragweed pollen and by Cole (1) with buckwheat. STUDIES IN SPECIFIC HYPERSENSITIVENESS. III 129 Since the clinical conditions of bronchial asthma, hay fever, urticaria, angio-neurotic edema and the erythemas have been definitely recognized as manifestations of human hypersensitive- ness and since it has been demonstrated that diagnoses can be made with the well known cutaneous reaction and that thera- peutic effects can be obtained by injection, a great deal of work has been done and the results have been published, but a careful review of this literature since 1915 shows only a few records of the general or constitutional reactions following the use of aller- gens specifically applied to hypersensitive individuals for the purpose of diagnosis or treatment. Rackemann (10) reported reactions in two cases following therapeutic injection xtract of horse dander. Walker (11) mentions but incidentally a gen- eral reaction following a therapeutic injection of an extract of flaxseed. Gustenberger and Davis (12) reported a reaction with egg protein, the intradermal method being used for testing. The writer (13) has called attention to the dangers of the consti- tutional reactions on injection and has noted their frequency with pollen extracts as being 3.75 per cent in 4192 injections in 339 hay fever cases up to January,1916. The occurrence of many other such reactions is personally known to the writer but they have not been reported in the literature. SYMPTOMS OF THE CONSTITUTIONAL REACTION The symptoms of general reactions in allergic individuals are entirely distinct and apart from the symptoms that occur in the normal man, even though the latter be given one hundred thous- and times the amount of allergen (for example, horse serum). Allergic symptoms are as characteristic as those of a typical lobar pneumonia. The onset of constitutional symptoms may be immediate or they may be delayed up to five days. Dis- cussion of this point is made in greater detail in the section on “Varieties of the Reaction.” In general, the symptoms are those of the various clinically recognized allergies and in any individual case they are usually those from which the patient suffers, plus certain manifestations in tissues not reached by the allergen under ordinary exposure. Thus case 2185 has asthma, 130 ROBERT A. COOKE coryza and urticaria from eating fish, and asthma and coryza from handling fish glue (LePage glue). He showed identi- cal symptoms in the constitutional reactions after a test and after a therapeutic injection of the sterile solution of LePage glue. On the other hand case 2197, who has clinically only hay fever from timothy pollen and never asthma, developed coryza and in addition asthma and glandular swellings in the neck, following the subcutaneous injection of timothy pollen extract. Case 2180 had, in addition to the usual clinical symptoms, nausea and severe headache for a week following the first injection of horse dander extract. We can group constitutional symptoms as those that are usual and those that are infrequent. Usual symptoms Coryza. This term is meant to include the ocular symptoms of corneal injection, lachrymation and itching, as well as the nasal symptoms of discharge, sneezing, and edema of the mucous mem- brane causing obstruction. The term “Allergic Coryza’”’ has been employed by the writer (14) to designate all forms of vaso- motor rhinitis that are allergic in nature. Coryza is a particu- larly common symptom of general reactions and it may occur on the ingestion of allergens as foods and drugs, as well as when these substances are injected for test or treatment. It is more common, however, with substances such as pollens. that are natur- ally absorbed by inhalation. Asthma. This is somewhat more common in cases with clini- cal asthma but it does occur in clinically non-asthmatic cases, as no. 2197 mentioned above. Asthma is a symptom of the bronchial edema which makes the general reaction dangerous and which may be the cause of death from asphyxia. Urticaria. This is a very usual manifestation as would be in- dicated by the frequent occurrence of the immediate skin reac- tion on test. It is prevented from being a usual clinical symptom by the fact that a sufficient amount of allergen is not absorbed through the respiratory mucous membrane and carried by the systemic circulation to the skin. STUDIES IN SPECIFIC HYPERSENSITIVENESS. III 131 Erythema. In some cases the skin becomes scarlet without the appearance of urticarial wheals. Pruritus. This condition usually accompanies the urticaria or erythema but does exist without either. One case not included in the 1920 series regularly (at least six times) had a very marked pruritus ani within fifteen minutes after an injection of timothy pollen extract. Edema. The angio-neurotic type of edema may occur in any part of the body, and when once developed, it takes two to four days for the condition to subside. In cases of allergic coryza it is most commonly observed in the tissues of the upper and lower lids so that the eyes may be entirely closed. Occasionally the sclera is also involved and the edematous scleral tissue may protrude between the lids. Edema of the glottis has not been identified though it may have occurred in the fatal case cited later in this paper, and edema of the gastro-intestinal tract may explain some of the cases in which abdominal pain is a symptom. Cough. This may occur independently of asthma though it is often associated with it. It is violent and paroxysmal like the cough of pertussis and seems to be due to laryngeal irritation. Infrequent symptoms Glandular enlargement. This is noted twice in the cases included in the protocol, nos. 902 and 2197. It has not been noted by me in more than six cases. While it is noted as part of an immediate reaction in case 2197, it is certainly not itself a noticeable phenom- enon in less than three hours and the swellings usually last three days. The submaxillary and cervical glands, especially the anterior chain, are the only ones that have been observed, except in one case not included in the protocol, in which, follow- ing one of the injections, the preauricular glands were so enlarged that the condition was at first diagnosed as a parotitis. In this case following the next injection the same glands became swollen and tender and with them the cervical group. When the same reaction occurred a third time after injection, the time interval in all three instances being about twenty-four hours, 132 ROBERT A. COOKE there could be no mistaking the relation between injection and glandular enlargement. Headache. This may be of a mild type, frontal or occipital, and, like the glandular enlargements, does not develop immediately. It may persist, as in one case, no. 2180, fora week. On the other hand, headache may be excruciatingly severe and of a migrainous type. The writer experienced one such reaction in himself fol- lowing an injection of horse dander extract. The headache developed in ten hours and gave the sensation truly described as “splitting.” It lasted six hours. The writer is otherwise not subject to such headache. Fever. This, together with chilliness, has been complained of in a few cases but it has not been verified by actual readings except in the case with preauricular adenitis, when a temperature of 101°F. was observed with each of the three glandular reactions. In this case it began in six hours and lasted for twenty-four hours. Nausea. Nausea sometimes accompanied by syncope and vomiting is not very usual. With violent immediate reactions of the usual type, vomiting may occur inside of an hour and under these conditions it is so copious that there must be an extraor- dinarily large secretion from the gastric mucosa to account for the volume. In one case, not included in the protocol, in which a constitutional reaction with asthma took place after aspirin by ingestion, the vomiting followed in twelve hours and with it the attack ceased. ‘This was also the usual clinical course, ac- cording to the patient, whose attacks had for some time followed the use of aspirin taken for headache on the advice of his physician. Diarrhea. No cases in the 1920 series exhibited this symptom. I can recall two cases in which diarrhea followed the injection of ragweed pollen extract. The attacks started in six hours and lasted from twelve to twenty-four hours. The regularity of recurrence after injection in both cases is the sole reason for considering it a general reaction. There were no other constitu- tional symptoms in these two cases. STUDIES IN SPECIFIC HYPERSENSITIVENESS. III 133 Acute abdominal pain. This symptom has been noted in a few cases not in this series and has been attributed to an angio- neurotic edema of the gastrointestinal tract. Acute abdominal pain, cramplike in nature, over the lower abdomen, developed within one hour after the first ragweed injection in case 2227. The case was not observed during this attack and there is no record of its duration. It was not associated with vomiting or diarrhea. This may have been a type of reaction similar to those discussed under dysmenorrhea. The menses were absent in this case on account of the fact that this patient, a girl, was barely thirteen years of age and the menses had not yet been established. Dysmenorrhea, or rather an untimely and scanty menstrual flow, following acute cramplike pains in the lower abdomen is recalled in the case of two women aged twenty-six and thirty- eight, respectively. In both of these cases the symptom was part of an immediate reaction with asthma, coryza and urticaria, the menstrual flow itself not being apparent until three hours later and lasting only one day. Among many pregnant women tested and treated no such symptom has ever appeared, but extreme caution is always taken in such cases, for abortion might be induced. Syncope. This was not noted in any of the 1920 series. It does occur as an immediate effect, namely within one hour, and is usually associated with nausea and vomiting. I recall but one case in which it occurred alone. Reference is not made here to those occasional cases of syncope due to mental or nervous in- stability in which the patient will give a history of similar attacks following the sight of a hypodermic needle or a drop of blood. Cardiac collapse. I recall but one case in my entire experience in which this condition took place as a primary symptom. Of course cardiac dilatation and vasomotor collapse do occur secon- darily in the fulminating types of reactions with extreme dyspnea and partial asphyxia. But in the case referred to above, there developed, within one-half hour and with an absence of all usual general symptoms, a profound weakness and prostration, pallor and sweating without the loss of consciousness of syncope. 134 ROBERT A. COOKE The heart rate was 140 and the pulse was imperceptible, with cold skin and marked sweating. One milligram of strophanthin was given intramuscularly with adrenalin chloride 15 minums. Except for the adrenalin effect the condition was normal within one hour. From such a categorical list of the symptoms of a general reaction, one draws but a hazy idea of the pictures of the actual reaction in the individual case, but, as case histories will be cited later illustrating the various points to be brought out, no attempt will be made here to fill this deficiency. VARIETIES OF THE CONSTITUTIONAL REACTION The constitutional reaction can conveniently be considered as occurring in two forms, the immediate and the delayed. 1. The wmmediate general reaction The immediate reaction occurs within one hour after the aller- gen is introduced. This more or less arbitrary time limit has been adopted by the writer because in practical clinical work one sees a very large group whose reactions fall well within this limit of time, in fact well within one-half hour. In some cases it ensues on the instant (within one minute) and the severity of the symptom is in direct proportion to the brevity in time of onset. The sooner the symptoms begin, the greater their intensity and the greater the danger of a fatal result. On the other hand if cases do not react within the hour, the reactions are usually delayed for a period of at least six hours, and may not make their appearance for five days. No definite reactions have been identi- fied after a longer interval though there seems to be no reason why they should not so occur. The protocol of the constitutional reactions accompanying this paper shows only 3 cases, nos. 2364, 2381, 2473, with 4 reactions, in which the reaction appeared after one hour and under six hours, out of a total of twenty cases with 44 delayed reactions; i.e., after six hours. While this differentiation of immediate and delayed reactions is, for the present at least, based upon confessedly arbitrary grounds, it STUDIES IN SPECIFIC HYPERSENSITIVENESS. III 135 serves a definite clinical purpose in that it separates the intense and dangerous reactions from those that are merely subjectively disagreeable. Allergens may cause immediate constitutional reactions by whatever path they may be introduced systemically; that is, after test, after injection or on ingestion. Since these immediate general reactions may occur when the cutaneous reaction is negative as well as when positive, it will be well to discuss separ- ately the general reactions occurring with allergens which give a positive skin test and those occurring with allergens which give a negative skin test. a. When the cutaneous test 1s wmmediately positive. The imme- diately and genuinely positive intradermal test is pathognomonic of a cutaneous hypersensitiveness and indicates an accompany- ing clinical hypersensitiveness of the mucous membranes of the respiratory tract in approximately 95 per cent of the cases re- acting by test to pollens and animal epithelia. In other words, the intradermal test reactions agree with the clinical histories or can be clinically substantiated in 95 per cent of the cases reacting to the pollens and animal epithelia. This has been well shown by Vander Veer in the following table based on studies of cases in our clinic. Z A a AZ Zima o AS 8p z Bein PO By Bo Fp Bivined zy (aS an 98 Bet yeas SEs poe ee eGnbe | suese | stents | ease | 8.88 mn nD DQ Dm m nN nf D ° Qe ~ RarFAH QA”AH QHAA RZryA Leeann | zeezm | zezenm | Qeeazn | 2 onS per cent IRollenghe secant cists: 155 29 4 4 96 Animal epithelium ...... 20 17 2 0 95 When 0.01 cc. of extract is injected intradermally for the test according to the usual technic, a general reaction may follow within a few minutes. Ten such cases are cited in the protocol. One had two reactions with different allergens. Thirty-one cases in the protocol gave 42 immediate general reactions on therapeutic injection of the allergen. This is due directly to an overdose and the time of onset and symptoms are exactly similar to reactions from the test. Following is the history of the case in which death resulted from the test. 136 ROBERT A. COOKE Cas: 2673. Boy, three years of age, developed an attack of asthma at the age of eighteen months. The attack started with cough, then dyspnea and vomiting followed. There was no coryza. The dyspnea became increasingly severe until he passed into collapse and a semi- conscious state with a pulse rate of 160 and marked cyanosis. From this time on he was constantly asthmatic with exacerbations. These severe attacks lasted from one to seven days and occurred about once a month. As he would vomit immediately after ingestion of egg, no eggs had been used for over a year. He had mild urticaria at the age of one year. Eczema had been present for the first two years of life. Physical examination. The patient was an undersized, poorly-nour- ished boy. Respiration was labored. He appeared anemic. There was nothing else of note except the sibilant sounds and rales through- out both lungs and a marked double Harrison’s groove. On the first visit he was tested with milk, egg and cereal preparations, eight tests in all. Tests with very dilute egg protein were only sugges- tive but the casein preparation gave a marked reaction. The cereal preparations were negative. ‘Two days later he was tested with more concentrated egg preparations. Ovomucoid gave a moderate reaction and egg white globulin a marked reaction. The meat extracts were negative. The next day he was tested as follows: (The decimals indicate milligrams of nitrogen per cubic centimeter of solution; 0.01 cc. was used to test.) OETIS TOGt Ace eGor 0.1 negative Chicken epithelium.. 0.5 marked Dustiextracticce ose negative Horse epithelium.... 0.04 slight Dog epithelium...... 0.1 negative Horse serum......... 0.1 marked Cat epithelium ...... 0.1 negative Rabbit epithelium... 0.2 marked No untoward results had followed these tests. The reactions all subsided within twelve hours. Two days later, November 29, 1920, he was better than usual. The following tests were made: Tests HUA WCU Esco t ls: tres. act PCeR, Pao ae eee 0.1 Vanilla Maariv). Sh. at) S5. Pea Ee eee 0.1 Chocolate PI Grse eG BEIM | 95... < doce Mee 5 See Beek oe 0.4 Peanut Weave elena eee: ccc so. has Pes Re ee 0.1 marked Cocoanut Within two minutes it was noticed that the reaction at the site of the LePage glue test had spread up and down the arm and there were many STUDIES IN SPECIFIC HYPERSENSITIVENESS. III 137 fine urticarial spots appearing all over the arm. The boy then suddenly broke out in a general rash, his face began to bloat with an edema, until his eyes were closed. Cough and dyspnea were marked for a minute, he was deeply cyanotic and respiration ceased, though respira- tory efforts continued for a minute longer. Artificial respiration was attempted but no air could be made to enter or leave the chest. The heart continued to beat for a minute after respiration ceased. He had been given 1.0 cc. of adrenalin (1:1000) intravenously at the onset of the attack and strophanthin 0.125 mgm. In my opinion this was a genuine allergic death from asphyxia and not due to any associated condition of status lymphaticus. I have attributed this death to the LePage (fish glue) solution because the patient was not very sensitive to the horse epithelium as shown by the test on a previous day and because the marked local reaction and urticaria started about the site of the LePage test. In this case the LePage solution used contained 0.1 mgm. of nitrogen per cubic centimeter; 0.01 cc., at most it can be supposed 0.02 cc., was injected. This means that death in this case was caused by a dose containing only 0.001 to 0.002 mgm. of nitrogen. On ingestion of the allergen in its natural form, where the cutaneous reaction has been positive, a reaction may occur immediately and, if the buccal, esophageal and gastric mucous membranes react as well, either the substance cannot be swal- lowed at all or it is very quickly vomited. This is the condition in many of the egg allergies of children. Such mucous membrane reactions are in reality local manifestations but in certain of these cases absorption of the allergen into the systemic circulation may take place through the hypersensitive buccal or lingual mucous membrane. In a child of eight years, exquisitely hypersensi- tive to egg, a mild urticaria developed fifteen minutes after a piece of cake containing egg had been placed against the tongue for a minute. The saliva was not swallowed and the tongue remained protruded until she rinsed her mouth out thoroughly with water at the end of the experiment. The tongue itched and was very red and slightly swollen. 138 ROBERT A. COOKE There are cases in which the gastro-intestinal mucous mem- brane appears not to be sensitive to a substance which may produce a general reaction when the substance is eaten. An example of such an occurrence is presented in the history of case 2826. This individual, a woman of thirty-three years of age, exhibited the symptoms of edema of the lips and face within a few minutes after the ingestion of two hazel nuts and after twenty minutes she experienced a severe attack of asthma that lasted for two hours. She had had such a clinical reaction on two other separate occasions. Urticaria and pruritus had never been a part of this clinical reaction. ‘There were no gastro- intestinal symptoms either at the time or subsequently. Other nuts besides hazel nuts could be eaten with impunity. b. When the cutaneous test is negative. Occasionally, even when the skin test is negative, a general immediate reaction ensues after cutaneous test, after subcutaneous injection, or after in- gestion, that is unmistakable as to its cause, as will be shown in the case 1766 to be cited. This can be explained on the assump- tion of a complete absence of skin allergy with a hypersensitive- ness limited to the respiratory mucous membrane. In my ex- perience this has occurred only with the drugs and particularly with aspirin. The clinical history bears out the above assump- tion, for in the cases of aspirin allergy, if urticaria has been one of the clinical symptoms a positive cutaneous reaction is ob- tained but otherwise the cutaneous reaction is negative. The only difference between the general reaction that occurs after ingestion of aspirin and that following its injection is that twenty to thirty minutes elapse before the onset of symptoms when the drug is ingested and only two to five minutes when it is given by injection. This seems to be merely a matter of rapidity of absorption. In all the aspirin allergies observed, with or with- out positive skin test, there has been no immediate reaction of the alimentary mucous membrane when the drug is ingested but the symptoms of coryza and asthma are the same as when the drug is injected. Vomiting has occurred in several cases eight to twelve hours after ingestion and this act usually terminates the attack. This indicates a central effect. The following case STUDIES IN SPECIFIC HYPERSENSITIVENESS. III 139 illustrates the general reaction occurring on test, on injection and ingestion of the drug where the cutaneous test itself was negative. Case 1766. Ageforty,female. This individual has had three attacks of asthma in the last three months but each one has been extremely severe. The first attack started at 3 p.m. She had been in usual health and was on the street at the time and remembers being just able to crawl to her room. She described the sensation of strangling as though a rope were being pulled tight about her neck. She evidently be- came unconscious for she found herself on the floor at 1:00 a.m. She did not recover for two weeks during which time her chest was sore, as though she had been beaten. There was cough and nausea for several days, no headache, no fever and no urticaria. In the next attack eight weeks later she was seen by a physician within one-half hour after the onset of the attack. Adrenalin was given and she was taken to a hospital. The symptoms were the same as in the first at- tack. The attack was controlled and lasted altogether three-quarters of an hour but she was ill for two days with weakness and prostration. The third attack, exactly similar, occurred four weeks later. It started at 10 p.m., and had lasted for three-quarters of an hour when morphine was given. This attack incapacitated her for three days. Each of these attacks had occurred on the third or fourth day of the menstrual period. Physical examination was practically negative. Closer ques- tioning then revealed the fact that she had been given a prescription for menstrual headaches and that she had taken one of these capsules twenty to thirty minutes before each of the three attacks of asthma, They contained strychnin, quinin and aspirin. Quinin bisulphate and aspirin in solution were injected intradermally. The amount intro- duced was less than ;}y grain of each. At the site of test there was ab- solutely no reaction, either immediate or delayed. Within fifteen minutes after the test the patient began to cough violently, complained of a sensation of filling of the throat and thickness of the tongue, numbness of the hands with itching and dryness of the lips and throat. There was no erythema or urticaria. These symptoms disappeared in one-half hour. Six days later a test was again done with zy55 grain of aspirin. There was no reaction at the site of test. A similar but milder constitutional reaction followed immediately. Five days later quinin was tested with negative local and constitutional reaction. Then toby grain of aspirin was given by subcutaneous injection and the same immediate general reaction ensued in five minutes, with no reaction at the site of injection at the time or later. | 140 ROBERT A. COOKE 2, Delayed constitutional reactions Under ‘‘Immediate Reactions” the reasons were given for the establishment of the time limit of one hour for such reactions. Reactions occurring after this time are considered as delayed. a. When the cutaneous test is immediately positive. The writer has observed a number of instances in which constitutional symptoms have occurred after an incubation period of one hour or more following a subcutaneous injection of the allergen. Tn no ease was this ‘‘delayed” general reaction observed to follow the preliminary test or the first subcutaneous injection but always after a number of injections (usually 8 or 10) had been given. In the protocol appended to this report there are 20 cases in which 44 delayed general reactions occurred. It will be seen that these delayed reactions were always observed after repeated previous injection had been made. In striking contrast with these observations are the 10 cases in which constitutional reaction followed the preliminary test. In all of these the general reaction occurred immediately. It must not be inferred from these statements that an im- mediate general reaction cannot follow the later injections. Indeed such an occurrence has been frequently observed. The length of the time interval between the injection and the onset of symptoms in the treated cases seems to depend entirely upon the amount of allergen injected, for a sufficiently large amount of the allergen may at any time produce an immediate reaction. Rackemann’s (10) reported cases bear out this observation. In my own series, case 2292 of the protocol had given delayed general reactions as shown by urticaria, thirty-six hours after each injection from the tenth to the twentieth. But with the twentieth dose (change to new extract) an immediate coryza developed. Another interesting phenomenon in connection with delayed constitutional reactions in which the delay is induced by therapeutic subcutaneous injections is that a local reaction at the site of injection may be entirely absent if injections have been given in approximately the same site, but here again a sufficiently large or a sufficiently concentrated injection of the STUDIES IN SPECIFIC HYPERSENSITIVENESS. III 141 allergen will suffice to produce an immediate reaction at the in- jection site, as well as an immediate constitutional reaction. The phenomenon of the delayed general reaction appears to be an expression of the purely relative insensitiveness which the writer in another article (15) in this Journal has chosen to designate by the term ‘‘Hyposensitization.’”’? In such cases a marked positive immediate cutaneous reaction is always ob- tainable with sufficiently concentrated extracts. This induced delay of the constitutional reaction in hyposensitive cases may obtain whether the allergen is absorbed from a subcutaneous injection or from an intracutaneous test. It likewise holds when the allergen is absorbed by ingestion as shown by the following history. Case 1107. Age two years when seen in 1917. The first time egg was ever given, about one year previously, his face, lips and tongue became swollen, almost at once. He choked, coughed and became dyspneic and cyanotic, vomited several times and complained of ab- dominal pains. The father has early hay fever and the mother attacks of urticaria. On skin test the boy reacted markedly to the proteins of egg white but not to egg yolk. He was given injections of egg proteins. After seven injections he could eat egg in pudding in small amounts without trouble. After ten injections he could take five teaspoonfuls of soft boiled egg without trouble. Four more injections were given in this year and he could then eat one egg every other day without trouble until April, 1919, when he developed cough and hoarseness. Eight more injections were given. After this it was noted that he could eat an egg every other day without symptoms. When he was fed one egg every day he developed a harsh hollow cough by the second day, thus showing the delayed constitutional reaction. When he was put on two eggs a day he developed edema of the face about eight hours later—again a delayed constitutional reaction. All symptoms of cough and edema disappeared forty-eight hours after egg was discontinued. b. When the cutaneous test is negative. Delayed constitutional reactions occurring to allergens to which cutaneous reactions are negative belong to an entirely different group from the hyper- sensitive cases under consideration. ‘The discussion of them is therefore reserved for a later paper. 142 ROBERT A. COOKE CAUSES OF CONSTITUTIONAL REACTIONS 1. Mode of introduction of the allergen All the clinical studies in allergy here presented have been carried out by means of the intradermal test and the subcutaneous injection. Resort is never had to intravenous injection. The rate of absorption from the skin and subcutaneous tissue is approximately the same. General reactions from intravenous injection would undoubtedly be more immediate and more severe. In making the cutaneous and subcutaneous injection it is per- fectly possible that at times the point of the needle should lie in a small vein or lymphatic vessel and undoubtedly this does occur at times and may account for some of the constitutional reac- tions resulting, but such an occurrence is rare and seems un- avoidable and is of no practical account in comparison to the causes to be discussed under 2, 3, and 4. 2. The reactivity of the individual There are all grades of reactivity varying from those that give very marked positive reactions with the weakest dilutions to those that barely can be said to be positive with the most con- centrated extracts. Naturally those reacting violently to the weakest dilutions—with the pollen extracts a concentration of 0.0005 mgm. of nitrogen per cubic centimeter—are most sus- ceptible to the constitutional reactions even when tests and in- jections are properly carried out. An appreciation of this vari- ation in reactivity is of the utmost practical importance in the therapeutic management of the cases. 3. Activity of the allergens Great difficulty was experienced in the past by the instability of the extracts, notably the pollen extracts. As will be observed by a glance at the protocol, many of the constitutional reactions followed a change of extract; that is, from one that had been in use a few weeks to a new and freshly prepared one. Even by repeating the tests and by diminishing the dose accordingly it STUDIES IN SPECIFIC HYPERSENSITIVENESS. III 143 was not possible always to take sufficient account of the deteri- oration that had resulted and general reactions ensued. In 1921, with the use of extracts prepared by Coca (16) as given in another article in this issue this great disadvantage has been overcome and we now have preparations that have remained practically stable for over ten months. 4. Concentration and dosage of the extracts The concentration of the allergen used in the test and the total amount given by injection are of paramount importance. Dis- cussion of these questions is reserved for a later paper. 5. Cumulative effects a. With the same allergen. Case 2286, cited in the protocol, reacts clinically and by test to a num- ber of curious and unusual allergens, such as celery, skin of peaches, raspberry and in addition has hay fever and asthma from the pollens of daisy, dandelion and ragweed. On May 6, 1920, she was tested in- tradermally with extracts of daisy, dandelion, ragweed, celery, timothy, apple, lilac, wistaria. The first four extracts gave marked cutaneous reactions. She was then given by injection daisy and dandelion, 0.001 mgm. of nitrogen of each. There was considerable local reaction for twenty-four hours. No constitutional reaction developed. Two days later the same dose of the same extract of daisy and dandelion out of the same bottle produced a constitutional reaction within one hour with asthma, coryza and urticaria as the symptoms. It is difficult to in- terpret this reaction on the second injection except on the assumption of a cumulative effect, for on the first visit the patient had actually re- ceived more than on the second as the tests were done only a short while before the first injection was given. b. With different allergens. If a hypersensitive patient is tested with ten to sixteen different extracts and happens to be acutely reactive to four or six of them it is the writer’s impression that a general reaction is much more likely to ensue as a result of the sum total effect of all the allergens exerting their influence upon the same reacting mechanism. I have no case that affords 144 ROBERT A. COOKE absolute proof of this assumption. The suggestion is offered because it has a definitely practical bearing and will serve as a warning against the carrying out of too many tests at the same time. 6. Allergens which have caused constitutional reactions A list is here given of all the allergens that, in the writer’s experience have definitely caused constitutional reactions, either on test, injection, or ingestion: LePage Dog epithelium Orris Ovo-mucoid Dog saliva Dust extract Egg albumin and globulin Flaxseed Aspirin Pollens, all kinds Linseed Quinine Horse epithelium Cottonseed Ipecac Horse serum Mustard Chicken epithelium (feathers) Rabbit epithelium Pepper Duck epithelium (feathers) Rabbit serum Buckwheat Goose o Cat epithelium Wheat Nuts (peanut, hazelnut) Having thus outlined the causes of the general reaction it is obvious that they may be avoided by proper care. Granted than an occasional injection or a test may be accidentally de- livered in a venule a serious reaction will not occur if; first, extracts used are sufficiently dilute; second, injections are not repeated too frequently; third, tests, not to exceed six or eight, are made at one time. It is not the purpose of this paper to outline instructions for the use of extracts made according to the directions of Coca in another paper in this issue. The object here is to call attention to the dangers inherent in their use. Specific directions for their use will be given in a sub- sequent paper. FREQUENCY OF CONSTITUTIONAL REACTIONS It may be said that in the series presented general reactions have ensued with great frequency and this is true for the year 1920. But it may be added in extenuation, that these results occurred in a serious attempt to determine the limits of diag- STUDIES IN SPECIFIC HYPERSENSITIVENESS. III 145 nostic reactivity and to define the limits of therapeutic dosage. Furthermore, many hitherto unknown allergens have been dis- covered and in the positive identification of these reactions have occurred. The justification, if such is needed, lies in the fact that the study and correlation of the facts has enhanced our knowledge of the nature of allergy, as well as of the technic of diagnosis and treatment. The following table is based upon our series of 473 consecutive allergic cases studied in 1920. All of the cases giving constitutional reactions are included in the appended protocol. TREATMENT OF CONSTITUTIONAL REACTIONS A thorough knowledge of the treatment of the reactions should be had by every one attempting this type of study. First of all, one should quickly recognize the onset of the symptoms usually shown by the extensive urticarial wheals developing about the site of the injection or test or by a beginning erythema or short paroxysmal cough or increasing dyspnea. At once a tourniquet should be tightly applied about the arm above the site of the tests or injections in order to prevent the transporta- tion of more allergen to hypersensitive tissue through the systemic circulation by means of the lymph or blood stream. Adrenalin 1:1000, 1 cc. in adults, in children 0.4 to 0.6 ce., should be given at once subcutaneously or, if the reaction is severe, an intravenous injection of the same amount. If there is a continued increase of symptoms the dose should be repeated in two to five minutes. The writer has never seen any ill effects from these large doses of adrenalin and is confident that if it is used often enough and in large enough dose, the serious results of a reaction can be avoided. Adrenalin is the very best drug available. In the presence of a cardiac dilatation from violent respiratory effort or in the pres- ence of vasomotor collapse, strophanthin, 1 mgm., intravenously, (in children, a proportionate dose) should be given without de- lay. When the attack is controlled and has passed its peak of severity, morphin in proper doses may be used. The writer has never seen any advantage in the use of atropine in these attacks and the use of this drug militates against the best effect of strophanthin, should it be required. THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 2 146 ROBERT A. COOKE REFERENCES -(1) Smrra, Henry Lee: Arch. Int. Med., 1909, 3, 350. (2) Brown, AARon: Jour. of Immunol., 1922, 7, 97. (3) Larsen, H., Pappock, R., ALEXANDER, H. L.: Jour. Immunol., 1922, 7,81. (4) GorrsteIn: Therap. Monatsch., 1896, 10, 269. (5) Rosenav, M. J., anp ANDERSON, J. F.: Hygienic Lab. Bull., No. 29, Apr., 1906, p. 79. (6) Gittertre, H.F.: N.Y. State Med. Jour., 1909, 9, 373. ~ (7) Parx, Wm. H.: Trans. Assoc. Amer. Phys., 1913, 28, 95. (8) BuaxeLey, C. H.: Exp. researches on the cause and nature of hay fever, London, 1873; New observations on hay fever, London, 1889; British Med. Journal, 1898, 1, 867. (9) Dunpar: Berl. klin. Woch., nos. 26, 28, 30, 1905; Zeitsch. f. Immuntaets- forsch, 1907, 7; Deutsch Med. Woch., 1911, 37, 578. (10) RackEMANN, Francis M.: Jour. Amer. Med. Assoc., 1917, 69, 889. —(11) Waker, I. CHanpterR: Jour. Med. Research, 1917, 36, 459. (12) GeRTSTENBERGER, H. J., anD Davis, J. H.: Jour. Amer. Med. Assoc., 1921, 76, 721. (13) Cooxs, R. A., AND VANDER VEER, ALBERT, JR.: Jour. Immunol. 1916, 1, 231. (14) Cooxs, R. A.: Tice’s Practice of Med., Vol. 5, Page 481. (15) Cooxs, R. A.: Jour. Immunol., 1922, 7, 219. (16) Coca, A. F.: Jour. Immunol. 1922, 7, 163. STUDIES IN SPECIFIC HYPERSENSITIVENESS IV. NEW ETIOLOGIC FACTORS IN BRONCHIAL ASTHMA ROBERT A. COOKE From the Department of Bacteriology and Immunology, Division of Immunology in Cornell University Medical College, New York City, and the First Medical Division of the New York Hospital, New York City Received for publication December 27, 1921 The cutaneous test is accepted today as a diagnostic procedure in the study of human hypersensitiveness. The marked im- mediate cutaneous reaction that can be confirmed repeatedly is the evidence of the hypersensitiveness of the skin. In the writer’s article “On Constitutional Reactions,” in this issue of the Journal (page 000), it is shown by a table of cases that the typical cutaneous reaction is likewise indicative of clinical hyper- sensitiveness in at least 95 per cent of the cases reacting to the extracts of such airborne substances as pollens, powdered root of orris and animal danders. Bronchial asthma was early recognized as the chief clinical manifestation of the hypersensitive state in human beings and many of the cases were readily diagnosed by testing with extracts of such substances as those mentioned above, easily conceived and long known clinically to be important excitants of an attack. For those cases not thus easily diagnosed an explanation was sought along two lines, first that it was due to bacterial proteins acting as allergens and secondly that the paroxysm of asthma was a reflex effect. 1. BACTERIAL PROTEINS AS ALLERGENS The thesis has been advanced that the asthmatic paroxysm can be induced by the absorption of bacterial proteims which, acting as allergens, produce the symptoms just as any other 147 148 ROBERT A. COOKE foreign proteins in the specifically hypersensitive man. This was perhaps a natural corollary to the work already done on foreign proteins in hypersensitive humans by Noon, Schloss, Cooke, Goodale, Longcope, Talbot and others. Since Walker is largely responsible for the term ‘Bacterial Sensitization” or ‘Bacterial Asthma,’ meaning thereby that the bacterial protein, acting as an allergen, is a basic etiologic factor in one of the clinical allergies as bronchial asthma, let me review his work for the facts by which he concludes the thesis has been sustained. He began with a study (1) of the bacterial flora of the nasal and bronchial secretions of asthmatics and recovered principally Staphylococcus albus and aureus, Micrococcus tetra- genous and catarrhalis and a new diphtheroid organism. He carried out agglutination tests with Staphylococcus aureus, using the serum of 80 asthmatics. In only 3 cases was the test carried out with the serum of the individual from whom the organism was obtained. In 2 of these cases there was agglutination and in 1 there was not. Out of the 80 sera tested, 54 agglutinated, some in dilutions of 1:50 or less and some 1:100 or higher up to 250. No mention is made of any control with the serum of non-asthmatic individuals. He concludes (page 379) that no inferences can be drawn as to the relations between the isolation of Staphylococcus pyogenes aureus from the sputum and nasal secretions of patients and the agglutination tests of the sera of those patients. In his Study III (2) on the sensitization of patients with bronchial asthma to the bacterial proteins, as demonstrated by the cutaneous reactions, he reports on 100 asthmatics. He used a bacterial powder obtained by centrifuging the bacteria from normal saline in which they were washed. They were then washed twice in alcohol, afterward in ether and then pulverized. The powder was put on a scratch and dissolved in 4%} sodium hydrate. He observed five types of reaction, three of which he calls positive. One of these positive types is the urticarial wheal which we recognize as characteristic of the positive allergic cutaneous reaction, the second a small papule with surrounding erythema, the other consists solely of hyperemia. These reac- STUDIES IN SPECIFIC HYPERSENSITIVENESS. IV 149 tions develop in one-half hour at which time the reading is made. There is no reason given for considering them positive except that 67 cases in the group were negative, that is did not give the same hyperemia. Ninety-seven cases were tested with Staphylococcus aureus and 19 were considered to be positive. Nine of the 19 gave an urticarial wheal as the evidence of a reaction. There were no controls on non-asthmatic cases re- ported. His conclusion is that “In patients with bronchial asthma, positive reactions with the protein of Staphylococcus aureus are more common than with the protein from other bacteria.” In the next study (3) dealing with the subject Walker turns to “complement fixation and precipitin reactions with the serum of bronchial asthmatics who are sensitive to the proteins of wheat . . . . and bacteria, using these proteins as antigens, and the cutaneous reaction as an indication of sensitization.” Six cases of asthma are presented which gave a cutaneous reac- tion to Staphylococcus aureus that was called positive. Three of them showed a positive complement fixation and 3 were negative; but 2 of the 3 negative cases gave positive precipitin tests, while the third case, designated as M. 8. was negative. Speaking of this case, Walker says: “Since M. 8. was relieved of asthma by Staphylococcus pyogenes aureus vaccine, and the case, J. H. N., was greatly improved during treatment with desensitizing doses of S. pyogenes aureus protein, the asthma in these 2 cases would seem to be caused by staphylococcus pyogenes aureus.” The organism is not reported as having been sought in the nasal or bronchial secretions of this case M. 8. In con- clusion, Walker states (page 265) that it is not possible to cor- relate the results obtained with complement fixation, specific precipitation and cutaneous tests. But he then asserts in Study XIII (4) that ‘the cutaneous reaction has proven to be of great value in determining the cause of asthma from bacterial protem as from other proteins.”’ He then proceeds to a discussion of “the relationship between cutaneous reactions, serum agglu- tination tests, and bacterial examination of the sputum and nasal secretions in determining the part Staphylococcus pyogenes 150 ROBERT A. COOKE aureus and albus may play in the cause of bronchial asthma.” The author states that 30 cases form the chief basis of this paper, whereas only 21 appear accounted for in the text, 5 under protocol I, 2 under protocol IJ, 5 under protocol III, 3 under protocol IV, used as controls, and 6 other cases without pro- tocols. In conclusion he says: “The cutaneous test has proven to be the safest and best test for determining the bacterial cause of bronchial asthma.” Let us examine these cases and see what constitutes the proof. We must eliminate the 3 control cases, leaving 18 for consider- ation, 13 of which gave a positive cutaneous reaction to Staphylo- coccus aureus. In only 4 of the 13 cases was the organism recovered. Nine of the 13 were treated with Staphylococcus aureus vaccine and the asthma relieved at least temporarily, but in only 3 of these 9 was the organism recovered. In contrast to this there are 5 cases with a negative cutaneous reaction in 3 of which the organism was recovered and all did just as well under treatment with aureus vaccine. On the other hand, 12 of these 18 cases gave a positive agglutination test and the organ- ism was recovered in 7 of the 12. Eleven of the 12 were treated with aureus vaccine and relief of asthma was obtained in 10 of the 11 cases. In other words, in would be very difficult for any one to attempt on the basis of such figures to maintain that the cutaneous reaction was any more efficient than the agglutination test. From this point on all of Walker’s papers are based on his belief that he has proven the cutaneous reaction in bronchial asthmatics to be as etiologically diagnostic with bacterial pro- teins as with other proteins, such as egg, wheat, pollens and animal danders. This critical review of Walker’s studies can lead only to the conclusion that he has not brought forth any proof to show that his so-called positive reaction with bacterial proteins, as used by him in asthmatics, has any bearing upon the bronchial con- dition or is etiologically diagnostic in any individual case cited. He has only shown that asthmatics treated with a vaccine were relieved, irrespective of the reaction obtained with bacterial protein. His published results must be due to what may be STUDIES IN SPECIFIC HYPERSENSITIVENESS. IV 151 considered as a non-specific effect obtained by vaccine, thus differentiating the bacterial proteins very sharply from the other foreign proteins concerned in allergy with which non-specific results are not obtained. Rackemann (5) has supported this thesis of intrinsic bacterial asthmas, but he finds the non-hemolytic Streptococcus more prev- alent, having isolated this organism in 60 per cent of 40 cases. He used the carbolized bacterial suspension for the intradermal test and he describes positive early reactions occurring within a half hour and characterized by the typical allergic urticarial wheal as well as a late twenty-four-hour reaction with redness, swelling and tenderness. ‘Twenty out of 39 cases tested with autogenous vaccine gave a positive cutaneous reaction, both early and late reactions being used as criteria. In a group of 56 cases he obtained 60,7 per cent of positive reactions, which is considerably higher than Walker’s figures of 15.7 per cent, but Rackemann’s tests were made by the intradermal method and his cases were selected as probably infective in type. Walker’s tests were made by the scratch method and his cases were part of a general group of unselected asthmatics. Rackemann states that ‘Treatment was successful in fairly close accordance with the presence of a positive skin test.” In this work also the therapeutic results are used as a criterion by which this writer concludes that the cutaneous reaction demonstrates the importance of the bacteria in certain types of asthma. The importance of Rackemann’s work lies in the fact that in a majority of his cases he demonstrated the presence of the specific organism in the particular individual. We cannot compare the figures of Rackemann, that is 60.7 per cent of 56 selected cases considered as bacterial asthmas on the basis of a skin test, with his later study (6). These later figures show only 108 cases classed as bacterial asthma in a total of 590 cases, that is 18.3 per cent. In this later work the method of diag- nosis of bacterial asthma, that is whether by skin test or not, is not stated. Is there any way in which this question of the bacterial asthmas can be settled? Two years ago the writer (7) stipulated two 152 ROBERT A. COOKE postulates which must be fulfilled in order to establish the proof of the causal relationship between the allergen and the clinical reaction and they hold for bacterial as well as for other substances acting specifically upon hypersensitive man. 1. Hypersensitiveness must be demonstrated by one of the following procedures: a. A typical local reaction either cutaneous or ophthalmic must be elicited or b. The original allergic manifestation must be repro- duced at will on the introduction of the substance, either inhaled, ingested or subcutaneously injected. 2. It must be shown that the individual has come in contact in some way with the suspected substance in order to permit it to act an an etiologic factor. The writer has studied this subject somewhat from the point of view of these postulates. First, bacterial powders were made of various organisms by the alcohol-ether method described by Walker. The powder was then dissolved in carbolized saline to the point of saturation, filtered clear, and used for intradermal test. The organisms included Pneumococcus, types 1, 2, and 3, Staphylococcus albus and aureus, the latter two each made from at least a dozen strains. Tests were made in aseries of fifty cases of bronchial asthma and only two, Case Nos. 2783 and 2879, reacted with a typical immediate urticarial wheal, these reactions being confirmed. ‘The first of these two reacted to the extract of Staphylococcus albus and Pneumococcus type 1, and the second to Staphylococcus albus. Cultures of the nasal and bronchial secre- tions of the two yielded only a Streptococcus viridans in the first case and Streptococcus viridans and Micrococcus catarrhalis in the second. The Staphylococcus albus and Pneumococcus type 1 were sought but could not be recovered in either case. The cases in which Pneumococcus of the fixed types has been recovered in the sputum have all been tested but never yet shown a positive skin test. There was no suggestion that the bacterial extracts would operate as the other known allergens. Secondly, it is always possible with extracts of pollens, danders, foods, and drugs, to elicit an immediate constitutional reaction STUDIES IN SPECIFIC HYPERSENSITIVENESS. IV 153 in a hypersensitive individual by the subcutaneous injection, or, in the case of food and drugs, by the ingestion of the specific allergen in sufficient dosage. With the bacterial preparations, however, no similar immediate allergic reactions have been obtained in my clinic and I find no record of such in the litera- ture. The mere increase in the degree of an existing asthma twelve to twenty-four hours later will not suffice as proof. We have tried by such measures to prove the possible allergic nature of the bacterial reactions and have failed. Further work might be done and certainly better extracts might be prepared, but from such work we were forced to conclude that the bac- teria if they operate as fundamental causative factors in bronchial asthma, do not act as allergens per se, but in some way not under- stood today, and that the cutaneous test with bacterial proteins: is of no value in the diagnosis of allergy. In cther words, the conception of a bacterial asthma has been: based solely upon analogy and the analogy is not upheld by proof. ‘Bacterial Asthma” has become a convenient term by which to designate many of those cases not reacting to the genu- ine allergens, but such cases should be classed as undiagnosed. even though resort be had to vaccine therapy with apparently good results. Since the diagnosis of bacterial asthma is not as yet susceptible to positive proof and is only arrived at by exclusion the writer has taken the attitude that it is wiser to continue to search for new factors which can be shown to be specific agents in accord- ance with the postulates laid down. In this way by increasing the percentage of diagnosed allergic cases the possible cases of bacterial sensitiveness are more accurately separated from the whole group and will lend themselves to more productive study. 2. REFLEX ASTHMA Most of the present day writers, discussing vasomotor rhinitis and asthma either together or separately, appear to believe that a vasomotor edema of the respiratory mucus membrane with a resulting rhinitis or asthma can be induced by some irritant 154 ROBERT A. COOKE acting reflexly upon the membrane. Walker (8) discussing this point says ‘“The causes of symptoms may be classified as mechani- cal, chemical, odorific and thermal. Among the mechanical causes any kind of dust is the most frequent cause, more espe- cially sweeping dust and hay dust.” He further says, “Some of these patients are sensitive to some type of protein which may have rendered their nasal mucus membranes sensitive to these irritants, others are not sensitive to the proteins.” This is in part an adoption of Goodale’s idea regarding what he has termed olfactory vasomotor rhinitis or pseudohay fever. Goodale (9, 10) mentions the fragrance of certain plants such as lily of the valley, lilac and hyacinth, as excitants of attacks of sneezing as well as asthma, in cases where the tests with the pollens themselves were negative. In other words both these writers conceive that nonspecific irritants acting through a reflex mech- anism may be fundamental causes of the asthmatic paroxysm in individuals who are not hypersensitive. This is a return to the idea so well summed up by Osler in the first edition (1892) of his Practice of Medicine. He says Briefly stated, then, bronchial asthma is a neurotic affection char- acterized by hyperemia and turgescence of the mucosa of the smaller bronchial tubes and a peculiar exudate of mucin. The attacks may be due to direct irritation of the bronchial mucosa or may be induced reflexly by irritation of the nasal mucosa and, indirectly, too, by reflex influences from stomach, intestines, or genital organs. The writer (11) has contended that non-specific irritants can only operate in those cases that are specifically hypersensitive. They are never fundamental etiologic factors. To be sure these non-specific irritants do produce attacks of asthma in cases that have not been diagnosed but certainly in no greater number than in the diagnosed group. The further proof of the conten- tion lies in the fact that these non-specific and reflexly acting irritants cease to operate and to be productive of attacks in diagnosed cases where the paroxysms have been absent for a period of time either from removal of the specific cause or after improvement and relief as a result of specific therapy. STUDIES IN SPECIFIC HYPERSENSITIVENESS. IV 155 It is the special object of this paper to show that the two supposedly reflexly acting mechanical excitants particularly selected by Walker, namely, hay dust and house dust, are genuinely specific factors and that they operate in the specifically hypersensitive individual just as do pollens, animal dander and the other well known allergens that demonstrate their clinical effect after absorption by inhalation. A. Hay dust The first case in which hay dust was shown to act specifically is here described. Case 2207. A man, forty-five years of age, was seen in March, 1920, when he complained of bronchial asthma. His father was one of three asthmatic children, in a family of ten children. The patient’s first attack of asthma occurred in 1900 and came on when he was visiting in a farm house to which a barn was attached. The next attack occurred one year later after he lay down in timothy hay. Subse- quent attacks have come on following the handling of objects that had been packed in timothy hay. The patient states very positively that he has no trouble (either hay fever or asthma) in the country or else- where during the months of May, June and July, when the grasses are in flower. Tests carried out with the strongest extract of timothy pollen resulted negatively. Tests carried out later in both the skin and the eye with an extract of the timothy hay itself resulted in both instances in definite positive reactions. Similar tests made with the same extract upon normal individuals resulted negatively. Further use of the extract of the timothy hay has verified the specificity of the reactions. Three of the 327 cases of asthma under consideration were found to be hypersensitive to the substance. These cases indicate that attacks of asthma on exposure to dusty hay are not to be considered as a reflex effect of a non-specific excitant, as Walker states, but as an expression of a specific allergic reaction. B. House dust Following is the history of the case that led to the discovery of the presence in house dust of a specific allergen. 156 ROBERT A. COOKE Case 1763, T. F., male, age twenty-six, had had frequent attacks of asthma for fourteen years. He then enlisted in the army in June, 1917, and while stationed in Texas had no trouble at all. He returned home in October for six days and had severe asthma all that time. On his return to Texas the attacks disappeared and he was again free until he returned home in December. He applied for treatment in Janu- ary, 1918, having been continuously ill and unable to work for a month. He was tested by the intradermal method with extracts of all our then known products including pollens, sachet powders, animal danders, foods, and drugs with negative results. He was advised to use ap army cot and air pillow. The attacks were lessened. Several times he slept away from home and was well, but the attacks returned regu- larly when he stayed at home. Finding all tests negative, a culture of the sputum had been made and an influenza bacillus and Streptococcus viridans had been isolated. A vaccine! had been made and the patient had been treated for over two months with vaccine injections with no improvement in his condition whatsoever. In April of that year, he was instructed to bring all the dust that could be collected from his room, going over it carefully with a vacuum cleaner. This dust was then extracted just as pollens, orris and other substances are extracted, made sterile by filtration and used for intradermal tests. For the first time in this case very marked positive reactions were obtained in the skin at two sites of test and this was followed in a few moments by the development of a mild constitutional reaction with asthma, coryza, general erythema, and pruritus as the symptoms. With proper dilution of the extract this local reaction could always be elicited. A number of other cases hitherto undiagnosed gave marked cutaneous reactions with this same dust extract injected intradermally. The patient’s home was then investigated. It was a clean, modern house. His mattress was made of long curled horse hair, the pillows were feather; there was only a small rug on the floor. He had been negative to extracts of feathers and to horse hair and continued so on repeated testing. Although it was not possible to discover the exact substance giving the reaction, the presence of a specific airborne factor was demonstrated. The patient then moved to California where he has been entirely free from symptoms. 1 The culture and vaccine were made by Professor John Torrey of the Depart- ment of Hygiene, Cornell Medical School. The examination was made from the fresh specimen. STUDIES IN SPECIFIC HYPERSENSITIVENESS. IV 157 We were then face to face with a new procedure in the diag- nostic study of asthma and with a possible new but unknown substance. Efforts to discover the active ingredient in the dust extracts by comparative tests has so far yielded negative results. Of course, in houses in which there is a dog or a cat, the extracts of the dust agree by test but in lesser degree with extracts of the dander of the respective animals. The same holds true of the agreement by test of the dust extracts of rooms in which orris root powders are used with the orris extract itself, and this is the fact in those houses where rabbit hair pillows and mattresses are used. But dust collected from homes where none of these TABLE 1 Reactions s Fy _ . a E g ie 5} ei a a a z =I n <2) a 2185|+++| 0 |—--—-|---— Ssasae) 4h 0 0 |---| O |4+++4+ eee te ntetes ee O47 n 0 0 0 LIN rst eB 2170|+++)] 0 0 O.McOit' 0 0 Oeste =F lastest ae leer 2L60l-- 4-4-1 |) 0 0 Os Olelecla=t- leiestiestnlaiaaie ie sO 0 0 0 2165\++-+)] 0 0 Cee o 0 WY Nsespaal. Ww jee 2196|++-+)| 0 0 0 ;0;) O |---| O 0 OPE eae 21s) 0 0 0 |0; O |---| O 0 Osler ++-+ = Marked positive. 0 = Negative. ——— = Test not done. articles can be found also gives an extract which produces a marked positive cutaneous reaction in certain cases, and is still an unidentified factor. The extract does not agree by test with any other known extract used. Table 1 clearly demonstrates this point, and in all of these cases a single extract known as Dust Extract No. 1 was used and all tests were made by the intracutaneous method. The cases in table 1 were selected from those in which marked dust reactions were obtained, but it will be seen that while other extracts are occasionally positive there is no agreement between them and the dust reactions. Many similar cases could be cited, but the table is abbreviated pur- posely as greater length would not add to the argument. 158 . ROBERT A. COOKE The preparation of the dust extract is carried out according to the suggestions found in Coca’s article in this issue of the Journal of Immunology on the preparation of extracts. In brief, we can say that the dust collected by means of a vacuum cleaner is treated first with ether to remove all fatty substances, and is then extracted with the standard extracting fluid. The extrac- tion is allowed to continue for two or three days when the solu- tion is filtered off, sterilized by filtration through a Berkefeld and put through all the sterility tests. This extract contains nitrogen, probably, in many chemical combinations so that these extracts cannot be standardized satisfactorily by the nitrogen content and the nitrogen determination has only a relative value. While the chemical studies of these dust extracts have not been completed and are still under way, it can be stated here that by dialysis a considerable quantity of nitrogenous substance appears in the dialysate, but that the dialysate does not con- tain the reacting substance. When the dust extract is sealed in sterile tubes and heated to 212°F. for thirty minutes there is a diminution in the activity of the extract, but under these con- ditions no precipitate forms. When heated to boiling in an open vessel so that the carbon dioxide of the extracting fluid is driven off, some precipitation takes place and the extract loses all power of reactivity in a very short time. The question that naturally arises is, what is the actual or relative importance of this new extract? It can be at once stated that as a diagnostic procedure the testing of dust extracts is of the utmost importance in demonstrating the presence or absence of environmental substances which act on absorption by inhalation, and whether or not this environmental factor is domiciliary or occupational. The writer has always been im- pressed by the importance of the respiratory tract as the chief path of absorption in all adult asthmatics, and the following table arranged for comparison illustrates this point by the large percentage of cases in the inhalation group. In table 2 the cases of Rackemann and of Walker (12) are arranged for com- parison with 327 of the writer’s cases of bronchial asthma studied in 1920. The figures for 1921 will be published shortly. Racke- STUDIES IN SPECIFIC HYPERSENSITIVENESS. IV 159 mann’s figures are those supplied by him in a personal communi- cation and are more recent than the figures appearing in his article (6). All those cases, such as bacterial, not diagnosed by means of a positive cutaneous reaction are placed in the general group of “intrinsic asthma” and classed as undiagnosed. In this respect the cases of all three authors are treated in exactly the same manner. It will be noticed that the percentages of cases occurring from pollen, danders, powders, and foods do not make a total of 100 in any of the three writers’ statistics. This is due to the fact that many cases are examples of multiple hypersensi- tiveness and belong in two, or even three or more, groups and are therefore counted more than once. Consequently, in order TABLE 2 EXTRINSIC CAUSES ppitarsriy 8 2 , A ‘S Absorption ! Bacterial Absorption by inhalation 3 Besimestion ce BETTE A AUTHOR z a ae Coe Dn : ‘g 28] nose S | Pollens | Gander |pewder'| = | Food | Drugs | 2 | cases [o} — be » be » =I » he Pe) be > » be alS/2/3/8/2|8|2/8| 2/8) € | 2/4/28 Gee Ee |e lee eee Vela dee eae a zag ak eel t= Veo) dO) PA lag SH ase pigeon lace Rackemann ..... §90}150/25.0} 55) 9.0) 9 | 1.5/27.8) 19} 3.2) 0 31.0/408/69 .0 Walker? o5.050.52 400} 92/23 .0] 78/19.5) 0 23.0} 68/17 .0} 0 40 .0}241/60.0 Cooke ees ota 327| 86/26 .0)138/42.0/52 |16.0/69.8] 12) 3.6) 4] 1.2 |73.4] 87/26.6 to arrive at the percentage of cases due to substances absorbed by inhalation it has been necessary to take the percentage of cases undiagnosed and by subtraction arrive at the percentage of diagnosed cases. From the percentage of diagnosed cases we subtract the percentage of cases of food reactions and so arrive at the figures for the percentage by inhalation. For example: Among the writer’s cases 26.6 per cent are undiagnosed; this means that 73.4 per cent were diagnosed. Only 3.6 per cent of the cases are diagnosed as due to foods, and by subtraction this leaves 69.8 per cent as due to substances absorbed by inhalation. In other words, practically 70 per cent of all the writer’s cases studied are due to substances that are conveyed in the form of 160 ROBERT A. COOKE dust and that act after absorption by the mucous membranes of the respiratory tract. It deserves some comment that the per- centage of cases of the diagnosed group are for Rackemann 31, for Walker 40, and for the writer 73.4 per cent. I believe that these differences can be satisfactorily explained on three grounds. First, the number of routine tests made with extracts of airborne substances. The greater the number of tests the greater the improvement will be in the percentage of positive diagnoses; inevitably this must result in a decrease in the percentage of the so-called bacterial and undiagnosed cases. ‘There are a number of known substances to which Rackemann and Walker appear to attach little or no importance, for either they are not men- tioned at all, or only casually mentioned in their writings. For example: Scarcely a day passes at our clinic on which we do not diagnose by positive cutaneous reactions several cases that are hypersensitive to the dander of rabbit and goat. ‘The impor- tance of these reactions lies in the fact that we are able in many cases to demonstrate the presence of these substances in the homes of the individuals reacting. The untreated and unster- ilized hair of both rabbit and goat, containing a large amount of nitrogenous substance from the attached dander, is used very extensively, not only in New York but in all large centers of population, by families of foreign birth. The Italians import and use goat hair for pillows and mattresses, while the Germans, Slavs, Poles, and Hungarians, and among these especially the Jews, use rabbit hair for the same purpose. Second, Rackemann and Walker both use the so-called scratch test of Schloss, whereas the writer uses the intracutaneous test which Brown (13), in another article in this issue, shows to be by far the more delicate and the more efficient. Third, the writer uses fluid extracts prepared according to the directions published by Coca (14) in this Journal. These extracts, according to Brown (13) give a greater percentage of positive reactions even when the scratch test is used than do the extracts prepared according to the directions of Wodehouse (15) and made commercially available and used by Rackemann. Walker does not state in his publi- cations what extracts were used by him. STUDIES IN SPECIFIC HYPERSENSITIVENESS. IV 161 In order to show the importance of the dust extract, actually and relatively, a list is given in table 3 which shows all of the individual substances found to be of diagnostic importance in the 327 cases under consideration. In this table is also put down the number of times each one of these extracts was found to react as a causative factor, and the dust extract was important in 33 per cent of the entire group. It is seen that the number of reactions obtained is much larger than the number of eases studied in spite of the fact that 87 were negative to all tests; TABLE 3 Total number of cases 327 oar | PER CENT Noa PERCENT EXTRACT CASES Or EXTRACT CASES OF REactine | CASES REacTinc| CASES PALO LE Ssiieraie Ciec8 st 12 3.7 Horse epithelium....| 41 12.8 FRAP WER Gs. c4.+00 .oh2s 74 22.0 || Cat f Se lp aly Bars Omissey ce cosnn. 47 14.4 Dog & ee 18 D.0 1s Tbe eg aa peat 10 3.0 Rabbit Ms ee 18 5.5 Wiheatrs: ats aeeen 8 sit Cow of Sega 2 0.6 SoTL nen ee ee a 11 Te a | Me fs) OF Aan oy prt ee 3 0.9 Buekwheat........2..% 11 3.4 Ipecaer yates sas 1 0.3 Cottonseed........... 2 OF6./)| Hig pemamee meet: frek t Pp Bae ree ae Seats e 4 | ae He) a ee gi al ae 2 0.6 HEE er ets eet, SA a's 3 OS" 1 \Celeryerass.ta. sot s 2 0.6 su NOn Meo ee. ssc. 109 So 20" \|\\Peme ies i meten oo, ae it 0.3 ORUHETS ecco tase 2 84 251 “|W RASHOEPIY | jnnitats a: 1 0.3 and this is because a very large percentage are examples of multiple hypersensitiveness. What has been said in this paper for the importance of dust extracts in asthma obtains in the same way and in the same degree in allergic coryza. CONCLUSIONS In this paper it has been shown that the group of substances absorbed by inhalation play a much more important part as specific causative factors of asthma than is generally considered to be the case by other investigators. Diagnoses arrived at in accordance with this idea may be made with a greater degree of assurance on account of the fact that they are based upon posi- 162 ROBERT A. COOKE tive findings and not upon negative findings, as is the case with the so-called bacterial asthma cases, which the writer insists are to be more properly classified as undiagnosed. The new pro- cedure of testing dust extracts has yielded valuable information in that it permits a study of the occupational or domiciliary environment of an asthmatic and establishes a positive diagnosis in certain cases not obtainable by any other means. Further, it has shown the presence of a substance in most house dusts that is in itself an important factor, but the nature and source of which is as yet unknown. The dust of hay, also, may act as a specific allergen and is not to be considered solely as a simple mechanical irritant. REFERENCES (1) Waker, I. CHANDLER: Jour. Med. Research, 1917, 35, 373. (2) Waker, I. CHANDLER: Jour. Med. Research, 1917, 35, 487. (3) WaLxKeEr, I. CHANDLER: Jour. Med. Research, 1917, 36, 243. (4) Waker, I. CHANDLER: Jour. Med. Research, 1917, 36, 295. (5) RackEMANN, Francis, M.: Jour. Immunol., 1920, 5, 373. (6) RacKkEMANN, Francis, M.: Am. Jour. Med. Sci., 1921, 162, 807. (7) Cooxn, R. A.: Tice’s Practice of Medicine, Vol. 5, Page 496. (8) WaLker, I. CHanptER: Jour. A. M. A., 1920, 75, 782. (9) Goopatez, J. L.: Boston M. & 8. Jour., 1916, 175, 181. (10) Goopatsz, J. L.: Boston M. & S. Jour., 1918, 179, 295. (11) Cooks, Ropert A.: Tice’s Practice of Medicine, vol. 5, pages 499 and 513. (12) Waker, I. CHANDLER: Boston M. & S. Jour., 1918, 179, 288. (13) Brown, Aaron: Jour. Immunol., 1922, 7, 97. (14) Coca, ArTHuR F.: Jour. Immunol., 1922, 7, 163. (15) WoprHouss, R. P.: Boston M. & S. Jour., 1916, 175, 195. STUDIES IN SPECIFIC HYPERSENSITIVENESS V. THE PREPARATION OF FLUID EXTRACTS AND SOLUTIONS FOR USE IN THE DIAGNOSIS AND TREATMENT OF THE ALLERGIES WITH NOTES ON THE COLLECTION OF POLLENS ARTHUR F. COCA From the Department of Bacteriology and Immunology, Division of Immunology, in Cornell University Medical College and the New York Hospital Received for publication February 1, 1922 From the beginning of his work upon the diagnosis and specific treatment of the allergies (hay fever, asthma, the urticarias), Robert A. Cooke (1, 2) has employed for both the diagnostic and therapeutic injections, fluid extracts of the various materials containing the exciting agents of these conditions. These ex- tracts were originally made with physiological salt solution to which was added the usual percentage of carbolic acid The preparations were standardized according to their nitrogen con- tent. Cooke and also Cooke and Vander Veer, and Vander Veer, have reported their successes in the use of these extracts as diagnostic and therapeutic preparations At the time when these investigators became associated with the writer in the organization that has been established in the New York Hospital for the diagnosis and treatment of the allergies, some of the ex- tracts that had been in use were subject to certain disadvantages. In particular, the pollen extracts were often quite unstable in their activity both as diagnostic agents and as therapeutic material; the extract of feathers, especially chicken feathers, seemed to possess little or no specific activity in cases of known hypersusceptibility to these materials. The preparation of fluid extracts was undertaken by the writer with the purpose of overcoming the disadvantage mentioned. It was necessary to provide a sterile preparation of not too low 163 164 ARTHUR F. COCA concentration which would remain stable for at least six months. The pursuit of this problem was attended with some serious diffi- culties which impeded its purpose. One of these difficulties lay in the fact that the sole test object was a human being suffer- ing from a distressing condition, which might or might not be relieved by the therapeutic use of the new kind of protein preparation. There was, therefore, a natural hesitancy to change from one kind of preparation to another for experimental reasons. Fur- thermore, the comparison of different preparations by means of the cutaneous or of the ophthalmic reaction is interfered with by the natural difficulty of obtaining consent of suitably sensitive individuals to submit to the experimental injections. The prob- lem was further complicated by the number of the possible factors concerned in the deterioration of the preparations; for example, light, temperature, chemical reaction, quality of glass in which the preparations were stored, and sterility. Indeed, it must be admitted at the outset that some of these factors have not yet been investigated. It is not known why the preparations in present use satisfy the requirements of potency and relative stability; whether on account of mere sterility or on account of a modification of the composition of the extracting fluid. In any case the preparations about to be described are to be looked upon as merely modifications, perhaps only slight ones, of those employed previously by Cooke and his associates. While the problem is still under investigation it has been found desirable to make the present preliminary publication on account of the thoroughly successful diagnostic and therapeutic use of which the preparations have been found capable. In deciding upon the composition of the extracting fluid, the fact has been borne in mind that the exciting agents often reach the sensitive mucous membrane in a dry state. Hence it could be assumed that the active principle in these materials was soluble in a slightly alkaline and saline solution. While it was known that certain vegetable proteins for example, gliadin, are soluble in aqueous media only in the presence of free alkali (sodium hydrate in hundredth normal concentration or stronger), STUDIES IN SPECIFIC HYPERSENSITIVENESS. V 165 it was believed that such proteins could have little significance in human hypersensitiveness since they are not soluble in the natural secretions of the mucous membranes. Free alkali was thought to be disadvantageous as an extracting agent on account of its known tendency to denature certain proteins. The de- sired alkalinity was therefore obtained with the use of sodium bicarbonate. This salt has the desirable faculty of neutraliz- ing both free acid and free alkali and is able, in consequence, to correct any change toward either reaction which they tend to develop in the course of preparation, or during storage. After some experimentation, the following composition of the extracting fluid was adopted. RPEEINEERY CHIPOLTOLE Re tee erect ovat: 2\s:4 = ela vale cvatetaierenidelete naman 0.5 per cent Sodium bicarbonate....NaHCoO; in such concentration that 10 ce. of the final fluid equalled about 3 ce. of 7p alkali Spar ali Gy AGIOS, 2p bomen ajo fis, 5 6 6 Oke In final concentration of 0.4 per cent The solution was made without the use of heat and with the avoidance of excessive shaking. When it was desired to dilute an extract, or other preparation with this fluid after the original extract had been sterilized, the diluting fluid itself was sterilized by filtration through a sterile Berkefeld filter. As a general rule this alkaline extracting fluid is used for all dry materials such as the cereals, the danders, the nuts, and the pollens. It is also used for certain vegetables that contain little juice, such as sweet potato, fresh beans, fresh peas and for the meats. The extracting fluid is used as a diluting fluid when dilution of the extract is desired. When the original material contains considerable fluid such as the fruits and most of the vegetables, it is advantageous, in order to avoid too great dilu- tion, to use a ‘‘preserving fluid’”’ containing the constituents of the ‘‘extracting fluid” in a higher concentration. Such a pre- serving fluid we have prepared containing 2.5 per cent NaCl, 1.25 per cent NaHCO;,2. per cent carbolic acid. It has been found that many of the fruit juices contain much more than enough acid to neutralize all of the NaHCO; in the added pre- serving fluid. We do not know that the resulting partial neu- tralization of the acid is of any advantage in the preparation of 166 ARTHUR F. COCA the acid juices. For several reasons it has not been possible to make an adequate study of the method of preparing these juices. We have used the preserving fluid referred to for both the neutral and the acid juices on the ground of convenience. The extraction of the dry materials is carried out at room tem- perature and usually this extraction is continued for forty- eight hours—sometimes for three days. In these conditions the concentrations of carbolic acid (0.4 per cent) in the extracting fluid is not always sufficient to present the multiplication of bac- teria in the mixture. This difficulty was met with the use of toluol, which is able not only to prevent bacterial growth but to kill nonspore bearing organisms. When the material to be ex- tracted is in the form of a powder or fine meal it is advantageous to mix the toluol with the powder before adding the extracting fluid as this makes certain that the toluol reaches all of the sub- stance. It also prevents the formation of clumps which are difficult to break up. If the material contains oily substances which may interfere with the infiltration of the extracting fluid, the oil is first removed with ether which does not denature proteins. We have not found it necessary nor convenient to use a special extraction apparatus such as the familiar Soxhlet apparatus. We have simply mixed the ether with the substance in a sedimenting jar and as soon as sedimentation was complete we have decanted the ether and made further extractions in the same way with additional portions of ether. As the different materials contain different percentages of oily substances, the number of changes of ether must be varied. With nuts and oily seeds, we have been guided in this respect by making a rough determination of the quantity of oil removed by the successive fractions of ether. This determination was made by evaporating a few cubic centimeters of the etherial extract in a beaker immersed in hot water and noting the quantity of the oily residue. With the pollens the extent of the oil ex- tracted could be judged by the depth of color imparted to the successive portions of ether. Complete extraction of the oily substances has not been found necessary for some materials, such STUDIES IN SPECIFIC HYPERSENSITIVENESS. V 167 as the nuts and the pollens. After the extraction of most of the oily substances the ether has been driven off from the material before it was mixed with the extracting fluid. In most of the extracts and preserved juices a precipitate forms upon standing. As precipitation continues even after the fresh extract has been filtered, it is necessary to wait until the precipitation is complete before carrying out the further steps of the preparation. In the case of some of the vegetables, this precipitation has been found to cause relatively little reduction of the nitrogen content of the extract. On the other hand, the precipitate in the extract of the meats and fish is doubtless wholly proteid. Conceivably some of the exciting agents of the allergy to the original material are in part or entirely lost by this precipi- tation. We have no evidence to offer on this point. The separa- tion of the precipitate from the fluid extract offers difficulties varying in degree with the different materials. If the precipitate is not too voluminous it can be removed at once with the use of the Sharples centrifuge. This instrument, which is almost in- dispensable in the preliminary clearing of larger volumes of fluid, is not adapted to the handling of quantities much less than five hundred cubic centimeters. The smaller quantities of extract may be filtered through paper, preliminary to the final steriliz- ing filtration through Berkefeld or similar filters. The precipitate that forms in some of the extracts is so volumi- nous that it is impossible to use the Sharples centrifuge for its removal. In such a case, a partial separation can usually be effected with the use of a fine mesh towel laid over a sieve. As the precipitate tends to form an impervious mat upon the cloth, it is necessary at intervals to scrape off the collected precipitate with a large spoon. During the period in which the sedimenta- tion is taking place in the extract it is advisable to keep the fluid covered with a shallow layer of toluol. To prevent evaporation of the toluol the container should be tightly stoppered. Many of the extracts are found to be sterile after having stood for three days or longer under toluol. Advantage has been taken of this circumstance in the preparation of some protein sub- stances, particularly the isolated globulins, which are not at all 168 ARTHUR F. COCA or with difficulty filterable through the Berkefeld type of filter. However, sterilization of the extract has been secured for much the greater part by Berkefeld filtration. The filtering flasks were sterilized as usual with dry heat. The filters were sterilized by one exposure in the autoclave to steam under a pressure of fifteen pounds for fifteen minutes. The connecting perforated rubber stoppers were boiled for ten minutes in 5 per cent carbolic acid and again for ten minutes in water. The main stock of the filtered extract was stored in sterile 16-ounce bottles. For clinical purposes extracts are distributed in vaccine bottles or homeopathic vials, which are capped with ‘‘No-air” stoppers. The greater part of the stopper must be cut off with scissors so that the cap may be more readily punctured by the syringe needle in removing the contents of the bottle for test purposes. The bottles and caps are sterilized separately by boiling for twenty minutes in plain water. The caps are held firmly on the vials with the use of narrow rubber bands. Sterility tests have been made by introducing about 0.25 ce. of a filtered preparation at the bottom of a 12- to 15-em. column of neutral nutrient agar containing 1 per cent of dextrose. Another portion of the preparation was deposited at the surface of the agar in the same tube. The agar was quickly solidified by placing the tube in cold water. The nitrogen content of all the preparations was determined by the Kjeldahl method and generally adjusted by dilution to 0.5 mgm. or less per cubic centi- meter. Further dilutions of these ‘‘stock solutions’? were made for diagnostic and therapeutic use. It is not in the province of this article to discuss the dosage of the different preparations nor the dangers attending their use. These questions are considered in the various publications of Cooke, Vander Veer and Brown. It may be stated here, however, that the nitrogen content of the usual dilutions of all the pollen extracts was: 0.1, 0.05, 0.01, 0.005 and 0.001 mgm. With many of the vegetables and fruits it was thought advantageous to leave the nitrogen content at its original concentration. While there are some general principles that apply to all of the members in each of the several groups of materials, the method STUDIES IN SPECIFIC HYPERSENSITIVENESS. V 169 employed for each group must be modified to meet certain pecu- liarities of some of the members of the group. It is intended to refer to some of these peculiarities, although it is not possible in the limits of this communication to describe the procedure followed in each of the individual extracts that have been pre- pared. It will be useful to describe in some detail the method employed in several instances under each group. Dry materials Whole wheat flour: One kilogram of the flour was thoroughly mixed with 150 ce. of toluol and then with 2200 ce. of extracting fluid. After twenty-four hours at room temperature in which the mixture was once thoroughly stirred up the supernatant fluid was drawn off into a filter- ing flask with the use of negative pressure. With the sediment 2000 ce. of fresh extracting fluid were mixed and after a further period of twenty-four hours the supernatant fluid was drawn off as before and mixed with the first extract. Three liters of fluid were thus obtained. A slight sediment settled in the fluid over night, which was not greatly increased in the next few days. The supernatant fluid was filtered through a Berkefeld filter. Filtration was slow, but not so slow as it was with an extract of another sample of whole wheat flour which had been prepared previously. One cubic centimeter of the filtrate was found to contain 175 mgm. of nitrogen. Rice polish:! 12 pounds of rice meal were mixed with 150 ec. of toluol and this mixture was stirred well first with one liter of extracting fluid; to this mixture was then added a second liter of extracting fluid; the mixture was left at room temperature until the following day when 300 cc. of supernatant fluid were drawn into a filtering flask. 750 cc. of extracting fluid were mixed with the sediment and on the following day an additional 750 ce. of supernatant fluid were obtained and mixed with the first extract. After a first filtration through the Berkefeld filter the nitrogen content per cubic centimeter was 1.4 mgm. A slight 1A generous’ quantity of this material, which is also called rice meal, was kindly donated by the Louisiana State Rice Milling Co., 100 Hudson Street, New York City. It has been assumed that the allergen of rice is a protein which is the same in all parts of the grain. In accordance with this idea the rice preparation has been made from the materials (“rice bran” and ‘rice meal’), which are removed from the grain in the process of polishing, because these materials contain about 14 per cent of protein. 170 ARTHUR F. COCA precipitation made a second filtration necessary. The nitrogen con- tent after this final filtration was practically the same as it was after the previous filtration. Dry “‘navy”’ bean: The dry beans were ground in a meat chopper and the coarse meal was then ground in a coffee mill. The powder was mixed with toluol and extracting fluid. After twenty-four hours the fluid was pressed out through a clean towel and centrifuged in a Sharples centrifuge. The extract was covered with toluol in a tightly stoppered flask and allowed to stand at room temperature. After five days a voluminous precipitate had formed and settled. The precipitate was removed in the Sharples centrifuge. After a further five days under toluol a second precipitate had formed which was likewise removed by centrifugation. Precipitation continued for several months. A clear extract was obtained at the end of six months in which no further pre- cipitation had taken place one month later. One cubic centimeter of extract contained 5 mgm. of nitrogen. Horse dander: A quantity of dander obtained by currying and con- taining few hairs was mixed with three volumes of ether by stirring. After the sediment had completely settled the ether was decanted and discarded. The material was again extracted with another equal por- tion of ether. The ether was completely removed from the sediment by stirring the latter in a beaker which was immersed in hot water (50 to 60°C.). Fifty grams of the resultant material were mixed with 100 ce. of toluol and then with 1000 ce. of extracting fluid. The mix- ture was allowed to stand over night in a stoppered flask. During this period the mixture was shaken once. On the following morning the entire mixture was shaken up and thrown on a hardened filter paper. Eight hundred cubic centimeters of a clear brown filtrate were obtained after six hours, this quantity being increased to 800cc. over night. The filtrate was immediately filtered through a Berkefeld filter. One cubic centimeter of the filtrate contained 0.5 mgm. of nitrogen. The filtrate remained clear. Pollen: 90 grams of ragweed pollen were treated with ether in the same way as was the horse dander until the decanted ether showed only a slight yellow color. Four or five extractions suffice for ragweed pollen. The grass pollens contain less oil than the ragweed pollen. After the ether had been removed from the pollen as in the case of the dander, the entire quantity of pollen was mixed with 2700 cc. of ex- tracting fluid. The mixture was covered with toluol as usual. After four days, during which time the sediment was shaken up once or STUDIES IN SPECIFIC HYPERSENSITIVENESS. V 171 twice daily the supernatant fluid was decanted and the sediment was mixed with a second portion of 450 ec. of extracting fluid. As soon as the sediment had settled the supernatant fluid was decanted and mixed with the first portion. The combined decanted fluid, amounting to 3000 cc. was filtered through a Berkefeld filter. One cubic centimeter of filtrate contained 0.3 mgm. of nitrogen. The filtrate remained clear. One cubic centimeter of an extract of timothy pollen made in the same way contained 0.42 mgm. of nitrogen. Some idea of the rate of deterioration of the pollen extracts was obtained in the following experience: Four different preparations of the extract of timothy pollen were made with the extracting fluid in the manner above de- scribed. Preparation Ta was made March 4, 1921 and kept in the ice- box. Preparation Tb was made December 9, 1921. Preparation Te was made November 10, 1920, and kept at room temperature (15 to 32°C.). Preparation Td was made February 12, 1921, and kept at room temperature. The nitrogen content of the four preparations per cubic cen- timeter was as follows: Ta—0.4 mgm., Tb—0.36 mgm., Tc— 0.28 mgm., and Td—0.336 mgm. All of the older preparations had been preserved in their original concentration. A few days after preparation Tb was made, the four preparations were diluted to an equal nitrogen content in three different concentrations and compared in the clinic as to their specific activity by in- tradermal injection in an individual subject to early hay fever. The results are shown in table 1. These results indicate that the pollen extracts deteriorate some- what in nine months if they are kept in the ice-box and that this deterioration is considerably greater if the extracts are kept at room temperature. Feathers: Feathers are washed once with enough ether to wet them well. The ether is wrung out by hand and the feathers are spread upon clean newspapers to dry. The removal of the ether is hastened by occasionally turning over the mass of feathers and by compressing 172 ARTHUR F. COCA them. The feathers from which most of the oil has been thus removed are passed in successive batches through 2 one-liter portions of extract- ing fluid, each batch being allowed to remain about ten minutes in each portion of fluid. Naturally the successive batches of feathers carry a certain quantity of fluid with them after the immersion, and this cir- cumstance places a limit on the quantity of feathers that can be treated in this way by the two liters of extracting fluid. On the other hand, the more feathers used the greater will be the concentration of the extract- able protein in the resulting extract. After the last batch of feathers has been carried through, the two portions of fluid are combined and cleared in the Sharples centrifuge. As precipitation sometimes takes place in the extracts of feathers it is well to allow them to stand for a few weeks under toluol before the final Berkefeld filtration. Hair and wool are treated in the same way as feathers. TABLE 1 Showing the relative specific activity of the different extracts of timothy pollen on intradermal injection 0.001* 0.005 0.01 (OSS S Ee: io ce pac Sa Geen ae =E=5 +++ t44+ Mica ipsa & anlar ce Ps cere eae +444 ttt4 UNG Toso Roos pene aoe ee ee ze ape Dane= eee one Aaa unbeas on abe aioe trai Shoe jt * Milligrams of nitrogen per cubic centimeter. Moist materials Moist materials, from which little or no juice can be expressed by hand, such as meats, sweet potato, green pea, turnip, cauliflower, and lima bean, are passed twice through a meat chopper, mixed with ex- tracting fluid and covered with a thin layer of toluol. The extraction is generally interrupted after twenty-four hours and the fluid is obtained by pressing the mixture in a stout towel by hand. The amount of extracting fluid used varies with the nitrogen content. For the meat three or four volumes of the fluid (figured on the weight of the material) are used. For the vegetables, one or two volumes. Some of the shell fish yield a juice in addition to the meat. To the juice one-quarter of its volume of preserving fluid is added. With the chopped meats three or four volumes of extracting fluid are mixed. These two portions are then thrown together and allowed to stand under STUDIES IN SPECIFIC HYPERSENSITIVENESS. Vv 173 toluol for at least two weeks. With some of the fruits and vegetables that can be peeled, such as peach, tomato, orange, lemon, and grape fruit, it is advantageous to make Separate preparations of the pulp and the peel. The pulp in such case is carefully separated from the remain- der of the fruit and the juice is obtained from it by squeezing through a towel; the peel is ground in a meat chopper and mixed with a quantity (not too great) of extracting fluid. Juices of fruit and of some of the vegetables, such as string bean, celery, cabbage, lettuce, spinach, cu- cumber, and white potato, are mixed with one-quarter of its volume of preserving fluid and allowed to stand in well stoppered containers under toluol with which the fluid should be once thoroughly shaken. Miscellaneous preparations In general, heat has been avoided in the preparation of the extracts, although according to Wodehouse this precaution is not necessary in the case of some of the vegetables. In the preparation of fish glue, the glue has been diluted with ten times its weight of 0.5 per cent sodium chloride. The diluted material has been sterilized by boiling. E9q preparations As in the early experiments of Schloss, different proteins of the egg have been isolated for separate testing. The first step in all the methods of separation of the egg white proteins consists in dissolving the egg white in water. This process is expedited by forcing the whites through @ moist towel into the diluting fluid. The eggs used for the purpose were strictly fresh (not more than a few days old). Euglobulin was obtained by saturating a 30 per cent solution of egg white with sodium chloride. The globulin was purified by repeated precipitation. The final precipitate was dialyzed against 1 per cent sodium chloride in the presence of toluol. The diluted material was kept under toluol until the solution was found to be sterilized. This method of sterilization was made necessary by the fact that the globulin could not be filtered through the ordinary Berkefeld filter, The use of the preparation was. discontinued because no way was found to prevent continued precipitation of the substance. Pseudoglobulin, crystalline egg albumin and Ovo-mucoid were sepa- rated in the usual way, except the final stages of precipitating and washing the coagulable proteins with alcohol. This step was omitted because of the resulting denaturing of the proteins. These proteins can be sterilized by filtration. 174 ARTHUR F. COCA In carrying out the cutaneous test upon individuals believed to be sensitive to egg white proteins it is convenient to make a preliminary test with the whole egg white. This procedure, however, suffers from the disadvantage that the several component proteins of egg white are not present in equal concentration. At the suggestion of Dr. Cooke we have attempted to overcome this difficulty by mixing equal volumes of solutions of globulin, albumin, and ovo-mucoid. These solutions contain the respective isolated proteins in the same concen- tration; namely, 1.5 mgm. of nitrogen per cubic centimeter. The globulin had been separated by repeated precipitation with half satur- ated ammonium sulphate. This preparation no doubt contained albumin. It included both euglobulin and pseudoglobulin. The ammonium sulphate had been removed by dialyzing against changes of 1 per cent sodium chloride. Toluol was used as an antiseptic during the dialysis. The albumin fraction was obtained from the filtrate after the separation of the globulin by the addition of a little acetic acid and by increasing the (NH4)2SO, saturation to a little over 7 per cent. Both crystalline and non-crystalline albumin were thus included in this fraction. Dilutions of the egg and milk proteins have been made with 0.4 per cent carbolic acid in 1 per cent sodium chloride. This diluting fluid was sterilized by Berkefeld filtration. Carbolic acid should not be mixed in concentrated solutions, such as 5 per cent, with the diluted egg proteins, as this causes precipitation of some of the proteins. Yolk proteins: For the separation of ovo-vitellin we have used the method of Levene and Alsberg, up to the point of extraction with alco- hol. This latter step was omitted. The final precipitate was dis- solved in 10 per cent sodium chloride and mixed with an equal volume of distilled water which was added slowly with constant stirring. Since the substance is insoluble in physiological saline solutions it is possible that even a skin potentially sensitive to it would fail to react to its injection. No diagnostic reactions have been obtained with the prep- aration. Yolk proteins soluble in physiological salt solution. The following method has been used to obtain yolk proteins that are soluble in physiological saline solution: The yolks were separated from the whites and placed in 1 per cent sodium chloride solution. Each yolk was lifted out of the fluid in a large spoon and the chalazae and film of egg white were removed with the use of a soft towel and scissors. lod STUDIES IN SPECIFIC HYPERSENSITIVENESS. V 175 After a further rinsing in a large dish of 1 per cent sodium chloride solution each yolk was again lifted out of the fluid on the spoon and the surface was dried with a towel. The yolks were broken up in a large separatory funnel and treated with several portions of ether, then with about 3 to 2 volume of 10 per cent sodium chloride. To the resulting solution were added 5 to 6 volumes of distilled water which precipitated the ovo-vitellin. The precipitate was removed by paper filtration. The filtrate was acidified with a few drops of 40 per cent acetic acid and mixed with somewhat more than an equal volume of saturated ammonium sulphate. The resulting precipitate was collected on a paper filter, placed in a dialyzing bag (fish skin) and dialyzed against changes of 1 per cent sodium chloride until no trace of sulphate remained in it. Toluol was depended upon for sterilization. Milk proteins: Casein is precipitated as usual with acetic acid and separated from the fluid by straining through » stout towel. The towel containing the casein is immersed in changes of water in which the casein is thoroughly washed. With the aid of a pestle, the casein is forced through a wire kitchen strainer, while an assistant keeps the casein well moistened with distilled water. To this suspension of the precipitated casein in water 4 per cent sodium hydroxide is slowly added, while the mixture is continually stirred. The addition of the alkali is continued until early all of the casein is dissolved. Stirring is continued some time beyond this point and if some casein still remains undissolved the solution will be neutral to litmus. The mixture is filtered through paper and the filtrate is kept under toluol. Before filtration through a Berkefeld filter, the solution is further diluted with 0.4 per cent carbolic acid in 1 per cent sodium chloride to a nitrogen content of 2 mgm. per cubic centimeter. After filtration, the nitrogen content per cubic centimeter is further reduced to 1 mgm. From the original acetic acid filtrate the other proteins are precipi- tated with the addition of 13 volumes of saturated ammonium sulphate and the precipitate, after its collection on a paper filter is freed of the ammonium salt by dialysis against changes of 1 per cent sodium chloride. COLLECTION OF POLLENS In the collection of pollens, two principal requirements have been laid down. The first of these has been to obtain the pollen as free as possible from other material (dust from the soil and 176 ARTHUR F. COCA from other parts of the plant); the second equally important requirement is that all moisture be eliminated from the collected pollen before it is stored. The method of collection is different with the different groups of plants; that for the grasses is as follows: the heads are cut off with scissors and spread, in not too thick layer, upon strips of glazed paper. In order to obtain the pollen as free from extraneous matter as possible it is important to shake the collected heads vigorously over a sieve or coarse wire netting before spreading them on the paper. The room used for this purpose should be warm, dry, well lighted and protected from dust. The attic rooms of the ordinary house are generally suitable. Ordinarily the greater part of the pollen obtainable in this way drops out of the heads within twenty-four hours. It has been found, however, by transferring the heads to a fresh sheet of paper at the end of twenty-four hours, that an additional quantity of pollen can be obtained in the second twenty-four hours. For this reason alone it seemed profitable to leave the heads on the glazed paper for at least forty-eight hours. A more important advantage, how- ever, is offered by the observance of this longer period. This advantage is the more thorough drying of the pollen which is so necessary to its proper preservation. Pollen that has been stored in tightly stoppered bottles without first being allowed to dry tends after a time to form clumps and may even become united in a single mass, perhaps under the influence of bacterial growth. On the other hand, if it is sufficiently dried, the material remains indefinitely in the form of a fine yellow powder. If the pollen in this condition is protected from moisture as in a tightly closed fruit jar, the activity of the extracts made from it after at least a year’s standing corresponds as in the begin- ning with the nitrogen content. There is thus no evidence of deterioration. At the end of forty-eight hours the heads are taken up from the paper and rubbed in a large wire sieve in order to obtain most of the pollen that still adheres to them. The pollen together with the dried flowers of the grass, which likewise have dropped off or are rubbed off in the sieve, are collected in one place by raising STUDIES IN SPECIFIC HYPERSENSITIVENESS. V LTZ the ends of the paper strips and tapping the under surface of the paper. The whole mass is then emptied into a clean enameled pan. Most of the dried flowers cam be removed from this mass by preliminary sifting with a fine domestic sieve of about four inches diameter. The final sifting should be carried out with a 200-mesh copper wire sieve with the use of a soft, fine, hair brush of one to one and a half inch breadth. The season for collection is different with the different grasses. That for the most important member of this group, the timothy grass, begins in the neighborhood of New York City about June 20 and ends practically July 6. The heads are not ready for picking until some stamens have appeared on them. The most favorable time of day for collection is in the late afternoon and evening. We have an impression that the yield of pollen is somewhat greater when the heads are collected soon after a shower. The season for collecting ragweed pollen begins about August 20 and ends about September 5. The method used for the grasses is not applicable here. It was thought necessary, therefore, to shake off the pollen as it ripens naturally in the field. The pollen of a single plant does not ripen all at one time but over a period of a week or longer. Several of the sessile anthers on a single bract open each day and the contained pollen is released. This process takes place usually in the morning soon after the sunlight has reached the plant. It is profitable to begin the collection as soon as the anthers have opened because after this has taken place the pollen begins to drop out by its own weight or is carried away by the wind, which generally begins to blow after sunrise. The pollen is collected by bending the tops of the plants down so that they can be inserted into the collect- ing bags and gently shaken. The pollen is readily loosened in this way. Too vigorous shaking loosens the pollen on other parts of the same plant. It requires some dexterity to get the tops into the bag before the pollen has been jarred off. The bags are made of the same glazed paper as was used for the collec- tion of the timothy pollen, and they measure about 11 by il inches. THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 2 178 ARTHUR F. COCA While the foregoing method of collecting ragweed pollen will always be useful for small-scale collection, it will be superseded for collection on a large scale by the method about to be described. After the bracts have begun to open the stems that bear them are gathered and allowed to dry. When the bracts are thoroughly dried they are stripped off the stem and ground in a mortar or in a wooden ball mill in order to loosen the pollen. The ground mass contains not only the pollen but also other portions of the bracts. It was observed that the pollen sediments rapidly in ether but floats in carbon tetrachloride and it was suggested by Olin Deibert of this division that in a mixture of these two fluids the pollen alone might float whereas most of the other parts of the plant might snk. After some experimentation by Deibert a mixture of 75 parts of carbon tetrachloride and 25 parts of ether was found to effect the desired result. The pollen recovered by this method may be washed with two changes of ether, then dried and sifted in a 200-mesh wire sieve and pre- served in tightly stoppered bottles. REFERENCES (1) Cooks, R. A.: Laryngoscope, 1915, 25, 108. (2) Cooxr, R. A. and VanpER VEER, A.: J. of Immunol., 1916, 1, 201. STUDIES IN SPECIFIC HYPERSENSITIVENESS VI. DERMATITIS VENENATA W. C. SPAIN From the Department of Bacteriology and Immunology, Division of Immunology, in Cornell University Medical School, and the First Medical Division of the New York Hospital, New York City Received for publication January 6, 1922 Among the various forms of human hypersensitiveness, the susceptibility to ‘poison ivy” presents certain peculiarities which make it particularly interesting. One of these peculiari- ties is the constancy and characteristic features of the lesion. These features are the vesicular form, the erythema and the pruritus. Some preliminary observations by my colleagues, Drs. Cooke and Coca, seemed to indicate two other striking peculiarities of this condition. Cooke’s experiences (unpublished) had resulted in the impression that a relatively high percentage of adults possess a sensitiveness to the active principle of ‘‘ poison ivy” as shown by experimental contact carried out with a method which will be presently described. A short series of tests car- ried out by Coca (1) with the same method seemed to indicate that a considerable difference of susceptibility to “poison ivy”’ existed between adults and children under five years of age. Both of these impressions demanded further investigation before a final decision regarding them could be reached. Such an investiga- tion was undertaken and the results of it are here presented. There is considerable confusion in the literature as to the nomenclature of ‘‘poison ivy.’”’ In the present paper we have employed the term Toxicodendron radicans (L.), to designate “poison ivy,” “climbing ivy” or ‘‘three-leaved ivy,” thus fol- lowing the classification of Kuntze as adopted by the United States Herbarium at Washington, D. C., the New York Botani- cal Garden, and other institutions. This classification appears 179 180 WwW. C. SPAIN to be the most logical, since the word ‘‘Rhus,” the term used hitherto for the genus, is applied to the members of the Sumac family, that are not poisonous, being replaced by the genus term “Toxicodendron” (Greek, “poison tree”). Under this system of terminology, Rhus toxicodendron L. (Rhus quercifolia Steud.) “poison oak,’ eastern species, becomes Toxicodendron toxi- codendron (L.) Britton. Rhus vernix (L.), Rhus venenata D. C., “poison sumac,” ‘‘swamp sumac’ or “‘poison-wood,” a plant that is not a true sumac, becomes Toxicodendron vernix (L.) Kuntze. We are indebted to Mr. Percy Wilson of the New York Botanical Garden for his assistance in the problem of proper terminology. Many explanations have been offered as to the specific cause of Toxicodendron poisoning.! Burrill (2) believed that poisoning of Rhus toxicodendron |Toxicodendron radicans (L.), Kuntze,] was due to bacterial growths, but later (3) admitted that his hypothesis had not been proven. Khittel, as quoted by Warren (4), claimed to have found a volatile alkaloid as the cause, while Maisch (5) isolated a volatile acid which he called “toxicodendric acid,’’ and to which he ascribed properties simi- lar to formic and acetic acids. Pfaff (6) proved that toxicoden- dric acid was in reality acetic acid, and stated that the toxic agent was a non-volatile oily substance which he named “toxi- codendrol.’”’ Acree and Syme (7) considered the poisonous principle to be a glucoside, yielding rhamnose, fisetin and gallic acid upon boiling with dilute acids. Warren (4), working upon the poisonous principle of Rhus vernix [Toxicodendron vernix (L.), Kuntze], and McNair (8), studying the poisonous principle of Rhus diversiloba, [Toxicodendron diversilobum (T. and G.), Greene] were unable to obtain results comparable to those of Acree and Syme. Warren and MeNair isolated an acid resin. Stevens and Warren (9) consider the active principle of Rhus vernix [Toxicodendron vernix (L.), Kuntze] to be a resin which is 1 The term poisoning is used in this paper as a convenient designation of Dermatitis venenata due to “‘poison ivy,’’ although this condition is obviously not a true poisoning, since the active principle by injection has no effect upon lower animals. STUDIES IN SPECIFIC HYPERSENSITIVENESS. VI 181 a “clear, amber, oily, red, non-volatile liquid” and “a powerful escharotic.”’ With the use of a 95 per cent ethyl alcoholic extract of the fresh leaves of Toxicodendron radicans, Cooke (10) was able, in 1916, to reproduce the typical vesicular lesion of ‘‘poison ivy” in susceptible individuals. Cooke employed the following method of applying the extract for test purposes. The extract was made by mixing 95 per cent ethyl alcohol with fresh leaves of Toxicodendron radicans, which first may be run through a meat chopper. This mixture was allowed to stand for a few days, and the clear extract was obtained by paper filtration. A deposit of some resinous material upon the wall of the flask indicated that the extract represented a satu- rated solution of the active principle. In our experience, such an extract shows no tendency to a lessening of its activity within at least six or eight months. The extract employed in this investigation was at no time more than four months old. In the center of the gummed surface of a square of adhesive tape, 5 by 5 em., there was placed a bit of white blotting paper 0.5 by 0.5 em., which was then saturated with the alcoholic ex- tract of Toxicodendron radicans. The gummed surface of the square of adhesive was then applied to the skin, thus bringing the bit of blotting paper, while still saturated with the alcoholic extract, into intimate contact with the epidermis. We have termed this the ‘‘patch test.”” The flexor surface of the forearm was chosen because of the more delicate texture of the skin of this region, and because of its accessibility. There was no previous preparation of the surface to be tested. After a period of three days the patch was removed, and the area was washed, first with ether to take away any remaining portions of the adhesive, and then with alcohol to remove all traces of the active principle. The skin was inspected for the vesicular lesion typical of ivy poisoning. Every forearm that had been tested was examined twice a week, and differences in the findings were noted. A reaction was considered positive when a typical vesicu- lar lesion of poison ivy was reproduced beneath the patch. No case was considered positive unless there was vesiculation with 182 W. C. SPAIN erythema and local itching. Plus signs, with the usual interpre- tation, are employed in the tables to describe the degree of reaction. The solid residue of a chloroform extract was also employed. This material was applied directly to the skin over an area about 0.5 cm. in diameter. This area was covered with a square of adhesive tape to the center of the gummed surface of which was placed a square of glazed paper about 1 by 1 cm. to prevent absorption of the residue by the adhesive material. TABLE 1 Comparison of the intracutaneous and patch methods DAY OF OBSERVATION Third Seventh Tenth Fourteenth Seventeenth oe SEX | INITIALS a 3 3 2 z | =] = =} | 3 = 5 S < = <= S a = 4 668523). Penis: Bsa|.0 2s Oy arse OW arpae UG seae |) Ses Gr00O1)) FE 0 == 0 ae 0 ae 0 IF 0 a5 2205. Be Ge 80 ataate 0 ee 72294 | M| A. K.| 0 |++++4] 0 |++++l0 |++++] 0 |+++44] 0 /4+++ 69407 | F |C.S.} 0} +++ ]}0] +++ |o | +++ ]0] +4+4+/0}) +++ 74003'| F | TE. D.| 0 0 0 0 0 0 0 35 0 al qeGos |G. O.9) Ooh e-e- | Oo eee On dott At the beginning of this study a comparison was made between the intracutaneous and the patch methods of testing. The usual intracutaneous method was employed in performing these tests. A fine hyperdermic needle was introduced into the layers of the skin, and a small wheal was produced by the injection of about 0.05 cc. of the alcoholic extract of Toxicodendron radicans. A control of 95 per cent ethyl alcohol was injected in each case. It was found in a selected series of seven cases (table 1), that whereas all the cases were positive to the patch test, three being markedly so, in none was there a positive result with the intra- cutaneous method. Im a series of thirty cases, using the ender- STUDIES IN SPECIFIC HYPERSENSITIVENESS. VI 183 mal (intracutaneous) tests, Strickler (11) asserts that ‘‘a positive reaction is indicated by a papule, redness and tenderness at the site of injection,” at the end of forty-eight hours, and claims to be able to differentiate between ivy and sumach poisoning with this test. Strickler mentions a list of thirty cases which were successfully diagnosed by this method. His method of pre- paring the poison ivy extract to be used in the tests, consists of ‘“‘gathering the fresh leaves of poison ivy and extracting with absolute alcohol, filtermg and precipitating. The precipitate is dried and extracted in Soxhlet extractors for ten hours. The extract obtained is dried at low temperature. The toxin is carefully weighed, and dissolved in absclute alcohol, to which a certain amount of sterile water is added to make itnon-iritating.”’ This is apparently Syme’s (12) method. In our experience the injection of the alcohol alone produced the same effect as did the injection of the extract itself. In both instances, only a papule about 0.5 em. in diameter with erythema was produced, a lesion quite different from the vesicular lesion of Dermatitis venenata. Hence, we considered the intracutaneous test of no value. DIFFERENCES IN INCUBATION PERIODS One of the outstanding features of ivy poisoning is the ineuba- tion period. In even the most sensitive individuals this feature is present. In the published records of the clinical manifesta- tions of ivy poisoning, there are indications of a difference in the incubation period of this condition in different individuals, but this point has not been systematically investigated. MeNair (13) notes that ‘‘within twenty-four hours after exposure (to Rhus poisoning) patients frequently break out with a rash. This latent period or period of incubation, is dependent on the slow diffusibility of the poison into and within the skin, as well as in the predisposition of the patient.’’ White (14) has never seen ivy poisoning develop clinically after the fifth day of incuba- tion and in his experience the lesions usually appear within twenty-four to forty-eight hours. Schamberg (15) also thinks that twenty-four to forty-eight hours is the usual clinical incuba- tion period. 184 W. C. SPAIN As the individuals that were used in the present investigation were kept under continued observation, it was possible to study systematically the differences in the period of incubation. The incubation period was found in some instances to be unexpectedly prolonged. In one individual, in fact, no effect was observed until the twenty-fourth day when a one plus reaction developed at the site of the patch test. This reaction consisted of an area of erythema in which were several elevations not distinctly vesicular. There was definite itching. In the absence of char- acteristic vesicles there may be some doubt as to the specific nature of this lesion. The tested individuals have been divided into two groups. Group 1 is composed of those individuals that were observed for a minimum of ten days. The results of the tests of these cases are presented in table 2. Group 2 is composed of those indivi- duals that were observed for only seven days. The results of the tests upon these individuals are shown in table 3. No case was included in either group that was not seen on each consecu- tive observation day. No cases were recorded that were not observed for at least two consecutive observation days. In group 1 there are 80 cases, 28 (or 35 per cent) of which were found to be insusceptible to the tests during the entire observa- tion period of 10 to 17 days. These 28 cases were not included in the table (table 2), as they had no incubation period and add no information. If all of the 80 cases had been judged by the results recorded on the third day, 48 cases (or 60 per cent) would have been classed as negative, an error of 25 per cent. If all of the cases had been judged by the third and the seventh day readings only, 40 cases (or 50 per cent) would have been classed as negative, an error of 15 per cent. If these cases had been judged by the third, fifth and seventh day readings, 33 cases (or 41 per cent) would have been judged negative, an error of 6 per cent. From these figures it is seen that the degree of error in the determination of susceptibility by this method drops from 25 per cent on the third day of observation, to 6 per cent on the tenth day of observation. This emphasizes the importance of not judging the reaction from a patch test until after a ten day observation period. ——— STUDIES IN SPECIFIC HYPERSENSITIVENESS. VI TABLE 2 Group 1. Cases arranged according to length of incubation pertod DAY OF OBSERVATION oO | Twenty-first Sean AGE | SEX INITIALS = 3 = 3 s z a re 5 | a ; 5 g sa R sal al vp) 72543 | 18 | M| J.C. 0 0 0 0 0 76625 | 48 | F | R.K. 0 0 0 oT 70432 | 21 | F | BE. K. 0 0 0 + + faded lee ee. W. 0 0 0 te + 70745 | 42 | M| P. V. 0 0 0 + 70137 | 44 | F | A.B. 0 0 + + — Go139 | 14 | M | A. P. 0 0 ote + 4 74534 | 47 |M| J. K. 0 0 + + 0 66997 | 10 | F | C.B. 0 0 + + — 70728 | 26 | F | E.K. 0 0 + + a 73666 | 35 | M| A. D. 0 0 + + 73063 | 40 |M/| S.N. 0 0 + + 70531 | 20 |M| M. H. 0 0 + + qoeds..\ 29 | Mo) M. P. 0 =- + — _ mettO) | 2o- | El. MoG, 0 ae + + + 71090 | 49 | F | M.L. 0 oF a. aa + 69954 | 40 |} F | C.G. 0 ata + 4 44-4 65417 | 27 | M| A. J. 0 == > - + 70255 | 31 | M| §.S. 0 te 7 a 0 66107 | 62 | M| L.M. 0 a 0 0 Soe temieto ei WE We Ge ee | ete ey Reread PMU CAC MG: asaiesetel+ habe igi PE ois Gee e eM EBS Boy lee |. eb + =e ob SL ZEA MIU NSIS Os a Vl ei WN ++ 4 SETS ade ak td SE ea al Ol g(r | 69407 | 10 | F | C.S8. Tp eee ae inate ct peste TAG es UTES INS) PS Si es te pe cl Wee ee te a a 70256 | 31 | M| P. K. ++ ++ ++ Jose toe! 2 69154 | 51 | F | L. B. ++ -- ae = fe 75040 | 45 | M| C.L. ++ ++ ]/4+4+] 4+ aEeE 66852 | 44 | F | S.B. ++ ae t+ eese pe G00) | 26 | F | HL. ++ +e se on 70827 | 44 | F | HN. arate ak ++ + a 62424 | 40 |M/} M.M. | ++ t+ + = oe 66616 | 40 | F | LF. a +4 + = ab 73507 | 40 | F | J.S. ++ + oe in 0 62626 | 12 | M/A. O. ++ ++ ++ + 73906 | 30 | M| G.B. ++ -- a8 ai Bi + | Twenty-fourth of. 186 W. C. SPAIN TABLE 2—Continued ‘ | DAY OF OBSERVATION = Rica a AGE | SEX INITIALS = = e iE 3 : g g | 3/3 3 2 z 5 2 | 8/8 63620 | 10 | F | C.R. + + = 2 + 15660 | 12 | F | L.S. ae 2a 0 0 0 T4iz3 (245. | oP Se Gs. =F + == =F + 66103 | 22 | M| W.S. 4 + Je eee 73514.| 3\F |M.R. | + 2 fe + - 6940 |34|M] R.A. + + + + + 67508 | 28 |F | A.M. | + ae 0 0 0 69572 |26 |F | E.D. | + 4 + + 4 73504 | 36 | F eS: =5 0 0 0 0 71623 | 28 | M| M.S. + aa - + st 196 | 43 | F | F.P. me Ze xe 71615 | 33 | M| J. K. aa <- + Sse TB Seen 65656 | 60 | M| P.C. “= a 0 0 0 71827 | 44 | F H. M. + + + + In group 2 (table 3) there are 24 cases, 13 or 54 per cent of which are negative when classed by the reaction that had devel- oped by the third day only. When judged by both the third and the seventh day observations, 10 or 46 per cent were nega- tive. It will be seen that the results from the third day obser- vations and from the third and seventh day observations in groups 1 and 2 are very similar. DIFFERENCES IN SUSCEPTIBILITY It is evident in the results of the tests in all of the tables, that the different individuals are not equally susceptible. In fact, every degree of susceptibility was observed from the least to the greatest. All of the markedly reacting cases had developed the typical lesion by the third observation day. Where the incubation period was longer than three days, the degree of the local test reaction was slight, being one plus in every instance. In three of the most susceptible cases, in addition to the typical lesion produced beneath the patch, lesions upon the arms, hands STUDIES IN SPECIFIC HYPERSENSITIVENESS. VI 187 and ankles, with general pruritus, had developed within twenty- four hours, that is, before the patch was removed. Close ques- tioning had revealed that these individuals had not been exposed to poison ivy other than by the patch tests. In these individuals, the adhesive plaster was still in place upon the third day, and the area was washed with ether immediately upon its TABLE 3 Group 2. List of cases observed for 7 days only DAYS OF OBSERVATION CASE NUMBER AGE SEX INITIALS Third Seventh 70528 36 M M. P. eae 5 ea 73766 30 F ACR: Se4- staatn 75356 34 M Bok: qPSes= ar deat 75401 28 M 10%, Ves aps acme 72295 30 F EeaGr aRse teats 72879 16 M A.A SPP Sate 72554 16 F E. D. aPa5 sh 73279 24 F As Ar Siaaie =e 65949 45 M M. B. ae =i 70816 11 M VG: ar = 62629 47 M Wests aR =P 69568 53 M L. S. 0 0 71086 48 M E. C. 0 0 70830 45 M I. W. 0 0 66106 29 F Sule 0 0 73579 43 M Wireke 0 0 68120 48 M CBE 0 0 73586 50 M As Lee 0 0 73416 42 F A. W. 0 0 74938 7 M 1310p 0 0 75357 57 M In, IDE 0 a 75399 29 F D. L. 0 == 753994, 36 F A es Oe 0 a 68809 67 M Je: 0 0 removal. The alcoholic extract had been used in the tests and in impregnating the bits of blotter with the extract, an excess was avoided, because such an excess would have prevented the proper adhesion of the patch. These cases, therefore, suggest the possibility that the typical dermatitis of poison ivy develops not only at the point of original contact with the active principle 188 W. C. SPAIN but also at times upon skin surfaces distant from the primary lesion, by transference of the active substance through the blood and lymph. While this explanation appears plausible, the possibility must be considered that some of the active substance diffused through the adhesive material, in which it is readily soluble, and thus arrived, either at the edges of the patch, or on the external surface of the canvas, from which sites it could be transferred to other parts of the body by external routes. A correlation of the different degrees of experimental suscepti- bility noted above, with the degree of clinical susceptibility in the TABLE 4 Tests upon group of boys DAY OF OBSERVATION CASE CLINI- NUM- AGE CLINICAL HISTORY CAL BER ONSET Third Seventh Four- teenth 109 | 14 | General, face and body 2 j+-4+4+4+)/4+4+44] ++ 112} 15 | General, face and body 2 + + 0 117 | 20 | Severe on hand and feet 2 a + 0 114 | 16 | Severe on hands 2 }+++;)4+4++) + 115 17 | Severe on feet 2 |+4++4+4\/4++4+44+) ++ 116 | 18 | Slight Z SESE = ae ) 110 | 15 | Slight 2 aE se shai 0 111 | 15 | Slight 3 oo ee 0 108 | 14 | Slight 2 aPar + 0 106} 13 | Slight, on extremities 3 + + 0 105 | 13 | Negative 0 ae 2 0 107 | 14 | Negative 0 0 0 0 113 | 15 | Negative 0 0 0 0 same individuals, could not be undertaken in these series since most of the individuals were of foreign birth, and were not ac- quainted with poison ivy or its natural effects. However the opportunity for making such a correlation was offered in a group of boys. The protocol of these comparative observations is presented in table 4. The group consisted of thirteen boys be- tween thirteen and twenty years of age, who, in midsummer, pitched camp at a site covered with poison ivy vine. Ten of the thirteen gave the clinical history of ivy poisoning as the result of this experience. The tests, in two of the three clini- STUDIES IN SPECIFIC HYPERSENSITIVENESS. VI 189 cally negative cases, were negative. The third clinically nega- tive case gave a faintly positive reaction (one plus) by patch test on the third day, but this reaction was as pronounced as in four other cases clinically positive, in two of whom the attack was mild, in two, severe. All of the cases giving severe test reactions had suffered severe poisoning while in camp. Two | other cases that had been severely affected while in camp gave only faintly positive reactions. No individual that had exhibited any degree of clinical susceptibility was negative on test. The clinical incubation period of all the cases was practically the same, the dermatitis appearing upon the evening of the second day in all but two of the cases. In these two cases the lesions occurred upon the morning of the third day. In general it can be said that, with the kind of extract employed, there is a degree of parallelism between the results of the tests and the clinical history of the cases. AGE INCIDENCE Of the eighty individuals included in group 1, 52, or 65 per cent, responded with a positive reaction to the patch test. All but one of these individuals were eight years of age, or over. The number of tested individuals in the different age periods is too small to permit a definite conclusion as to a possible differ- ence in susceptibility in poison ivy among individuals over eight years of age. For purposes of comparison however, we may take 65 per cent as representing approximately the average suscepti- bility of the individuals over eight years of age with the technic, and with the use of the alcohol or chloroform extracts described above. Through the courtesy of Dr. F. H. Bartlett oppor- tunity was afforded to apply the patch tests in 18 infants be- tween five weeks and eighteen months of age, all patients in Dr. Bartlett’s service at the Babies’ Hospital, New York City. The results of these tests are shown in table 5. Two tests were carried out in each case. On one arm the alcoholic extract was applied by means of the patch test. On the other arm the residue of the chloroform extract was employed in the manner previously described. It is seen that no positive reaction occurred by either test. 190 W. C. SPAIN TABLE 5 Tests upon tnfants DAYS OF OBSERVATION CASE NUMBER AGE SEX Third Seventh Tenth months 118 i F 0 0 Home 119 2 F 0 0 0 120 43 M 0 0 0 121 5 M 0 0 Home 122 5 M 0 0 0 123 6 M 0 0 0 124 6 F 0 0 0 125 63 M 0 0 Home 126 7 F 0 0 0 127 7 F 0 0 0 128 8 F 0 0 0 129 8 M 0 0 Home 130 103 M 0 0 0 131 11 M 0 0 0 132 11 F 0 0 0 133 12 F 0 0 0 134 12 F 0 0 Home 135 16 F 0 0 0 136 18 M 0 0 0 SUMMARY 1. The typical vesicular lesion of Dermatitis venenata can be produced by means of an alcoholic or chloroform extract of the fresh leaves of Toxicodendron radicans, applied to the skin surface. 2. In our hands, the typical vesicular lesion of Dermatitis venenata could not be produced by the intradermal injection of an active alcoholic extract. The lesion thus produced was not different from that caused by the intradermal injection of the solvent. 3. With the patch test, differences can be demonstrated in the susceptibility of different individuals to poison ivy, and in the incubation period of the lesion. 4, With the technic used, infants could not be shown to be susceptible. STUDIES IN SPECIFIC HYPERSENSITIVENESS. VI 191 REFERENCES (1) Coca, A. F.: Jour. Immunol., 1922, 7, 193. (2) Burritx, J. T.: Am. Assoc. Adv. Sci. Proc., 1882, 31, 515. (3) Burritt, J. T.: Garden and Forest, 1895, 8, 368. (4) Warren, L. E.: Pharm. Jour., 1909, 83, 531. (5) Maiscu, J. H.: Am. Pharm. Assoc. Proc., 1865, 13, 166. (6) Prarr, F.: Jour. Exp. Med., 1897, 2, 181. (7) Acreg, 8S. F., anp Symz, W. A.: Am. Chem. Jour., 1906, 36, 301. (8) McNair, J. B.: Jour. Am. Chem. Soc., 1916, 38, 1417. (9) StevENS AND WARREN: Am. Jour. Pharm., 1907, 79, 518. (10) Cooxe, R. A.: From unpublished clinical studies. (11) Srrickier, A.: Jour. Amer. Med. Assoc., 1921, 77, 910. (12) Syme, W. A.: Some constituents of poison ivy. Johns Hopkins Univ. Dissertation, 1906. (13) McNatr, J. B.: Jour. Infect. Dis., 1916, 19, 419. (14) Wuits, J. C.: Jour. Cutan. Dis., 1906, 30, 280. (15) ScoamBera, J. F.: Jour. Cutan. Dis., 1906, 30, 280. ‘Sg eg Fie | | puinbippaeekseniiia a alae Hiss eee tak phe Daa SARL cht dV Aa ARR: sinner rainopsn : Be, a ; St 2 COAS re ae ie eee eee 7 eal AAD IRR ats TS 8 ROE ki ay 2 ae ena Dd ER eT Fine ek eA cr ak y ; A P . : ; > ; fab tt | helt es oh - an \ Me Hy OS whe ‘Hesek ‘nkelt tds, 2 BO aS ee Re se Sith BOLT ak opgeld aed. deg ale pose : ba ae TOE Ce ed sol ark tee ne as Te fa tA Roc lied inks the omit retin et AK, ae | aie ie TD TOS \ eas Boe ak pile, ar. ES a, Wie entatell cpakil av b sondeyy bes bhamidfidenurs (wi ot Re ee aly ea Wes Po GH. a tee} a ; ‘eo vie Cl OF eee Avatal ent) ae ees ; + CALA DG, OBE UD rete coy ND Soe eS i Gee OE POE aR eR ap 5 Onna a be, eat , , oa ae . i Ai i ‘5 & ; ae jae Se : ay m a ‘ i . ‘ re 4 ‘ r x sit " cores | STUDIES IN SPECIFIC HYPERSENSITIVENESS VII. THE AGE INCIDENCE OF SERUM DISEASE AND OF DERMA- TITIS VENENATA AS COMPARED WITH THAT OF THE NATURAL ALLERGIES ARTHUR F. COCA From the Department of Bacteriology and Immunology, Division of Immunology, in Cornell University Medical College and the New York Hospital Received for publication December 15, 1921 Any classification of the various forms of hypersensitiveness must be considered tentative so long as the mechanism of these phenomenona remains unknown. In the absence of this know- ledge, however, it is useful to group the phenomena according to the facts at hand. Thus, anaphylactic hypersensitiveness can be properly separated from all the other forms because there is no satisfying evidence that these latter are dependent upon an antibody-antigen mechanism, as is the former. In a previous publication the author associated the drug, food and animal idiosyncrasies, hayfever, asthma, serum disease and dermatitis venenata under the term allergy. Two of these conditions are distinguished from the others by noteworthy peculiarities. Serum disease presents the almost constant characteristic of the incubation period and dermatitis venenata differs from the others in which the skin is involved, in the con- stant and characteristic nature of the lesion. It has been re- cently found, also, that serum disease differs strikingly from most of the other forms of human hypersensitiveness in its percentage incidence. ‘The observations of Longeope and Mackenzie and of Rufus Cole supported by statistics collected in the Boston City Hospital (1). indicate that about 90 per cent of the white race are susceptible to ordinary serum disease upon intravenous administration of large quantities of horse serum. ges: ss) 4 , ¥ time) f ef iu ai teds a +185 airy Wl OY ‘, ¥ . rs a) can 03", Sothiadint avo, ee i ied eel f tt, Ae : aa By But hy cite fig Hohe nade LAP a ‘ ate i i 4 nes , Beh cot hs ties CT ak ees Fas 9) ats “ie f f : | er aft - ‘ be At ' Pumas), 12 ‘5 eNP tat Sena i ut hws ite Sea mn Cn) PMA) «4 yj é is , i. en eA ri, Pn sy cree F Bela Sh) aad Herat 2} a wid wae { nes, a , i i in ie = * Pert: 8 vRoe Vay ie Soe af ‘ een é PT tbl fe rh ye eR RIT Ve) Aa) ee > dey ; Re STUDIES IN SPECIFIC HYPERSENSITIVENESS VIII. ON THE RELATIVE SUSCEPTIBILITY OF THE AMERICAN INDIAN RACE AND THE WHITE RACE TO THE ALLERGIES? AND TO SERUM DISEASE ARTHUR F. COCA, OLIN DEIBERT anp EDWARD F. MENGER From the Department of Bacteriology and Immunology, Division of Immunology, in Cornell University Medical College, New York City and the New York Hospital, New York City, and the Department of Bacteriology in the University of Kansas, Lawrence, Kansas Received for publication December 15, 1921 The fact that serum disease generally appears after an incuba- tion period following the injection of an antigenic protein and is usually accompanied by the formation of specific antibody (pre- cipitin) would seem to separate this form of human hypersensi- tiveness from those which are exhibited immediately upon natural contact with the exciting agent in the entire absence of demon- strable antibodies. The former condition seems to be experi- mentally induced and dependent merely on a suitable previous contact with the exciting agent—serum protein: the latter, according to the studies of Cooke and Vander Veer, appear to be wholly subject to an hereditary factor and not dependent upon previous contact with the exciting agent. Our knowledge of these conditions is drawn from observations upon the white race. Certain publications, to which we shall refer again, seemed to indicate that some of the forms of hypersensi- tiveness in human beings (asthma, dermatitis venanata) are rare among the American Indians andit was thought that investigation of this apparent racial difference might produce some informa- tion regarding the nature of the different forms of specific human 1 By allergy is meant the inherited forms of specific human hypersensitiveness, such as hay-fever, asthma, angio-neurotic edema and urticaria. 201 202 A. F. COCA, O. DEIBERT AND E. F. MENGER hypersensitiveness as well as some indication of the relation of the American Indians to the other human races. Particularly, it seemed desirable to find out first, whether the American Indian is relatively infrequently affected by the in- herited forms of hypersensitiveness such as asthma and hay- fever and secondly, whether or not in such case the induced con- dition of serum disease is also infrequent in that race. If the incidence of serum disease were found to be equal in the Indian race and in the white race, this form of human hypersensi- tiveness would be, thereby, the more sharply differentiated from the inherited form. If, on the other hand, both forms of hyper- sensitiveness were found to be equally infrequent, then an identity in at least a part of the underlying mechanism of the two forms would be indicated. Through correspondence with physicians that had had exten- sive medical experience among American Indians we learned that hay-fever, asthma, urticaria and drug allergy are very uncommon in the full blooded individuals of that race. We are able, with the codperation of Dr. N. P. Sherwood, Professor of Bacteriology in the University of Kansas, and Mr. H. B. Pearis, Superintendent of Haskell Institute in Lawrence, Kansas, to observe the occur- rence of serum disease in twenty-six students in Haskell Institute, all listed as full blooded American Indians. THE OCCURRENCE OF THE ALLERGIES AMONG AMERICAN INDIANS A search of the literature has revealed but few references to the incidence of allergic conditions in the Indians.?__ In his report on Physiological and Medical Observations among the Indians of Southwestern United States and Northern Mexico (1), Ales Hrdlicka remarks on page 188, ‘‘No instances of severe asthma were encountered,” and on page 191, asthma is ‘‘rare among Southwestern Indians.” In his article on “‘Skin Diseases among Full Blooded Indians of Oklahoma” (2), E. S. Lain writes: 2 In this article the word Indian always refers to the American Indian. STUDIES IN SPECIFIC HYPERSENSITIVENESS. VIII 203 The extreme rarity of dermatitis medicamentosa, dermatitis venenata, may partially be accounted for by the lack of common usage of drugs, soaps, and other chemicals, though the Indians frequently apply quite freely on their faces and other exposed parts paints and dyes made from the wild plants, berries, ete. The skin of the Indian is apparently almost immune to the poison ivy and other plants which cause so much discomfort to the women of our race. Our own inquiry was made in the form of a questionnaire, answers to which were received from eleven physicians and one superintendent of an Indian school. The replies were based on extended medical observation of about 40,000 full blooded Indians. Those replying to the questionnaire were: Dr. A. J. Anderson, Lawrence, Kansas. Mr. Clyde M. Blair, Superintendent Chilocco Indian School, Chiloecco, Oklahoma. Dr. A. D. Lake, Gowanda, New York. Dr. Otis E. Lovelady, Ponea Agency, Whiteagle, Oklahoma. Dr. A. E. Marden, W. 8S. Indian Vocational School, Phoenix, Arizona. Dr. James M. Meason, Pima School, Sacaton, Arizona. Dr. Charles M. Ming, Okmulgee, Oklahoma. Dr. W. W. Rublee, Sherman Institute, Riverside, California. Dr. F. A. Spafford, Flandreau, South Dakota. Dr. A. J. Wheeler, W. S. Indian Vocational School, Phoenix, Arizona. Dr. Lawrence White, Keshena, Wisconsin. Dr. A. M. Wigglesworth, Albuquerque, New Mexico. The essential questions proposed were: 1. Have you observed the occurrence of asthma or hay-fever in full-blood Indians? 2. Do these conditions occur in the white population of the vicinities in which the observations upon the Indians were made? 3. Have you observed the occurrence of urticaria in full-blood Indians? In all of the replies question 2 was answered in the affirmative. 204 A. F. COCA, O. DEIBERT AND E. F. MENGER The following men could recall no instance of asthma, hay-fever or urticaria in a full-blood Indian: Mr. Blair, Drs. Lovelady, Meason, Anderson, E. F. Menger, Ming, Roblee, Spafford, Wheeler, and White. Dr. Marden has seen ‘“‘two or three cases of asthma,” and “numerous cases of urticaria” in full-blood Indians (about 10,000 individuals under observation). Dr. Menger has seen “hay-fever, asthma, urticaria and derma- titis venenata only in the breeds” (about 1500 individuals under observation). Dr. Wigglesworth has seen hay-fever and asthma in full-blood Indians and “urticaria many times.” (About 15,000 individuals under observation.) Dr. Lake has seen ‘‘a few cases of hay-fever and of asthma.” He states that “‘hives are common and poisoning from ivy is frequent.’”’ However, he writes that ‘‘very few’ of the Indians under his observation are full-blooded. ; These replies suggest that while the allergic trait is not absent in the Indian race this character is very much less marked in the Indian than it is in the white race. The observation of allergic symptoms in some pure Indians makes it inadmissible to assume the complete absence of the basic condition in all of the Indians that do not naturally present the outward signs of it. It may be that, exactly as in the white races, some individuals are hypersensitive in a degree too slight to allow the development of symptoms upon the natural contact with the exciting agent. In such individuals the injection of the agent would reveal the latent condition. From the foregoing facts it could have been anticipated that if serum disease is associated in some way with the allergies it should occur in full-blood Indians upon injection of serum, but much less frequently and possibly in milder form than it does in the white race. In fact, this was the outcome of the experiences which we are about to describe. 3 In the same inquiry the statement of Lain regarding the relative insuscepti- bility of the Indians to ‘‘poison ivy”’ was confirmed. STUDIES IN SPECIFIC HYPERSENSITIVENESS. VIII 205 We had learned from Dr. W. T. Longcope that in his experience symptoms of serum disease occur in at least 90 per cent of white individuals receiving 100 ce. or more of whole horse serum by intravenous injection. With this high percentage as a standard of comparison it was thought that at least some indication of the relative suscepti- bility of the two races to serum disease could be obtained from the results of the intravenous injection of about 100 cc. of normal horse serum into even a few of the Indians. The observations were made upon twenty-six full-blooded Indians, all students in Haskell Institute situated in Lawrence, Kansas. We consider it a privilege to record the fearlessness and patience with which these intelligent young men, most of whom were fully aware of the investigatory nature of the procedure, submitted to the injections. It is not possible to distinguish any individuals among them in the respect of courage or fortitude. Those who presented themselves in the first series faced uncer- tainty as to the possible pain and danger attending the injections. Those that came later came with the knowledge of the unpleasant symptoms which some of their predecessors had suffered. It should be recorded that the two who were the last to be injected had witnessed the effect of the injection upon Charlie Hutchison. These two were given ample opportunity to withdraw, but neither would take advantage of that suggestion. THE OCCURRENCE OF SERUM DISEASE IN THE AMERICAN INDIAN The material injected was fresh normal horse serum preserved with the addition of 0.6 per cent of a mixture of equal parts of “Three Cresols’” and ether. According to Krumwiede and Banzhaf (3), one of the unpleasant immediate effects (chills) of the injection of serum preserved with cresol is avoided by adding the preservative mixed with an equal volume of ether. It may be stated here that this symptom was absent in all but one of the individuals in the present series (Jesse King). The serum was prepared especially for this investigation in the laboratories of the H. K. Mulford Company and delivered 206 A. F. COCA, O. DEIBERT AND E. F. MENGER to us in double stoppered 110 cc. tubes together with a number of the H. K. Mulford Company intravenous outfits. With these convenient outfits we were able to make the injections at the rate of six in an hour.*® Previous to the injections cutaneous tests were made with dried horse serum prepared by the H. K. Mulford Company. All of these resulted negatively. The serum was brought to body temperature by placing the tubes in warm water. The effects of the injections in the different individuals were as follows: 1. William Frank (Creek). Received 110 cc. Slight headache on the day of injection and temperature of 99.8F. Presented no other symptoms at any time.® 2. Joe Bearhead (Cheyenne). Received 110 cc. Complained of headache on day of injection. Had some epistaxis on the third day. On the 10th day there were a few urticarial spots on the arms and body without elevation of temperature; these were present also on the 11th day still without fever, but they had disappeared by the 12th day. 3. Benjamin Osage (Cheyenne). Received 110 cc. Immediately after the injection a swelling 1 cm. in diameter appeared on the left forearm just above the wrist. The swelling was situated more deeply than the usual urticarial wheal. A smaller spot appeared on the fore- head and another 2 mm. in diameter above the one on the forearm. There was severe headache on the day of the injection with a tempera- ture of 100°F. There were no further symptoms. 4. Harber Johnson (Creek). Received 110 cc. There was im- mediate injection of the conjunctiva with some swelling of the eyelids and face. There was headache on the day of the injections, but no symptoms thereafter. 5. William Atkins (Otoe). Received 25 cc. There were no symp- toms at any time. 4We are under great obligation to Drs. John Reichel, F. M. Huntoon and J. A. Murphy of the H. K. Mulford Company for the care with which the serum was prepared and delivered to us. 5 Acknowledgement must be made of the skillful assistance of the nurses— all Indian girls—under the direction of Miss Anderson, superintendent of nurses in the Institute. 6 There was some coughing during the injection or immediately afterward in most instances and in some a slight sense of oppression in the chest. STUDIES IN SPECIFIC HYPERSENSITIVENESS. VIII 207 6. Thomas Wasson (Baiut). Received 110 cc. Within ten minutes after the injection two urticarial spots appeared on the eyelid and the neck. There were no further symptoms. 7. Delmar Scott (Mojave). Received 110 cc. There were no symp- toms until the 8th day when there was fainting and a temperature of 100°F. On the 9th day generalized urticaria appeared which lasted three days. There was no elevation of temperature on the 9th day nor thereafter. 8. Robert Leve Leve (Walapi). Received 110 cc. There were no symptoms until the 12th day when there were severe joint and muscle pains without temperature elevation or eruption. 9. William Hampton (Choctaw). Received 110 ce. Complained of joint pains on the 5th day and again on the 10th day when some urti- carial spots appeared. The joint pains continued on the 11th and 12th days but there was no urticaria after the 10th day. There was no ele- vation of temperature at any time. 10. Lawton Raymond (Navajo). Received 110 cc. There were no symptoms at any time. 11. Joseph Parnell (Assiniboin). Received 110 cc. There were no symptoms at any time. 12. Wallace Littlefinger (Sioux). Received 110 cc. On the 11th and 12th days there was urticaria without fever and on the 13th day there was a temperature of 99°F. with joint pains but without urticaria. 13. Edward Davenport (Sac and Fox). Received 110 cc. There were no symptoms at any time. 14. Emery Redbird (Ottawa) Received 110 cc. There were no symptoms at any time. 15. Morris Baken (Choctaw). Received 20 cc. On the 11th day there was swelling of the left parotid gland without fever. On the 12th day there were joint pains with a temperature of 100°F. No symptoms thereafter. 16. William Ruskin (Navajo). On account of the very small size of the vein a double puncture occurred causing a hematoma, which pre- vented the injection. 17. Marion Runsthru (Assiniboin). Received 110 cc. There were no symptoms at any time. 18. Andy Snap (Creek). Received 110 cc. On the 6th day there was urticaria without fever; on the 7th day the urticaria had disappeared but there were cramps and weakness with a temperature of 99°F. Urti- caria reappeared on the 8th day accompanied by edema of the face and 208 A. F. COCA, O. DEIBERT AND E. F. MENGER a temperature of 100°F. There were no symptoms on the 9th day nor thereafter. 19. Jesse King (Creek). Received 95 cc. There was immediate marked uneasiness with pain in the chest and about the eyes. Pulse at eight minutes 140; at thirteen minutes 118, and at eighteen minutes 104. At thirty minutes he was able to walk out. At one hour there was a chill. There were no further symptoms. 20. Mason Kawaykla (Apache). Received 110 cc. There was urticaria on the 10th and 11th days without fever or other symptom. 21. Walter Emarthla (Creek) Received 110 cc. There were no symptoms at any time. 22. James Foster (Creek). Received 110 cc. There were a few urticarial spots on the 10th day without fever or other symptoms. 23. Abel Archibald (Creek). Received 110 cc. There were no symptoms at any time. 24. John Alonzo (Pueblo). Received 90 ce. Vomited immediately after the injection, but presented no further symptoms at any time. 25. Charles Hutchinson (Arapahoe). Received 80 cc. Immediately complained of pain in the chest. There was edema of the eyelids and congestion of the conjunctivae; a general weakness and a rapid and weak pulse (116) five minutes after the injection. Had to be carried to bed There was severe vomiting after two hours. The edema of the eyelids and the weakened heart action persisted into the following day. There- after there were no further symptoms. There was a previous history, in this case, of a weak heart and a constitutional weakness. 26. Andrew Juan (Pima). Received 110 ce. There was headache and a temperature of 100°F. on the 2nd day, but no further symptoms at any time thereafter. 27. Edward Meeks (Arapahoe). Received 110 ec. There was a slight malaise immediately after the injection and on the 9th day urti- caria without elevation of temperature. In considering what symptoms were to be regarded as indica- tive of ‘“‘serum disease”? we have been guided, naturally, by the attitude of those upon whom we depended for our ‘‘control” series among the white race, that is Dr. W. T. Longcope and Dr. Rufus Cole. These observers have used only eruption and joint pains with or without fever as criteria of that condition, and as is customary they have ignored the immediate effects of the injections, excepting eruption. We have, therefore, omitted STUDIES IN SPECIFIC HYPERSENSITIVENESS. VIII 209 from consideration the effects noted in the cases of William Frank, Harber Johnson, Jesse King, John Alonzo, Charles Hutchi- son and Andrew Juan. In table 1 is given a summary of the symptoms that could be regarded as those of ‘‘serum disease”’ resulting from the injec- tions into the Indians. It is seen that the incidence of the condition among those injected was 46 per cent; the average duration of the symptoms was two days and the average elevation of temperature was 0.38°F. In comparing the course of serum disease as just described in the Indians with that in the white race it must be pointed out that the injections made in the Indians were given to healthy young men who were under orders to report to the physician in charge (Dr. E. F. Menger) the slightest ailment. For this reason the existence of serum disease was recognized in several of the Indians by symptoms (such as joint pains and edema and by the immediate appearance of a few urticarial spots) which are rarely noted in the usual medical history. It was necessary, therefore, in order to obtain comparable statistics among white individuals to seek them in institutions in which the occurrence of serum disease is given particular attention. These requirements were amply met in the records of a series of cases in the Presbyterian Hospital of New York City which were under the care of Dr. W. T. Longeope and Dr. George M. Mackenzie. We are indebted to these observers for permission to publish the data and to Dr. E. P. Maynard by whom the data were collected. It is evident that the relative susceptibility of the two races should not be judged solely by the relative percentage incidence. A fair comparison should include also the factors of duration of the disease and its severity. As an index of severity we have used the single feature of elevation of temperature. In a series of fifty-two individuals treated with serum (horse) injections in the Presbyterian Hospital and observed continuously for one month, serum disease occurred in forty-eight instances—an inci- dence of 92.4 per cent. THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 2 “1.66 SOATY “A001 oN 908] “ssou BuepiT -YBOM 0 ‘soarzy | ‘sdureig ea? es (OOKC)) CBUg “10986 “09°86 *\To9°86 "SOATY "SOATY "SOATY ON ON MOT *‘H.9°86 ‘sured ‘sured ‘sured ‘sured 0 qarol’ quroL JarLoL purof "***(MBqo04H) woydure yy “09°86 “(aey) 0 SOAT]T eo eevceces (9019) 19480,] a “Ho9°86 “09°86 “N 0 "SOATTT *SOATTT “ssss(oqoedy) vlyAemeyy “1066 ‘sured qurof “109°86 *109°86 "SOATY ON ‘SOATTT "SOATT] “see**(xnoIg) osUyoy} wT “HOOT | “09°86 ‘sured | ‘pryored 0 quior | Jo surjjomg sseeees(aeqooyO) wexeg Yel a WI 4101 416 428 qu. qe 381 NOILGLNI 40 Ava A ne Le Be ee ee ee eee ae suDIpUT papoo]g yynf UL asvasip wntas fo as4n09 JDIVUL}9 ay} Duxnoyy l GTavoL “H09°86 “H.9°86 “SOATY “SOATY 0 MOT MOT “1.986 *SOATY oN ‘sured 0 qurof *\1.9°86 *\109°86 0 "SOATH] “SOATTT 0 “09°86 "SOATY [e19uex “09°86 "SOATET *H.00T “ssou “489 AA *SUIO4 -duAs 1ayyinj oN ‘syods [8trv914.1n Moj SATO} CIPIUIUTT "SUl04 -duAs 10yjInj ON ‘sjods [BlivoTzIn Moyj ‘ATO}eIPOMIUTT eeeeree **(qnIeg) UWOSS® A\ “sees**(guuefeyD) o3usO “**(guuefeyD) pvoyiveg "****(1dB[e) 9AVT OAaT seeeeeess* (QABlOTA) 44009 "***-*(goqedery) sHIOP LLL Ls sss SSS 211 pal We A. F. COCA, O. DEIBERT AND E. F. MENGER In table 2 the duration of the condition in forty-two of the cases is given and for direct comparison the duration of the con- dition in the twelve Indians. The average duration in the white individuals was four and one-half times as great as it was in the Indians. As an index of the relative severity of the serum reaction, com- parison has been made of the average elevation of temperature above the normal level. TABLE 2 Duration of serum disease after intravenous injection of serum IN 42 WHITES (PRESBYTERIAN HOSPITAL) NEW YORK CITY ee a Paha Days Cases | Days Cases Days Cases Days Cases 1 1 9 1 15 2 1 5 3 3 10 1 16 1 2 3 4 11 11 1 20 2 3 1 5 6 12 1 De 1 4 3 6 2 13 i 24 2 Uf. 4 14 1 25 1 Average duration 9 days Average duration 2 days In determining the average elevation of temperature in the white individuals only those cases were considered in which the temperature had been normal before the symptoms of serum disease had appeared. As all of these individuals had been suf- fering from a bacterial infection they were, according to Dr. Mackenzie, who selected the cases, possibly more prone to tem- perature elevation than a normal individual. Dr. Mackenzie suggests, therefore, that this instability be accounted for in estimating the temperature elevation in these individuals by advancing the normal level somewhat. As the temperatures in these cases, were taken per rectum, Dr. Mackenzie suggests that 100°F. be taken as the normal temperature. This suggestion has been adopted. Following are the highest rectal temperatures observed in thirty-seven of the Presbyterian Hospital cases during the period in which the individuals presented symptoms of serum disease: STUDIES IN SPECIFIC HYPERSENSITIVENESS. VIII 213 104.0 101.0 104.8 100.6 99.8 101.6 103 .0 99.4 100.0 100.6 103 .6 101 .2 101.0 103 .0 102.2 103 .0 99 .6 100.0 100.0 100.0 104.0 100.2 102 .4 99 .6 100.2 100.0 100.0 99.8 100.2 99.2 100.2 99 .4 102.0 101.2 100.4 99.8 102.2 The total elevation of temperature over 100°F. in all of these individuals was thus 42.6°, or an average elevation of 1.15°F. The total elevation of temperature over 98.6°F. (by mouth) in the twelve affected Indians was 4.6°F., or an average elevation of 0.38°F. TABLE 3 Comparison of the serum reactions of the white and Indian races eee eee ee oe WHITE INDIAN Serum-treated cases followed completely.................--- 52 26 AMES On SETUN CISEERCs «hin o410 aniais 2 24 +See hi BEST 48 12 Percentage INCIDENCE, ...<, 2c. = 00 2. 2s sw oo ye sien ole ne seins s 92 .4 46 Average duration (days)..............0ceeee cere ence ee eees 9— 2 Average temperature elevation, degrees F............-..--. 1.15 0.38 DAE ERE Ie Sa nr orem EBERT S St Verde ae Eves etre ees a reese eae Ratio in which the two races are affected: 46 Dearie Cunt agit In table 3 is presented a summarized comparison of the occur- rence of serum disease in the two races. The ratio in which the two races are affected is properly estimated by multiplying to- gether the single ratios of the three different factors—incidence, duration and severity. By this calculation the susceptibility of the white race is found to be about twenty-seven times as great as that of the Indian. Objection must be made to the placing of much dependence upon the figures used in this calculation on account of the small- 214 A. F. COCA, O. DEIBERT AND E. F. MENGER ness of the number of individuals in both of the series that form the basis of comparison. In the circumstances, access cannot be had to further data as to serum disease in the Indian. However, we are able by the courtesy of Dr. Rufus Cole to supplement the observations in the white individuals with his own experiences in the Hospital of the Rockefeller Institute. It was important to obtain further data upon the white race because of the surprisingly high percentage of incidence observed in the Presby- terian Hospital—the highest, by far, that has yet been reported. We quote here from a personal communication from Dr. Cole: No intensive study of serum disease has been made at The Hospital of The Rockefeller Institute and any statistical conclusions drawn from our experience cannot be considered final. However, I have collected 223 cases in which amounts of serum varying from 30 cc. to 2000 ce. were administered. Only a few of the patients, however, received the very large amounts; most patients received from 200 cc. to 400 ce. during a period of two to three days. We have records concerning the occurrence of urticaria, skin rash, enlarged glands, red and painful joints, and fever. The statistics regarding fever will have to be studied much more carefully before much stress can be laid upon them, since in many pneumonia patients who have received serum, it is impossible to say whether a late elevation of temperature is the result of the ad- ministration of the serum or is due to some complication. Among the 223 cases, the occurrence of a skin rash, usually urticaria, at some time following the injection of serum, was noted in 158 patients; i.e. in 70 per cent. The rash appeared from one to thirty days following the injection of serum and varied in degree from a few urticarial wheals to urticaria covering almost the entire body. In a number of cases there occurred an erythematous rash; in a few cases it was scarlatini- form or morbilliform. Thirty-eight of the 223 cases ended fatally; in a number of instances, death occurred within a few days after the administration of serum before sufficient time had elapsed for a rash to develop. In only 9 of these 38 cases was the occurrence of a rash noted. If these 38 cases are omitted, 185 cases are left, of which 149 showed a rash at some period; i.e. 80 per cent. The occurrence of painful joints was noted in a considerably smaller number of cases, only in 68 of the 185 cases, and general glandular enlargement was noted in only 36 of the 185 cases. In some cases, however, enlarged glands, or painful joints, were noted where there was STUDIES IN SPECIFIC HYPERSENSITIVENESS. VIII 215 no skin rash, so that among the 185 cases there occurred 153 in which skin rash, glandular enlargement, or painful joints were present at some period, namely, in 85.4 per cent of the cases. When our former statements regarding the frequency of occurrence of serum sickness were made, we disregarded those cases in which urti- caria occurred immediately following the administration of serum and those cases in which the signs and symptoms were hardly noticeable. Employing, however, your definition of serum disease, at least 85 per cent of our cases may be said to have suffered from this condition, and considering the fact that in a few instances we undoubtedly overlooked very minor manifestations, it is quite probable that the statement that “90 per cent of patients following administration of foreign serum suffer from serum disease’’ is correct. Further support of Dr. Cole’s opinion was obtained in a study of a series of 367 cases receiving intravenous injections of serumin the Boston City Hospital. We are indebted for permission to publish these statistics to Dr. Edwin H. Place, Physician in Chief to the South Department of the Boston City Hospital. Ninety-nine of this series were cases of pneumonia that had received anti-pneumococcus serum. We are indebted to Miss L. M. Corcoran for assistance in compiling the data on these ninety- nine cases. | The average amount of serum injected in these cases was 272 cc. The incidence of serum disease recorded in the histories is 67.7 per cent. The remaining 268 individuals were cases of diphtheria in the South Department of the hospital and received an average of 150 ce. of diphtheria antitoxin serum. As antitoxin is adminis- tered in this department sometimes in the form of the pseudo- globulin fraction and sometimes as whole serum, it was necessary for us to inquire in each instance as to the nature of the prepara- tion used. This was readily done because the laboratory serial number of the preparation injected was always noted in the record. We have been able to obtain this information through the kindness of Dr. Benjamin White, Director of the Division of Biologic Laboratories of the Massachusetts Department of Publie Health, in whose laboratories the antitoxin was prepared. 216 A. F. COCA, O. DEIBERT AND E. F. MENGER The incidence of serum disease in the diphtheria cases was 70 per cent. It should be noted that the incidence in this series receiving an average of 150 cc. is about the same as that of the pneumonia series receiving 272 cc. Fifty-six individuals re- ceived less than 100 ec. and of these fifty-six, eight received less than 50 ce. These percentages are so close to the original estimated 70 per cent of Dr. Cole that they furnish strong support to his conclu- sion that some form of serum reaction follows the intravenous injection of horse serum in about 90 per cent of individuals so treated. The correction introduced by Dr. Cole in his original estimate is particularly applicable in the Boston series on account of the fact that all cases that were under observation for ten days or more are included in it. The evidence presented above seems to leave little doubt as to the approximate correctness of the figure, 92.4 per cent, used in our comparison of the two races to indicate the incidence of serum disease in the white race. This being the case, the esti- mated ratio of susceptibility of the two races to serum disease may be regarded with some confidence as indicating, at least, a wide difference in the relative susceptibility. ‘It is seen that a similar difference in the relative susceptibility of the two races is exhibited in both allergy (hay-fever, asthma, urticaria) and serum disease. This fact by no means proves these two conditions to be due to the operation of an identical mechanism. It merely suggests a similarity of mechanism in both conditions, which may not be complete. SUMMARY Through inquiry it has been found that the American Indian is apparently much less frequently affected by the allergies than is the white race. An experimental study of the occurrence of serum disease in twenty-six volunteer full-blood American Indians indicates that. the Indian race is much less susceptible to that condition than is the white race. STUDIES IN SPECIFIC HYPERSENSITIVENHESS. VIII PAW This similarity in the relative susceptibility of the two races to these two conditions suggests a similarity in the underlying mechanism of both of the conditions which, however, need not amount to a complete identity. REFERENCES 1. Hrdliéka Ales, Smithsonian Institution, Bureau of American Ethnology. Bulletin 34, 1908. 2. Lain, E. S8., Journ. Amer. Med. Assoc., July 19, 1913, page 168. QTV f#1HOUy UALSO of VIN LUV WOT VINAVMV SM ONIM dVNQ OUALSN AY NIMSQYT MOD V ¢ CULE YY DMOANWAV (CT UAONTAC TOT TH V WV UOLLTV AA NOSVJN af * XANY NOUV J WVITITA SRIMO YW AUGINGT Guy MAST GOWTVIV A MOU MOVE Nay NOSuHany SMO, Nvo NOSTHO“A TT ‘ GUVMGG HVNNVYET XOVGWING CLADINSOW NNVQl sindnq ANVUATA UP saArTANUoD auvM doy MOUANY SUTAVAS aval SSI va] XONVN VNNGYT ‘IMZvyET SUND WV SIONV4 iT MOU HTIAAIN TVIANUV GNOWAVY NOLINV]T GAT TAAT ILOOg NOSSV A SNIMLY NOSNHO(* aDVSO AVAHUV ET UNVUT Hdaso NOLMYV’] WVITTIA DUAL LO YY UVWIACT SVNOH J, WVITILAA UCAUVET NINVENGET aor WVITTIAA MOM LNOW STUDIES IN SPECIFIC HYPERSENSITIVENESS IX. ON THE PHENOMENON OF HYPOSENSITIZATION (THE CLINICALLY LESSENED SENSITIVENESS OF ALLERGY) ROBERT A. COOKE From the Department of Bacteriology and Immunology, Division of Immunology, Cornell University Medical College, and the First Medical Division, New York Hospital Received for publication December 1, 1921 In 1911, the writer began the application of the principle of desensitization to the treatment of the allergies (hay fever, asthma, urticaria, angio-neurotic edema). This principle was employed under the influence of the view of Wolff-Eisner and of Meltzer that the hypersensitiveness of human beings is an expression of anaphylaxis; that is, dependent upon the presence in the sensitive tissue of specific antibodies. Coca (1, 2), in a recent analysis of the phenomena of anaphy- laxis and those of allergy, has pointed out differences between these two conditions of such a nature as to make an identity of the two seem unlikely. One of these differences is a fact that had been observed by the writer in innumerable instances; namely, that although a certain degree of lessened sensitiveness can be obtained in allergy by administration of the exciting agent, either in its natural state by mouth or in extracts by injection, this effect in human beings has never been observed to approach the entire insensi- tiveness of the condition of ‘‘complete desensitization” in ana- phylaxis in the lower animals. In illustration of this principle, the following concrete obser- vations were drawn by Coca out of our unpublished records. In all of the individuals suffering from any form of allergy, who have been rendered clinically insensitive to the natural contact with the exciting. agent, the suitable administration of the agent, 219 220 ROBERT A. COOKE by intracutaneous or subcutaneous injection will demonstrate the persistence of the hypersensitiveness. In some of our cases, indeed, it could be shown that even the dnsensitiveness to the natural contact was only relative. For example patient A. K., No. 1107,! sensitive to egg, was brought by appropriate injections and oral administration of increasing doses of egg white to the point where one whole egg every other day was tolerated, but on several trials he was found unable to eat two eggs on one day without exhibiting symptoms. The phenomenon of desensitization is a constant characteristic of the condition of anaphylactic hypersensitiveness in all the iower animals in which the phenomenon has been studied. This must be true because the hypersensitiveness of anaphylaxis depends upon the presence of antiprotein antibodies (precipitins) in the tissues; and because it is always possible by a suitable manner of injection of the antigen to neutralize these precipitins so gradually as to avoid the physiological reaction of anaphy- laxis and to neutralize them so completely that the hypersensi- tiveness is entirely removed. It would seem, therefore, that desensitization, which is a con- stant characteristic of anaphylaxis, should be demonstrable in that condition in every animal in which anaphylaxis occurs. A conceivable exception might exist in an animal in which the antibody production was so rapid that the neutralized precipitins were immediately replaced. However, no such exception is known. : It would seem natural for one seeking to demonstrate desensi- tization in allergy to choose serum allergy as the first object of study; particularly, the ordinary form of this condition in which the clinical symptom of urticaria is only exhibited after an incu- bation period following the injection of serum and which is gener- ally accompanied by precipitin formation. However, no one has yet applied the procedure of desensitization to individuals under these circumstances. In the rarer form of serum allergy, in which symptoms appear 1 This case is described in greater detail in the article on constitutional reac- ‘tions; this Journal volume 7, p. 141. STUDIES IN SPECIFIC HYPERSENSITIVENESS. IX 221 immediately after a primary injection and in which the previous existence of antibodies has not been demonstrated (3), attempts have been made to reduce the sensitiveness by the injection of increasing quantities of serum, beginning with small doses. In his paper on ‘‘ Asthma Complicating the Serum Treatment of Pneumonia” (4), H. L. Alexander describes three cases in two of which the procedure just referred to was employed. In one of these the subcutaneous injection of 1 cc. caused urticaria, and the same quantity injected intravenously two and one-half hours later caused nausea, vomiting and asthma. One and a half hours after this occurrence 1 cc. was again injected and this again caused nausea and vomiting, but without an asthmatic attack. Two further injections of 2 cc. and 4 cc. produced no symptoms and a final intravenous injection of 65 cc. was accom- plished with no more serious effect than an attack of ‘‘hay fever’ and of asthma. Another individual was brought gradually to the intravenous injection of 1 cc. of serum without symptoms. ‘Then the intra- venous injection of 40 cc. of the serum produced asthma and subsequent injections of 40 cc., 60 cc., 65 cc., 70 ce., and 75 ce. at from six to eight and a half hour intervals all caused mild or moderate ‘‘reactions.” Alexander adopted the view, at that time unopposed, that he was dealing with a state of anaphylac- tic sensitiveness. He overlooked the difficulty attaching to this view by reason of the uniform absence of demonstrable antibodies in the form of serum sensitiveness that he was studying. In a paper entitled ‘‘Serum Desensitization” (5), George M. Mackenzie added two further cases to those of Alexander. In one of these 1 cc. of serum injected intravenously produced respiratory difficulty and marked urticaria. An identical in- jection twelve hours later caused no symptoms, but after a further interval of one and a half hours, 2 cc. again produced marked urticaria, this time with edema. In the second case 16 cc. of serum produced eruption, although 8 cc. administered three quarters of an hour previously had caused no symptoms. In all of these experiments the writer’s experiences referred to above have been duplicated. A lessened sensitiveness was 272 ROBERT A. COOKE attained in a practically important degree, but not complete insensitiveness. Moreover, this result was obtained under cir- cumstances—the repeated intravenous injection of relatively large quantities of serum—which warrant the conclusion that complete specific Insensitiveness is not attainable in allergy. The uniform failure to induce complete insensitiveness in allergy indicates, according to Coca, that the relative insensi- tiveness attained in allergic conditions is of a nature quite dif- ferent from that of desensitization. The objection may be raised against Coca’s conclusion that the difference between these two effects need not be a qualitative one; it may be only quantita- tive. That is, the clinical insensitiveness attainable in allergy may represent a partial desensitization. In fact, a superficial examination favors this latter view. However, facts are at hand to prove that the difference between the two effects is not quantitative but qualitative. They indi- cate, by the same token, that the allergic hypersensitiveness of human beings is not due to the influence of precipitin—it is not anaphylaxis. We are referring to the results published by Coca and Kosakai (6) in their study of the reactivity of partially neutralized pre- cipitin. ‘These investigations showed that the partial neutrali- zation of precipitin takes place according to a law which is quite different from that governing partial chemical precipitation. As they did not analyze their experiments from this point of view, we will do it here. In their table (1) it appears that in guinea-pigs passively sensitized with 0.4 cc. of a precipitating serum, only slight symp- toms were caused by the injection of 0.0025 ce. of the antigen solution. After a partial neutralization of the precipitating serum with a quantity of the antigen corresponding with 0.0008 ce. animals were sensitized with 0.4 cc. of the treated serum and it was found that in these animals 0.01 cc. of the antigen was required to produce slight symptoms; that is, twelve and a half times as much as was used for the partial neutralization. In a similar experiment (table 4) in which the partial neutrali- zation was carried out i vivo with 0.008 cc. of the antigen 0.1 cc. was required to cause symptoms in the sensitized guinea- Eo STUDIES IN SPECIFIC HYPERSENSITIVENESS. IX 223 pigs—again twelve and a half times the quantity of antigen used for the partial neutralization of the antiserum (partial desensi- tization in this case). A third experiment (table 8) with a different serum, resulted in a larger ratio between the dose of antigen used for partial desensitization and that required to produce symptoms. The ratio in 8a was 0.00025:0.0133::1:53; in 8b 0.0025:0.2::1:80. The discrepancy between these two ratios is due to experimental error. In the same paper Coca and Kosakai were able to demonstrate that the reactivity of partially neutralized precipitin, as judged in vitro by the criterion of precipitation, is quantitatively the same as it is in vivo as judged by the criterion of anaphylactic shock. The fundamental difference between the quantitative laws governing immunological and chemical precipitation is seen in a comparison of the experiments just described and the following: if to 10 cc. of a decinormal solution of silver nitrate 1 ce. of decinormal hydrochloric acid is added and if the resultant pre- cipitate is removed a further precipitation can be produced by the second addition of 1 ec. of decinormal hydrochloric acid; moreover, the second precipitate will be of exactly the same weight as was the first. The experiments of Coca and Kosakai reveal a simple but fundamentally characteristic peculiarity in the partial neutra- lization of precipitin which may be used as a test to determine the nature of human hypersensitiveness. If the latter is dependent upon the presence of precipitin, then the successive injection of identical or nearly identical quantities of the exciting agent should not cause repeated exhibition of symptoms. A multiple of the partially desensitizing dose is always required to cause symptoms upon a subsequent injection. Conversely, if symp- toms, even in slight degree, recur upon such repeated injections, then precipitin can have no part in the production of the symptoms. Applying this test to the observations of Alexander and of Mackenzie which we have referred to above, we must conclude that the serum sensitiveness which they were attempting to 224 ROBERT A. COOKE modify was not anaphylactic in nature and that the reduced sensitiveness which they established was therefore not a partial desensitization. In the first case cited from Alexander, two successive intrave- nous injections of 1 cc. of serum were followed in each instance by symptoms which were altered at the second injection but still marked. In the second case six successive injections of nearly equal and large quantities of serum each produced symptoms. In one of Mackenzie’s cases an intravenous injection of 1 ce. of serum was given without symptoms, yet one and a half hours later 2 cc. caused marked reaction. The conditions in Mac- kenzie’s second case were similar to those of the first and the results were the same. One of the outstanding and practically important features of allergic hypersensitiveness is the specific reactivity of the skin. The easy accessibility of this tissue invited study of the effect of injections of the exciting agent of the hypersensitiveness upon the cutaneous reaction. The writer had had numerous opportunities to make such a study following subcutaneous injections given in hay fever and asthma for therapeutic purpose. The effect of those injections has usually been a specific lessening of the general cutaneous reactivity, which, however, never approached extinction. Recently Mackenzie and Baldwin (7, 8) have investigated the effect of local application of the exciting agent upon the cutaneous reaction and they have concluded that after repeated applica- tions to the same site the reactivity of that site is specifically exhausted. In some instances the exhaustion did not persist longer than twenty-four hours; in one case, however, the intra- cutaneous injection of a 1:10 dilution of egg white caused a complete suppression of reactivity in the same site for three days. In several unpublished experiments we had been unable to demonstrate an exhaustion of the cutaneous reactivity after an interval of about twenty hours following one or more intradermal injections, although in some instances the solution used for the attempted exhaustion was several times stronger than that used for the final test on the following day. STUDIES IN SPECIFIC HYPERSENSITIVENESS. IX 225 In further experiments, which we are about to describe, we have made repeated intradermal injections in the same site at short intervals and we have succeeded in producing practically complete exhaustion of the reactivity of the site to the substance in the particular concentration in which it had been injected. However this effect could be shown not to be specific; it appears, thus, to be merely a temporary exhaustion or fatigue of the general power of reactivity on the part of the tissues to irritation. The experiments were carried out by the writer on himself. This circumstance permitted a continuous observation as to both the objective and the subjective phenomena during the considerable period of time occupied by the tests. The writer is highly sensitive to horse and rabbit dandruff and also to the serum of these two animals. The first experiment was carried out with extracts of the danders; in the second experiment the two sera were used. In both experiments the volume of fluid injected was always the same—0.01 to 0.02 cc. The concentration of the allergens in the different solutions is indicated by the quantity of nitro- gen in 1 cc. of fluid as determined with the Kjeldahl method. For example, horse epithelium 0.01 means an extract of horse dander 1 cc. of which contained 0.01 milligram of nitrogen. All injections were made intradermally and when repeated injections were made in the same site the needle was always introduced into the same puncture orifice. EXPERIMENT 1? July 23. 12:35 p.m. Horse epithelium 0.01 is injected into site 1 on the anterior surface of the left forearm. 12:50 p.m. The area of reaction is drawn (fig. 1, chart 1. 4:55 p.m. The wheal at site 1 has disappeared but the site is still red and there is an area of edema of the size of a quarter. Horse epithelium 0.01 is injected into site 1. 2 On July 25 at the time the tests were made with horse epithelial extract 0.05 milligram of nitrogen per cubic centimeter there were noted some constitutional effects—coryza, asthma and slight urticaria. About five days after these tests were made there developed further mild symptoms of asthma with coughing and 226 5:05 p.m. 10:15 p.m. 10:30 p.m. 8:15 a.m. 8:30 a.m. 2:10 p.m. 2:20 p.m. 4:10 p.m. 4:25 p.m. 6:20 p.m. 6:30 p.m. 10:30 p.m. 10:45 p.m. expectoration and an occasional hive. ROBERT A. COOKE The area of reaction is drawn (fig. 2, chart 1). The wheal is somewhat larger than that produced by the first in- jection and there is marked itching as at first. Site 1 still presents an area of redness and edema measur- ing about 2 by 3 inches. Horse epithelium 0.01 is injected into site 1. The area of reaction is drawn (fig. 3, chart 1). There is no increase in the zone of hyperemia. A wheal is again formed accompanied by considerable itching. July 24 The zone of redness and edema still persists (2 by 3 inches). Horse epithelium 0.01 is injected into site 1. The area of reaction is drawn (fig. 4, chart 1). The reaction as a whole is distinctly less than at the first injection.® There is very slight edema and no redness at site 1. Horse epithelium 0.01 is injected into site 1. The area of reaction is drawn (fig. 5, chart 1). The wheal is of about the size of the earlier ones but it is of shorter duration; the zone of hyperemia is strikingly decreased. There is practically no edema and no redness at site 1. Horse epithelium 0.01 is injected into site 1. The area of reaction is drawn (fig. 6, chart 1). The wheal is less sharply defined and hyperemia is almost absent. Horse epithelium 0.01 is injected at site 1. The area of reaction is drawn (fig. 7, chart 1). The wheal is ill defined, merging into a small zone of hyperemia. Horse epithelium 0.01 is injected into site 1. The area of reaction is drawn (fig. 8, chart 1). The reaction is distinctly more marked than it was at the last preceding injection. There is no itching. This necessitated the use of adrenalin in a dose of ten minims two to three times daily. The condition lasted for ten days and then subsided over a period of four days. After the second series of tests with serum to which the writer is less sensitive there was no constitutional reac- tion aside from an occasional hive during four to five days after the last test. 3 It must be borne in mind that the relative duration of the reactions and the relative intensity of the hyperemia are not indicated in the drawings. STUDIES IN SPECIFIC HYPERSENSITIVENESS. 130. 8.m. 7:45 a.m. 9:15 a.m. 9:30 a.m. 10:40 a.m. 10:50 a.m. 5:55 p.m. 6:05 p.m. 6:15 p.m. 6:30 p.m. 8:30 p.m. 8:35 p.m. 8:50 p.m. IX 297 July 25 1 There is a small papule 3+ inch in diameter about the puncture orifice. The papule is slightly tender and not red. Horse epithelium 0.01 is injected into site 1. The area of reaction is drawn (fig. 9, chart 1). The reac- tion is of short duration. There is no itching. The papule is slightly tender; there is no redness. epithelium 0.05 is injected into site 1. The area of reaction is drawn (fig. 10, chart 1). The wheal is larger than the preceding one; there is no hyperemia and no itching. The wheal is of short dura- tion. The papule is tender; there is no redness. thelium 0.1 is injected into site 1. The area of reaction is drawn (fig. 11, chart 1). The injection of this more concentrated solution caused an evanescent hyperemia and considerable itching which was felt also in the cubita fossa. Only the papule persists. Horse epithelium 0.1 is in- jected into site 1 and also into another place on the same forearm—-site 2. The two areas of reaction are drawn (fig. 12a, chart 1— site 1; fig. 12b, chart i—site 2). The reaction at site 1 was of very short duration without itching. That at site 2 was accompanied with itching, hyperemia and edema which persisted until 6:00 p. m. on the following day. Rabbit epithelium 0.05 is injected into the lower anterior surface of the right forearm—site 3. The area of reaction at site 3 is drawn (fig. 13a, chart 1). There is a lymphangitis which is visible at intervals from the injection site to the cubital fossa, where it is especially marked. The area of reaction at site 3 is again drawn (fig. 13b, chart 1). Both the wheal and the zone of hyperemia have increased. On the following morning edema and redness were still present. Rabbit epithelium 0.5 is injected into site 1. The area of reaction of site 1 is drawn (fig. 14, chart 1). There is a slight lymphangitis extending to the cubital Horse Horse epi- 228 © 10:45 p.m. 4:00 p.m. 4:15 p.m. 8:45 p.m. 9:00 p.m. 8:50 p.m. 9:05 p. m. 7:45 a.m. 8:00 a.m. 12:45 p.m. 1:00 p.m. 8:25 p.m. 8:35 p.m. 11:30 p.m. ROBERT A. COOKE fossa and a very slight lymphangitis from the fossa to the axilla. The reaction produced at site 1 by the last injection has subsided; no trace of it is left. July 26 Horse epithelium 0.1 is injected into site 1. The area of reaction at site 1 is drawn (fig. 15, chart 1). The resulting wheal was larger than that of 12a. It persisted for one-half hour and was accompanied with itching, which extended up through the cubital fossa into the axilla. EXPERIMENT 2 August 25 Horse serum 0.1 is injected into site a on the right forearm The area of reaction is drawn (fig. la, chart 2). At site x on the same forearm rabbit serum 1 was injected as a preliminary test of the cutaneous reactivity to that material. The area of reaction at site x is drawn (fig. 1b, chart 2). August 26 Site a is still red and slightly edematous. is injected into site a. The area of reaction is drawn (fig. 2, chart 2). The reac- | tion appears to be slightly greater than after the pre- ceding injection. This is doubtless due to the per- sistence of the effects of the first injection. Some of the wheal and of the hyperemia still persist from the second injection. Horse serum 0.1 is in- jected into site a. The area of reaction is drawn (fig. 3, chart 2). There is still some hyperemia at site a. Horse serum 0.1 is injected into site a. The area of reaction at site a is drawn (fig. 4, chart 2). There is some itching in the site, but no increase in the hyperemia and little in the wheal. Shght hyperemia still persists at site a. is injected into site a. Horse serum 0.1 Horse serum 0.1 STUDIES IN SPECIFIC HYPERSENSITIVENESS. IX 229 11:40 p.m. The area of reaction is drawn (fig. 5, chart 2). There is slight itching in the site but no real increase in the wheal nor in the hyperemia. August 27 7:45a.m. Horse serum 0.1 is injected into site a. 8:00 a.m. The area of reaction at site a is drawn (fig. 6, chart 2). There is very slight itching and faint hyperemia in the site. ‘The wheal is not increased in size. There is an urticarial spot on the right thigh. 9:00 a.m. Horse serum 1 is injected into site a. 9:10 a.m. The area of reaction at site a is drawn (fig. 7, chart 2). This reaction lasted only twenty minutes. There was a constitutional reaction—asthma, with generalized itching and coryza with itching of the eyes—which persisted for about two hours. 12:10 p.m. Horse serum 1 is injected into site a and into site b on the left forearm. 12:20 p.m. The area of reaction at site a is drawn (fig. 8a, chart 2). This reaction lasted fifteen minutes. There was a con- stitutional reaction with coryza and generalized itching which lasted about two hours. Five urticarial spots appeared on the body. 12:25 p.m. The reaction at site b is drawn (fig. 8b, chart Z)e,, Lis reaction persisted for twenty-four hours. 3:00 p.m. Rabbit serum 1 is injected into site a and into site c on the left forearm. 3:10 p.m. The area of reaction at site a is drawn (fig. 9a, chart 2). There is no itching and no definite hyperemia. The size of the wheal is not increased. The reaction at site is marked, though not so intense as that at site b. 3:15 p.m. The area of reaction at site c is drawn (fig. 9b, chart 2). The reaction persisted for about twenty hours, with edema, itching and hyperemia. Both of these experiments bear out the conclusions which we arrived at in our discussion of the observations of Alexander and of Mackenzie: namely, that there is no true desensitization in allergy. In experiment 1, three successive injections of the same quantity of horse epithelium 0.01 produced marked reactions 230 ROBERT A. COOKE (figs. 1, 2, and 3, chart 1). In experiment 2, two successive injections of horse serum 0.1 produced marked reactions. The second noteworthy feature of both experiments is that aiter a site has been made nearly insensitive to one concentra- tion of the material that is being injected, a vigorous reaction can be produced with a stronger concentration of the same material (fig. 11, chart 1; and fig. 7, chart 2). In this fact may be seen a possible explanation of the belief of Mackenzie and Baldwin that the local ‘“‘exhaustion” of the allergic reaction is specific. These authors may have tested the exhausted site with a second protein in a concentration higher than that of the protein used for the exhaustion. The third result of the tests, which is evident in both experi- ments, is that the local insensitiveness produced by the repeated injections is not specific, as Mackenzie and Baldwin thought. It is merely the well known nonspecific fatigue of the tissues to protracted irritation. Before this condition of fatigue has set in, the allergic mechan- ism continues to function upon repeated identical injections and even after the tissues have become fatigued to one concentration of the allergen the mechanism can be shown to be still intact by the injection of the allergen in greater concentration. In experiment 1, after site (1) had been rendered insensitive to horse epithelium 0.1, the injection of the stronger extract of rabbit epithelium 0.5 produced a distinct reaction. This reac- tion was, however, much weaker in intensity and duration than that caused by the same injection into the fresh site (3). It was probably no greater than one that would have been produced in the ‘“‘exhaused”’ site by horse epithelium 0.5. In experiment 2, the drawings of the final reactions are in themselves convincing evidence of the non-specificity of the local insensitiveness produced by repeated injections. The necessity of seeking further evidence to support the fore- going interpretation of our experiments has been recognized. It was necessary, first, to find out whether a non-specific local exhaustion of the skin is possible and, secondly, if this was true, to see whether any significant difference is demonstrable in the Se STUDIES IN SPECIFIC HYPERSENSITIVENESS. IX Pan development of this non-specific local exhaustion and of that induced by the repeated injection of allergens in specifically sensitive individuals. These two questions were investigated in some experiments that were carried out in this department by Dr. W. C. Spain and Dr. Ruth Guy with peptone and histamin. The susceptibility of the skin to peptone was first described by Philippson in the Giorn. ital. d. mal. ven. e. d. pelle in 1899; that to histamin was first described by Eppinger in the Wien. med. Woch. in 1913. Since the local effect of both of these substances is non-specific they lent themselves well to the pur- pose in mind. Mackenzie and Baldwin (8) state that Sollman found the histamin skin reaction inexhaustible and that they themselves found this reaction actually to increase in size and intensity with each successive application of the substance. A review of the experiments of Sollman (9, 10) shows that they are inade- quate to determine this question. In the earlier experiments with Pilcher (9) he states (page 313), ‘‘When repeated applica- tions are made (histamin) whether to the same or to the opposite arm, the wheals of the later applications appear smaller than the earlier.” In his later study (10) Sollman describes only one experiment (no. 18) in which a ‘‘repeated”’ application of hista- min was made at one site. In this experiment the second appli- cation of a jy per cent solution after a thirty-five minute inter- val produced a more vigorous reaction than had the first. No further applications were recorded. Evidently no information regarding the possible exhaustibility of the histamin reaction is obtainable from this experiment. Mackenzie and Baldwin (8), also, used the strong j per cent solution of histamin. They do not state how many applications were made nor what intervals of time were observed. In the experiments by Dr. Spain and Dr. Guy the concentration of the two substances in the solutions is given in percentage of the dry material. The quantity injected was the same as in the writer’s experiment (0.01 cc.). All of the injections were intradermal. In the first experiment by Dr. Spain all of the injections were made into the same puncture orifice on the left forearm as follows: 232 12:55 p.m. 1:15 p.m. 2:30 p.m. 2:45 p.m. 3:00 p.m. 3:20 p.m. 4:00 p.m. 4:20 p.m. 4:50 p.m. 5:10 p.m. 7:15 p.m. 7:35 p.m. 8:10 p.m. 8:30 p.m. 8:30 p.m. 8:30 a.m. 11:45 a.m. 12:00 a.m. 4:00 p.m. 4:30 p.m. ROBERT A. COOKE July 23 Histamin, 0.01.4 Area of reaction drawn (figure 1, chart 3). Histamin, 0.02. Area of reaction drawn (figure 2, chart 3). Histamin, 0.04. Area of reaction drawn (figure 3, chart 3). of the wheal are indistinct. Histamin, 0.04. Area of reaction drawn (fig. 4, chart 3). forms slowly and is indistinct. Histamin, 0.04. Area of reaction drawn (fig. 5, chart 3). The wheal is not elevated above the level of the surrounding tissue; its outline is indistinct. Histamin, 0.04. Area of reaction drawn (fig. 6, chart 3). the wheal is questionable. Histamin, 0.04. Area of reaction drawn (fig. 7, chart 3). The outlines of the reaction are very indistinct: the surrounding skin is indurated in a space about 3 inches in diameter. There is no definite wheal. Histamin, 0.2. No definite wheal resulted from this in- jection. No drawing was made. July 24 The area of induration still persists, with a shot-like central node. No injection made at this time. Histamin, 0.4. Area of reaction drawn (fig. 8, chart 3). A very indis- tinct wheal is formed which is only slightly elevated. There is little erythema. July 26 The outlines The wheal The outline of Histamin, 0.4. The outlines of the reaction are very indistinct reaction drawn (fig. 9, chart 3). Area of 4100 cc. of this solution contained 0.01 grams of histamin. The substance was obtained from Burroughs Wellcome & Co. under the name Ergamine Acid Phosphate. STUDIES IN SPECIFIC HYPERSENSITIVENESS. IX 233 In the second experiment by Dr. Spain, all of the injections excepting the final control injection were made into the same puncture orifice on the left forearm as follows: 9:00 a.m. 9:15 a.m. 10:00 a.m. 10:05 a.m. 10:10 a.m. 11:00 a.m. 11:15 a.m. 12:05 p.m. 12:15 p.m. 2:00 p.m. 2:15 p.m. 3:45 p.m. 3200) p.m. 5:00 p.m. 5:10 p.m. 9:00 a.m. 9:15 a.m. 9:30 a.m. 9:45 a.m. September 13 Histamin 0.1. Area of reaction drawn (fig. 1, chart 4). There is decided erythema and much ake after five minutes. Histamin, 0.1. The reaction caused by the first injection has subsided. There is immediate itching. There is a sensation of dizziness and ‘‘light headedness.”’ which lasts about one to two minutes. Blood pressure is 112/80. .(The normal pressure is 118/80 —-116/80.) Area of reaction drawn (fig. 2, chart 4) Histamin, 0.1. The wheal and most of the erythema have disappeared. Area, of reaction drawn (fig. 3, aie 4). The reaction is apparently equal to the previous one. Histamin 0.1. Area of reaction drawn (fig. 4, chart 4). There is very little reaction; the size of the wheal and extent of the erythema are not increased. Histamin, 0.1. Area of reaction drawn (fig. 5, chart 4). There is little erythema and little increase in the size of the wheal. Histamin 0.2. Area of reaction drawn (fig. 6, chart 4). There is some increase in the size of the wheal, though the reaction is not violent. The zone of erythema is comparatively small. Histamin, 0.2. Area of reaction drawn (fig. 7, chart 4). The reaction was slight. September 14 Histamin 0.2. Area of reaction drawn (fig. 8, chart 4). The reaction was slight; the hyperemia was hardly visible. Histamin, 0.2. Area of reaction drawn (fig. 9, chart 4). There is no in- crease in the size of the heal: the hyperemia is vague. 234 11:00 a.m. Jar bs amt 11:00 a.m. 11:15 a.m. ROBERT A. COOKE Histamin, 0.4. Area of reaction drawn (fig. 10, chart 4). The wheal remains the same and the hyperemia is very slight. Histamin 0.4 in right forearm. Area of reaction drawn (fig. 11, chart 4). The experiment by Dr. Guy was earried out on herself as follows: 10:29 a.m. 10:37 a.m. 11:55 a.m. 12:07 p.m. 2:07 p.m. 2:40 p.m. 4:00 p.m. 4:12 p.m. team. 11:16 a.m, 12:50 p.m. 1:04 p.m. 2:09 p.m. 2:25 p.m. 3:56 p.m. 4:01 p.m. 9:08 p.m. 9:15 p.m. 11:03 p.m. Wet 7:pan: 10:47 a.m. 10:56 a.m. September 22 Histamin, 0.02, site 1. Area of reaction drawn (fig. 1, chart 5) Histamin, 0.02, site 1. Area of reaction drawn (fig. 2, chart 5). Histamin, 0.02 site 1. Area of reaction drawn (fig. 3, chart 5.) The wheal was rather indefinite. Histamin, 0.02, site 1. Area of reaction drawn (fig. 4, chart 5). Wheal again not sharply defined. September 24 Histamin, 0.02, site 1. Area of reaction drawn (fig. 5, chart 5). A definite reaction. Histamin, 0.02, site 1. Area of reaction drawn (fig. 6, chart 5). The reaction was slight. Histamin, 0.02, site 1. Area of reaction drawn (fig. 7, chart 5). Slight reaction. Histamin, 0.02, site 1. Area of reaction drawn (fig. 8, chart 5). Slight reaction. Histamin 0.02, site 1 right forearm. Area of reaction drawn (fig. 9, chart 5). Slight reaction. Histamin 0.02, site 1. Area of reaction drawn (fig. 10, chart 5). Slight reaction. September 24 Histamin 0.02, site 1. Area of reaction drawn (fig. 11, chart 5). Very slight reaction. =e —— eo ‘SUOISe] Surpuodsar1o9 oy} Jo vsoy} J[vYy ATJOVXE ov SBUIAMvIpP a80q3 Jo S19JOWVIP OT, ‘T0'0 “untpouztds osxoy Suorjoofur yyUTU ST OIG 6 BI ‘TO ‘unrpoyyide asioy { uorzoofut YYUIEZINO} fT 9419 “eT “BIW "10'0 ‘umnipeyyide os1oy fuoro0lur yYYSIO {T o71g *g io) sf go “wantpeyyide yiqqea fuorqoolur yyuezATYY {T OIG “FT “BI ‘10°0 “umnrjeyytdo esioy ‘ worz00fur yyUAAOS {] ORIG *y “Bry *197B] SINOY OY “BET UI SB UOTJOVOI OUTYS {Ee O4IG “qeT ‘SIT "10'0 ‘amntpoyyida esroy ‘ uorzoalut YyQxIs fT ay1g *9 ‘Shy ‘gO TunTpouzde 4rqqus Suoroalur ysry fg oyTg “BET “BI "100 “uantpoyydo estoy Suoryoolur yIFy ST OIG *¢SIy ‘TO ‘wnrpeyytde esioy Suoroolur ysay fz OIG “qZr “SIT ‘T0°0 “umntpoyztdo essoy {uoryoofut yQANOJ {T OIG “F“SLT ‘TO “umrpeyztde osroy fuoryoolur yyyjoM9 $7 OFIG “BZ “SLT ‘10'0 “wmrpeyzde estoy ‘uoryoofur paryy fT ayIg *g “BET ‘TO ‘untjoqzide osioy { uorzoolur YyUOANTO ! [ OFIG “TT BUT "10'0 “wntjoyytdae osroy fuoryoalur puooas !{ o7IG *Z ‘BI,T "GO'0 “umnyyoyytda ossoy uorzoofur YZUO4 ST ORIG “OL SLT “T0'0 “umnrpouytde osioy Suoroolur qsry fy ORIG “TSI [T “Lavay 2 Sl V ZI COOKE ROBERT A. 236 ‘guoIse] BuUIPUodser109 oY} JO o80y} J[BY A[}O"Xe O1B BZAIMBIP 9804} JO S1OJOUUVIP OY, J ‘winsas 41qqva {uoryoofur ysay fo ay1g “G6 BIT “7 ‘wunses 41qqet {uoTpoolfUr YQUIU SB V4Ig °G ‘BIT “1 ‘umnsos osi0y {uoryoolut ysry 6q OFIG “Gg “BIT “| ‘unsas oss0y SuoTpOLUr YBYSIO !B OYIG “Bg BIT ‘] ‘unsos os10y {UoTpOaluT YIWOAS +B O4IG “2 SL ‘TQ ‘wmnszos osr0y {uoryoofur YYXIS (B OFIG *g ‘SLT ge A Ye vp ‘T'Q ‘tanses osi0y { uOTPOOfUT YAY SB oWIG “GBI ‘TQ ‘wndes osi0Y { UOTZDOLUT YPANO; !B OIG “FSET ‘TO ‘tunes oss0y SuoTZdeLUT paTyy !B OyIG “EAI ‘TQ ‘umsos osi0yY { UOLZoafuT puodeS ‘B 941g *Z “SI “, ‘umnsoes yIqqea fuorTyolut 4ysry fx OFIG “qT “SIT ‘TO ‘umes osr0Y fuoTZOo!uUr 4say {eB O4IG “VT BLT Z LuvHO VI STUDIES IN SPECIFIC HYPERSENSITIVENESS. IX 237 Z Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. 9. CHART 3 1. Site a; first injection; histamin, 0.01 per cent. 2. Site a; second injection; histamin, 0.02 per cent. 3. Site a; third injection; histamin, 0.04 per cent. 4, 5 6 i Site a; fourth injection; histamin, 0.04 per cent. . Site a; fifth injection; histamin, 0.04 per cent. . Site a; sixth injection; histamin, 0.04 per cent. . Site a; seventh injection; histamin, 0.04 per cent. 8. Site a; ninth injection; histamin, 0.4 per cent. Site a; tenth injection; histamin, 0.4 per cent. The diameters of these drawings are exactly half those of the corresponding lesions. 238 OOnronFrkwhnd ROBERT A. COOKE Cuart 4 . Site a; first injection; histamin, 0.1 per cent. . Site a; second injection; histamin, 0.1 per cent. . Site a; third injection: histamin, 0.1 per cent. . Site a; fourth injection; histamin, 0.1 per cent. . Site a; fifth injection; histamin, 0.1 per cent. . Site a; sixth injection; histamin, 0.2 per cent. . Site a; seventh injection; histamin, 0.2 per cent. . Site a; eighth injection; histamin, 0.2 per cent. . Site a; ninth injection; histamin, 0.2 per cent. .10. Site a; tenth injection; histamin, 0.4 per cent. 11. Site b; first injection; histamin; 0.4 per cent. The diameters of these drawings are exactly half those of the corresponding lesions. STUDIES IN SPECIFIC HYPERSENSITIVENESS. IX 239 7 4 4 wer Sage 45 25 : ; 3 (} CuHart 5 Fig. 1. Site 1; first injection; histamin, 0.02 per cent. Fig. 2. Site 1; second injection; histamin, 0.02 per cent. Fig.3. Site 1; third injection; histamin, 0.02 per cent. Fig. 1. Site 1; fourth injection; histamin, 0.02 per cent. Fig.5. Site 1; fifth injection; histamin, 0.02 per cent. Fig. 6. Site 1; sixth injection; histamin, 0.02 per cent. Fig. 7. Site 1; seventh injection; histamin, 0.02 per cent. 8 Fig. 8. Site 1; eighth injection; histamin, 0.02 per cent. Fig. 9. Site 1; ninth injection; histamin, 0.02 per cent. Fig. 10. Site 1; tenth injection; histamin, 0.02 per cent. Fig. 11. Site 1; eleventh injection; histamin, 0.02 per cent. Fig. 12. Site 1; twelfth injection; histamin, 0.02 per cent. Fig. 13. Site 1; thirteenth injection; histamin, 0.2 per cent. Fig. 14. Site 2; first injection; peptone, 5 per cent. The diameters of these drawings are exactly half those of the corresponding lesions. 240 ROBERT A. COOKE 11:47a.m. Histamin 0.02, site 1. Area of reaction drawn (fig. 12, chart 5). Questionable reaction. 12:22 p.m. Histamin0.2 Site 1. 12:30 p.m. Area of reaction drawn (fig. 13, chart 5). Questionable reaction. 12:45p.m. Peptone (Witte) 5.0, site 1. 12:55p.m. Noreaction. No drawing was made. 12:45 p.m. Peptone (Witte) 5.0, site 2, left forearm. 12:55 p.m. Area of reaction at site 2 drawn (fig. 14, chart 5). The materials used in the foregoing experiments by Dr. Spain and Dr. Guy produce their effect upon the skin by direct action and not by the mediation of an immunological reaction. This is evident from the fact that most individuals are susceptible to these substances upon the first administration of them. There are quantitative differences in this susceptibility which we may discuss in a future paper, but these differences do not contradict the principle just stated. If the effect of the injection of the proteins of horse and rabbit dandruff is indirect—due to an immunological reaction—it should be expected that the existence of the intermediate immuno- logical mechanism would be expressed in some difference in the results of the repeated injections as compared with those of the repeated injections of the directly acting histamin and peptone. However, a comparison of the five series shows them to be identi- cal. In all, there is either an initial increase in the degree of the reaction or at least no change in it followed, then, by a decrease which is non-specific. The evidence which we have presented seems to warrant the conclusion that the lessened sensitiveness induced in allergy by the injection of the exciting agent is a phenomenon which differs in its mechanism from that of desensitization in anaphylaxis. The former seems not to depend on the neutralization of precipi- tin as does the latter. There is, however, to be considered the hypothetical possibility that the reactions of allergy depend on the presence in the tissues of ‘‘antibody-like” substances or so-called natural antibodies. This explanation was offered by Cooke, Flood and Coca (8) STUDIES IN SPECIFIC HYPERSENSITIVENESS. IX 241 but it was later dismissed by Coca (1), presumably on the ground that those hypothetical antibodies could not be neutralized completely. We hold this ground to be for the present a valid one because there is no known natural antibody which cannot be completely neutralized. It is useless to assume in allergy the existence of a natural antibody lacking a cardinal character of the known antibodies. One may as well assume the intermediary action of natural antibodies to explain the specific physiological effects of drugs. On account of the confusion that must result from the use of the well defined term “desensitization” to designate clini- cally lessened sensitiveness in allergy it is suggested that the latter be referred to as a “‘hyposenisitization.” The etymologi- cal defect and also the convenience of this proposed term are the same as those of hypersensitization, which is in general use. SUMMARY 1. The quantitative relations in partial desensitization are contrasted with those that obtain in the state of lessened sensi- tiveness in allergy, the differences being such as to indicate a difference in the mechanism of these two phenomena. 2. The phenomenon of “local exhaustion” of the allergic cutaneous reaction described by Mackenzie and Baldwin is studied and found, in disagreement with these authors, to be non-specific. 3. It is proposed to distinguish the lessened sensitiveness induced in allergy from the state of desensitization in anaphy- laxis by designating the former condition as a state of hyposensitization. 242 ROBERT A. COOKE REFERENCES (1) Coca, A. F.: Hypersensitiveness. Tice’s Practice of Medicine. W. F. Prior Company, Hagerstown, Md., vol. 1, p. 107. (2) Coca, A. F.: Jour. Immunol., 1920,'5, 363. (3) Cooxs, R. A., FLoop, anp Coca, A. F.: Jour., Immunol., 1917, 2, 217. (4) Auexanper, H.L.: Arch. Int. Med., 1917, 20, 636. (5) Macxenziz, G. M.: Jour. Amer. Med. Assoc., 1921, p. 1563. (6) Coca, A. F., anpD Kosaxat, M.: Jour. Immunol., 1920, 5, 297. (7) Macxenzin, G. M., anp Batpwiy, L. B.: Proc. Soc. f. Exp. Biol. and Med., 1921, 18, 214. (8) Macxenziz, G. M., anp Batpwin, L. B.: Arch. Int, Med., 1921, 722. (9) SoumtMAN AND PitcueR.: Jour. Pharm. and Exp. Ther., 1917, 9, 309. (10) Sotuman, T.: Jour. Pharm. and Exp. Ther., 1917-18, 10, 147. IMMUNOLOGICAL STUDIES ON TYPES OF DIPH-. THERIA BACILLI! I. AGGLUTINATION CHARACTERISTICS II. PROTECTIVE VALUE OF THE STANDARD MONOVALENT ANTITOXIN WILLIAM H. PARK, ANNA W. WILLIAMS anp ALICE G. MANN Received for publication December 17, 1921 The subject of this communication would have been exhaus- tively investigated long ago if there had been practical evidence that a monovalent diphtheria antitoxin did not protect from diphtheria. It is true that in regard to B. tetant and B. botu- linus laboratory investigations have given contradictory results as to the efficacy of a monovalent serum. These investigations have shown us: first, that tetanus bacilli, although they are divided into several types by their response to agglutinins, nevertheless, all produce toxins which are neutralized by a mono- valent antitoxin; second, that the reverse is true of the bacilli causing botulism, in so far as the production of toxin is concerned. Among the strains of B. botulinus there are at least two different types producing toxins which respond only to homologous antitoxins. As to diphtheria bacilli, however, there are two series of ob- servations which have been made concerning the degree of pro- 1Read at the meeting of the American Association of Immunologists, March, 1921. This article combines the results of portions of two investigations. The first portion was a part of our studies on the respiratory infections and was designed to show the relative values of the absorption-agglutination method and other serological tests in determining epidemic strains. The second por- tion was a part of a series of studies on the similarity of the toxins produced by the diphtheria bacilli and the adequacy of a monovalent antitoxin. In the studies on the absorption-agglutination characteristics of the diph- theria bacilli, we were given valuable aid by technicians supplied by the Influenza Commission of the Metropolitan Life Insurance Company. 243 THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 3 244 W. H. PARK, A. W. WILLIAMS AND A. G. MANN tection afforded by monovalent diphtheria antitoxin. These observations have seemed to give all-sufficient evidence that the toxins produced by the different strains of the diphtheria bacilli are alike, at least, in their affinity for antitoxin. The first is, that for the last twenty-five years, ever since diphtheria anti- toxin has been used in practice, injections of a monovalent antitoxin in hundreds of thousands of persons known to be in contact with infection, have given practically complete protec- tion for two weeks—the period during which antitoxin is known to remain in appreciable amounts in a human being after injec- tion. The second series of observations is that routine virulence tests throughout the world have with few, if any, exceptions shown that a monovalent diphtheria antitoxin is able to pro- tect animals given a dose of culture fatal in others not given antitoxin. If any doubt has existed in regard to the universal worth of a monovalent diphtheria antitoxin, it is one that has led to ex- perimental investigation from time to time as to whether an antitoxic serum alone was as efficient as a combined antibacterial and antitoxic serum and whether from this standpoint a poly- valent serum might not be more advantageous. But the paper published by Havens (1) in 1920 seemed to show that diphtheria bacilli were divided into two agglutinative groups and that each group produced a toxin that was quantitatively at least, a very different one from that produced by the other group, so that a monovalent serum for group I had no appreciable effect on group II toxin unless the serum was given in enormous amounts. ‘These findings, if correct, are of so much practical importance that reference will be made to them in some detail in connection with our own work. I. AGGLUTINATIVE TYPES OF B. DIPHTHERIA A number of investigators have reported, the obtaining of an agglutinating serum that differentiated between “‘true” and ‘‘nseudo” diphtheria bacilli (Schworer (2), Sander (3), Mason (4)), but no one had described definite agglutinative types among the toxic strains until Durand (5) published his observations in STUDIES ON TYPES OF DIPHTHERIA BACILLI 245 1918. Durand announced that he had obtained four agglutina- tive types. In 1920 he published a further report in which he stated that on examining 252 strains by this test he found that they were divided as follows: Type A contained 16 strains Type B contained 8 strains Type C contained 31 strains Type D contained 76 strains Type E contained 51 strains Unclassified 71 strains These types were strictly specific in that each absorbed only the respective agglutinins from the homologous serum. Early in 1920 Dr. Durand visited our laboratory, demonstrated his methods, presented us with his type strains which he now designated with figures instead of letters and gave us the follow- ing facts in regard to their sugar fermentations: All ferment dextrose and levulose. MALTOSE | DEXTRIN | GLYCEROL ae ee Type I—Americaine (‘‘ourNo.8’’)| + + —_ - - Type? Te—Durand 2.0. 6.242. . 6 - - — = a ype TWAS -Nodeb sh ee... cdke. gene + + + - = Type IV—Benjamin................ + + + + - Type V—Sirbeaux................ + + + + _ He also gave figures to support the following observation: There was less paralysis and attendant mortality in cases of diphtheria treated with diphtheria serum produced by the in- oculation of both toxin and bacilli; i.e., an antitoxic and ‘‘anti- bacterial” serum, than with serum produced by the use of toxin alone. As Durand used all of the type strains in making this bacterial serum he had no evidence as to the effect of a mono- valent antibacterial serum. Our own work on agglutination. We immediately began inocu- lating a series of horses? with Durand’s type strains and with our own strain of ‘‘no. 8,” as well as with several strains from each 2 This part of the work was done by C. R. Tyler and D. W. Poor. 246 W. H. PARK, A. W. WILLIAMS AND A. G. MANN of two apparent epidemics* of diphtheria, one in an institution in New York City and the other in a village near the City. The horses were inoculated intravenously every day for three day periods, with suspensions of cultures prepared as follows: The bacteria were grown for eighteen hours on Loeffler’s coagulated blood serum medium at 37°C. The growths were then scraped into 0.85 per cent salt solution and transferred to tared Purdy tubes. They were sedimented in the centrifuge for twenty minutes and the superna- tant liquid was removed. The weight of moist sediment wasdetermined and the sediment was suspended in 10 cc. salt solution. The required weight of sediment (in suspension) was measured off and made up to 15 cc. for injection. TABLE 1 Direct agglutination of the strains of B. diphtheriae used for torin-antitorin comparison HORSE SERUM TITER STRAINS 3 Z No.8 | Nodet | ,Sit- | Beni | 11 | 3362 Moped. t Our uss... 0 ccc. ce coc 1600} 0 0 0 0 Dype 1.* Americsines.:../osor. 2. 6 oe 1600 0 0 0 0 Pype Lil.; (Nodets ick 6. coe. ae 0 | 1600 0 0 0 0 PypelV.. Benjamin:. 2... 2. o%2.. ict 0 0 0 800 | 800 EGG LW 5 Picts siacccfthe ee atere n Sees one aco 0 0 0 800 | 800 ype V2 ‘Sisbeaux. os. . s)..se2 3. ae 0 0 400 0 0 0 LN BIR Te 1 a Oe EE OF Pe 0 0 0 0 0 400 The first dose was 1 mgm.; each succeeding dose was doubled unless the condition of the horse contra-indicated it. After each three day period, the horse was allowed to rest for seven day. ‘The first three doses of the bacillus emulsion were heated at 55°C. for one hour before inoculation. Later, living cultures were inoculated. From four to six series of inoculations were given each horse. In table 1 are given only the strains which were used by us to test the protective action of our monovalent serum. 8 Reported in Arch. Ped., 1921, 38, 329. STUDIES ON TYPES OF DIPHTHERIA BACILLI 247 With the serums obtained, we corroborated fully the work of Durand showing that there are at least five agglutinative types of diphtheria bacilli. Havens kindly sent us four strains of his group II and we found that none of them agglutinated with our ‘‘no. 8’’ serum. Havens did not mention Durand’s work in his article. He reported the results of his study of 206 strains. These were from acute cases, convalescents and carriers. All varieties of morphologic cultures were tested but he says that ‘‘morphology bears no observed relation to agglutinating properties.” A certain strain was chosen to inoculate rabbits for the production of agglutinating serum. Havens does not give the dosage or the course. He simply states that the injections were made intra- venously in increasing amounts, and that an agglutinating serum was obtained with a titer of 4860 for the homologous strain. Of the series of 206 strains, 169 agglutinated in this serum to the same titer as the homologous strain, while 37 failed to agglu- tinate in any dilution. One strain of this second group was chosen to inoculate a rabbit for the production of agglutinating serum. This serum gave a titer of 4800 or higher. The other 36 strains of this group agglutinated to the same titer. He concludes as follows: Judging by the evidence furnished by the agglutination test there are two biologic groups of the diphtheria bacillus. No evidence of cross agglutination was found among the members of this series [obtained in Iowa City]. All members of the series fell into one of the two groups with no differences observed in the degree of agglutinability. As judged by the results of agglutination, two groups exist which include all strains of the diphtheria bacillus. Havens said that the virulence in animals was approximately the same in the two groups. When we consider the results of Havens in connection with Durand’s and those of our own, it seems fair to infer that in the comparatively small city from which he drew his material, only two types of bacilli were present. 248 W. H. PARK, A. W. WILLIAMS AND A. G. MANN There can be no doubt whatsoever that in all large communi- ties there are many types and at least several dominant ones. II. PROTECTIVE ACTION OF DIPHTHERIA ANTITOXIN (STANDARD NO. 8), AGAINST DIFFERENT AGGLUTINATING TYPES OF B. DIPH'FHERIAE* The question of the protection results is so important that we quote Havens’ experiments in detail. In attempting to show “the protective properties of diphtheria antitoxin with reference to biologic groups,” he reports first the following tests with various cultures and diphtheria antitoxin no. 8. Havens’ tests CULTURE GROUP AMOUNT peed ste ie tee RESULT ee a a ee ee 210 I 1 None Died, 48 hours 210 I 1 100 i Lived 221 1 1 None Died, 36 hours 221 II 1 100 I Died, 36 hours 141 II 1 None Died, 48 hours 141 II 1 100 I Died, 48 hours The results of these tests of Havens indicate that group I antitoxin has no neutralizing effect upon the toxins produced in the animals by group II cultures. Similar results followed when he injected cultures intracutaneously. Six strains of group I bacilli were inoculated into guinea pigs that had been injected with 100 units of group I antitoxin. No cutaneous lesions were produced. Eighteen strains of group II bacilli were inoculated into guinea pigs that had received 100 units of group I antitoxin. The reactions in 15 were identical with the controls, while those in three were slightly less than the controls. Before discussing Havens’ results with toxin, we give in table 2 our findings with cultures of at least five different agglutinating types including three of Havens’ ‘“‘group II” strains. The cultures were heavy eighteen hour growths in veal broth. 4 We were assisted in this part of the work by Jane L. Berry, Harriet L. Wilcox and Charles Greenwald. TABLE 2 Life-saving effects of type I antitoxin in animals injected with bacilli of various types ee ic GG asrrtoxi nesuut grams ce. 248 2.0 None | Died 3 days 230 3.0 None | Died 2 days 355 3.0 None | Died 2 days page wae no, 82 Ney 230 3.0 0.2 | Died 2 days y 230 3.0 0.5 | Died 2 days 360 3.0 1.0 Remained well 355 3.0 2.0 Remained well Type III. (Durand’s Nodet, 225 0.2 None | Died 2 days. France) 225 0.2 0.5 Remained well 235 0.2 None | Died 9 days Type V. (Durand’s Sir- 400 0.75 None | Died 4 days beaux, France) 235 0.2 0.2 Died 8 days 405 0.75 1.5 Remained well Type IV. (No. 11, New tat 242 2.0 None | Died 3 days City) [| 245 2.0 0.5 Remained well 240 0.05 None | Died 2 days mus eh Ae ree etBenya-)}") “955 0.1 0.2 | Died 6 days ; 235 0.05 0.05 | Remained well No. 3362, unclassified (New oie vie i ae 4 nal 2 York City) 0.2 0. ie ays 255 0.2 0.5 Remained well B, unclassified (Near New 235 0.1 None | Died 6 days York City) 245 0.1 0.2 Remained well 237 1.0 None | Died 3 days No. 151 (Havens’ group II, 262 1.0 None | Died 2 days Iowa City) 255 1.0 0.2 Remained well 245 1.0 0.5 Remained well 225 0.2 None | Died 2 days No. 184 (Havens’ group II, 232 0.2 None | Died 2 days Iowa City) 245 0.2 0.2 Died 6 days 230 0.2 0.5 Remained well 217 0.1 None | Died 10 days 440 0.3 None | Died 2 days No. 8847 (Havens’ group II, 235 0.3 0.5 Died 3 days Iowa City) 225 0.3 1.0 Died 3 days 215 0.3 5.0 Remained well 255 0.3 5.0 Remained well 249 250 W. H. PARK, A. W. WILLIAMS AND A. G. MANN The results given in the table do not show any striking dif- ference in the protecting value of the type I antitoxin, when used in animals inoculated with a type I strain or with strains of other types. All who have worked extensively with living cul- tures have discovered that the results are much more irregular than when toxins are used. The cultures vary from day to day, according to the culture medium used, in the abundance of their growth and in the vigor and apparent toxicity of the bacilli; a much larger number of units is required to protect against a fatal dose of living bacilli than against a fatal dose of toxin. There is nothing in our results to indicate to us that there is any marked difference in the pathogenic power of the different types of bacilli. The last three cultures are of unusual interest since they were sent us by Havens as examples of his Group II strains. All produced toxin of moderate strength when grown in bouillon. It is to be noted that our antitoxin produced by type I no. 8 bacillus was capable of protecting the animal which had received a subcutaneous injection of an amount of culture which produced death in the control animal within two days. Indeed, the type I antitoxin appears to be as efficient against two of the cultures sent by Havens as against strain ‘‘no. 8.” Although with the third culture a considerably larger quantity of antitoxin was required, it is noted below in table 3 that the toxin produced by this culture was neutralized by the usual amount of antitoxin. The experiments of Havens showing the neutralizing results of mixtures of his group I antitoxin and group II toxin are given in the following tabulation: Havens’ tests STRAIN MINIMAL oneuE NUMBER TOXIN GROUP LETHAL DOSES aga RESULT 53 II 2 10 Died, 72 hours 53 II 2 _ Died, 72 hours 74 101 2 5 Died, 48 hours 74 II 2 20 Lived, marked virulence 74 II 2 25 Lived, oedema 74 II 2 50 Lived, no lesions 34 I 2 5 Remained well 34 I 2 Died in 48 hours STUDIES ON TYPES OF DIPHTHERIA BACILLI 251 The group II toxin required far larger amounts of group I anti- toxin to neutralize it than did the group I toxin. Neutralizing experiments with mixtures of group II antitoxin and group I toxin. STRAIN MINIMAL SOUR TOXIN GROUP ANTITOXIN RESULT NUMBER LETHAL DOSH UNITS 34 I 1 5 Died, 96 hours 34 I 1 Died, 96 hours As seen in the above tabulation, one minimal lethal dose of group I toxin was not neutralized by five units of group II antitoxin. The fact that in Havens’ tests very large amounts of group I antitoxin usually neutralized 2 minimal lethal doses of group II toxin, indicates that the group I antitoxin contained the same TABLE 3 Toxin results TOXIN NUMBER OF pobre Me cae Ale de From culture Type DOSES a grams No. 8 i 250 1.0 0.02 Lived Durand II 230-260 1.0 0.02 Lived Nodet III 230-260 1.5 0.02 Lived Benjamin IV 230-260 2.0 0.02 Lived No. 11 IV 230-260 1.0 0.02 Lived 3362 Unclassified 230-260 2.0 0.02 Lived 151 Havens, group II 230-260 1.0 0.02 Lived 8347 Havens, group [I 230-260 1.5 0.1 Lived 8347 Havens, group II 230-260 | 100.0 10.00 Lived Rr a ee ee neutralizing substances for group II toxin as for group I toxin but that there were quantitative differences in the proportion of the different antitoxic substances. Havens stated ‘‘the results may throw some light on those cases of diphtheria which are not benefited except by large amounts of antitoxin.”” He alluded also to the death rate re- 252 W. H. PARK, A. W. WILLIAMS AND A. G. MANN maining at 10 per cent in cases of diphtheria. He suggested adding his group II toxin to horse injections, to the Schick test toxin and also to toxin-antitoxin mixtures. Table 3 gives the results which we obtained with neutralizing mixtures of our own antitoxin no. 8 and the types of toxin indicated. It is evident from this table that with the strains obtained from New York, from France and from Jowa no such difference has been discovered, as Havens believes he found. It would take repeated experiments and many animals to prove whether or not there is any minor difference either quantitatively or qualitatively between these toxins produced by the different agglutinative types. CONCLUSIONS The group of diphtheria bacilli contain strains belonging to several agglutinative types. The toxins formed by these different types are, however, quali- tatively alike and, from the practical standpoint, quantitatively so. Whether or not there are slight quantitative differences, further studies will be necessary to show. Strong toxin from any diphtheria bacillus strain is suitable for the Schick test and for immunization of man or animal. A monovalent antitoxic serum is suitable for protective and curative measures against all diphtheria strains. REFERENCES (1) Havens, L. C.: Jour. Inf. Dis., 1920, 26, 388. (2) ScowoneR, J.: Wien. Klin. Woch., 1902, 15, 1274. (8) Laneer, H. Centralbl. f. Bakt., 1916, Orig., 78, 117. (4) Mason, E. H.: Mil. Surgeon, 1919, 45, 560. (5) Duranp, PauL: Compt. Rendus Soc. Biol., 1918, 81, 1011 and 1920, 83, 611 and 613. THE RELATIONSHIP OF LIPOIDS AND PROTEINS TO SERUM REACTIONS IN TUBERCULOSIS W. RAY HODGE anv M. F. MACLENNAN From the Connaught Antitoxin Laboratories, University of Toronto, Canada Received for publication January 3, 1922 While working with sera of tuberculous patients! it became necessary to note the effect of (a) the removal of lipoids and (b) the precipitation of globulins on certain serum reactions. The results obtained seemed sufficiently clear cut for publication in the present paper. PART A. LIPOIDS IN TUBERCULO SERUM REACTIONS The relationship of lipoids to immunity reactions has been a subject of great interest in recent years. Krumwiede and Noble (1) have recently been unable to confirm the claim made by Stuber that agglutinins are lipoidal in nature. Levaditi and Yamanouchi (2) have claimed that the substances in syphi- litic sera responsible for the Wassermann reaction, can be ex- tracted with alcohol. Noguchi (3) was unable to confirm this observation; he concluded that the reacting substance in syphi- litic serum was non-lipoidal in character. Kolmer and Pearce (4) found that ether and chloroform narcosis tended to diminish the concentration of non-specific fixation bodies in normal, in- activated dog and rabbit serum. Observations on the serum of humans, withdrawn during and immediately following ether narcosis, show that there may be a profound alteration of the 1This work was carried out through the codperation of facilities at the Research Division of the Connaught Antitoxin Laboratories, the Department of Soldiers’ Civil Reéstablishment Chest Clinic, and the Department of Medicine, University of Toronto. It forms part of the investigation of clinical and labora- tory data in tuberculosis which Doctor Caulfeild and his associates are con- ducting, the results and conclusions of which will be published later. 253 254 W. RAY HODGE AND M. F. MacLENNAN Wassermann reaction under these conditions. It seems likely, then, that alterations in the lipoidal content of serum must have an important bearing on this reaction. Kolmer (5) further showed that extraction of normal rabbit and dog serum with ether tends to remove a large portion of the serum constituents responsible for non-specific fixation with lipoidal and bacterial antigens; he also found that feeding lecithin to these animals was followed by increased power of the inactivated serum to adsorb complement. Noguchi (6) has shown that only sera containing lecithin, fatty acids or soaps are capable of activating cobra venom (rendering it hemolytic). Venom hemolysis indi- cates the possible important relationship of lipoids to hemolytic complement, venom containing the hemolytic amboceptor and the complement probably being derived from corpuscular lecithin (Kolmer). Calmette, Massol and Breton (7), studying the ac- tivating power of different sera on cobra venom, found that inac- tivated sera of tuberculous patients very frequently showed this characteristic. Calmette (8) attempted to use this reaction in the diagnosis of tuberculosis. In 77 sera from tuberculous patients the results obtained were as follows: turban 1, reaction + in 76 per cent; turban 2, reaction + in 57 per cent; turban 3, reaction + in 70 per cent. In 26 normal sera: reaction + in 37 per cent. He also observed this reaction in the serum of patients suffering from syphilis, in cerebrospinal meningitis, in Addison’s disease, in general paralysis of the insane, and in severe maladies associated with more or less profound affection of the nerve cells or suprarenal capsules. Free lecithin was also found to occur in the serum of anaesthetized subjects. Finally the inactivated serum of certain animals is normally active; the horse, dog, rat, goat and rabbit are in this class. Kolmer (5) has drawn attention to the frequency with which non-specific complement fixation bodies are present in normal inactivated dog and rabbit serum and he has shown that extraction with ether tends to remove these bodies. Calmette’s observations show that the serum of these animals normally contains large amounts of lecithin, fatty acids or soaps; this goes to confirm SERUM REACTIONS IN TUBERCULOSIS 255 Kolmer’s opinion that the substances responsible for the non- specific fixation obtained with these sera are lipoidal in character. It is evident from even such an incomplete review of the literature that lipoids have a not inconsiderable influence on the Wassermann reaction (although the weight of opinion considers the reacting substances non-lipoidal in nature), that the lipoidal content of certain animal sera favors conditions for the develop- ment of non-specific fixation and finally that the lipoidal content of most tuberculous sera is higher than normal. Our own work has been an attempt to show in what way the lipoidal content of tuberculous sera is related to the complement fixation reaction in tuberculosis and to Caulfeild’s (9) inhibitive reaction. Methods The object of this part of the work was to extract the lipoids from the serum as completely as possible and to note the effect of this removal on the reactions under examination. In the earlier work an attempt was made to extract the lipoids by thoroughly mixing the serum with petroleum ether (4 volumes of petroleum ether to 1 volume of serum). After extraction the serum was separated by centrifugalization. The results with this method, however, proved quite indeterminate in a series of 90 sera tested. It was thought that the method employed might remove the lipoids only very incompletely, accordingly Friedemann and Herzfeld’s method (10), slightly modified, was used. Five-tenths cubic centimeters of serum were spread on a filter paper and dried in an electric oven at 45°C. for one hour. 1 volume dried in this way was reserved for a control; to the other volume in a test tube was added a mixture of equal parts of absolute alcohol, ether and chloroform. The tubes were shaken half an hour after which the mixture was poured off and the filter paper dried at 37°C. Saline was then added to the dried filter paper to make a dilution 1:5; the filter paper was worked into a pulp, the liquid pressed out, poured into another tube and centrifugalized to re- move the particles of filter paper. The control portion of the serum was treated in the same way except that the extraction with alcohol, chloroform and ether was omitted. 256 W. RAY HODGE AND M. F. MAacLENNAN 1. Effect of extraction of lipoids on the complement fixation reaction Table 1 contains the results obtained, when fifteen sera from tuberculous patients, showing varying degrees of fixation,? were were extracted in this way. Comment. ‘The results in the control untreated fraction (dilu- tion 1:10) and in the control dried fraction (dilution 1:5) were quite comparable, there being only a moderate loss of fixation bodies in the latter. The dried and extracted sera, however, in TABLE 1 SERUM DRIED AND REDISSOLVED SERUM DRIED, EXTRACTED AND UNTREATED SERUM, 1:10 IN SALINE, 1:5 REDISSOLVED IN SALINE, 1:5 *3-1-0-0 2-1-0-+ C.H. 3-+-0-0 3-2-+-1 1-0-0-+ 4-2-+-0 4-3-2-2 2-1-0-1 4-3-2-0 4-3-1-2 1-0-0-1 4-3-+-0 4-3-0-0 C.H. 4-3-1-0 4-4-3-1 C.H. 4-2-0-0 44-2-2 C. Ho. 4-2-+-0 4-3-1-1 C.H. 44-3-0 44-2-0 Ga 4-4-3-0 44-3-+ CoH. 44-3-0 44-3-0 C.H. 444-0) 444-2 ©. H. 4-3-2-0 4-44) 2-1-0- 3-+-0-0 3-2-+-0 2-+-0-2 4-2-0-0 4-1-0-0 C2. *4 = no hemolysis; 3 = 25 per cent hemolysis; 2 = 50 per cent hemolysis» 1 = 75 per cent hemolysis; + = almost complete hemolysis; 0 and C. H. = com- plete hemolysis. First three tubes contain 0.1 cc. of serum and antigen, and two» two and a half and three units of complement. The fourth tube is a control containing 0.2 cc. of serum and 2 units of complement. almost every instance, showed a marked loss of power of fixation, In many instances the loss of power of fixation was complete. It is evident then, that extraction in this manner either destroys or removes* substances wholly or partially responsible for the 2 Petroff’s whole bacillus antigen was used throughout for the complement fixation tests. 3 An attempt was made to test the alcohol, chloroform, ether extractives for complement fixation bodies without success on account of the high anti-comple- mentary power of this fraction. SERUM REACTIONS IN TUBERCULOSIS 257. development of the complement fixation reaction in sera of tuberculous patients. It was thought that comparable results might be obtained by extraction of sera of animals immunized with the tubercle bacillus. The results of extraction of such fixation guinea-pig and rabbit sera, however, failed to demonstrate a definite loss of power of fixation in these instances. Assuming that the fixation bodies are of similar chemical constitution in human, guinea-pig and rabbit sera, these results are puzzling. The fact that the sera of the artificially immunized animals showed a higher concentration of fixation bodies than the human tuberculous sera may be a possible explanation, as here the extraction of lipoids may be incomplete. It is true that the method employed does not completely free sera of lipoids. This has been demonstrated by Suranyi (11) who followed the technic of Friedemann and Herzfeld closely and then extracted the supposedly lipoid-free serum in a soxhlet apparatus. He found that frequently as much as 25 per cent of the serum lipoids remained after extrac- tion by the method of Friedemann and Herzfeld. 2. Effect of the extraction of lipoids on the inhibitivet reaction Thirty inhibitive sera were extracted by the method described. There was no evidence to show that extraction removed the ac- tive substances. It was found however, that there was a quite marked loss of inhibitive power in the sera that had been dried and redissolved without extraction. The extracted sera showed a similar, but not greater, loss of inhibitive power. It is evident 4The technic of this reaction has been fully described by Caulfeild (9). Briefly, it is found that certain human sera, combined with anti-complimentary doses of antigen (AE extract of the tubercle bacillus) and two and a half units of complement, will free the complement from its non-specific adsorption by the antigen. In our tests four tubes are used, the first three tubes containing varying doses of antigen, two and a half units of complement and 0.1 cc. of inactivated serum from which the natural amboceptor has been previously removed by extrac- tion with sheep cells, the fourth tube is the ordinary serum anti-complementary control. A very strong inhibitive serum will show hemolysis in all tubes. A strong inhibitive serum will give a 4-0-0 reading; a weak inhibitive serum will give a 44-0 reading and a serum containing no inhibitive substance will give a 4-4-4 reading. 258 W. RAY HODGE AND M. F. MacLENNAN then that the components of the dried sera were only very in- completely redissolved in saline, the inhibitive fraction remain- ing largely insoluble on the filter paper. Caulfeild (9) observed a similar loss of inhibitive power when antigen-serum mixtures were dried on filter paper and redissolved. The researches of Jobling and Petersen (12) in connection with the increased antitryptic power of guinea-pig serum following anaphylactic shock, suggested an indirect method of determining the effect of lipoids on the inhibitive reaction. These authors believe that the increased antitryptic power of guinea-pig serum immediately following anaphylactic shock, especially if pro- tracted, is due to an increased concentration of the unsaturated lipoids in the serum. The increased antitryptic power is ex- plained by the liberation of lipoids following cellular destruction. Rusjnjak (13) has shown that the antitryptic power of serum following anaphylactic shock is frequently raised as much as 100 per cent; the optimal conditions for the development of this increase are obtained when an incubation period of from fifteen to thirty minutes intervenes before shock symptoms appear. 3. Effect of anaphylactic shock on the inhibitive reaction In the following experiments anaphylactic shock of varying severity was induced in guinea-pigs and the inhibitive power of the serum! before and after shock was determined. It is not possible to give the protocols of these experiments in detail. The most important single fact is the length of the latent period before the onset of shock symptoms. In table 2 the results are shown of the study of these sera before and after anaphylactic shock; the sera from animals showing a latent period of fifteen minutes or more before the onset of shock symptoms are grouped together. It is here that one would ex- pect to find the inhibitive power increased if it is due to un- saturated serum lipoids. * Guinea-pig serum was ideal to work with here because normal guinea pig serum often gives a weak inhibitive reaction, hence we know that conditions are favorable for the development of the reaction and for the perception of varying amounts of inhibitive substance. SERUM REACTIONS IN TUBERCULOSIS 259 Comment. Sera 1 and 4 show an increased inhibitive power following anaphylactic shock, but this occurs nowhere else in the series. Sera 9 to 13 should develop ideal conditions for in- creased antitryptic power in the serum as the latent period was fifteen minutes or over in each of these animals ; yet these sera show no definite increase in inhibitive power. In this series of TABLE 2 INHIBITIVE REACTION BEFORE SHOCK INHIBITIVE REACTION FOLLOW- ING SHOCK NUMBER REMARKS Latent period less than fifteen minutes eee ee eee eee Me 444-0 4-4-4-0 Moderate shock 3 4-4-4-0 4-4-4-0 Moderate shock 5 4-4-4-0 4-4-4-0 Sudden, fatal shock 6 4-4-4-0 4-4-4-0 Mild shock ai 4-4-4-0 4-4-4-0 Moderate shock 8 4-4-4-0 4-4-4-0 Severe shock Latent period fifteen minutes or more 1 4-4-4-0 4-2-0-0 Very mild shock 4 4-4-4-0 4-4-0-0 Mild shock 9 4-4-4-0 4-4-4-0 Mild shock 10 4-4-4-0 4-4-3-0 Mild shock 11 4-4-4-0 4-4-3-0 Mild shock 12 4-4-4-0) 4-4-4-() Mild shock 13 4-4-4-0) 4-4-4-0 Mild shock 14 4-44-09 4-4-4-0 Moderate shock 15 44-4-0 4-4-4-0 Moderate shock 16 4-44-09 4-4-4-0 Moderate shock 17 4-4-4-0 4-4-4-0 Moderate shock 18 4-4-4-9 4-4-4-0) Moderate shock 19 4-4-4-0 4-4-4-0 Moderate shock a ee ee animals anaphylactic shock of all degrees was encountered, from the severe and immediately fatal type to the almost impercep- tible type. The animals which showed the increased inhibitive power were only very mildly shocked. As this increase occurred definitely in only two cases in twenty it seems very unlikely that it can be associated with the increased lipoidal content which is supposed to occur so frequently under these conditions. THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 3 260 W. RAY HODGE AND M. F. MacLENNAN PART B. PROTEINS IN TUBERCULO SERUM REACTIONS The réle of proteins in certain immunity reactions is well recognized, in others less definitely so. Huntoon (14) and others, in an extensive study of the chemical nature of agglutinins, bactericidal and protective substances in immune sera, concluded that such antibodies were colloidal in nature; they were not affected by trypsin over considerable periods, were not in the euglobulin or pseudoglobulin fractions and were not soluble in ether. The relationship of proteins to complement fixation has not been definitely determined. Noguchi (3) found that sera of untreated syphilitics almost universally showed an increased globulin content; there was, however, no complete parallelism between this increased globulin content and the positive Wasser- mann reaction. Rowe (15) found an increased globulin content in sera in all infections with the exception of acute tonsillitis, typhoid fever and chronic bronchitis. He also found that the Wassermann reaction is not dependent on a quantitative in- crease of serum globulins. Kapsenberg (16) has recently shown that the fixation bodies responsible for the development of the Wassermann reaction are completely removed when the globulin fraction is precipitated with ammonium sulphate and that they can be demonstrated in the globulin fraction. Nishida and Petroff (17) in a study of fixation bodies in tuberculous sera concluded that these substances are either globulins or that they are adsorbed by the globulin fraction of the serum when it is precipitated by ammonium sulphate. Calmette (18) working with the serum of cows immunized with the tubercle bacillus, precipitated the euglobulin fraction by bubbling carbon dioxide through the diluted serum and found that the fixation bodies remained in the supernatant fluid. In the present work euglobulin was precipitated by bubbling carbon dioxide through serum’ diluted 1:10 with distilled water till a heavy turbidity developed. The sera were allowed to stand 6 Sera were inactivated at 56°C. for half an hour and when the inhibitive reac- tion was being investigated all sera were previously extracted with sheep cells to remove the natural amboceptor. SERUM REACTIONS IN TUBERCULOSIS 261 for two hours till a flocculent precipitate was formed; they were then centrifugalized. The supernatant fluid was poured off and rendered isotonic. The precipitate was redissolved in saline to give the original 1:10 dilution. 1. Effect of precipitation of euglobulin on the complement fixation reaction in tuberculosis The results of the precipitation of euglobulin in human sera showing various degrees of fixation and in the sera from a rabbit and a guinea pig which had previously been immunized with the tubercle bacillus, are shown in table 3. TABLE 3 TREATED SERUM SERUM UNTREATED SERUMI koe Sn ee ee Euglobulin fraction} Supernatant fluid Control tube 2 Control tube 2 Control tube 2 Complement units: core eee eee es: 2-3-4-5-6-7 2-3-4-5-6-7 2-3-4-5-6-7 Enrram nos VO es re et 4-3-2-0-0-0-0 3-0-0-0-0-0-0 4-4-3-0-0-0-0 INNADEN OAS. : . dertisanterertle «5 «6s 4-4-4-4-4-4-0 4-2-0-0-0-0-0 4-44-444-0 immunized rabbits se... oe 4-4-4-4-4-4-0 | 2-0-0-0-0-0-0 | 4444444-0 Immunized guinea-pig.......... 4-4-4-4-4-4-0 | 4-3-0-0-0-0-0 | 444444-0 Comment. The results are definite and uniform. ‘The fixa- tion bodies in the supernatant fluid are in practically the same concentration as in the untreated serum. There is an occasional and slight amount of fixation in the euglobulin fraction. As the control tubes of the supernatant fluid hemolyzed much more rapidly than those of the euglobulin fraction, the slight and irregular fixation in this fraction is likely due to the antilytic power of the euglobulin plus the antilytic power of the antigen. Renaux (19) separated the euglobulin from syphilitic sera using this method. He found that the fixation bodies remained in the supernatant fluid and appeared only irregularly in the euglobulin precipitate. Renaux in addition made a very interest- ing observation. He separated the diluted serum into two por- tions to one of which he added a small amount of lipoid emul- sion (cholesterinized heart extract); the portions of diluted serum 262 W. RAY HODGE AND M. F. MacLENNAN were now treated with carbon dioxide in the ordinary way. Tests with the supernatant fluid and euglobulin for fixation bodies now showed that in the case of the serum previously treated with lipoids the fixation bodies had been completely transferred to the euglobulin sediment; in the serum free from lipoids the fixation bodies of course remained in the supernatant fluid. We were able to repeat Renaux’s experiment with the same result; a sample protocol is shown in table 4. Comment. It is seen, that, by the addition of lipoidal emulsion (Wassermann antigen) to the serum before the precipitation of euglobulin, the fixation bodies are transferred to the euglobulin sediment. There seem to be two possible explanations of this phenomenon: (a) The lipoids added are either adsorbed by the TABLE 4 EUGLOBULIN wees Without tipoid jemulstons 22720 } Pee eee «o's *0-0-+-1-2-0 0-244-+4-0 Lipoid emulsion added before precipitation of euglobulim®.. .. (2-tcGck ... tet: Peete. 4). 0-244444 0-0-0-0-0-0 * The first five tubes contain antigen, 3 units of complement and varying amounts of 1,10 serum; i.e., 0.05, 0.1, 0.8, 0.5 and 1.0 cc. The sixth tube is a control containing 2.0 cc. of serum and 3 units of complement. euglobulin sediment or actually combine with it to form a lipo- protein, or (b) The increased concentration of lipoids following the addition of antigen to the diluted serum may result in the precipitation of the fixation bodies with the euglobulin fraction. If explanation (a) is correct, then the lipoids in a state of emul- sion are simply adsorbed by the flocculating euglobulin and the fixation body is carried down in combination with its antigen. If explanation (b) is correct, then the fixation body is primarily precipitated and the antigen is pulled down in combination with the fixation body. Assuming explanation (b) to be possible it is at once apparent that a similar phenomenon might be obtained by the addition of lipoidal emulsion to sera containing tuberculo- complement fixation bodies. Accordingly human, rabbit and guinea-pig fixation sera were tried out in this way. It was found that the addition of lipoidal emulsion (Wassermann anti- SERUM REACTIONS IN TUBERCULOSIS 263 gen and AE extract’ of the tubercle bacillus) to these sera, before treatment with carbon dioxide, has practically no effect on the supernatant fluid. Either, then, we are dealing with a different type of fixation body in tuberculous sera which is not precipi- tated, or the fixation bodies in syphilitic sera simply combine with the antigen immediately and are precipitated after the combination has occurred. 2. Effect of precipitation of euglobulin on the inhibitive reaction A further explanation of the inhibitive reaction seems essential at this point. Caulfeild described his inhibitive reaction as present in human sera in various degrees, but a strong inhibitive reaction (4-0-0 or stronger) was considered pathological and in cases of tuberculosis was associated with a favourable prognosis. Simultaneously Calmette described an inhibiteur reaction in the serum of animals hyperimmunized to the tubercle bacillus. The method of demonstration of Calmette’s reaction (20) differs from the method described by Caulfeild. It will be recalled that Caulfeild used as his antigen an alcohol-ether extract of the tubercle bacillus (AE extract). In this reaction four tubes are used for each serum, the first three containing varying doses of antigen and the fourth being the customary serum control for anti-complementary power. Of the tubes containing antigen, the third contains the smallest dose which alone just completely adsorbs two and one-half units of complement, the second contains an amount which adsorbs five units of complement and the first contains sufficient antigen to alone adsorb ten units of complement. The first three tubes each receive 0.1 cc. of serum (previously treated with sheep cells to remove natural ambocep- tor) and the fourth 0.2 cc. of serum. Each tube receives two and one-half units of complement. The periods of incubation are the same as those used for the tuberculo-complement fixation test. When 0.1 cc. of human serum reverses the antilytic power 7 Fixation with this antigen is only very rarely obtained with sera from tuber- culous sources. The sera employed in the above experiment all gave negative fixation reactions with this antigen; thus the factor of the combination of the antigen and fixation body is eliminated in this experiment. 264 W. RAY HODGE AND M. F. MacLENNAN of the smallest dose of antigen so that complete hemolysis results in the third tube (4-4-0) this is considered a normal inhibitive reaction. When, however, the inhibitive power is strong enough to produce complete hemolysis in the second and third tubes (4-0-0) or partial hemolysis in the first tube (2-0-0), this is con- sidered a pathological reaction. Calmette used as his antigen an aqueous extract of the tubercle bacillus (B;); as this antigen is not anticomplementary in as low a dilution as 1:5 it is at once apparent that it could not be used by Caulfeild’s method. Cal- mette demonstrated his reaction by combining sera of hyper- immune animals with his B, antigen. The inhibiteur serum and antigen were given a preliminary incubation of half an hour and then complement and the fixation serum were added and the complement fixation test proceeded with as usual. He found that in such a combination the fixation was completely masked.® We have recently been able to demonstrate a similar reaction® in human sera. Many sera of tuberculous humans which give a strong positive fixation with Petroff’s antigen will also give a positive or strong positive fixation with Calmette’s B, antigen. When these sera are combined with certain human sera the fixation is found to be inhibited. Sera giving a strong inhibitive reaction with Caulfeild’s method almost always give a strong inhibiteur reaction with Calmette’s method also. The evidence thus far, then, points to the identity of these two reactions. Calmette separated the euglobulin from his hyper-immune sera and found that the inhibitive substance was contained in this fraction. We precipitated euglobulin from human sera, from the serum of normal rabbits and guinea-pigs and from the serum of rabbits and guinea pigs immunized with the tubercle bacillus. The results are shown in table 5. Comment. It is evident from this table that the inhibitive substance in sera is almost completely precipitated in or with the 8 Caulfeild (21) in his earlier work encountered an occasional serum which gave a positive fixation with an alcohol ether extract of the tubercle bacillus. Com- binations of such sera with inhibitive sera in the presence of AE extract antigen showed the inhibitive sera were, when strong enough, able to mask the fixation. 9 These results will be published in a separate paper. SERUM REACTIONS IN TUBERCULOSIS 265 euglobulin fraction. This is true not only for the inhibitive substance present in pathological sera but also for that present in normal human and normal guinea-pig serum. The pathologi- TABLE 5 ARTO S Wns heer Sha ah a a ee ee TREATED SERUM ~ UNTREATED adie SERUM Euglobulin Supernatant fraction fluid Se ee ee a 44-0-0 4-4-0-0 4-4-0-0 ivormalsnumeanes soccer ee ES 4-4-0-0 4-4-0-0 4-4-4-0 4-3-0-0 4-3-0-0 4-4-4.0 4-2-0-0 4-0-0-0 4-4-9-0 4-2-0-0 3-0-0-0 4-4.3-0 4-2-0-0 4-0-0-0 4-4-0-0 4-0-0-0 2-0-0-0 444.0 4-1-0-0 1-0-0-0 4-4-0-0 4-1-0-0 4-0-0-0 4-4-1-0 4-4-4.0 4-440 4-4-4-0 4-4-4.0 4-4-4. 4-4-4. Mormalarabebrtest oi cet. 13 Lo boc ore 4-4-4.0 444-0 4-4-4-0 444-0 4-4-4-0 4-4-4. 444-0 4-4-4-0 4-4-4-0 4-4-4-0 4-3-0-0 4-44-09 4-4-4-0 4-4-0-0 4-4-4-0 I foadeadiisti 4-4-4-0 4-4-2-0 4-4-4-0 PUMUBIZEGET ADDI, oe es ERE, Peres cnet fw ey 4-3-0-0 4-4-4-0 4-4-4-0 4-4-0-0 4-44-09 4-4-4-0 4-3-0-0 444-0 Normal pumes pig fens nk 4-4-1-0 4-4-0-0 444-0 Normal guinea pig 3...../) 0. 2"... 4-4-2-0 4-4-0-0 444-0 Normal guinea pig 5.......4:............ 4-4-1-0 4-4-0-0 444-0 Normal guineas Dig 6 n Godin 4 24 : 4 40 INC EbICsaiise eect ctianteeerSehela= oie = 0 24 * The upper figure refers to the percentage of phagocyting leucocytes. + The lower figure refers to the percentage of leucocytes containing 10 or more cocci. The question may arise in the reader’s mind as to why leucocytes treated with lactic acid pH 4 were injured so that phagocytosis was com- pletely inhibited in experiment 9, but only partially inhibited in experi- ment 10. The reason is that in experiment 10, a greater number of leucocytes were exposed to a given quantity of acid than in experi- ment 9. The leucocytes washed with solutions in which lactic and acetic acids had been neutralized with sodium hydroxide, showed no injury. 292 ALICE C. EVANS This observation was confirmed in another similar experiment for which the data are not given. The effect on leucocytes of the salts of organic acids is again considered in experiment 14. Experiment 11. Experiments 9 and 10 did not demonstrate the relative toxicity of hydrochloric and citric acids, or indeed, whether they are toxic at all. Experiment 11 was planned to compare the effects of these two acids. The protocol is presented in table 8. The leuco- cytes were washed as in the previous experiments, and then suspended TABLE 8 Ciiric acid is shown to be slightly (if any) more toxic than hydrochloric acid of similar H-ion concentrations HYDROCHLORIC ACID CITRIC ACID Leucocytes washed with saline Leucocytes washed with saline solution of pH solution of pH pH of antigen pH of antigen 7.0 5.6 4.6 4.0 7.0 5.6 4.6 4.0 70 60* | 60 60 72 70 56 60 72 68 is S27) 32 | 36. | 44 44 | 32 | 48 | 44 58 68 64 64 60 58 48 72 (2 40 ; 32 48 48 48 ; 28 36 44 20 53 52 60 68 68 46 64 60 28 12 ; 36 40 44 48 : 36 32 16 4 A8 68 56 48 44 36 8 8 os 36 48 40 32 Sk - 12 12 0 0 * The upper figure refers to the percentage of phagocyting leucocytes. + The lower figure refers to the percentage of leucocytes containing 10 or more cocci. The figures showing inhibition of phagocytosis are indicated by the heavy line. in buffered saline solution of pH 7. But in this experiment further variations were introduced by preparing series of antigens with the use of the same test solutions as were used for washing the leucocytes. Thus the leucocytes were twice exposed to acid solutions. Since the dissociation constants of the two acids differ considerably at the two higher concentrations of H-ions used in these experiments, the buffer action of the bacterial cells reduced the acidity of the hydrochloric acid solutions more than that of the citric acid solutions. Hence the antigen which was prepared with the HCl solution of pH 4 is more comparable TOXICITY OF ACIDS FOR LEUCOCYTES 293 with the antigen prepared with the citric acid solution of pH 4.6 than with the antigen prepared with the corresponding citric acid solution. This fact must be considered in interpreting the results of the experi- ment. The leucocytes subjected to the greatest amount of hydrochloric acid were scarcely injured by it. Those washed in the most highly acid solution and mixed with the most highly acid antigen gave slightly lower figures than the other tests, but the difference is not significant. The figures for all of the other tests of hydrochloric acid solutions are as nearly alike as would be expected in comparative tests if all conditions were the same, signifying that the leucocytes were uninjured by the acid solutions. The leucocytes washed in citric acid solution of pH 4 and mixed with the antigen of pH 5.8 appear to have suffered some injury. Those washed in the citric acid solution of pH 4.6 and mixed with the antigen of pH 4.6 were definitely injured. The results of the experiment indi- cate that citric acid is slightly more toxic to leucocytes than hydro- chloric acid of equivalent H-ion concentrations. Experiment 12. The preceding experiments have not shown that hydrochloric acid is toxic for leucocytes. Experiment 12 was planned to demonstrate the toxicity of hydrochloric acid. The protocol is presented in table 9. The leucocytes were washed as in the preceding experiments, but instead of making the final leucocyte suspension in neutral saline solution, they were made in the same test solutions as were used for the washing. As in experiment 11, a series of antigens was made up in the various test solutions. The pH values of the anti- gens after the modification of the H-ion concentration of the solution by the buffer action of the bacterial cells is given in the table. In this experiment, then, the test acid solutions were of pH values similar to those used in experiment 11, but the leucocytes were sub- jected to them once more than in that experiment. The toxicity of hydrochloric acid for leucocytes is well demonstrated. Since hydro- chloric acid is completely dissociated in solutions of low concentration it can be assumed that the toxicity of this acid is due to its free H-ions. Experiment 13. The toxicity for leucocytes of butyric acid, a common product of pathogenic bacteria, has not yet been considered. Prelimi- nary experiments showed that it was of about the same toxicity as acetic acid, possibly a little more toxic. The protocol for a detailed experi- ment, similar to experiment 11 for showing the relative toxicity of hydrochloric and citric acids, is given in table 10. 294 ALICE C. EVANS Equal quantities of leucocytes were washed with the various test solutions, the test solutions were washed away, and then the leucocytes were suspended in buffered saline solution of pH 7. Antigens were prepared in the same test solutions. The dissociation constants of ace- tic and butyric acids are similar, hence the buffer action of the bacterial suspension had the same effect on the corresponding solutions of the two acids. It increased the pH value of both 5.6 solutions by 0.1. The 4.6 solutions were not perceptibly changed by the addition of the bac- teria. The figures in the corresponding positions in the table are there- TABLE 9 The toxicity of hydrochloric acid for leucocytes LEUCOCYTES WASHED WITH BUFFERED SALINE SOLUTION AND SUSPENDED IN THE SAME SOLUTION pH OF ANTIGEN : pH of solutions 7.0 5.6 4.8 4.0 mie 60* 84 72 76 60t 40 48 44 eter 80 72 56 36 48 44 20 8 oe 76 40 24 20 52 24 12 0 80 36 20 16 aii 64 20 12 0 * The upper figure refers to the percentage of phagocyting leucocytes. 7 The lower figure refers to the percentage of leucocytes containing 10 or more cocci. The figures showing inhibition of phagocytosis are indicated by the heavy line. fore exactly comparable. The data show very little difference in the toxicity of acetic and butyric acids. There is, however, a slight indi- cation that butyric acid may be somewhat more toxic than acetic acid. Experiment 14. This was planned to obtain data supplementary to that given in table 7, which showed that toxic acid solutions were not harmful to leucocytes after neutralization. The effect on phagocy- tosis of one per cent of sodium citrate in the final suspension is shown in the protocol presented in table 11. The test solution was a one per TOXICITY OF ACIDS FOR LEUCOCYTES 295 TABLE 10 Butyric acid is perhaps slightly more toxic than acetic acid ACETIC ACID BUTYRIC ACID Leucocytes washed with saline Leucocytes washed with saline pH of antigen solution of pH Peorantines solution of pH Po data ee sag oe fk gir] -lbeeaal i e e Ot Bet ef oe ls ala’ * The upper figure refers to the percentage of phagocyting leucocytes. { The lower figure refers to the percentage of leucocytes containing 10 or more cocci. TABLE 11 Duplicate tests showing that one per cent concentration of sodium citrate of pH 7.0 does not inhibit phagocytosis CONTROLS. FINAL SUSPENSION IN BUFFERED FINAL SUSPENSION IN ONE PER CENT CITRATE PHYSIOLOGICAL SALINE SOLUTION OF pH 7.0 SOLUTION OF pH 7.0 64* 64 68 80 52T 44 44 64 * The upper figure refers to the percentage of phagocyting leucocytes. { The lower figure refers to the percentage of leucocytes containing 10 or more cocci. TABLE 12 Leucocytes absorb H-ions from weakly acid solutions pH pH AFTER ADDITION OF BEFORE LEUCOCYTES ADDITION OF TLE GEN ae te eG EE oe eee HCI | Lactic | Acetic | B nse f P 5.0 6.7 Saline solution unbuffered................ 56 66 68 4.0 Onl 4.0 Saline solution buffered................... 4.8 5a toa 5.6 ye ech 5 ; 4.0 4.8 46| 4.6 Saline solution buffered................... 46 52 47| 47 296 ALICE C. EVANS cent solution of commercial alkaline sodium citrate in physiological saline solution. The reaction was adjusted to pH 7 by the addition of hydrochloric acid. The treatment of the washed sensitized bacteria and the leucocytes in the control suspensions and in the test suspen- sions was exactly the same, except for the different solutions used in making up the finalsuspensions. The differences in results, slightly in favor of the citrate solution, are within the limits of experimental error. Thus it is shown that a solution of sodium citrate in one per cent concentration having a H-ion concentration of pH 7 is not toxic for leucocytes. This result is of special interest, as it will be pointed out in the later discussion. TABLE 13 A summary of the literature shows a similar order of effectiveness of various acids on various kinds of living cells. The effectiveness increases in the descending column a PRN ET ESTIONOE TOXICITY TO B. couit TOXICITY TO LEUCOCYTES GIANT CLAM* BROWALLIA PETALS} HCl HCl HCl Citric Citric Hl Citric Lactic Lactic Lactic Lactic Acetic ‘ 3 Acetic Butyric } Acetic Acetic Butyric * Harvey, 1914. t+ Haas, 1916. t Wyeth, 1917. DISCUSSION General conclusions Incidentally in connection with experiments designed to obtain knowledge on other points in question, many observations were made on the buffer action of leucocytes. In table 12 certain data are presented, selected by virtue of the freedom from red blood corpuscles of these particular leucocyte suspensions. The table shows that when leucocytes are placed in unbuffered solu- tions of very slight acidity, they absorb H-ions until the pH of the solution is slightly below 7, if their volume is sufficient. (Leucocytes added to a neutral or weakly alkaline unbuffered solution will bring it also to a reaction slightly below pH 7.) TOXICITY OF ACIDS FOR LEUCOCYTES 297 When leucocytes are added to a weakly buffered saline solution such as was used in these experiments, the quantity of H-ions which they absorb depends upon several factors: (1) The total number of H-ions removed from a solution by a given quantity of leucocytes depends upon the nature and quantity of buffer substances in the solution. (2) The dissociation constant of the acid concerned is also a determining factor. A greater quantity of H-ions was absorbed from HCl solutions than from solutions of the tested organic acids of equivalent concentration. Similar quantities of H-ions were absorbed from acetic and butyric acids, which have similar dissociation constants. (3) It was also observed that leucocytes which have already absorbed H-ions have a reduced capacity for further absorption. Hag- gard and Henderson, and C. L. Evans made similar observations on the buffer action of red blood corpuscles. Their capacity for absorption of H-ions depends upon their previous history (4) It would be expected that a dense suspension of leucocytes would absorb more H-ions from a given quantity of any solution than would a thinner suspension. The experiments showed that to be the case. These observations are in agreement with those made by Gray upon the eggs of trout, from which he drew conclusions for living cells in general. The data presented in the tables show that leucocytes which have absorbed H-ions have thereby suffered an injury which affects their capacity for phagocytosis. Koltzoff reported that the phagocyting capacity of Carchesium was restored when removed from an acid to a neutral solution. Such was not the case with leucocytes in these experiments. The cumulative effect of repeated exposures to acid solutions was demonstrated many times. It has already been mentioned in connection with the data presented in table 4. It may also be observed in tables 8, 9, and 10. For example, in table 8, leucocytes whch had been washed once in citric acid solution of pH 4 showed no injury; and leucocytes which were placed in a citric acid solution of pH 4.6 for the final suspension showed no injury if they had not been previously exposed to acid; but when leucocytes which had been previously washed in citric 298 ALICE C. EVANS acid solution of pH 4 were placed in citric acid solution of pH 4.6 for the final suspension, their capacity for phagocytosis was almost completely destroyed. In agreement with many other investigators who have studied the relative effect of inorganic and organic acids on various kinds of living cells, these experiments show that the organic acids studied (possibly with the exception of citric acid) have a specific toxicity in addition to the toxicity of the free H-ions, as deter- mined by comparison with the toxicity of solutions of HCl, which undergoes practically complete dissociation in the weak dilutions used in these experiments. It will be recalled that Harvey compared the results of his own investigations and those of several who had preceded him, and that he found little agreement in the order of the various acids relative to their effect on the different living cells studied. From the data given by Harvey, and that given by Haas and Wyeth in more recent publications, it was possible to tabulate the order of effectiveness on various kinds of living cells of those acids which were used in this study. The comparison, made on the basis of degree of dissociation, is given in table 13. It shows acetic butyric except that Harvey found citric acid definitely more toxic than HCl. It may be added here that some observations on the relative effect of these acids on red blood corpuscles indicate that this order would not hold for them. Citric acid showed an unques- tionably specific toxicity, greater than that of lactic acid. It has been noted that several investigators have reported that leucocytic activity is stimulated by minute quantities of acids. Hamburger presented data to show that lipoid-soluble substances, including butyric acid, stimulate phagocytosis when present in minute quantities. In the protocol showing the effect of butyric acid on leucocytes (table 10) figures are given for 3 tests in which the leucocytes were treated with weaker solutions of varying strength of butyric acid than the experiment showed to be toxic, yet there was no evidence of a stimulative action in any of them. However, it happened in several experiments a perfect agreement in the series HG} < lactic citric TOXICITY OF ACIDS FOR LEUCOCYTES 299 that higher figures for phagocytosis were obtained when the leucocytes had been treated with sub-toxic acid solutions than when they had been treated with neutral solutions. Such ex- amples may be observed in tables 2, 4, 8 and 9. Since, however, such indications of a stimulative action of acids occurred irregu- larly, and never to a significant degree, it seems more reasonable to regard the mentioned figures as variations due to experimental error. Practical applications The experiments show that when phagocytic tests are made mm vitro it is important to protect the leucocytes against exposure to acid solutions. ‘This is accomplished by the use of a buffered saline solution. It may appear that such a precaution is un- necessary if proper care is taken for the cleanliness of glassware. It is true that the tables show that the leucocytes withstood one washing in an acid solution of pH 4.6 (see tables 8 and 9) without any evidence of an inhibition of phagocytosis. But they showed the injurious effects of much weaker acid solutions when the exposures were repeated. As a matter of fact, before the sensi- tiveness of leucocytes to acids was taken into consideration it happened rather frequently that the routine quantitative tropin test for the potency of commercial anti-meningococcus serum failed. And here again it may be emphasized that the tropin test has been generally condemned as unreliable. This is what may happen when all due care is taken for the cleanliness of glassware. The citrate solution used for taking up the pleural exudate may be of an acid reaction. The leucocytes are exposed to a comparatively large quantity of it. Then they are washed in a large quantity of saline solution. If it is un- buffered, and has been allowed to stand with no protection from the air of the laboratory other than the cotton plug, it may have absorbed enough CO, from the atmosphere to bring the reaction to as low as pH 5.8. The leucocytes are again exposed to the saline solution in smaller quantity when it is added for making the final suspension. Thus they may be subjected three times to the effects of acid solutions. 300 ALICE C. EVANS The results obtained with the use of acid citrate brings up the discussion of a controversy in the literature in regard to the effect of citrate on phagocytosis. In their early studies on phagocytosis-promoting antibodies Wright and Douglass mixed the blood with sodium citrate to prevent clotting. Their technic has been widely adopted, but Snapper, Ouweleen, Hamburger and Hekma, Radsma, Fenn, and Wolf have all reported that citrate is injurious to leucocytes. On the other hand Stuber and Riitten reported that they could not demonstrate a harmful effect of citrate. Under the conditions of the experiments reported here, a one per cent solution of sodium citrate in the final suspension showed no inhibition of phagocytosis. It seems probable that some of those who reported that citrate was injurious to leucocytes may have been working with an acid product. Ouweleen’s statement that the injurious action of sodium citrate was neutralized by serum emphatically suggests that this investigator was working with an acid citrate. The demonstration of the injurious effect of acid citrate on the phagocytic activity of leucocytes, together with the finding of commercial sodium citrate preparations of an acid reaction, lead to the practical conclusion that when sodium citrate solu- tion is used for preventing coagulation of body fluids containing living cells, it is a matter of importance to ascertain that the solution is not of an acid reaction. On the other hand the leucocytes were apparently unaffected by solutions of slightly alkaline pH values. = — =~ = = ~~ a =~ = ~~ = ms = = =~ = ~~ = =~ — = = = bt l= bh bo i) SASAHINGUVd NI HAREWON LIGGVU GNV ISuld HAAWON NIVULS WOKS !VUaSs DIDOD0NOD suzuyn)b6o snobojowoy fo uord.sosqo aBnjuariad ay} aonpur sainbyf ayJ, “ur.s4s snav000buruau [ pun surv.ys snov090u06 P9}99198 BI YJUN dias 9199090U0B Lg fo suorjd..osqy 8 WIAVL 337 THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 4 338 JOHN C. TORREY AND GEORGE T. BUCKELL basis, then, the highest possible score for the twenty-seven sera would be 54. This graph shows clearly the fact that of these particular strains, 34 possesses the widest and most generalized absorptive capacities, effecting a marked absorption of the homologous agglutinins from twenty-four of the twenty-seven sera. This 100 90 80 70 60 50 40 30 20 Absorption value in percent S41 30° 1 «15 49 «FPS Se) 16°49 4 ene Gonococcus Strains © Oeee CuHaRrT 1 strain, as has already been mentioned, is one of the set obtained from the New York Health Department Laboratories under the designation ‘‘S.’”’ Strains 11 and 30 proved to be next in effi- ciency in absorptive capacities with the others grading off to those with the least generalized absorptive properties, namely, strains 41 and 49. We do not wish to convey the impression that strain 34 is unique among our collection of cultures in its generalized affinities for it seems highly probable that certain other strains, such as 8 and 42, might have proved quite as effective. A SEROLOGICAL STUDY OF THE GONOCOCCUS GROUP 389 General observations in reference to table 8 Tolluch (17) in discussing the importance of the application of standardized agglutination methods in a study of immuno- logical types of a given bacterial species, mentioned the desira- bility of using sera with as uniform a titer as possible. Other investigators, such as Gordon (18) for meningococci and Hooker (7) for the typhoid bacillus, have reported the presence of greater specificity in sera from rabbits after a few inoculations than after many; or as Gordon has expressed it, in the “first born ag- glutinin.”’ Although these general principles may apply to the gonococci, it was not found feasible either to restrict the number of inoculations to a small number or to attempt to employ sera with a nearly uniform titer. This was due to the fact that the rabbits responded very unequally to inoculations with various strains. With some strains a fairly high titer could be obtained after six inoculations given in a period of two weeks, but with other strains a prolonged course of immunization was necessary before a serum could be obtained with a titer perhaps only one- half or less as high as with the former. As has already been mentioned, relatively inagglutinable strams were encountered more frequently than was anticipated. In- agglutinability, however, does not necessarily mean the absence of agglutinogenic or agglutinin-absorbing capacities as was demonstrated by McIntosh and McQueen (19) some years ago in a study of an inagglutinable typhoid strain, and more recently in reference to typing meningococci by Mathers and Herrold (20) and others. On the other hand Bemains (21) has shown that a strain of B. dysenteriae, which was artificially rendered in- agglutinable, also lost its agglutinogenic and agglutinin absorbing functions. It might be expected that the relatively inagglutinable strains of gonococcus would also prove poor absorbers of agglu- tinin, but this, although true in some instances, was not always the case; certain poorly agglutinating strains, such as 25 (table 8) absorbed agglutinins strongly and consistently. Another similar point, which has already been ‘discussed, is further illustrated in table 8, namely, that the agglutinogenic and the absorptive capac- 340 JOHN C. TORREY AND GEORGE T. BUCKELL ities of a strain may notrun entirely parallel. For example, strain 38 produced a serum from which the homologous agglutinins were absorbed in greater or less degree by all these twelve selected strains and yet strain 38 removed none of the homologous agglu- tinins from sera produced respectively with strains 1, 18 and 30. Other instances are provided by strains 27 and 49, both of which absorbed agglutinin poorly and were apparently unrelated to other larger groups of gonococci but which, nevertheless, pro- duced agglutinins readily absorbable by regular types. In de- termining serological relationships among the gonococci, then, it is not sufficient alone to absorb other sera with a given strain but a serum should be produced with this strain and absorbed in turn with other typical strains as it is evident that in some instances the function of binding agglutinin may have become inactivated or suppressed in some degree without affecting the agglutinogenic propensities. Teague and McWilliams (22), in a study of the effect of spon- taneous agglutination on the absorption of agglutinins, found that in the case of a typhoid strain spontaneously agglutinating sub- cultures absorbed less agglutinin than did the normal sub-cultures. They inferred from this finding that physical aggregation may play a part in effecting the amount of agglutinin absorbed by various strains. Gonococcus strains with tendencies to spon- taneous clumping have been encountered not infrequently in this work but this tendency has not seemed to interfere with agglutinin absorption, perhaps because the aggregations were more readily broken up than is the case with flagellated bacilli such as B. typhosus. Vulvovaginitis strains This series of gonococcus cultures included thirteen strains isolated from cases of vulvovaginitis in young girls. The ages of these patients ranged from two to eleven years, with ten under six years. These strains were studied with particular care in order to determine if they constitute a type immunologically distinct from those recovered from adult gonorrheal infections. Louise Pearce (23), from a comparative serological study of six A SEROLOGICAL STUDY OF THE GONOCOCCUS GROUP 341 infant strains and nine strains from adult cases, came to the con- clusion that there are inherent differences in the types of gonococci causing these two classes of infections, although not to an extent which would warrant separating them into two distinct immuno- logical groups. Her methods of investigation included agglutina- tion tests without absorptions and complement fixation experi- ments. The results with these two methods were harmonious. It is obvious that if the gonococcal strains recovered from these two classes of infection could be shown to be representative of serologically distinet types, it would clarify our ideas in regard to their epidemiological relationship. In table 9 the results of absorptions with sera immune to two of these vulvovaginitis strains are reported and in table 8 the absorptions of these sera (108 and 239) with twelve adult strains. In table 5 are detailed further the absorptions of a serum (104) immune to an adult strain (15) with these infant case strains. These several cross absorptions indicate clearly that no definite serological distinction may be drawn between infant and adult strains. Among these infant strains, as is demonstrated in table 9, one encounters the same lack of homogeniety as to type as would be observed among a similar series of strains isolated from adult cases. We also find among the infant strains cer- tain ones (48, 60, 61, 63, 64) which absorbed strongly the homol- ogous agglutinins from an adult strain serum (table 5) and also others (55, 56, 57, 62, 75) which removed little or none of such agglutinins from this adult, regular strain serum. One fairly definite characteristic, however, was noted in the study of these infant strains and that was that their agglutinogenic potential- ities were not strong. No serum with a titer above 1:800 was obtained with any of the four strains employed in rabbit immuni- zations. With one strain (48) the titer could not be forced above 1:250 in spite of a prolonged series of inoculations of large dosage. It should be noted, however, that the production of low titer sera was not confined to infant strains as certain adult strains, especially of the irregular type, exhibited the same ten- dency, but it appeared to be more of a distinguishing trait of the infant strains. These infant strains also, as a class, tended 342 JOHN C. TORREY AND GEORGE T. BUCKELL to absorb agglutinins rather less strongly than most of the adult strains. This weakness in antigenic properties may be a factor in determining the mild yet chronic course which most of TABLE 9 Two sera immune to vulvovaginitis strains (children cases) 64 and 56, absorbed with other strains from the same type of cases. TITER HOMOL- TITER ABSORB-| TITER HOMOL- PERCENTAGE TITER WITH = OGOUS STRAINS UNABSORBED ING STRAINS OGOUS STRAIN STRAIN WITH ABSORPTION SERUM WITH AB- WITH UNAB- ABSORBED OF SPECIFIC SORBED SERUM | SORBED SERUM AGGLUTININS SERUM Strain 64 serum, titer 800; absorbed 1:25 55 50 <50 800 100 88 56 <50 <50 800 800 0 57 <50 <50 800 400 50 58 200 <50 800 200 75 59 400 50 800 100 88 60 400 100 800 200 75 61 200 50 800 100 88 62 100 <50 800 400 50 63 800 100 800 800 0 64* 800 100 800 100 88 65 100 <50 800 800 0 75 400 <50 800 400 50 Meningococcus <50 <50 800 800 0 Strain 56 serum, titer 500; absorbed at 1:25 48 100 <50 500 50 90 54 750 100 500 250 50 56* 500 500 58 500 100 500 100 80 60 1000 <50 500 250 50 61 1000 <50 500 250 50 62 100 <50 500 250 50 64 750 100 500 100 80 Meningococcus <50 <50 500 500 0 * Homologous. these cases run in that insufficient stimulus is provided for the production of an adequate amount of bactericidal antibody on the part of the host. It should be observed that the majority of these vulvovaginitis strains were isolated from patients under treatment at a clinic A SEROLOGICAL STUDY OF THE GONOCOCCUS GROUP 343 (Vanderbilt Clinic, through the courtesy of Dr. B. Wallace Hamilton) and it was thus entirely unlikely that the infection could have been received in any instances from a common source. On the other hand, if such strains are isolated from cases occur- ring in an institution, there is always a possibility that a single strain may have been spread through contact and, accordingly, a study of such cultures might give an unwarranted impression of serological homogeniety as existing among gonococci causing these vulvovaginitis infections. DISCUSSION Within recent years much time and effort has been devoted by various investigators to the detection and formulation of distinct immunological types among different species of patho- genic bacteria, the members of which are culturally identical or nearly so. Without doubt these efforts, in some instances, have served a very useful purpose. There is, however, the lurking hazard in extending this method of grouping to other bacterial species that one may depart widely from a natural system of classification. A type or group, if established on a rational basis, should represent a distinct serological entity. The antibodies of a serum produced with a member of a given type should be absorbed completely, or nearly so, by all strains referred to the same type but should not be affected by strains assigned to a different type. Or, as Eastwood (24) has expressed it in an illuminating discussion of this question, “‘if strains belonging to the species fall into distinct groups without cross-division, when tested by a number of monovalent sera, grouping is indicated; but if such tests produce marked cross-division, grouping is not justifiable.” To quote further, ‘‘where each main group is so elastic that its margin of separation from the others is small, one begins to raise the question whether the adopted system of grouping has turned out to be artificial and arbitrary, and whether the species under consideration is really amenable to subgrouping.”’ In these remarks, Eastwood was referring particularly to certain proposed groupings of the meningococci. Dopter’s orig- 344 JOHN C. TORREY AND GEORGE T. BUCKELL inal division of the meningococci into the true meningococci and the parameningococci was a natural one in that we have here two groups which are generally recognized as being serologically distinct, but in the further division of the meningococci into four groups in accordance with Gordon’s classification (18), two groups have been added which, perhaps, are not founded on as sound a basis. In fact, a considerable number of investigators of the meningococci have reported that the Gordon group III is not sharply separated from I, nor IV from II; affinities which Gordon, himself, has recognized. Griffiths (25) and also Scott (26), from a study of spinal and nasopharyngeal strains derived from much the same sources as those of Gordon, both concluded that only two main groups should be recognized; a conclusion which Eastwood has considered justifiable. Griffiths has also inferred that ‘‘the two groups are not fixed types but may be further sub-divided by means of absorption experiments into sub-groups which are probably ‘centers of variation’ of the differ- ent stages of evolution of the meninogococcus antigens.” It is the opinion of the writers that there is much less justi- fication for the formulation of distinct groups in the gonococcus than in the meningococcus family. Among the gonococci there are no two groups as sharply separated as the normal meningo- cocci and the parameningococci. In fact, if agglutinin absorp- tion and agglutinogenic properties of each strain are taken into consideration, as they really should be in an attempted formula- tion of types, we would find so many strains exhibiting individual immunological variations and also inter-relationships between what might, at first sight, be considered representatives of distinct types that the line of demarkation between the proposed groups would tend to become completely obliterated. We have, accordingly, on the basis of the evidence already presented, distributed our gonococcus strains under three general headings, namely, regular strains, the most generalized as regards antigenic properties, intermediate strains, which are quite closely related to certain of the regular types, and the irregular strains which exhibit marked individualistic variations. In employing these general terms we are adopting a mode of classification similar A SEROLOGICAL STUDY OF THE GONOCOCCUS GROUP 345 to one which has been extensively applied to the meningococci, but in view of our findings, especially as regards the marked antigenic lability inherent in the gonococcus group, we feel that it is as definite a one as the conditions justify. This phenomenon of antigenic lability, as exhibited by the gonococcus, is most interesting. As has been mentioned, Her- manies noted a marked degree of lability on the part of certain members of his type 2 group. During the course of our experi- ments we have observed numerous instances in which the anti- genic constitution of strains have undergone some change under cultivation. By way of example reference may be made to the results obtained with two sera (217 and 236, table 8) immune to strain 29. An interval of over a year occurred between the times of production of these sera against a strain which had become stabilized. It may be noted, however, that strain 1 has, during that interval developed affinities very similar to the serum strain 29, a member of the regular group, although soon after isolation this strain 1 showed a close relationship to strain 41, a member of the irregular group (table 8). Strain 25, also, during this period has developed evidence of a close relationship to this serum strain, and strain 7 has passed from an intermediate position to a close affinity to the regular group. These are only a few instances, among many, in which there has occurred a change in antigenic responses to various sera. Beside such changes in these comparatively recent strains, we have the still more striking evidence of antigenic lability as presented by the changes which have occurred in the ‘‘ Torrey strains” during the fourteen years of cultivation (table 7). In the face of such a tendency to mutation in antigenic constitution on the part of gonococci, one would have no assurance that strains selected as type representatives might not change entirely in character after a period of cultivation. We believe, in fact, that the whole tendency under conditions of artificial culture is for reversion to our regular group and that a strain having attained that dis- position of its antigenic components remains in a comparatively stable condition. Changes in type under cultivation have also been noted by 346 JOHN C. TORREY AND GEORGE T. BUCKELL various observers as occurring in certain strains of the meningo- cocci. Griffiths, in fact, in view of the variations in antigenic characteristics observed during sub-cultivation concluded that ‘‘meningococcus antigens are not precisely fixed or stable sub- stances, but are liable to modification under the influence of environment.”’ He believed these changes might take one or the other of two directions; toward either increased or diminished complexity of the structure of the receptor apparatus, which would mean an increase or diminution in the range of binding capacity. Scott has likewise reported temporary variations in agglutination between different sub-cultures of the same strain of meningococcus, and also variability in the absorptive power of one strain. Both of these phenomena, however, he considered due to the presence of two varieties of cocci of differing sensitivity to agglutinin and absorptive powers within a single strain with sometimes one and sometimes the other predominating and not to antigenic changes within the substance of the cocci. These peculiarities, however, in the absorption of agglutinin and also the large number of serological varieties which he encountered, caused him to doubt the practical value of this mode of classi- fication in defining types among the meningococci. Walker (27) has concluded in connection with a discussion of the signi- ficance of meningococcic types that ‘‘fixity of these (Gordon’s) is not proved.”’ Butterfield and Neill (9), although believing Gordon’s type strains to be stable and finding sera prepared with them very effective in the classification of their series of meningococci, reported that certain strains during a year’s time apparently changed from one type to another. A very interesting observation in this connection is that of Eberson (28) on the effect of ultra-violet rays on the antigenic properties of the meningococci. He found that by exposure to these rays, regular and irregular meningococci could be so altered that they developed the property of producing mutual agglutinins, but parameningococci could not be so altered. He further concluded from similar tests that regular types of meningococci include with- in their protein molecules the elements of the parameningococci but that the latter possesses no agglutinogenic radicle common to the regular and thus constitutes a distinct type or species. A SEROLOGICAL STUDY OF THE GONOCOCCUS GROUP 347 In attempts to explain the development of serological variants among certain groups of bacteria, a number of different hypoth- eses have been offered. The theory of Meinckeandhiscolleagues of the loss of avidity on the part of certain receptors of the cholera vibrio has already been mentioned. Eastwood (24) has suggested that a more reasonable explanation of these changes in antigenic behavior lies in the conception that antigen is ‘‘a chemical substance which may exist in one or other of several different chemico-physical phases, demonstrable in vitro. And the same conception would apply to antibody.” He further suggests that these various chemico-physical phases are closely related to changes in stereo-chemical structure of the molecules of the bacterial protoplasm. This theory, of course, is at present highly speculative, but it may find confirmation through later advances in physical chemistry. Hermanies, in a discussion of the origin of the sub-groups and races which he encountered in his group 2 of gonococci, proposes the theory that his four types were originally derived from a common type containing alone the agglutinogen x, but under different environmental conditions this x moiety has tended to retrogress and has been replaced in greater or less degree by the antigenic elements a, b, c and d, according to the particular race to which a given strain is related. ‘This process, he suggests, may be carried so far that the original « element becomes lost and a new species characterized by one or other of the four acquired antigenic elements may arise. The new species would seem to be fixed in type because, as the antigenic element 2 has been entirely elimi- nated, there is no possibility of reversion to that originaltype. In view, however, of our experimental results it seems to us that this theory assumes a greater degree of acquired antigenic stability than is exhibited by the gonococcus group. It is our opinion that as a result of adaptations of gonococcal strains to different en- vironments, including the defensive agencies of the host, there have arisen varied molecular configurations or physical-chemical phases—to use Eastwood’s term—of the same specific basal substance, none of which, however, attain a state of complete stability and thus do not give rise to new fixed types. Further- 348 JOHN C. TORREY AND GEORGE T. BUCKELL more, as none of the original elements have been lost or elim- inated, there may occur in time following exposure to an en- vironment of fairly uniform type, such as is presented by a culture medium, a molecular redistribution of the antigenic sub- stance until a pattern resembling that characteristic of our regular and generalized group is attained. Griffiths (25) has suggested that enhanced virulence and in- vasive powers of the meningococcus may be correlated with increased complexity of the receptor apparatus; that is with an increase in range of combining capacity. From this point of view our most complex gonococcie strains are those designated as regular, with the intermediate next in order and the irregular types exhibiting the least degree of complexity. Although the majority of our strains were isolated from more or less acute cases of urethritis and accordingly their invasive propensities could not be evaluated, there were eleven strains in the series recovered from cases showing complications. Among these there was one case of ophthalmia, three of chronic prostatitis, one epididymitis, four joint infections and two septicaemias. Ten of the eleven strains isolated from these cases showing complications were of the regular or intermediate type. Although the number of these cases is too few for a definite conclusion, the tentative suggestion may be made that the irregular variant strains are less likely to give rise to complications than are those resembling the regular types. Jétten (16) found that his more virulent strains fell into two of his four groups, but as to whether the members of these two groups are serologically related to our regular types, we, of course, can offer no opinion. COMPLEMENT FIXATION In these complement fixation experiments with the gonococcus our purpose has been not so much to confirm the conclusions derived from the agglutination work in regard to the serological relationships of our strains as to apply the facts disclosed to the selection of strains most useful as antigens for the diagnostic complement fixation test. Incidentally, however, we have noted a marked degree of correspondence between the strain affinities revealed through these two methods. A SEROLOGICAL STUDY OF THE GONOCOCCUS GROUP 349 One of the great obstacles to the general use of this test in the past has been the difficulties associated with the preparation of the antigen, especially as the delicate nature and apparent fastidious cultural requirements of the gonococcus have made their maintenance a troublesome matter and also because it has seemed necessary to employ a considerable number of strains in preparing the antigens. Commercial distributers of biologi- cal products, also, have been unsuccessful in supplying this antigen in an entirely satisfactory way because of its tendency to develop anti-complementary properties. It would seem, then, that if one or two readily cultivatable, representative strains should prove to be as efficacious as antigens as the combinations of the ten or more strains, which have been used generally here- tofore, the test would be much simplified and hence more generally available. TECHNIC Antigen. The following serum-free solid medium has been used for growing the cultures for antigens. This medium has a growth accessory principle and is prepared according to the Huntoon (4) method, although the formula has been modified. It has generally yielded a luxuriant growth with the gonococcus strains employed. Salt-free, 1.5 per cent peptone, “vitamine” agar. Five hundred grams of fresh chopped beef heart, free from fat, one whole egg and 1 liter of distilled water are placed in a double boiler over a free flame and the temperature is maintained at 60°C. with constant stirring for five minutes. Fifteen grams of peptone (Difco) and 18 grams of flaked agar are now added and the temperature raised until the medium assumes a brownish color. It is then made slightly alkaline to litmus (with a 10 per cent sodium carbonate solution) and is transferred to a flask, or better to a coffee pot, and heated in the Arnold sterilizer at 100°C. for two hours. The medium may now be cleared through centrifuging or by filtering through glass wool; if the latter procedure is used, the meat residue should be deposited on the glass wool in a funnel and the fluid portion allowed to perculate through; several times, if necessary. No cloth, cotton or other absorbent material 350 JOHN C. TORREY AND GEORGE T. BUCKELL should be allowed to come in contact with it. After clarification, 2 per cent glycerine is added, the reaction is adjusted to pH 7.2, and the medium reheated. It is then distributed in potato tubes, sterilized in the Arnold and slanted. ‘The slants should be laid down not longer than the day before they are to be used. The large slants of this medium should be seeded from twenty- four hour growths on thesame medium. ‘They are then incubated for from twenty-four to forty-eight hours. In preparing the antigen the method of M. A. Wilson (29) was followed. The growth was well washed with 50 per cent and then 95 per cent alcohol, being allowed to stand one-half hour with each in a water bath at 37°C., with frequent shaking. After centrifuging, the sediment is covered with ether and allowed to stand at room temperature for one-half hour with frequent stirring. It is then centrifuged, the ether decanted off and the sediment placed over night in a dark place. The dry powder is then emusified in 0.85 per cent saline to the standard density and the proper dosage determined. It should then be heated at 80°C. for one hour, although in our experiments this heating was omitted as the antigens were prepared freshly at frequent intervals. The antigen may be preserved with 0.1 per cent phenol. Antigen prepared in this way yielded results superior to those obtained with filtered autolysates or with an antigen prepared according to Thomson’s (30) method. Gonococcic sera. Most of the immune sera were the same as those used in the agglutination experiments and, for the most part, had been preserved for a longer period of time. They were inactivated shortly before each test was performed. Complement. ‘The sera from three or four guinea-pigs were pooled and preserved by freezing. There was usually little or no loss in potency for four or five days. Only active comple- ment, fixable by gonococcus, was used. Hemolytic system. The anti-sheep system was used. One hemolytic unit of complement and two hemolytic units of ambo- ceptor were employed. The complement was titrated imme- mediately before the test and the amboceptor at frequent intervals. The complement unit was read at the end of thirty A SEROLOGICAL STUDY OF THE GONOCOCCUS GROUP 351 minutes in the water-bath at 37°C. Washed sheep corpuscles in a 2.5 per cent suspension were used, and the amboceptor and cells added separately. The test. The total volume of the mixed reagents was ad- justed to 2.5 cc. One unit of antigen was used and this was determined by titration with a polyvalent gonococcic serum or one with generalized affinities. This unit was the smallest amount of antigen which caused complete fixation of comple- ment with 0.1 cc. of a 1:100 dilution of the gonococcic test serum. The antigen was diluted so that this unit was contained in 0.1 ec. This unit, of course, was always less than one-half the anti-complementary dose and generally much less. This titration was usually made just before the fixation tests were started. Controls were always prepared for the hemolytic sys- tem, for anti-complementary action of the immune sera and also the antigen, and also for the stability of the sheep cells. The tests were incubated at 37°C. for the usual periods and readings made at the end of one hour-incubation of the completed test and also after the tubes had stood over-night in the ice-chest. The last reading is the one which is recorded. In accordance with the usual custom, the symbol 4 indicates complete fixation; 3, almost complete fixation; 2, the hemolysis of about one-half the cells; and 1, only a slight degree of fixation. COMPARATIVE TESTS In table 10 are recorded the results of comparative tests of the fixation with twenty-seven monovalent gonococcic sera, with a polyvalent sera which was also used in titrating the anti- gens, and with a meningococcic serum. The four antigens em- ployed in these tests consisted of the following single strains or combinations of several strains: First, strain 34; second, strains 15, 18, 34 and 41; third the ten ‘‘Torrey strains;” fourth, strain 42. In order to give full comparative value to the results, the four antigens were always tested on a given serum simultaneously and thus under exactly similar conditions. The object, as stated, was to determine whether such generalized strains as 34 and 42 yielded as good results as combinations of several 352 JOHN C. TORREY AND GEORGE T. BUCKELL TABLE 10 Complement fixation with 27 monovalent and 1 polyvalent gonococcic sera and 1 men- ingococcic serum, using the following four gonococcic antigens: (1) Strain 34; (2) Strains 15, 18, 34, 41; (3) Ten “‘Torrey strains’’; (4) Strain 42 IMMUNE SERA DILUTIONS ANTI- SERUM STRAINS GENS 0.05 | 0.01 | 0.004 | 0.002 | 0.001 | 0.00066 | 0.0005 | 0.00025 | 0.000125 1 “eat 3 yeaa et ee = Ss 2 2 Sib Sake i> Ieee t i see = & (246) 3 A) SS te Gene “hy: = & 4 a hrg |. oes usa Da = = = 1 7M Weer ta eed er Pe 3 3 2 1 5 2 ES RT aa 3 2 ss = (121) 3 4) 44 0 ai 3 2 = 4 PUM Rr etd! 1 en Len ae z. za 1 7s eg Ne a Ph | NY Pm Ss _ 7 2 Da. Ada aioe 2 2 1 o (231) 3 1d WG eal a ta ete = za 4 Aibiasl alll Tak bes 3 2 1 = 1 rae Whee a ha oe 3 3 1 1 8 2 AN a: Aci eer 3 2 1 = (243) 3 ZOpcg \rig fg) Dash) Qt Ws BS 2 o 4 Mahia) + ieBaes 3 2 as 2: 1 aA 4, eS 3 2 1 = 11 2 BR ah SEARO ES Wiis 1 1 1 = (118) 3 4) 2.b Sol compas 1 1 = = 4 4 irae acy rae es 2 2 - = 1 Ah el Aa 2 1 2 2 14 2 Ah AN Ads | Adephe® 1 1 = =: (103) 3 de] AL So) see 2 2 1 1 4 7p ae We: sae fa | a 4 3 2 1 1 delat ode toate 4 4 3 3 15 2 hol vA lA Aiba A 4 3 3 1 (104) 3 Mahe hr ee ts 3 3 3 3 4 4 | ere) ee Se 3 3 2 2 1 3 kaSily 2 val Bates Bib we = a 16 2 eR bales ip PAE Praga |) = = (87) 3 Sg) oat a ee ee = = 4 TUM Mm eee as mess bers 6) = = 393 0.00066 | 0.0005 | 0.00025 |0.000125 IMMUNE SERA DILUTIONS 0.004 | 0.002 | 0.001 TABLE 10—Continued 0.01 0.05 GENS A SEROLOGICAL STUDY OF THE GONOCOCCUS GROUP ANTI- SERUM STRAINS lo | N JOHN C. TORREY AND GEORGE T. BUCKELL 354 TABLE 10—Continued IMMUNE SERA DILUTIONS SERUM STRAINS 2 1 2 3 1 1 1 3 il 0.00025 |0.000125 3 1 2 2 2 3 3 3 1 3 1 2 0.0005 3 2 1 2 1 2 2 3 3 3 3 3 1 3 1 2 0.00066 3 2 3 3 1 3 2 2 2 4 3 3 4 3 2 2 4 2 3 3 3 4 3 2 3 2 2 1 3 2 2 4 3 3 4 3 2 1 2 4 2 3 0.001 A SEROLOGICAL STUDY OF THE GONOCOCCUS GROUP 355 TABLE 10—Concluded i IMMUNE SERA DILUTIONS SERUM STRAINS Bae ig See 0.05 | 0.01 | 0.004 | 0.002 | 0.001 | 0.00066 0.0005 | 0.00025 |o.000125 1 4 4 2 1 1 1 1 1 _ 2 1 2 —_ _ _ — _ _ Si 3 3 3 3 1 _ _ — - — 4 4 3 3 3 _ _ _ _ - 1 4 3 _ — — — _ _ 2 4 3 _ _ _ _ — — - me 3 4 2 1 _ _ — - — _ 4 4 3 1 _ _ _ _ — _ 1 4 4 4 4 4 4 3 3 2 Polyvalent 2 4 4 4 4 4 3 3 3 2 221 3 4 4 4 4 4 4 3 3 3 A 4 4 4 4 4 4 3 3 3 y) = = = — — = Meningo- z 2 < ie = ¥: i = is ve coccus x ee ee.) | es loch oa ee ee 9 a) 4 3 7 Cree (Nabe cra, ag ee ha oe strains. The second combination was selected as these four strains seemed at the time to be fairly representative in their affinities of our series of gonococcus cultures. As the results show, however, they failed to cover the variations exhibited by our collection of strains. The results in table 10 substantiate fully the conclusion which we reached as the basis of our agglutination experiments that our strains 34 and 42 are both markedly generalized in their affinities. Further it appears that an antigen prepared with either one of them yields on the average better results than those obtained with the combination of the ten ‘‘ Torrey strains” (antigen 3) and also than with the four strains combined in antigen 2, although the latter was expected to exhibit a wider range of affinities. In the case of no serum was fixation obtained with multiple strain antigen 3 in which the single strain antigens 1 or 4 failed; on the other hand with nine sera stronger fixation occurred with antigen 1 than with antigen 3, and with eleven sera stronger with 4 than with 3. 356 JOHN C. TORREY AND GEORGE T. BUCKELL In view of the above results, if a single antigen is employed for diagnostic tests, we would advise using either strain 34 or 42 alone, or perhaps in combination. It seems very questionable if better results would be obtained by combining a large number of regular and irregular strains in an attempt to produce a poly- valent antigen. In the first place it is questionable if such an antigen covering all the probable variants could be prepared; and in the second place when a large number of strains each with more or less limited affinities and representative of an equal number of so-called types are combined in a single antigen, then the antigenic elements for each of these types becomes so diluted that, in the dosage permissible, the effectiveness for each type would be greatly curtailed. Perhaps the best procedure would be to use two separate antigens in each test; one prepared with one or two representative strains with generalized affinities and good combining qualities, and the other with selected irregular strains. Selected strains for diagnostic and therapeutic applications For the preparation of a stock vaccine for use in cases for which an autogenous vaccine is not available, we would recommend a combination of our strains 15, 34 and 41. It seems probable, also, that these strains would serve as well as any in connection with the production of a curative serum. Strains 15 and 41 in combination may be used to advantage in the production of an agglutinating serum with wide affinities which might prove helpful in the identification of gonococci. For the preparation of an antigen for complement fixation we advise the use of our strains 34 and 42. This antigen should exhibit generalized and strong fixation properties and prove effective except in the case of the relatively few infections due to certain irregular strains of the gonococcus. With sera giving negative results with this antigen it would seem advisable to conduct a test with an antigen prepared from a combination of irregular strains such as 16, 22, 28, 41, 49, 62 and 77 before giving a negative report. Without doubt at times inconsistencies be- tween the results obtained with this complement fixation test A SEROLOGICAL STUDY OF THE GONOCOCCUS GROUP 357 and clinical and bacteriological findings have been due to the fact that the infecting strain was immunologically unrelated to those used in the antigen. Through the use of these two antigens we believe the chance of this occurring is considerably lessened, but even so no definite assurance may be offered that one or the other would combine with the antibodies produced by all possible variants of the gonococcus. SUMMARY 1. An analysis by agglutinin absorption methods of the serolog- ical relationships of seventy-seven gonococcus strains, isolated from cases of acute and chronic gonorrhea and its complications, indicated that they may not be distributed among a number of clear-cut immunological types. These gonococcus strains were representative of those occurring in widely separated geographical localities. 2. Although this investigation has not demonstrated the ex- istence of distinct immunological types to each of which a con- siderable number of gonococcus strains might be referred, it was found feasible to classify our-strains under the three general headings of (a) regular, (b) intermediate and (c) irregular strains. 3. Evidence is submitted which indicates the existence of a marked tendency to antigenic lability on the part of the gonococ- cus. We believe that instability of antigenic constitution is a general characteristic of the gonococcus group and that the strains exhibiting this tendency may not logically be segregated in one particular type, as has been suggested by Hermanies. 4, Among our regular strains, we have found certain ones which are highly generalized from the antigenic standpoint, and which appear to be representative of a large part of the gonococcus group. 5. Cross absorption experiments have indicated clearly that no definite serological distinction may be drawn between strains isolated from vulvovaginitis cases in children and those from gonorrheal infections in adults. 6. A few representative strains with generalized relationships and good antigenic properties have been designated as suitable 358 JOHN C. TORREY AND GEORGE T. BUCKELL for use in a stock vaccine, and also for the preparation of a poly- valent antiserum. 7. A marked degree of correspondence was noted between strain affinities as revealed by agglutinin absorption and com- plement fixation tests. 8. Two strains of gonococcus have been selected, each of which covers a large part of the group, and which may be used to ad- vantage in preparing an antigen for complement fixation diag- nostic tests. It is also suggested that a second antigen, prepared from a number of selected irregular strains, be employed in conjunction with the one prepared from the two generalized strains. 9. By the use of the simple sterilizable media, which have been described, these strains may be carried in laboratories with limited bactrieological facilities and the antigens prepared quickly, when- ever required. REFERENCES (1) Torrey, J. C.: Jour. Med. Research, 1907, 16, 329. (2) TraausE, Oscar AND Torrey, J. C.: Jour. Med. Research, 1907, 17, 223. (3) Scuwartz, H. J. anp McNetL, A.: Amer. Jour. Med. Sciences, 1911, 141, 693. (4) Huntoon, F. M.: Jour. Infect. Diseases, 1916, 23, 169. (5) Perry, M. W. anp Kotmer, J. A.: Jour. Immunol., 1918, 3, 247. (6) Sanns, J. E.: Jour. Immunol., 1920, 5, 97. (7) Hooxsr, 8S. B.: Jour. Immunol., 1916, 2, 1. (8) McCuintock AND CuLark: Jour. Infect. Diseases, 1909, 6, 127. (9) BurrEeRFIELD, C. T. anp Neriu, M. H.: U.S. Hyg. Lab. Bull., 1920, 124, 9. (10) Exser, W. J. anp Hunroon, F. M.: Jour. Med. Research, 1909, 20, 377. (11) Warren, S. H.: Jour. Path. and Bact., 1921, 24, 424. (12) VALENTINE, EvGENIA, AND Cooper, Greoraia M.: Jour. Immunol., 1919, 4, 359. (13) Hermaniss, Jonn: Jour. Infect. Diseases, 1921, 28, 133. (14) Hermantiss, JouHn: Jour. Infect. Diseases, 1921, 29, 11. (15) Mertnicksr, JoFFE AND FLEMMING: Zeitschr. f. Hyg., 1906, 52, 416. (16) JorrEn, K. W.: Munch. med. Wochenschr., 1920, 67, 1067. (17) Tutuocu, W. J.: Jour. Hyg., 1918, 17, 317. (18) Gorpon, M. H.: Grt. Britain Med. Research Cmm., Series 3, 1917, 99. (19) McIntTosH anp McQuEEn: Jour. Hyg., 1914, 14, 409. : (20) Matuers, G. AnD Herro.p, R. D.: Jour. Infect. Diseases, 1918, 22, 523. (21) Bemarns, T. H. C.: Jour. Path. and Bact., 1920, 23, 171. (22) Treacuez, O., anD McWitutams, H. I.: Jour. Immunol., 1917, 2, 383. (23) Pearce, Louise: Jour. Exper. Med., 1915, 21, 289. A SEROLOGICAL STUDY OF THE GONOCOCCUS GROUP 359 (24) Eastwoop, ArtHur: Jour. Hyg., 1918, 17, 64. (25) Grirritus, F.: Jour. Hyg., 1918, 17, 124. (26) Scort, W. M.: Jour. Hyg., 1918, 17, 191. (27) Waker, E. W. Anstey: Jour. Hyg., 1918, 17, 380. (28) Exserson, F.: Jour. Immunol., 1920, 5, 345. (29) Wiuson, M. A.: Jour. Immunol., 1920, 5, 499. (80) THomson, Davin: Lancet, 1918, 195, 42. (31) Coox, M. W. anp Starrorp, D. D.: Jour. Infect. Diseases, 1921, 29, 561. / : cee Vahl one AP A a, | J \ 0 eas | a Nee 9 a Ue §: ; nt : bs ' fF ii : i 08 iG CE" suo BUNyWOGNh: ate CUD CR OTe 4 Phe 4 wy een, Sooe a Ue eeb alt olan eaio wad A omen a3. a Cee 1 Ay 10 vlan, 24 geen ae? (ERI yeh AL MIRE hygtl wati RE ancora) ae IM ik bP one 'y ate: 9 Ra Oe re a a Cie? tt a ei 0 Ae A omen: pee LBa hg ro LE COOT. Torun ital, ok BA a iy | 105 BIOb denna perth | inten oP h 5 A nr eT ici bs 1 aie ae WW. oe ce) ae rere } Y irr j aes ce a rg i a Bir) Nien ty: 7 7 * X ns : j ; i] a 7 * a 7 ie STUDIES ON ACUTE RESPIRATORY INFECTIONS XI. A SEROLOGICAL STUDY OF ALPHA STREPTOCOCCI FROM THE UPPER RESPIRATORY TRACT" AGNES GOLDMAN AssIstep BY Betty 8. KoLcHIn Received for publication January 24, 1922 In a previous article of this series there was reported by Doc- tor Williams (1) and her co-workers a detailed study of the colony formation on standardized blood agar plates of Smith and Brown’s (2) alpha type streptococci isolated from the upper respiratory tract. The sugar reactions of these organisms were also given. The results of these studies showed a multiplicity of varieties of streptococci. In a continuation of this investigation, we have tested the serological relationships of a number of these strains by the agglutination and agglutinin-absorption methods. Very little detailed study on the serological relations of the green produc- ing streptococci seems to have been made and no minute study of the power of colony strains? to absorb agglutinins has been reported. Krumwiede and Valentine (3) have reported the results of direct agglutination tests made with members of the streptococcus viridans group obtained from both pathogenic and normal sources. These workers came to the conclusion that types pathogenic for human beings constitute, for all prac- tical purposes, a heterogeneous group. 1 One of a series of studies carried out under the direction of William H. Park, Anna W. Williams and Charles Krumwiede. The previous studies in this series were reported in the Journal of Immunology, vi, no. 1, January, 1921, and no. 5, September, 1921. This investigation was made possible by a grant of money from the Metropolitan Life Insurance influenza fund for a part of the expense. 2 By strain we mean, in this paper, a fishing strain, i.e., the direct descendants of any colony isolated from a plate made from a nasopharyngeal swabbing. 361 362 AGNES GOLDMAN In the following discussion, when we speak of the colony type of strains (I, IJ, III or IV) or of the sugar fermentation type, we refer to the classification in the previous report of Williams. SCOPE OF WORK AND DESCRIPTION OF STRAINS Study of the relationship of alpha streptococci by means of direct agglutination and also absorption of agglutinins, was made with the object of: First, detection of the presence or absence of groups which might indicate an epidemiological relationship to influenza and other respiratory infections. Second, detect- ion of possible correlations between colony, sugar fermentation and agglutination types. Third, comparison of strains from normal, cold and influenza cases. Fourth, making of a detailed study of one case (see case 154 below) to determine the close- ness of relationship between the strains from the nasopharynx of the same individual. Fifth, investigation, incidentally, of the more commonly encountered difficulties that limit the value of tests for the differentiation of the particular type of organisms studied. Work was done on strains from 33 cases classified as follows: 16 influenza patients, 2 doubtful influenza patients, 6 cold cases, 6 normal individuals, 2 measles cases, (cultures from Tunni- cliff (4)) and an influenza case studied by Doctor Rosenow (5). Altogether, tests are reported on 112 strains isolated from the above sources. Serum §stratns The sera studied were obtained by immunizing rabbits against each of seven strains. All seven serum strains except one, showed type I colonies on Smith and Brown blood plates; the exception 140-(1 complex) showed type III colonies The serum used for the greatest number of tests was 154-25. Case 154 (typical influenza) was the case selected for intensive study. Strain 25 of this case was one of a group of fishings (19 to 25 inclusive) of the mitis type. These fishings gave the ALPHA STREPTOCOCCI FROM UPPER RESPIRATORY TRACT 363 same colony formation on Smith and Brown (2) plates and identical sugar reactions. Planted on chocolate agar, these organisms did not change the color of the medium except that after long cultivation one of the group, strain 154-(21), did show a very faint green. On the Smith and Brown plates they gave small, faint but definite green zones in forty-eight hours at 37°C.; these zones did not change at icebox temperature nor on further incubation at 37°C. These fishings thus approached Smith’s gamma type of colony. This group was further characterized by the homogeneous suspension which resulted when the growth from solid media was rubbed up in 0.8 per cent saline. This was in marked con- Strains used for production of sera STRAINS SOURCE 154-(10) Nasopharyngeal swabbings from: 154-(12)8 ; : 154-(18) Typical influenza cases 154-(25) 140-(1 complex) 108-(9) Doubtful influenza case 67-(2)3 Normal individual trast to the results obtained with some other strains from the same case and from different cases as described later. The stained films of members of this group after twenty-four hours’ growth on chocolate medium, showed Gram-positive, medium-sized cocci with a tendency toward amphophile colora- tion. The organisms, in contrast to the majority of strains studied in this series, were morphologically comparatively regu- lar. In salt-free phosphate broth*t they grew in short chains. The length of these chains varied considerably on different occasions; sometimes there were many single organisms or groups of two and three scattered among short chains; at other times 3 These two sera were produced by Florence Bittman. 4 Salt-free beef broth containing 0.2 per cent sodium phosphate; reaction, pH 7.8. 364 AGNES GOLDMAN there were very few single organisms and the chains were some- what longer. Strains from two other cases, namely, 108-(9), 108-(14), 109-(1 and 2) resembled group 154-(19 to 25) in all the particulars (mentioned so far). Strains 100-(2), 100-(3) and 90-(11) were also similar to this group except that they had a tendency toward spontaneous clumping in saline and long chain formation in beef broth. In contrast to the group (19 to 25) of case 154 all the other colony strains of this case, produced a marked green color on chocolate agar and showed great variability in certain other properties. For instance, in the films of 154-(18) there were often present many irregular organisms which assumed diptheroid shapes. At times these forms were almost absent, at other times they were present in large numbers. Strain 154-(10) also varied in its susceptibility to suspension in saline; sometimes the sus- pension was quite homogeneous, again it showed a tendency toward spontaneous clumping. Irregular organisms were found occasionally in films of a strain from another case, 152-(6). This strain was also green producing on chocolate agar. Strain 140-(1 complex)® was a fishing from strain 140-(1) (influenza case). On the original Smith and Brown plate, com- plex colonies of the rosette type were interspersed with a cer- tain number of whetstone colonies. A year ago fishings were made from both types. Plates made from these sub-strains were practically homogeneous as to the type fished. After a year’s cultivation on chocolate medium, 140 complex showed both complex and simple colonies. A still later poured plate showed only simple colonies. In our tables we have retained the name 140-(1C) in order to distinguish it from the original mixed strain, and from the sub-fishing 140-(1 simple). The suspension of 140-(1) in saline was homogeneous, and the film showed Gram-positive cocci growing in chains. Strain 67-(2) was from a normal individual, a member of the Metropolitan Life Insurance force. The film showed Gram- 5 Complex refers to Smith and Brown complex types. ALPHA STREPTOCOCCI FROM UPPER RESPIRATORY TRACT 365 positive cocci growing in chains and the suspension in saline was thoroughly homogeneous. DIRECT AGGLUTINATION Our sera were obtained by inoculating rabbits with graduated doses of living twenty-four-hour cultures grown on chocolate agar and suspended in 0.8 per cent saline. Nine to twelve doses were given as follows: One dose on each of three consecutive days; then, after an interval of three to four days a second dose and so on. The largest dose was usually about half the growth on a chocolate agar slant ina 6 by 1 inch tube. The sera usually caused complete agglutination through dilutions of 1:600, 1:800 or 1:1000 and partial agglutination at considerably higher dilutions. Two methods were used in preparing the bacterial suspensions: First, the culture to’ be tested was transplanted daily on fresh chocolate agar (with some water of condensation) for several days before the test. Finally, an eighteen to twenty-four growth on the chocolate agar was scraped off and carefully rubbed up in 0.7 or 0.8 per cent saline solution. Sometimes, for reasons given below, salt-free beef broth containing 0.2 per cent sodium phosphate was added to the suspension. Effort was made to have the density of the suspension such that, even with some precipitation in the saline control which frequently took place after refrigeration, there remained a definite but not dense cloud in the agglutination tube. Second, when spontaneous or self- agglutination occurred, the growth (as above) was suspended in freshly prepared salt-free phosphate broth instead of salt solution. The serums, preserved with 0.3 per cent chloroform, were diluted with 0.7 or 0.8 per cent saline. The tests were set up with 0.1 cc. serum dilution and 0.9 cc. bacterial suspension; each strain was set up also with normal rabbit serum and a saline suspension as controls. The tests were incubated for two hours in a waterbath at 45°C. for the earlier tests; at 50°C. for the later ones. They were then put in the icebox over night and read the following morning. 366 AGNES GOLDMAN Difficulties encountered The greatest difficulty encountered throughout the tests was that of controlling the tendency toward spontaneous agglutina- tion in many strains of the streptococci studied. Most of the strains chosen for the production of serum did not agglutinate spontaneously, but this was far from being the case with many of the other strains tested with the serum produced. ‘The organ- isms appeared to be extremely capricious. A method that gave fairly satisfactory results at one time was useless at another. An organism which sometimes gave in salt solution a stable homogeneous suspension, at other times would gradually floccu- late on further standing. Whether this irregularity was due to slight differences in the suspending media, solutions and manip- ulations, or to more or less spontaneous variations in the or- ganisms or to other causes was never satisfactorily determined. To overcome the tendency to spontaneous agglutination, we tried various methods of suspension. On the whole the most satisfactory menstruum for these self-agglutinating organisms was a freshly prepared salt-free phosphate beef broth with a reaction of pH 7.8. Sometimes organisms which had once formed large clumps in saline solution, would remain evenly distributed throughout the phosphate broth after being shaken and allowed to stand over night. At times a reading was found possible with many refractory organisms by growing them over night in the phosphate broth and using the decanted cloudy upper por- tion for the test. However, the suspension in this portion was often too thin to make a satisfactory reading; at other times when it was heavier, a very slight cloud seemed to persist in the test tubes, even when there was reason to believe that complete agglutination would under other conditions have taken place. For this reason the saline suspension was preferred whenever possible. In certain suspensions it was found advantageous to use a salt solution with the addition of a small portion of broth, just enough of the broth being added to prevent self-agglutina- tion but not enough to interfere with the final reading of the test; one-third phosphate broth was usually sufficient to prevent flocculation in suspensions where this was apt to occur. ALPHA STREPTOCOCCI FROM UPPER RESPIRATORY TRACT 367 Notwithstanding all the difficulties in reading the reactions, after considerable experience with these strains some criteria of differentiation could be established. It has been said that readings were usually made on the fol- lowing morning, because the results were much more clear-cut after this interval. With some of the self-agglutinating strains grown in broth, however, just as others had found with self-agglu- tinating bacteria, a fairly accurate reading could be made after incubation in the water-bath, while on the following morning the controls showed complete agglutination. The character of the glassware seemed to be an important factor, as the following experiences show. A series of absorp- tion tests was unsuccessful because precipitation took place in all the controls, even with organisms usually giving satisfactory suspensions. ‘These irregularities occurred with one particular batch of saline, which was full of crystals. As other workers reported no irregularities with either the NaCl or distilled water, we concluded that the spontaneous agglutination might have been produced by the formation of an electrolyte from the dis- integration of the glass. Readings In the tables, complete agglutination with a perfectly clear supernatant fluid is represented by +, while +, x, and tr. represent increasing degrees of cloudiness, and — indicates the absence of all agglutination. Frequently sedimentation took place, but after shaking the tubes it was found that there was no macroscopic sign of agglutination. Because of this sedimenta- tion, whenever there was any cause for doubt, the tubes were shaken up after the preliminary reading and the reading often had to be modified in consequence. On account of the tendency toward spontaneous agglutination of so many of the organisms, it was believed that slight distinc- tions were important, so that the sign < has been used after a symbol as a modification to indicate that the result was less sharp than the symbol alone would suggest. 368 AGNES GOLDMAN Interpretation In recording our results we have taken as a positive cross- agglutination, a reading that gave a + or + < at 1:100 and +, or less, in the normal serum and saline controls. It is probable that complete agglutination with a highly sen- sitive antigen at 1:100 is no indication of a closer relationship than partial agglutination at the same dilution with an antigen that flocculates with difficulty. This fact somewhat vitiates general conclusions based on direct agglutination tests with alpha streptococci. Table 1 shows the results of direct cross-agglutinations with strains of case 154 and of other cases chosen for study on account of their clinical diagnosis or their colony type. It is seen that sera 25 and 18 of case 154 both agglutinate the same strains of the case with the exception of 154-(15 and 16), spontaneous agglutinators and therefore capricious. Another exception is 154-(7) which gives a strong reaction with serum 18 and a weak one with serum 25. In the few tests made with other cases, serum 18 also agglutinates the same strains as serum 25. Serum 12 (case 154) tested only on strains from its own case, aggluti- nates none of the same strains as sera 25 and 18. Table 2 sums up the results with the principal sera with which direct agglutinations were carried out. We note that 154-(25), a type I strain from a typical influenza case, produced a serum that gives direct agglutination with about one-third of all of the strains, from one-half of all the cases tested. Strain 140- (1 complex), a type III strain from a typical influenza case, pro- duced a serum that agglutinated about one-eighth of the strains from one-fifth of the cases tested. The serum from 67-(2) a type I strain from a normal throat agglutinated only one strain from a different case, 100-(3) in table 2. In table 5 are listed the number of cases which have strains tested with each one of the three principal sera and the number of cases which have strains agglutinated by the same sera. ‘These cases are listed acording to their clinical diagnosis. We are justified in concluding that no definite relationship is indicated ALPHA STREPTOCOCCI FROM UPPER RESPIRATORY TRACT 369 between clinical diagnosis and agglutination type if we keep in mind that more influenza than cold and normal cases were tested, especially with sera of strain 154-(25). Table 4 shows the clinical distribution of those cases in which the sera 154-(25), 140-(10), or 67-(2) agglutinated fully one or more strains from a case, but did not agglutinate all the strains which were tested from that case. It can be seen that although 18 cases had one or more strains agglutinated by one of the three test sera, 11 of these cases had other strains that were not agglu- tinated by these same sera, thus giving evidence of a multiplic- ity of agglutinative types. Table 2 shows us that the influenza case 154, 22 strains of which were tested with sera 154-(25), 154-(18) and 154-(12), and the influenza case 140, 8 fishings of which were tested with serum 140-(1¢e) each have at least two entirely distinct agglutination types, while the normal case 67, of which 4 fishings were tested with serum 67-(2), has only one type. As far as comparisons can be drawn from so meagre a number of strains and cases, these results together with those just cited from table 4, corres- pond with those obtained from the cultural study of strains in this series of cases; that is, although the strains isolated from cold and influenza patients may be of several types, those from nor- mal subjects are more apt to be all of one type. On the basis of colony type no clear-cut relationship is indi- cated in table 1. Serum 154-(25) (type I) agglutinates some strains of type I, II, and III and a doubtful type IV, though more of type I. This is to be expected, as the type I strains were found to be far more numerous and were present in a ereater number of cases. Serum 140-(1c) (type III) agelutinates several types also. The one strain from a different case (100) agglutinated by serum 67-(2) (type I) is likewise of type I. It is a strain that was also agglutinated by serum 154-(25). It is important to note that the sugar types in this series do not seem to run parallel to the serological types. However, it is a striking fact that with all the strains previously discussed, which resemble somewhat the gamma type, namely 154-(19) to (25), 108-(9), 108-(14), 109-(1 and 2), 100-(2), 100-(3), 90-(11), the sugar reactions are identical. THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 4 STR REO Se “He -|- +/+ ]+ -|- +/+ | + -|-|-o}x]+]/4+}+]-]- +}+ ]+ -|- -j}-/|-/]-|-|-}s=}+]+]4+]-]- =| =| + =| X Srelleeiiene (oc [Wert Sra het iste Poets [ees ile Seal ae -|- -|-|-|-|-]-|x]4+]4+]4+]} | - =| +] + + |x x|x{|xf—}—fualx]|tl+ 4) x fo +) +] + -|- yl }o}—}—| =} e}+]4+}+]-]- +.| +] + -|- -}-|-|-|-|]x]+]4+]4+]4+]-| +}+ | + =—}— —f—]—]—] ele Pelee eae a md oe es st slbes x} x] x}-f-} +] +} at) +] — apf} oo] a] x ed - = ye =a ae Te a og = a - -— =F - - - 4} 9 x| x1 x] + +J+{/+}+]+}+]/4+]}/4+]+]4+ ]+ -—|- XIX) x] —}—]ao) x] we] te] te] do 4) X| = Sails ae (bord ae Sals Sih Sal GaP ea lecai Sp ea eal) al ee =|) Se Sahara Wesco SeleG SIE Se sll oil feared Reeameel arb ume ec farce Peta 14 | “39 Xp el el x] xi +l ty +i t+] 4+)—f-f+]4+]4i+ 13) 9 x| xX] xX}—putel +t} 4t+i]4+]4+]/-[-] x] =] =| = -| 4 yyy] a} — | — 1x] +] elt yey} pp} x | x] = 9] 9 thei tl] — fof} oo} alse} x} —}—|—-} yey = +J/+]4+]x]}—-|x]x]|x{[x]x]+}+}4+]4+]4]+ 44] ‘14 om NF AE YT OO 1s Na | >< VN Ge [| tae <> a Pal | +{|+ +] + | + +|=]/+]+]+]+ +|+ +] -+ | + +/+] +] +] +] + MD a, 22] Ls tT! Fa | 8 Bs E | § ® |B gales Piel e 008 | 009 | 00¥ | 002 | 00T = | 008 | 009 | 00% | 00¢ | 001 e 008 | 009 | 00% | 002 E 5 5 ees ZI-PST WouGs eek, SI-PST WaUds pps Cg-poy] WaAUNaS DLIS JUQQDL JUaIvAOUOUW YPN WI0d0}da.ys DYdzo fo UoyDuUrIN BBD 199LUuT a 1 @Tavt Lytttaytt++y ttttt+ettts+it+ 001 “AYBS “ANSS “AILRS “ATIBS st STW SIT SIV SIN STH. “ATIBS “ATT’S “ANTES “AITeS “ATTBS “AYBS SnABUST SNABUB] “AITES BATTS “ARS SNABUST SNAVUST “AIRS “ATTRS “AITES WdAD uyoos _ OO Ot St | | WdAw ANOTOD Bzuonguy ezuenguy Bzuenguy Bzuenyuy BZUON UT BZUON BUY B2UON GUT RZUON GUT BzUON uy vzuonyuy Bzuenguy BzuEnyUuy BZUONGUT BZUON UT wzuonguy uzuenyuy wzuonyuy BzuONn Buy Bzuen guy ezuonguy ezuenyuy BZUON UT BzuOnyUy BZUON Uy wzuenguy uzuenguy SISOND VIC NIVUIBS TST 0&1 PST aBVO 370 Re 2-19 WaUaS ol — = ~ “1 +)+ Ty | 1} | “19 +) + ]+ +{+]+ +/+) + pear (esis I} ] = x |X 1} | ‘14 "000 || "19 q} | ‘44 x |X OT-0FT WaUas Stelle to liects al ae ar +++4+1 x 4 s+1+ 4 + XX 14 “17 fap |! ar 24 Ss +4+X+ 14 Sr Be eck SIT “ATT’S SOI STL “ATIBS STW “ATS “ATTBS *ABUST *ABUSI STU SIT SILA SOTA STUN SOTA “ATIBS SIFT “ATIBS, “ATTI®S “AITBS STULL SIQTTA, “ATTRS SIAL SIT snumbiay II *[(OULOFY-UON, SI “ATTRS “ATICS (3) STA “ATTES TSS ATTES II — _ ezuonguy Bzuonguy BzueNguy ezuonpuy ezuonguy ezuonyuy ezuonguy ezuanguy ezuonguy ezuonguy ezuonguy ezuanguy ezuonguy ezuonpuy ezuonpguy ezuenguy ezuenguy ezuonguy ezuonguy ezuonguy ezuonguy ezuenpuy ezuenguy ezuonyuy ezuonguy ezuanguy ezuonyuy *ezuonguy ezuonyuy ezuonguy ezuonguy ezuon uy ezuonguy ezuanguy ezaonguy 61 ae at 6ST SST OFT L¥I 8c1 SOT sI9T TOT 371 el OS oN Sal core PSH Pail leeSal le side am S| DSP oma ones Pea — | Gall ae = —|X)| = 008 | 009 | OOF | 002 | OOT XXX | xX X I XXX | xX xX x x ++ ++ 4 xXXX I 008 | 009 | OOF | 002 | OOT | | | | | —-|/—-|-|-|—-|x]+ -|-|-|-|-|x|x UR SEG ea oN eel teallh ate Se || See fee aiee al Salsa oe ‘ml—|x|x}|x| x] = sr |osc tam: |e >+| + TT | ste se) se ee AN a al ig are Ge "23 | | + Se | a este oe Se oor | am eee i= ete |e tlre =| — |e) epee] Stee [ asses te: |) Stel cctea teste] lio 4+)4+]4+]+4+ I [BuION 8 - —-}-}-/]-}-/]-]}-]-/]-]-]-]-|]-]-]-|-3}x]4+])+ I [euI0N g - —J— f=] =f) — pap Ht a] Hf He — | — pe] KY se te snuinbqy | [I puey [8uION z wf - >H>t +f ti +}—]-}-};-}-}-|-]7]-]-f-}-]-}-(- SHI | I [euro N Ir = Seth st |store | | a || SIH] I [BuLION z - >+4+]4]}/4]4+]-]- -|-|-/}-}]-|-]-|-|]-|]-]|- sIuW| I [eurI0N, I 19 Seber eseiese |p|) Ged ee SHIN | I [BuION I 06 - -{-|-|-|]-]-]- —-{/-|-|-|-|-|-]-|]x]+] + SH | = IT PICO é ~ -|-}|-/-]|-|-]- —-|-|-/]-]-|]-|‘3]-o]/x]x] = sw} IT PICO I 69 - —-{|-|-|-|x|/-|- rap fy Sar litre Neca eal fe | leerl (RG ea syeoag | = PICO € = —|—|—|xX] =] -|- ap} 0 fad] ae] — | — [>t] >4+] 4) 4) 4+ SIH) =I PICO I 1% ete a0 ear op Pease bart eset ae aed acd eet et el [ eet) ee] me || aa! Seale ‘Ayes | oT Le) Or - —-|-/-/]-]-/]-/}]-/]-}-]-]-]-}-|}-]-]-/]-|]-|- sw | PIOD 8 373 374 AGNES GOLDMAN TABLE 2 Summary of results of direct agglutination tests with sera 154-25, 140-1C, 67-2 NUMBER OF STRAINS SHOWING NUMBER OF STRAINS POSITIVE Se AGGLUTINATION AT 1: 100 RPS SITE AE PE aE ee EN he ee es | ae oa aaa or cases | HAVING ANY STRAINS Number AGGLUTINATED T not from TeSTay ar 1:100 otal naeRtOt Total Number not from number sonra number | case of serum strain strain number per cent number | per cent 154-25 103 82 39 26 Shed 32 15 46.9 140-1C 58 50 10 6 12.0 20 4 20.0 67-2 40 36 5 al 2nd. 18 1 oEe TABLE 3 Clinical distribution of those cases having one ormore strains agglutinated at 1:100 by sera 154-25, 140-1C, or 67-2 INFLUENZA COLDS NORMALS MEASLES Cases Cases Cases SERUM Num- ° um- = Num- : Number of cases Cases having __ | ber of baying ber of Baya ber of beving: tested agglutinating strains| cases ager cases |288! cases | 88" tested nating | tested nating | tested nating strains strains strains 154-25) -16 | 2,D.1-* 5 2 Da 6 4 6 3 1 1 140-1C} 8 1 6 2 5 1 1 0 67-2 Se DE. 1 5 0 5 0 0 0 *D.I. = Doubtful influenza. TABLE 4 Showing the number of cases in which a multiplicity of agglutination types was found among the strains of the same case NUMBER OF CASES TESTED AS NUMBER OF CASES IN WHICH TO SEVERAL STRAINS, maar tees eacace ICES SOME BUT NOT ALL HAVING ONE OR MORE AGGLU- = STRAINS TESTED WERE TINATING STRAINS ee ee ee AGGLUTINATED BHRUAH yi| fe foe ee ee = a a Influenza 2 /é& Influenza an Influenza 3|é& ae S| 8 |e 54a eon |e er BAR 2D eet 151° 3.) SSD ale oo tk 140-1C| 2 1 1 0 Oo ees? ig ft 67-2 0 Os) Ci co oO} 0) 0 0; 0 Total He yn 4D eet ial aA t ay eal An *D.1. = Doubtful influenza. + Serum strain. 1:50 +++++ +++4++ ++++ +++++ DIRECT AGGLUTINATION 1:100 ++++ +++4++ aril begin il ams 5s) Wee es 7S teuied ar ae tes ag haste |oce ath feel Ma ae bales sa Wa a “ip bea Mere Feely lege ramet cath oe “ae gual he = + |+<| X et arattar new Cm oS 37 ee bs ae tr. Saline tr. tr. TABLE 5 AGGLUTINATING STRAIN 154-25 154-19 154-21 154-10 154-18 154-19 154-25 154-21 154-25 154-10 154-25 154-18 Agglutinin absorption tests with sera of rabbits inoculated with strain 154-25 AGGLUTINATION BY ABSORBED SERUM i=) asd — a (b) Rabbit ¢ 154-25 157-3B 108-9 100-3 157-3B 154-25 108-9 154-25 100-3 154-25 (c) Rabbit e 154-25 109-1 and 2 152-6 157-3B 154-10 109-1 and 2 154-25 152-6 154-25 109-1 and 2 |+< 157-3B 154-10 154-25 154-18 154-18 154-25 375 tr. tres Xx a tr. a tr. + bir a +X+14+X81X1 1: 100 +574+14+X81X1 ae Bit ae a +15X1# I 1: 200 1:400 tr: 1:800 Absorbing strain (a) Rabbit ¢ 154-25 154-19 154-21 154-10 154-25 154-19 154-25 154-21 154-25 154-10 154-10 154-25 157-3B 108-9 100-3 154-25 157-3B 154-25 108-9 154-25 100-3 154-25 109-1 and 2 152-6 154-25 109-1 and 2 154-25 152-6 152-6 152-6 152-6 (d) Rabbit a 154-25 154-18 154-25 154-18 376 AGNES GOLDMAN TABLE 5—Continued DIRECT AGGLUTINATION AGGLUTINATION BY ABSORBED SERUM = rs] AGGLUTINATING| "=| laced 4 bal eam I= STES Sl Sl Sls sealetbeortmne Snes se lass) |e a | ost SiS jes dks ie strain = 3 an = = — =, = Zi NM i = = — i (e) Rabbit a et Se ee a ea ee ie Ue os ad ees ee hea Ae ees = =P se bag jae yl es |i = 13(-5 tr.|tr.} —|—] — 130-5 130-5 tr.} —}; —|]—-—|]-—- 154-25 154-25 = Sloss: |< lane 130-5 (f) Rabbit a Sei) se Pam lleae lem: ieee fale 154-25 —-;)}—-}]—-}]-)]-— 154-25 =) ics (all Mae nA Mkt Mise Uo lant ee oo Pass —-};—-}—-)}-—-|]— 154-25 194-25 |+/+/+/+]/+] JS. Discussion There seem to be related agglutinins present among a number of strains of alpha streptococci. These agglutinins must be present in considerable quantity, since no account has been taken of any reactions but those complete at 1:100 dilution. It is possible that with sera produced against a sufficient number of strains, a classification of all alpha streptococci according to agglutination type might be possible. The presence of an epidemic strain among the organisms iso- lated from influenza cases, is not indicated for several reasons. In the first place not enough strains were agglutinated by any one of the five sera tested to suggest that any of the serum strains represented an epidemic strain. However, a sufficiently marked relationship was present to rule out the possibility that the serum strains chosen were chance invaders which happened to persist in the presence of an epidemic strain. In the second place, the strains showing relationship with the serum strains, were distributed among all the clinical types studied, that is, influenza, measles, colds and normal. Two influenza sera agglutinated a number of normal strains and, therefore, sera against these normals might have agglutinated ALPHA STREPTOCOCCI FROM UPPER RESPIRATORY TRACT 377 influenza strains which makes less significant the lack of rela- tionship between the normal strain 67-(2) and all other strains tested with the exception of one, 100-(3). Nevertheless, it is worth noting that Krumwiede and Valentine found that several strains from normal tonsils were agglutinated by sera produced by pathological strains, while the sera produced by these ton- sillar strains did not agglutinate pathological strains. Krum- wiede and Valentine attributed this phenomenon partly to the greater tendency of the normal strains towards spontaneous agglutination. Although many of our pathological strains were self-agglutinating, it is a fact that out of the four normal cases which had strains agglutinated by our two influenza sera, three cases had only self-agglutinating strains; however, in the tests recorded on these strains, it was possible to control the antigen successfully. On the whole, there may be a real significance in the absence of cross-agglutination by the normal serum. In the third place, both influenza cases, that is, 154 and 140 showed the presence of at least two serologically unrelated al- pha streptococci. Some, but not all of the strains from several other influenza cases, when tested, agglutinated with the same serum. ABSORPTION OF AGGLUTININS Absorption tests were carried out with a number of different strains which had showed marked direct agglutination reactions. Besides the five sera recorded in the direct agglutination tables, two additional sera were used, one from strain 154-(10) and one from strain 108-(9); the latter as previously stated, was very similar to 154-(25) in a number of biological characteristics, as well as in its high direct agglutination by serum 154-(25). Case 108 had been diagnosed as one of doubtful influenza. Method With influenza strains grown on solid media it was difficult to get a sufficient number of packed cells from young growths. The organisms were, therefore, grown overnight at 37°C. in 378 AGNES GOLDMAN bottles containing 200 cc. of salt-free phosphate beef broth. The broth cultures were centrifuged at high speed in 50-cc. containers. ‘The organisms were then transferred to a graduated tube and again centrifuged at high speed after which the super- natant fluid was decanted; the right amount of serum was added to the packed cells to permit complete absorption to occur with the homologous strain whenever possible. Saline was now added to give a serum dilution of 1:10 when the serum titer was 1:800 or over, and of 1:5 when the titer was 1:600. The suspension was then incubated for three hours with frequent shaking at 50°C. and put in the icebox over night. The following day, agglutination tests were carried out with the centrifuged supernatant liquid. In adding the serum to the packed cells, we calculated as Valentine and Cooper (6) had suggested, that one half the amount of the reading might fairly represent dried organisms. Difficulties encountered In carrying out the tests it was frequently difficult to get com- plete absorption in the controls. This may have been due to technical shortcomings, or to the self-agglutinating nature of the antigen. The commercial centrifuge tubes are often care- lessly graduated so that the readings on the packed cells and on the saline added are inaccurate and, therefore, not comparable in the different tubes. When the quantities are small the error may be significant. One of the most difficult factors to control is the varying sensi- tivity of different antigens. For instance, a certain amount of antigen A may absorb all the agglutinins of the homologous serum; the same amount of antigen B may leave unbound agglutinins which in the subsequent agglutination test, combine with B. Of course it is possible in such a case that serum A has more agglutinins for B than for its homologous strain, but from our tests it seems far more likely that the result is due to the greater sensitivity of antigen B to all flocculating agents, bringing about an agglutination in the presence of a very few unbound agglu- ALPHA STREPTOCOCCI FROM UPPER RESPIRATORY TRACT 379 tinins which would be insufficient to affect A. It is only by a very careful comparison of controls and by cross-absorptions, that a correct interpretation of such contrasting results is possible. As was pointed out by Valentine and Cooper (6), the danger of non-specific absorption with too large a volume of packed cells is not great. We made the reverse test by treating our packed cells with increasing doses of serum and found that our largest and our smallest serum doses gave similar readings well within the limits of error. Results In table 5, a, b, c, d, e, and f represent six different sets of absorptions with serum 154-(25) obtained from two different rabbits. Some of the tests with serum of rabbit A were among the first carried out; the lowest dilution was 1:100. Afterwards, it was found better to make the lowest dilution 1:50 if the serum had a titer of 1:800 or more, and 1:25 if it had a titer of 1:600. Strains 154-(10), 154-(18), 152-(6), 157-(3B), 130-(5), J. S., (measles) were all green producing on chocolate agar. Strains 154-(19 to 25) 108-(9), 109-(1 and 2) and 100-(3), as has been previously stated, did not produce green on chocolate agar. The following tests were recorded on each strain tested, for example, ‘‘strain x”: 1. Agglutination by unabsorbed serum. 2. Agglutination by serum absorbed by ‘“‘strain x.” 3. Agglutination by serum absorbed by a serum strain. 4, Agglutination of the serum strain by serum which had been absorbed by ‘‘strain x.” The following facts were noted: With all but one of the strains tested there was some absorption of agglutinins shown by the fact that the titer of the absorbed serum when tested with its homologous strain, was greatly reduced. This reduction could not have been due merely to mass absorption, for serum 154- (25) when absorbed by a non-agglutinating strain and then tested with strain 154-(25), gave complete agglutination in a dilution 380 AGNES GOLDMAN of 1:600 and a marked reaction in 1:1000. Therefore, the ab- sorption was probably specific. When entire mutual absorption was used as a criterion, identity seemed to be complete for the two strains, 154-(19) and 154-(21) belonging to the same group as 154-(25) (table 5, a). Tests on other strains of case 154, namely, (20), (22) and (24) (not recorded), gave similar results. Serum 154-(25) absorbed by 154-(10) and 154-(18) which were green producers giving a full direct cross-agglutination, brought 154-(18) down completely through a dilution of 1:100, and 154- (10) through a dilution of 1:50 (table 5, a and d). This serum absorbed by strains from other cases, similar to 154-(25) in not producing green on chocolate agar, namely, 108-(9), 100-(3) and 109-(1 and 2), agglutinated these strains almost completely at a dilution of 1:200, (tables 5, b and c). With cross-agglutinating strains from other cases producing green on chocolate agar we obtained a gradation in the results. J. S., a measles strain, though completely agglutinated by serum 154-(25) did not ab- sorb any agglutinins from this serum. Serum 154-(25) agglu- tinated its homologous strain almost completely through 1:400 after absorption by 157-(8B). Absorption of this serum by strain 152-(6) left agglutinins to about the same extent as absorp- tion by 154-(18), while absorption by 130-(5) removed most of the agglutinins (table 5, f, b and c). We observed further, that although absorption of serum 154- (25) by 154-(10) removed the agglutinins from 154-(18), absorp- tion by 152-(6) removed them scarcely at all for 154-(10), only partially for 109-(1 and 2) and almost completely for 157-(3B). Another important point is that although this serum 154-(25) when absorbed by strain 154-(10), lost most of the agglutinins for the homologous strain except in the lowest dilution, it lost them only partially for the absorbing strain, and that the serum, fully absorbed by its homologous strain, 154-(25), retained suff- cient agglutinins to cause almost complete flocculation of strain 154-(10) (table 5, a). This is another illustration of the greater sensitivity of some of our antigens to traces of agglutinins in a serum, or in fact to any agency that favors flocculation. The results with serum 108-(9) are given in table 6. We see ALPHA STREPTOCOCCI FROM UPPER RESPIRATORY TRACT 381 that both strain 154-(25) and strain 152-(6) are agglutinated as completely as the homologous strain. We also observe a very pretty grading in their capacity for absorbing the specific agglu- tinins of the serum strain. Absorption is practically complete with one strain of the same case, 108-(14); with 154-(25), simi- TABLE 6 Agglutinin absorption tests with serum 108-9 (rabbit h) DIRECT AGGLUTINATION AGGLUTINATION BY ABSORBED SERUM Fev top | =< Preto Ss ere z 2 amamw! |olelslele eee ene NeLS lend Sle de sit sulees Womens 7) arog pee eg nt Uc Ue ee i Ml (UU Rai ere ene EN Ne ht) gogd4)tes te | (|| tee 4 a] 4] 4) 4 [eel] — | — | 154-25) tr.| — | -— | — | — | — | 15425 Ripe el ely c]— | — | 52-6) | | 4) a) a) | 162-6 TOSSA beet ee Ne es eae aT eBeO mre hs one (dea ese |e Pea mac feces Fret hog Mg Sia fa PAP pe Teco, ei ee ee) a en de Teo B te, oe eee ee Os 9 eg WE MPN Eis 1 a 5G lar in several respects, it is less complete; while with 152-(6), more markedly green-producing, there is no absorption apparent in the lowest dilution of the absorbed serum, but complete absorp- tion from 1:200 upwards. Table 7 records a set of absorptions of serum 140-(1 complex). In this table is recorded the only instance which we have ob- served of a practically complete absorption of a serum by a strain from another case. The organism in question 170-(3A) had a slight tendency toward spontaneous clumping. It may be worth while to quote from our notes on this test: Strain 170-(3A) was grown in phosphate broth. After incubation » and a night in the refrigerator, there was almost complete precipitation in all the tubes with a very faint haze in the controls. After shaking, 6 When this serum was used for absorption tests it was already several months old and was greatly reduced in titer. 382 AGNES GOLDMAN neither the serum nor broth control showed any vestige of agglutina- tion, nor did any of the tubes containing absorbed serum; but agglutina- tion was marked with the unabsorbed serum, so that the reading was absolutely clear-cut. We observed that this serum absorbed by 170-(8A) and cross- agglutinated with strain 140-(1 complex), showed a trace of agglutination in the lowest dilution tested. Because of the deterioration of the serum and of the fact that we had no serum produced by strain 170-(3) with which to control our results, TABLE 7 Agglutinin absorption tesis with serum 140-1 complex (rabbit e) DIRECT AGGLUTINATION AGGLUTINATIGN BY ABSORBED SERUM AGGLUTINATING | o o ° g 3 2 ae ra S So S Absorbing S S = & = AEN a 3B = R = strain S = Si = Se a a “4 = 5 = = el A il bp aes WS cael P22 ee 140-1C xX) XxX |r} —] = 140-1C ea? iva ct eee ep 140-4 —-}/-—-/;/—-}]}—-|- 140-4 140-4 —-!;—-]-/]-|- 140-1C 140-1C =X st. | — a 1404 ates ete eo Se 140-1C —-/—-}-|-|]- 140-1C =e Ger eee = eel ee ie 170-3A —-}/-—-;-|-|- 170-3A 170-3A —/—-;}—-}]-f- 140-1C 140-1C tr.| —|—|—-—-|] - 170-3A we do not attach great importance to this single case of identity of two strains from distinct sources. It is interesting to note, however, that 170-(3A) shows a tendency to produce type III colonies on a Smith and Brown plate, thus resembling strain 140-(1 complex). Comparison of strains from the same case We stated at the beginning that one part of our problem was the detailed study of a single case. This is of value because the degree of relationship among cross-agglutinating strains of the same case, Judged by the criterion of absorption, helps us to interpret results with strains isolated from different cases. Such ALPHA STREPTOCOCCI FROM UPPER RESPIRATORY TRACT 383 study is of assistance in distinguishing between an organism that tends to preserve fixed biological characters and one that is constantly subject to slight modifications. We chose for this study case 154 because it was a well marked case of influenza and all of the fishings were of the same colony type on blood agar plates. We found that the group 154-(19 to 25) absorbed completely the agglutinins of serum 154-(25). Absorptions of serum 154-(25) by strains 154-(18) and 154-(10) were incomplete. To make further tests we also obtained sera from strains 154-(18) and 154-(10). The full results of our various absorption tests are recorded in table 8. The results are somewhat difficult to interpret because the strains represent antigens of such varying sensitivity. All three strains are un- doubtedly very closely related. For instance, as previously noted, although strain 154-(25) absorbs all of its own agglutinins from the homologous serum, the serum so absorbed remains able to agglutinate the more sensitive strain 154-(10); that is, the same volume of packed cells which is sufficient to absorb all its own agglutinins from a certain serum dose when we use cells of strain 154-(25), is inade- quate when we use packed cells of strain 154-(10), to achieve the same result. If, however, the serum did not contain agglutinins in larger number or more specific for 154-(25) than for 154-(10) the same volume of cells of the latter strain should remove all agglutinins from serum 154-(25). Since it does not do this, we conclude that strains 154-(25) and 154-(10) are very similar but not identical. The serum 154-(10) at first suggests a clear-cut interpretation. Strain 154-(25) shows a low direct agglutination with serum 10; strain 154-(10) absorbs all the agglutinins for strain 154-(25) and strain 154-(25) only partially absorbs them for strain 154- (10). But our conclusions are modified when we observe in table 9 that strain 154-(10) has a higher titer with serum 154-(25) than the homologous strain, and also higher than it has with its own serum. We might, therefore, expect that serum 154-(10) would agglutinate strain 154-(25) in low dilution and that its agglutinins would all be absorbed by strain 154-(10). On the 384 AGNES GOLDMAN TABLE 8 Agglutinin absorption tests with sera 154-26, 154-10, 154-18 DIRECT AGGLUTINATION +|+ +|+ AGGLUTINATION BY ABSORBED SERUM AGGLUTI- NATING tr. irs Absorb- ing strain 154-25 154-10 154-25 154-10 Serum 154-10 (rabbit J) Serum 154-25 (rabbit C) Serum 154-18 (rabbit —| —| 154-18) X| X X| tr.| 154-10} +) K | X Serum 154-18 (rabbit D) cE 2 | sTRAIN aE 2\2/3islelg Pa rale lire Ry) ee FP Se PR Serum 154-25 (rabbit C) +! —| —} 154-25 —-|}—-|-|- +] tr.| tr.| 154-10 == |ya= |) Sex 154-10 +<|4+<} +) + 154-25 +)/x}]-| - +} tr.| tr.| 154-10 TO OMI nbEe =| =| 2) Teel ie lees 154-25 —;};-j;-|- 154-10 +i+]xXi xX +) —| —| 154-25 Sel) Si +! —| —| 154-18 - 154-18 is i tie 154-25 tr. D) —| —| 154-25) X| X | —}] -— = 154-25) —| — | —]}] -— — | 154-18] +] + | X | tr tty, Serum 154-10 (rabbit J) tr.| tr. Ee le Shee eens | ere 154-18} —| — | —| -—| - 154-10} +4] + | tr. | tr.) tr. alah eaeel Li) So bne locmanin a ts Sh ath THAI Ol |e etal elo a 154-10 SF) Seal ctor tee x 154-18] +) X | X | tr. - TABLE 9 Agglutination of strain 154-10 by sera 154-10, 154-25, 154-18 SERUM DILUTIONS CONTROLS SERUM TITER SERUM met ae nl ee pie 1:600 | 1:800 | 1:1000 | 1:1200 | 1:1400 | 1:1600 |Ner™2l] saline | “°srmain 154-10 a5 + == Ss =E = x tr. 1-1000 154-25 + SF so oe nr =f x tr. 1-1000 154-18 F + i fe i x tr. 1-600 TABLE 10 Summary of agglutinin absorption tests NUMBER OF STRAINS TESTED SHOWING COM- PLETE AGGUTINATION AT 1:100 — WOW re PRR NN FPN NN HWW NUMBER OF STRAINS TESTED FOR AGGLUTININ ABSORPTION WOW FN RK ND KY BR BB RP ee STRAINS TABULATED ACCORDING TO THE DEGREE OF AGGLUTINATION OF HOMOLOGOUS STRAIN BY SERUM ABSORBED BY INDICATED STRAINS : a Sa “4 A c 3 Eis, Woe eee atte [ee = ° 2 5 oo & 2 a 7 - em oo a a x 2 Sal ect aapes BNO eg =| x a 5 5 4 oO oo 8 3 io} era | ~-sS op = Ss a 5 = 50 oa a >= SERUM 2 e Sea eer menes aid a 4 | 8/33) 8S [Bel 2 he a “eo | Sd 5 a 2 ES ° ® lag SB 54 3 eS Be os Hara o.5 or as AS SNe Sel ae bs Bi ape 3S | oe @§ ood 3 LS ae 33 a | 28 as Ao] AS se <5 s | §2 ER Ba) gs a O ice] 3} oO of or os higy ‘d cai a iS Lm fiscal ee din a ‘cal +f, er say he Nat ee EA fer aL +9 ai | . Ties pasee ye: a Ps eee Ay roe ee ts iho Saft ON THE PHOTOLABILITY OF SERUM COMPLEMENT! EK. G. LUNDBERG From the State Serum Institute of Denmark, Copenhagen Received for publication January 31, 1922 The purpose of this investigation was to discover the laws governing the destruction of complement (alexine) in normal serum by light. While the inactivation of the complement by other means, such as heat, shaking, and the chemical influences of acids and alkalis, have been already quantitatively studied and described, photo-inactivation has apparently been studied only qualitatively (8, 9, 10, 11, 12). It seemed to the writer, there- fore, that the latter form of inactivation was worthy of a more detailed and exact investigation. While the work was in progress, the thorough researches of 8. C. Brooks (1, 2, 3, 4, 5) on the same subject, though with somewhat different technic, came to my attention. Although we have covered much the same ground, I believe it may be of interest to compare the results ob- tained independently by different workers and by different methods. TECHNIC Complement. The complement used was exclusively undiluted normal swine-serum. The blood was caught in sterile glass cylinders and left for twenty-four hours at room temperature; the serum was then decanted and frozen at about —14°C. and left so until required for use. On those occasions when the serum has been employed at once, without being previously frozen, it is marked in the table as “‘fresh ser.’ Before use, the bottled frozen serum was melted in water at about 25°C. 1T beg here to express my sincere thanks for, and appreciation of the kind advice and all other assistance constantly given me by Dr. Th. Madsen, Director of the State Serum Institute in Copenhagen. 389 THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 5 390 E. G. LUNDBERG and immediately afterwards was put into an ice-bath until wanted for the experiments. Radiation. For radiation there was employed a mercury-gas quartz lamp (from the ‘“‘Quartzlampen Gesellschaft,’ Hanau), burning at 140-volt tension at the poles of the lamp and 4.3 amperes. It was estimated that the efficiency of the lamp would remain constant as long as the two current meters showed con- stant values; a state which was always reached within fifteen minutes after lighting and was usually maintained automatically, and seldom with the aid of a rheostate. The absolute light- strength was never measured, but a constant range of 150 mm. distance was maintained between the lamp and the serum, when not purposely varied in order to change the relative light- intensity. The serum was exposed in a flat round dish, with a bottom of mirrorglass and a frame of the same material, covered with a plate of Jena uviolglass, 1 mm. thick, 12 em. in free diameter and giving a clear space of 4.5 mm. between the two glasses. The box was held together by picin, a cement of asphalt and rubber, and in the circular frame there was an opening 7 cm. wide for the insertion of a glass rod, acting as the mixer of the fluid, and for the filling and cleaning of the interior and also to allow of successive samples being rapidly taken out by means of a pipette. The dish had a capacity of about 80 cc. but never held more than 60 cc. of serum. The whole was immersed to two-thirds of its depth in a water-bath with a motor-driven mixer, contain- ing 30 litres, and was kept at a suitable constant temperature by means of a gas-flame or a current of cold water. The regula- tion was usually carried out by hand. The degree of accuracy has been noted in each experiment; the variations never exceeded +0.4°C., remaining, as a rule, within 0.10°. The mixing-rod inside the serum-holder, 3 mm. thick and curved like a lancet- blade on a shaft, was moved to and fro by the motor at therateof about 20 oscillations per minute. It left no part of the fluid at rest. As no foam was formed and the apparatus moved very gently, all destruction of complement by rocking may be re- garded as excluded (Scaffidi, 12). The dish was fixed at an PHOTOLABILITY OF SERUM COMPLEMENT 391 angle of 45°, with the surface of the water somewhat higher than that of the serum. The lamp was placed with its tube hori- zontal, parallel to the edge of the free serum surface and as near to this as possible. The distance was measured from the nearest point of the tube to the center of gravity of the front serum surface. The degree of accuracy for this distance may be estimated at 0.3 cm. Before radiating the serum in any experiment, the lamp- current and the bath-temperature were made constant. After this, serum was poured into the dish, which was then protected by means of an iron plate in front and a black paper behind, and the mixer started. A small quantity of serum (5 to 10 ec.) was put into a Jena-glass bottle in the same bath, light-sheltered, as a control of the degree of destruction by temperature (‘‘heat control”). After exactly ten minutes—in less than which time, the serum was found, on testing, to have reached the tempera- ture of the bath—the iron-plate was removed and the first sample, [O] was taken. After varying intervals successive samples were taken (each of about 2 cc.) and, when higher temperatures were reached, control samples were taken now and then from the bottle. These samples were immediately put into an ice-bath, where they remained, sheltered from light, until titration. Titration. The titration was performed at once or some few hours afterwards. The method was a slight modification of that adopted by the Institute for hemolytic measurements (Madsen-Boas, 6), and was carried out principally as follows: There was employed for each sample a series of 10 to 12 tubes, calibrated with an exactitude of less than 5 per cent. In each tube there was put 0.5 cc. “‘sensitized”’ sheep-blood corpuscles, a variable quantity of the serum sample (complement) and 0.9 per cent NaCl solution, up to 1.25 cc. of total fluid. The blood-corpuscle fluid (‘‘red cell suspension’’) was prepared from blood that had been centrifugated for ten minutes and washed with the same salt solution twice, after which 5 cc. of the cen- trifugate was diluted to 100 cc. This red cell suspension was sensitized with equal parts of a rabbit-sheep-hemolytic ambocep- tor dilution, just so strong that it contained 2.5 units. The 392 E. G. LUNDBERG amboceptor titer had been tested beforehand with the aid of a guinea-pig complement, which gave quicker andmoreappreciable results than did swine-serum. Each tube was first filled with the salt solution, then with the serum (complement) or serum dilution; finally was added the sensitized red cell suspension. The tubes were immersed in an ice-bath during the pipetting of serum and blood, and afterwards were put into a water-thermo- stat at 37°C. for exactly two hours, after which they were placed in an ice-box (temperature 2-7°C.) until the following day, when the hemolysis in each tube was colorimetrically measured in daylight by means of a Madsen hemoglobin scale. A minor error occurred during the course of the earlier experiments in these color measurements, namely, that a layer of red solution often came next above the intact blood corpuscles, and a brighter layer topmost. On these occasions the hemolysis was read approximately. Comparisons made with a better technique showed that the mistake did not exceed 10 per cent. In the later experiments, the tubes were shaken before being placed in the colorimeter. This apparatus consisted of a black box with three vertical cells, separated by thin partitions. By three pairs of square holes the light passed through the fluid and tubes in the cells. But this shaking often clouded the tubes and made any reading impossible. Still later on, I had the tubes thoroughly shaken before their being placed in the ice-box for the night, with the result that the fluid was almost completely homogeneous in the morning. I attribute the better conformity of the re- sults from the last period entirely to this technical detail. The fluid-quantities used being one-fourth of the doses in the original method of Boas, the real amounts of serum (comple- ment) in the glasses have always been multiplied by 4 when shown in the following tables. These last ‘‘whole dose’? amounts make up the titer figures, p. They were always chosen in some suitable part of the following series of p = 3, 2, 1.4, 1, 0.70, 0.50, 0.30, 0.20, 0.14, 0.10, 0.07, 0.05, 0.03, 0.02, 0.01. The results are entered graphically on a codrdinate system, the p’s as abscissae, the hemolysis percentage, h, as the ordinate. These last have been corrected before being entered, viz., as regards PHOTOLABILITY OF SERUM COMPLEMENT 393 spontaneous hemolysis in a control O-glass, and also with respect to the serum’s own color which, in the concentration p = 1, was found to be equal to 2-8 per cent hemolysis. I have also supposed that this color, which greatly resembles that of diluted hemoglobin adds to the value of the true hemolysis color. The error—if it be one—is probably not of any great importance. For the determination of the complementary strength by the aid of such a series of glasses, I have employed the following method: With the help of the above-mentioned hemolysis curves (fig. 1) I look for the h-values, i.e., the corrected and most probable hemolysis percentages, for each of the employed serum concen- trations, p, commencing at p = 0.02 and ending at h = 100 per cent. Each h is divided by its corresponding p. This fraction, h/p, indicates the efficiency-value of the serum mixture in question, i.e., the produced effect (h) per unit of agent (p). If, taking one curve, you sum up these fractions commencing, for instance, at p = 0.02, and divide each sum by the number of pay) ae its terms, you obtain a mean value, /P ; which is, however, n not constant. It increases as you follow the calculation along the curve, and finally reaches a maximum value. I have em- ployed this as a measure of the complement strength, S. Being a function of the relation between effect and cause, and a greater number of points being used on the curve, this formula is more rational than the one I first employed, which was very similar to that of Brooks (4), which took the mean of the p- values for only five different hemolysis percentages. And, by comparison, it gives more equal values and smoother curves than the latter. Furthermore, I have not adopted the relative values proposed by Brooks who takes a cold serum control as the unit strength for each separate experiment, because of the variations in strength this control exhibits during the time of experiment. It appears in general to increase in strength for some hours. The S-values thus obtained are employed for drawing the destruction-curves (fig. 2), both for the radiated samples and also for the dark heat-controls, the S as ordinates, the time as abscissa. 394 E. G. LUNDBERG (x) _ i201 —— eee a ee Radiated samples Fie. 1. Exprrtmment XVI. Hemotysis Curves PHOTOLABILITY OF SERUM COMPLEMENT 395 The reaction-constant, k, is then calculated for each curve out of the observed values, employing the formula for monomo- lecular reactions: eae €Xloge where, in this instance, a indicates the original complement force S,; a-« the remaining force S,, after the lapse of t minutes. ry 129 80°C. Fie, 2. Experiment XVI. Reaction Typr AND SPEED (e is the basis of the natural logarithms.) The symbol k thus gives a measure of observed destruction-speed, being the mean of the values for all the successive times. In order to discover the true photo-destruction speed, k:, I have subtracted the destruction-speed of the thermo-control, ky, from that of the radiated portion, k,. Wishing to find the influence of temperature on these reactions, three series were made on one day (experiments XIII, XV, 396 E. G. LUNDBERG XVII) in order to find the k-values in the van’t Hoff-Arrhenius formula: where 7,, T, are the absolute temperatures in the two reactions running at the speed k; and ks, and yu is the desired temperature- coefficient. Together with these two investigations—the search for the reaction, law and speed, and for the temperature-influence— I have also made some experiments in order to see if the altera- tion of the lamp-distance really produces the effect one might expect beforehand: a change in k-magnitude inversely propor- tional to the square of the distance. Secondly, I have tried the influence of dilution. In all other experiments, I have radiated whole serum in spite of its great opacity to chemical rays which has caused such a slow destruction-speed, even in this thin and well-stirred layer, that the exposure necessary has been from forty-five minutes to two hours. Even if the dilution were made with balanced salt solution, it might essentially alter the com- plement, and could not, a priori, be regarded as a non-destructive operation. A fairly large number of experiments were carried out, but only a few are reproduced here and of these only one is given in its entirety in order to show the method employed and the degree of error and uncertainty attending these extremely delicate in- vestigations. The other experiments of the same kind have given the same or similar results, and the same method has been followed as that described here. In the following tables, [0], [10], etc., mean samples taken from the dish 0, 10, etc., minutes after the commencement of exposure; [V10], [V60], etc., are samples from the heat-control bottle, taken 10, 60, etc., minutes after immersion in the bath. [VO] consequently, means the non-handled stock-serum kept at some few degrees above the freezing-point. The value of [V10] thus should be equal to that of [0]. PHOTOLABILITY OF SERUM COMPLEMENT 397 The figures in the primary tables (experiments XVI, XIII: ‘‘hemolysis in per cent’’) show the observed hemolysis percentage without any correction, 100 or >100 means complete hemolysis (no cell sediment on the bottom); 100* means a color equal to 100 in the seale, with a cell sediment on the bottom. All figures below 100 indicate incomplete hemolysis. The values above 100 per cent are attributed to the color of serum itself. EXPERIMENTS Out of the several tests made to find out the reaction-type and speed, I have chosen one as a specimen, it being of considerable length, and showing a very regular course of the photo-inactiva- tion. Here the primary data are given in full, so as to fully eluci- date my procedure. 1. Reaction type and speed Experiment XVI. Temperature 3.01° +0.15°. Distance: 15 em. Fresh serum. Amboceptor titer: 0.20 From table 1 there have been drawn the hemolysis curves in figure 1, with corrections made for own-color of the serum, ac- cording to the correctionline, topmost in the figure. From each of these curves there has been calculated the complement-strength according to the formula, explained above: g = 2h nas, n These values are reproduced in the following table 2 as Sus. and in the graphic representation of the reaction course in fig- ure 2. From them there is calculated the reaction-constant, k, in accordance with the monomolecular formula. The mean value is given for each of the radiated and the light-sheltered heat- control portions, k, and k,, respectively. Sj... shows the values calculated for a reaction with this mean value as reaction- constant, and the same initial value as the observed one. 398 E. G. LUNDBERG TABLE 1 Hemolysis in per cent (direct readings) FADIATED HEAT CONTROLS ON a ee A ig ee {0} [10] [20 [40] [60] [90] | [120] | [200] | [V0] | [V10] | [V30] | [V90} | [V210] 10) 80 | c40 | e25 2e0 >100 | — | c45 | c25 1.4 >100} *100 | — | c40 | c22 1.0 100} 65)| — 40 | 20 0.70 |>100} 100 |>100; 50; 60/38 40 | 14 |>100)>100)>100) 100 | 100 0.50 | 100} 100 80) 40) 45) 3 30 6} 100; 100; 100) 90} 90 0.30 75| 65 55} 37) 40] 3 22 fo idl) ZO Von eGo 0.20 55} 50 AD oO Zones D5} 05|) OOOO meds 2 1 0.14 40} 40 35 4 1 40| 45) 40} 40] 40 0 RP Nw ONONN _ tx 8 0.10 30} 30 22 2} 10 1 35| 30) 35) 35] 40 0.07 25) 22 18 1 2 0 30] 25) 30) 22] 20 0.05 12; 20 6 0 1 20). 18)" 20) AGHis 2 0.03 44 2 1 0 8 4 4, 4 4 0.02 LUN 0) 1 2 1 2 |i 2 0.01 0} 60 0 0 0; 0 0 Controls with: (1) no serum, only sensitized blood = 0 per cent hemolysis (2) serum only, corresponding to p = 1.0, from sample [V0]: 4 [20]: 4 [200]: 2 [V 210]: 4 * Indicates incomplete hemolysis. c Indicates approximate reading, because of opalescence and brownish tint. TABLE 2 Complement strength IN RADIATED PORTION IN HEAT CONTROL gfe lela le (ee dae eerie = 8 8 = 8 5 & & § | g =| & o STS (SLi S aS Se eS ei Seine o8 | o8]52| 52/23] 8| 3] 2] 28|e8l 2] 28| a8 aA SP (SSP PSF/St (S/S (S47 (S-/e2(Srisesiaes < < a {4 ja [4a [a |e [4 [ae |e Jae d4 Sobsss ocdaOtbios-. 2.29 | 1.95)1.49|0.62/0.73/0.41/0.30)0 .07|3 .17|2.97/2.76|2.43)2.24 Sealontti. d.csiitas 2.29 | 1.86)1.52)1.02/0.67/0.36/0.20/0.04)/3 .17|3.05/2.84|2. 28/1 .46 Ky, = 0.0203 mean; K, = 0.0037 mean. PHOTOLABILITY OF SERUM COMPLEMENT 399 A comparison between the observed and the calculated values show that the light-destruction within the limits of experimental error can be expressed by the monomolecular formula. A number of other experiments of the same kind all show the same monomolecular type with the same degree of exactness. Out of a number of experiment-series, where complement has been destroyed at various temperatures, I have chosen one (No. XIII) where the curves are specially smooth and easy to calculate. It should be observed that experiments with different temperatures, made on different days, are not comparable. The freezing of the stock serum during the night and other fac- tors, may cause alterations which we cannot account for. Only those experiments, therefore, are used for determining the influ- ence of temperature on the reaction speed which are carried out on the same day and with the same serum, and in quick succession, and under as similar conditions as possible. The following ex- periment shows a triple series, carried out at the temperatures: 46°, 36°, and 26°C. 2. Temperature influence Experiment XIII. Observed hemolysis 46° = 0.3° Dp [0] [10] [20[ [35] [50] [V0] [V10} [ V60} 2.0 65 95 1.4 90 60 95 1.0 65 55 95 0.70 100* 55 50 100 100 95 0.50 >100* | >100* 80 50 40 95 100 90 0.30 100* 70 55 40 35 95 95 90 0.20 65 60 40 39 22 70 95 90 0.14 50 45 35 22 16 40 65 65 0.10 35 30 22 16 6 40 40 35 0.07 30 22 18 4 2 30 30 30 0.05 18 14 10 1 22 18 16 0.03 8 6 3 10 12 10 0.02 4 4 1 5 10 8 0.01 1 0 1 1 1 400 E. G. LUNDBERG Observed hemolysis—Continued 36.0° + 0.10° Pp [0] [10] [24] [40] [60] {V10] }V70) 2.0 80 1.4 65 1.0 80 50 >100 0.70 > 100 > 100 100* 70 = >100 >100 0.50 100 100 100* 55 45 >100 >100 0.30 100 90 60 45 30 >100* 100* 0.20 75 55 45 35 25 80 75 0.14 50 45 30 30 18 55 55 0.10 45 40 25 18 4 50 45 0.07 40 35 22 16 1 40 35 0.05 25 20 12 6 i 25 25 0.03 12 8 4 2 14 12 0.02 8 2 2 0 6 0.01 0 0 0 0 26.0° = 0.15° P [0] [10] [20] [40] [80[ [V0] [V10] [V90] 2.0 80 1.4 55 1.0 100 50 100 0.70 100 100 100 85 45 100 100 100 0.50 100 100 100 65 40 100 100 100 0.30 100 70 70 50 30 100 100 100 0.20 45 55 55 40 22 55 70 70 0.14 55 50 45 25 16 50 55 60 0.10 30 18 40 22 8 25 40 50 0.07 30 30 25 14 2 30 30 30 0.05 _ 16 12 6 1 22 20 25 0.03 = 4 = 4 0 14 14 20 0.02 = 0 1 1 12 10 8 0.01 = 0 0 1 ft 0 Control with no complement: 0 hemolysis. Control ~vith serum only: Sample [V0]: 1. Sample [80]: 1. * Indicates incomplete hemolysis. >100 indicates deeper color than in scale tube ‘‘100.”’ PHOTOLABILITY OF SERUM COMPLEMENT 401 From the titration curves, drawn in accordance with these figures, the following mean values are obtained and the curves of figure 3 drawn. Titration results [0] [10] [20] [40] [80] [VO] | [V10] | [V90] 26° Sopa cuore sees DLO 219102295) ele son Oro on O neon On| so0 Wealeitses caus sees 5.0: 3.78 | 2.86 | 1.63) 0.58 | 5.0) 5.0 | 5.0 Ky, = 0.0280 mean; Ky = 0 [0] [10] [24] [40] [60] [V10] [V70] 36° peers sete eases 3.93 2 Sat SESS LEAT PyOP RT? |) "4 1S S87, WS ceilemicyormaqre be psibiest sic 3.93 2.387.) ACSb SIS OLGOL | AL LS cd ed Ky = 0.0311 mean; Ky = 0.0011 [0] [10] [20] [35] {50} [V0] {[V10] | [V60} 46° shy eee eae Ce 3.0 | 2.33 | 1.82 | 0.80 | 0.53 | 4.30 | 4.24 | 4.00 Peale ae seeinacds 3.0 | 2.22 | 1.63 | 1.04 | 0.66 | 4.30 | 4.25 | 4.05 K;, = 0.0305 mean; Ky = 0.0010. ““S.ace’” Shows the complement-strength values calculated under the presupposition that the reaction follows the monomolecular formula and that it has the reaction-constant k,, obtained as the mean of the observed complement-values S...2 ky gives the reaction-constant for the thermo-destruction at each tempera- ture. If this constant is subtracted from the corresponding light destruction-constant k,, we obtain the following values for the pure photo-reaction speed, /,: at 26° 0.0280; at 36°, 0.0300; at 46°, 0.0295. As these values show a very irregular course in relation to the temperature, no certain temperature coefficient ean be calculated from them. If there is any numerical value of such a coefficient, it must be low, or even negative. The same result is gained from two other experiments, re- produced here, although only as regards their final figures: 2In the experiment at 26°, the 10-value, being evidently due to one single mistake in titration, has been altogether neglected in the calculation. 402 E. G. LUNDBERG EXPERIMENT XV EXPERIMENT XVII ky hy ky ky ky hy SEAN Maer (AT call yea ane eet Aa | 15 0.0242 0.0 0.0242 15 0.0233 0.0 0.0233 30 0.0193 | 0.0 0.0193 30 0.0190 | 0.0006} 0.0184 45 0.0420 0.0060 0.0360 45 0.0330 0.0010 0.0320 Even here, the differences being small and irregular, it may be concluded as a whole that the temperature-influence on these reactions is very small or negative. 5 REBE BERR BREESE ES EE BERD TT het I \GEBEe EERE Gn —— HERS PeRP essa Beeoees Lt AVGSHR0ERR BO Sne ne BEE SEeooRe one Sa pt BEECH EEC =~ BEES SEER Pr tteet PEt fel 1 g@ eae Sess esteegecatocectasts a EEESES a ~~~ Na POA PN $044 Oe eka poy sald ata PERCE EEE H+ COREE CEH co PEELE EEE HERA Em a is A) a ae ea i sf fa sf ot ao fa REGRADESRESR ERR ARLEEEP Eee i | fA Yc fi a ia i fc Bai PCS 4 | (A | a a eo a elo APNG TaN BECKERS EEE Er Bee eee QHREK ASAE a ee SeRee pea ee PECANS ERESSNehepsane HoaRRSSaa pt LS eee w EEE EE Fig. 3. TEMPERATURE INFLUENCE PHOTOLABILITY OF SERUM COMPLEMENT 403 The following experiment is for the purpose of discovering whether a change in light-intensity alters the reaction-speed in the same proportion. For this purpose I have chosen to change the radiation-range in the proportion 1:2, which signifies an alteration in the light-intensity from 1 to 3. The temperature was, all the time, constant, and the same serum was employed. 8. Influence of light-intensity Experiment XVIII. 45° + 0.2°. Fresh serum. 1. Distance = 18 cm. (Light-intensity = 1) RADIATED MENGES Pes otate san loco aie ss 0 10 20 30 45 60 Dphaee catches eter = dub 2.20 1.47 1.06 0.74 0.51 0.36 WS calle savorteie tect aed ova eneue ace 2.20 1.62 1.16 0.81 0.50 0.30 K, = 0.0336 mean. 2. Distance = 21.2 cm. (Light-intensity = 4) 2 RADIATED AVEC CER eo os cisieiare;sie oyeseeveers 0 15 30 45 60 80 100 120 NR ie ee 1.57 | 1.31] 1.01] 0.80] 0.60 | 0.39 | 0.22 | 0.15 ister onherstk seule: 1.57] 1.23] 0.97] 0.76 | 0.66 | 0.43 | 0.31 | 0.23 EIT UIEE Eso orate cincieie crises oe 0 130 245 ems he MEE 2.03 1.18 0.63 (SGAIECACH BEES CHA nere 2.03 1.14 0.68 Ky = 0.0045. The observed values agree fairly well with those calculated according to the monomolecular formula. If the thermo-inactivation speed-constant k,, is subtracted from the photo-inactivation speed-constants k,; and k:, the true light-destruction speeds are obtained, viz.: at light-intensity 1:k, 0.0291; at light-intensity $:k: 0.0117. 404 E. G. LUNDBERG The proportion 2 is equal to 2.49 instead of the expected 2. The difference may possibly be regarded as lying within the experimental error. Finally, I wished to determine whether it was feasible to dilute serum before radiating it, and to reach the same results as with whole-serum, only at a higher speed, in proportion to the amount of dilution. From the phenomenon of hydro-inactivation of complement, I divined that even dilution with a balanced salt solution might alter the nature of complement, thus causing not only a proportionally lower strength of the fluid, and a propor- tionally more rapid destruction, in consequence of the greater transparency of the dilution, but also some unexpected change. For this purpose, whole-serum, a 1:3 dilution, and a 1:10-dilu- tion, were radiated on the same day and under entirely similar conditions. 4. Influence of dilution Experiment XIX. 39.95° + 0.20°. Fresh serum. A. Undiluted serum [0] [10] [25] [40] [60] {V0} (V70] Stal i 7.91 4.65 | 2.88 1.76 12000) 7EoH 4.65 Seana: ote 7.91 5.10 2.61 1.35 | 0.56 Ki = 0.0442 mean; K = 0.0078. B. Serum diluted 1:3 {0] [5] {101 [15] [25] [V0] (V35] He ee 0.88 0.65 0.53 0.44 0.26 1.08 0.83 Ch ee 0.88 0.68 0.53 | 0.40 0.24 Ky = 0.0518 mean; Ky = 0.0071. C. Serum diluted 1:10 {0} [2] [4] [6] [8] [12] [V0] [V22 SGheaseeeeie 0.22 0.17 0.16 0.138 Ot 0.08 0.26 0.19 Scale ocpanhorc 0.22 0.18 0.15 0.12 0.10 0.07 Ky = 0.0975; Ky = 0.0148. PHOTOLABILITY OF SERUM COMPLEMENT 405 Note: In (B) the directly found S-values are multiplied by the dilution-cipher 3 before being entered in this table and on figure 4. In (C) they are multiplied by 10. The destruction- curves thus show the real strength of complement. In spite of this operation, it is found that the diluted sera are dispropor- tionally weaker than the whole-serum. But all of them follow satisfactorily the monomolecular formula, as found by the cal- culated values. The proportion between the reaction-speeds, after subtract- ing the thermo-constants, is found to be: REACTION-SPEED ky EXPRESSED IN PER CENT Wace desenwme caste fc caccos kes s.cscle sm artes 0.0364 100 DERUMIUCIULE otter ee. sles ee eel. oe aes 0.0447 123 HerumpanatedalslO: eee... ke ees. See 0.0832 228 There evidently exists no parallelism between dilution and reaction-speed, although the latter increases with the former. 267-994 g Destruotion curves et varying dilutions (40 6.) "| [2/3] : whole-serum [2/2] : serum dilution 1:3 fa/ig: n 1:10 Observed values are multiplied by respective dilution figure, 1, 3, 10. Uppsr broken lines: ~~] : heat-controls. 7™'" 69 Fie. 4. Exprriment XIX 19 20 30 40 50 THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 5 406 E. G. LUNDBERG SUMMARY There has been shown by these experiments that: 1. The light-destruction may be fairly well expressed by the monomolecular formula. 2. The rate of the destruction is low, k varying from 0.019 to 0.050. 3. The temperature-influence is small or none. 4 certainly does not exceed two or three thousandths—which signifies an increase in k of 1.1 to 1.7 per cent for 1°C. increase of tempera- ture at 27°C. 4, Variation in lamp-distance causes changes in reaction-speed, which do not differ essentially from inverse proportionality to the light-intensity. 5. Dilution of serum before radiating causes not only the expected decrease of complement strength, but also alters very irregularly its magnitude. It causes, too, an increase in reaction- speed, which, however, is not parallel to the dilution-amount. 6. The color of serum changes obviously during radiation, from orange to dull yellow, the opalescence increasing at the same time. This is accounted for in the colorimetry. 7. As regards the exactness of the measurements made with this method, the primary figures have a possible error of about + 10 per cent. The graphically found values may be regarded as more exact (about 5 per cent possible error). The k- values are chiefly found to differ within + 12 per cent from the mean, which may be regarded as quite satisfactory. The temperature- variation in each experiment—being within + 0.4°—may be disregarded in consequence of the insignificant temperature- influence. REFERENCES (1) Brooks, S. C.: Jour. Med. Research, 1918, 38, 345. (2) Brooks, 8S. C.: Jour. Med. Research, 1918, 38, 345. (3) Brooks, 8S. C.: Jour. Med. Research, 1920, 41, 411. (4) Brooks, 8. C.: Jour. Med. Research, 1920, 41, 399. (5) Brooks, S. C.: Jour. Gen. Physiology, 1920, 3, 169. (6) Boas, H.: Wassermann Reaktion, Ref. in Communications de 1’Inst. Sero- therap. de l’Etat Danois, 1909 and 1911. PHOTOLABILITY OF SERUM COMPLEMENT 407 (7) MapsEN AND WatTaBIkI: Oversigt af Kong. Danske Vidensk. Selsk. For- handl., 1915, p. 125. (8) Baront, V., anp Jonesco-Minatssti, C.: Compt. Rend. Soc. Biol., 1910, 68, 393; 69, 273. (9) Courmont, P., Nocrzr anp Durour, A., Compt. Rend. Soe. Biol., 1913, 74, 1152. (10) Dorrr, R., anp Motpovan, J.: Wiener klin. Wochenschr., 1911, 24, 55. (11) Asettn, S., anp Stiner, O.: Zeitschr. f. Immunitatforschung, 1913, Orig., 10 ale (12) Scuarrip1, V.: Biochem, Zeitschr., 1915, 69, 162. ro itape leo SAO AY A Ae ee a ee ie LAO La vy he a” Pa i ik Ke ‘ ‘nits AP ee aol reper i iy " aw Sale nA Vast 0 ok a pics s Miter ve. wh wert ae - FOr 1 a x maT etelaurarorow?. -10°% aliboet araebiy ‘adoast iba le Peer Muar: Vt th conse Wid. .4 OFet Pod EGR Ts Avie nis Heed figgtic TO S02 5 Hints Pale eaten ae i ovr 4 Ri eae a anit Hath aie "heen! Seite YK Atty ole ert eae ss 390 esha JOtt WP a GA RINT. cualberieu let f!, yaclgh mare Wb? Ue AS RS Ri 6107 itinerant Gin veel.) aeoetias >.0) | Mae SRA, pi ‘nak aes eo cA Votieatel? len) ae Vibes jee 5 é 5 7 is Pel hy a PROPHYLACTIC TREATMENT FOR RABIES BY MEANS OF STANDARDIZED GLYCERINATED VIRUS JAMES McILVAINE PHILLIPS Pasteur Institute of Columbus, Ohio Received for publication March 18, 1922 In a previous article (1) we compared the various antirabic methods, discussed their gradual evolution from the crude pro- cedure employed by Pasteur to a much more accurate and, judging from our experience, more efficient technic—that is, the use of a ‘‘standardized” virus in such a form that it can be pre- served with practically unchanged virulence, for a long period of time. The Harris (2) method of desiccating and preserving fixed virus is almost ideal from the standpoint of therapeutic efficiency, but it is cumbersome and complicated by many manipulations which afford opportunitites for accidental contamination of the virus during its preparation. The number of living infectious units in the finished product is also greatly affected by any varia- tion in the rapidity of drying; as a result, after standardization and sterility tests have been made, much of the material pre- pared with so great effort, must be discarded. Some persons working with this method have had difficulty in maintaining the virulence of the virus. After many generations of storage the incubation period seems to have a tendency to lengthen. We determined, therefore, to try to find an equally efficient method, free from these objections, of preserving the virus in a living state. An analysis of the many papers on the subject suggests thai heat, oxygen, light, moisture and various chemicals are the agents which cause rabies vaccines as they are usually prepared, to lose their virulence; that when these factors are not present the virus remains active for amazingly long periods of time. 409 410 JAMES McILVAINE PHILLIPS Pasteur himself (3) sealed rabies virus in glass tubes by means of a blow pipe and noted that it maintained its virulence for several weeks in the summer heat. Because the oxygen in the tube was very soon reduced by the fresh tissue an anaerobic con- dition resulted. This was the first demonstration of the anaerobic preservation of rabies virus. To Roux (4) belongs the credit of the discovery that fixed virus retains its virulence when placed in glycerine: Calmette (5) put the discovery to practical use for preserving the cords used in the classical Pasteur method, so that their usefulness was extended. This method is still in use in a large number of labora- tories in this country and on the continent of Europe. We believe that the preservative action of glycerin depends chiefly on the fact that oxygen is only slightly soluble in it, as shown by the following figures compiled from Miiller (6). PER CENT OF GLYCERIN i ert Pera 848 BE OT EO Obie ci acoso tan, 86: = shade ose d Gch Sppnisteile sie caol> Sees sige ee 279 J WaT nd eae a OS 26 RI ROE Ey nt ean ai Bee A, A aa OO KOC 25 Nodataonpcamaplo TS soho: al! SPA See hott ntelen ele area ene 172 TABLE 4 Diagnosis of biting animal Laboratory diagnosis showing Negri bodies...............-0eeeeeeecceevees 626 Clinical sympconis’ OL TADICS... «8. Soe. tee he aie ows Ba eos oa le iain clea toe 584 Notiaade ; (animal eseaped)... 2250. 422095. Va te oie. 8st Se eerie tee ee 158 In addition to these human cases, we have given like treatments to twenty-nine dogs. One of these died of rabies on the fifteenth day after commencing treatment, or the twenty-first day after having been bitten on the head. None of the others contracted rabies. On account of the disastrous results which may follow a failure of the Pasteur treatment in a dog, we have always dis- couraged it. We consider the use of shorter time treatments especially reprehensible, as the almost invariable result is that PROPHYLACTIC TREATMENT FOR RABIES 419 the dog is freed from quarantine and thought to be safe before the real danger begins. Our experience in immunizing other animals with short treat- ments in the period between November 12, 1912, and January 1, 1922, is shown in the following table. We gave these animals just ten times the usual daily human dose, for ten days. One ease of treatment paralysis developed in a horse which promptly recovered. As two cows treated in this way developed rabies it would seem even with this intensive treatment that ten days is rather less time than should be used to secure the greatest possible efficiency in the treatment of animals. However, this is probably the longest duration of treatment in animals which is possible from an TABLE 5 EIND OF ANIMAL NUMBER RESULT Horses 39 Protected to date Mules 1 Protected to date Hogs 26 Protected to date Cattle 101 2 died of rabies, others protected economic standpoint. When dealing with human life there is no economic standpoint. We believe that our technic has solved the question of an econom- ical and permanent standarized supply of a potent virus. The remaining problems are to be found in the dosage and the duration of the treatment. As the dosage which we have given has been so efficacious and has not proved dangerous, we have hesitated to make a change. Certain clinical observations of the results of treatment in men and animals have deterred us from shortening the treatment. The one fatal case (table 2) was a severe lacera- tion above the eye, so large as to require several stitches to close it. The physician to whom we sent the treatment neglected to answer any inquiries and we did not receive any information until symptoms had appeared. This boy, of ten years, had been bitten four days before treatment was commenced. The wound had been treated only with nitrate of silver. The inoculations 420 JAMES McILVAINE PHILLIPS lasted eighteen days, 15 mgm. of living virus being the daily dose after the usual first three days treatment with dead virus. Symp- toms commenced one week after completion of treatment and the boy died three days later. Examination of his brain showed Negri bodies. We believe that a more prolonged and more in- tensive form of treatment might have saved this child in spite of inadequate cauterization and the fact that the wound was sutured. TABLE 6 ANIMAL ea DATE OF BITE aan RESULT Cow 1 Nose November 22 | November 24 Died December 15 Cow 2 Nose November 22 | November 24 Died December 17 Cow 3 Nose November 22 | November 24 Died December 17 Cow 4 Nose November 22 | November 24 Died December 19 Cow 5 Nose November 22 | November 24 Died December 19 Cow 6 Nose November 22 | November 24 Died December 19 Animals all bitten by the same dog, and were in their stations at the time. Treated by Drs. Failor and Morris, Veterinarians, Lima, Ohio. TABLE 7 ANIMAL LOCATION OF BITE DATE OF BITE | 7RE ATED RESULT Horse Nose March 4 March 6| Died March 25 Dog Hind leg March 4 March 7] Died March 29 Dog Hip March 5 March 8} Died April 12 Dog Fore leg March 5 March 9! Died April 7 Dog No visible marks March 5 March 10 | Died April 20 Two hogs and three dogs bitten by the same dog but not treated also died of rabies. Animals treated by Dr. Wise, Veterinarian, Medina, Ohio. When thinking of shortening the treatment in man, the oc- easional failures which occur in animals which have been given a six-day intensive treatment should be considered carefully. As an example of these we are giving two tables, in each of which animals were treated with virus furnished by different commercial houses respectively. Many attempts to measure the degree of protection which can be conferred on rabbits by the use of varying numbers of PROPHYLACTIC TREATMENT FOR RABIES 421 “minimal infectious doses” inoculated intracerebrally, have convinced us that these results are too variable to be dependable. The same can be said of the titration of the serum of protected animals for rabicidal properties. Therefore the clinical use of a method in a large series of cases over a long period of time, is the only reliable test, and changes in the length and intensity of treatment must be undertaken cautiously. REFERENCES (1) Puruurrs, J. M.: Ky. Med. Jour., 1921, 19, 278. (2) Harris, D. L.: Jour. Amer. Med. Assoc., 1913, 67, 923. (3) Pasteur’s System, Hydrophobia, Chatto and Windus, 1887, 46. (4) Roux, E.: Annales de I’Institut Pasteur, 1887, 1, 87. (5) Catmurrs, A.: Annales de l’Institut Pasteur, 1891, 5, 633. (6) Miuuur: Zeitschr. fiir Physikalische Chemie, 81, 483. (7) RemurncerR, P.: Annales de l’Institut Pasteur, 1919, 33, 28. Comptes Rendus Soe. de Biol. 1918, 81, 20. (8) Stimson, A. M.: Bull. No. 65, Hyg. Lab. U. S. Pub. Health and Mar. Hosp. Serv., Wash., p. 57. (9) AnpERSoN, R. P.: Jour. Ind. Eng. Chem. 1915, 7, 587. (10) Prrone, R.: La Riforma Medica, 1920, 36, 782. (11) Harris, D. L.: Jour. Inf. Dis. 1913, 13, 155. (12) Fermi: I] nuovo metodo Italiano per la cura antirabbica. Rome, 1916. Ann. d’igiene, 26, Suppl. (13) Cummine, J. G.: Jour. Inf. Dis. 1914, 14, 33. (14) Sempxe, D.: Scientific Memoirs, Officers of Med. and San. Depts. of Gov. of India, 1911, 44, 32. (15) Puiuurrs, J. M.: Berry, F. and Snook, J. H. Jour. Inf. Dis., 1921, 29, 97. (16) D’Aunoy, R.: Jour. Inf. Dis., 1921, 29, 261. THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 5 eM G SAREE ALS ORE eral de et SANE PROD Sone eiinaatitrs sai ted i tpacie tes tad wipes Tike Bre abkithieaetretligh ie iace nes (LORE S Ree ee Vitel dog LTT edly cee en ee ew PRA Ue ays sk PO CRY ie POP Hy eed ble, (cay Tg eae ae aera matey 3 ee ly Fe ed ode al aby hel _bitove ane aS eee eu! patents ie AV eK VED Baan a ab) Raed a ada by, . i codartne sant perpen 2 hat (i > 4 i ; . Cr Oe: 08 ih f igh yan obey me | j ‘ Yk Gy Mat i tA wd, oe | wha a| j ' ‘ ’ i iJ i i } y ii ne, ss ' * + ra “iH ae wets oe ee i ' ( ‘ ' Nie ; ; 4 4 2 hs f ’ ’ é hay Perey vc A : eer) er) | ‘ A Hae " weg p 7 me Cy | Til ers | ay Stee TLeP Ate, « jive) Th te aA 1) te alah ae fc ov oe? woe ee! o> 20 ee q A STUDY OF THE PRECIPITIN AND COMPLEMENT FIXATION REACTIONS WITH TUBERCULOUS EXU- DATES WITH SPECIAL REFERENCE TO TUBERCU- LOUS PLEURITIS ISAMU OGAWA From the Dermatological Research Institute of Philadelphia and the Pathological Laboratories of the Graduate School of Medicine of the University of Pennsylvania Received for publication April 22, 1922 This investigation was undertaken for the purpose of deter- mining whether the immunological reactions of precipitation and complement fixation occur with pleural exudates of tuber- culous origin, as additional means and aids for the differential diagnosis of pleural effusions. The necessity for diagnostic aids of this character has been impressed upon the writer since 1913, when he found large numbers of Japanese soldiers and many civilians in Mukden, South Manchuria, with mild pleuritis and exudates of unknown origin. At least 52 per cent of a group of forty-two of these individuals yielded negative v. Pirquet tuberculin skin reactions; cultures of the fluids from twenty-one were negative and tubercle bacilli were not found by smear methods. It is commonly believed that these “‘light pleurisies” (1) are of tuberculous origin but this has not been proven and the exact etiology is unknown. Since the demonstration of tubercle bacilli in pleural fluids of tuberculous pleuritis is frequently impossible and no other exact laboratory test being available, studies have been made with the precipitin and complement fixation reactions, and the results are summarized in this article. MATERIALS AND METHODS Pleural fluids have been obtained from patients in Philadelphia with tuberculous pleuritis, pneumococcus pleuritis, cardiac insufficiency with pleural transudates and from one case of carcinoma of the pleura with 423 424 ISAMU OGAWA effusion; also from a large number of guinea pigs and rabbits with ex- perimental tuberculous pleuritis previously described by Kolmer and the writer (2). The latter were particularly useful in this study be- cause means were afforded for determining the approximate time re- quired for the development of complement fixing antibodies in pleural exudates, which was not possible in the human cases of tuberculous pleuritis because of the indefinite histories in relation to the probable duration of infection. In the majority of instances of experimental tuberculous pleuritis in the guinea pigs, effusions were found in both right and left pleural sacs and frequently in the pericardial and peritoneal spaces; precipitin and complement fixation tests were conducted with all fluids. Precipitin tests were conducted with pleural fluids as precipitinogens and tuberculosis immune goat and calf sera, kindly furnished by Mr. Glenn, Dr. Paul Lewis and Dr. Aronson. These tests were conducted in the classical manner by laying clear pleural fluids over 0.1 cc. of im- mune serum in test tubes of appropriate size and reading the results after one and again after twenty-four hours. The complement fixation tests were conducted according to the tech- nic of Dr. Kolmer’s new method for bacterial complement fixation tests (8). Each fluid was first titrated for its anticomplementary activity and employed in 4, 4, 7s, st. ax, etc. of this amount with 4 the anticomplementary unit of an antigen of human tubercle bacilli pre- pared after a method described by Petroff (4). The primary incuba- tion was eighteen hours in a refrigerator at 6-8°C. plus 10 minutes in a water bath at 38°C., this method having been found by Kolmer to greatly increase the sensitiveness of the tuberculosis complement fixation test over the usual method of water bath incubation for one hour. The secondary incubation was one hour in a water bath, the readings being made a few hours later and recorded after the well known scale of ++++4+,+++4+,-++4+, + and — reactions. Each human and experimentally produced exudate was also studied bacteriologically, cytologically and the majority for albumin content. The results of these studies with the fluids from experimental lesions are given elsewhere (1); with the ftuids from human cases the data was employed with clinical data for establishing the diagnoses. RESULTS a. The results of precipitin and complement fixation tests with human tuberculous pleural exudates. These are summarized in EE REACTIONS WITH TUBERCULOUS EXUDATES 425 table 1. Of the fluids from twelve cases, five or about 42 per cent yielded positive precipitin reactions; the reactions, how- ever, were quite weak and usually discernible only after twenty- four hours. All of the fluids but one, however, yielded well defined com- plement fixation reactions (92 per cent). As shown in this table some of the fluids were hemolytic (nos. 5, 10, 11 and 13) so that the smaller amounts yielded positive reactions while the larger amounts produced partial or complete hemolysis. In every test the exudate controls with all amounts, antigen and hemolytic controls, yielded complete hemolysis and were satisfactory. Table 2 summarizes the results observed with five pleural exudates of non-tuberculous origin. All yielded negative pre- cipitin reactions; likewise all but one yielded negative comple- ment fixation reactions. | The one fluid reacting weakly positive was from a syphilitic with cardiac decompensation (no. 16); I believe that this result was due to the presence of syphilis ‘‘reagin’”’ fixing complement with the lipoids of the tuberculous antigen (5); for this reason the antigen of tubercle bacilli should be freed of lipoids in order to remove the possibility of these cross complement fixation reactions. b. The results of precipitin and complement fixation tests with pleural, pericardial and peritoneal exudates in expervmental tuber- culous pleuritis of guinea pigs. These animals were infected with injections of virulent bovine bacilli! into the right pleural sacs. Well defined pleuritis was produced in the majority of animals on both sides; in many instances tuberculous pericarditis with effusion also resulted. Tests were made with effusions developing four to twenty-seven days after infection and the results are shown in table 3. Of eighteen exudates from the right pleural sacs, four or about 22 per cent yielded positive precipitin reactions; eleven or about 61 per cent yielded positive complement fixation reactions. No 1 Strain H, kindly furnished by Dr. F. Boerner of the Pennsylvania State Livestock Laboratories. ISAMU OGAWA 426 = > = - + + |90°0| — Se lee ete ae | ott = ie ela toe eS | = | oto, — —} - | = = | + | ++ | ro] + — |4+4+4]44+4+4]44+4+4+) -— | - [900] + — |+t+/+++] - | -— | - Joo] + Se ee a tt ltt | wo). + — | t+ tet ++ti+++4i+++4] 90] - — | +H j +44 4+44l+++4l+++4] 90] ee eS + | gto) = — — — 5 — — 9°0 — —- | - | - | + | +4] ++ |] to] = joayuog | mf ax ¢ af nM snow NOILOVAU NOILVXId INGWATAWOO “anxo0 “ata ‘oqo ‘ping [eanoyd jo yun Arvyuouto;durooryuv oyy Jo F y oUON ouoN oUuON qUISOI ouoN ouoN ouoN oUuON ouoN quosoig SUVGAWS NI IWova @TOUTAOL siytanoyd “qT, stytameyd *g J, siytaneyd “eT, stytinoeyd *g J, stplineyd *g siyiinopd “qT, stytmneyd *g J, Ca Atquqord) siytane[g siytnoyd “qT, siylanoyd “gq TL, siynoyd "gL, siytineyd ‘gL, SISONDVIC TIVOINITIO ul 6G | 02 A LG | 8T INI sy ROG Slick W | 6f | ot I OS si) Tt W | &@ | OT uf Ié 6 W | &@ L ul VG 9 W L ¢ W | 98 p W |.¥é I sunah wad xs IDV | -WON asvo sayppnxe yoinayd snopndwaqn} Upiuny YyyUN UoYoxY puamajdwmod pun Uijrdiao%Ed fo s}nsay T WTaViL REACTIONS WITH TUBERCULOUS EXUDATES 427 positive reactions were observed with fluids removed earlier than twelve days after infection; after the fifteenth day follow- ing infection, nine out of ten fluids or 90 per cent yielded positive complement fixation reactions. Eighteen left pleural fluids were examined; of these one or 6 per cent yielded positive precipitin reactions but eleven or 61 per cent yielded positive complement fixation reactions. Of nine fluids removed fifteen days or longer after infection of the right pleural cavity, eight or 88 per cent yielded positive comple- ment fixation reactions. TABLE 2 Results of precipitin and complement fixation tests with human non-tuberculous transudates and exudates PRE- |, oun COMPLEMENT CASE CIPITIN we") FIXATION REACTION NUMBER AGE SEX CLINICAL DIAGNOSIS REAC- ee TLON S| conn 4/4 | 4] 24}as| Control years cc 2 47 F | Cardiac insufficiency _ OI |Se 22h 3 14 M | Pneumococcus pleuritis — 0.6 |—|—|—|-|-| — 14 /adult} M | Pneumococcus pleuritis _ 0.06;—|/—|—|—|—| — 16 26 M | Cardiac insufficiency* — 0:90 ee 17 40 F | Carcinoma of pleurae — 0.6 |—|—|—|-—|— = * Patient, a syphilitic with aortic insufficiency; blood serum gave positive Wassermann reactions. No evidences of tuberculosis at necropsy. Two pericardial fluids were examined; both yielded nega- tive precipitin reactions while one (removed eighteen days after infection) yielded strongly positive complement fixation reactions. Five peritoneal fluids were examined; all yielded negative’ precipitin and complement fixation reactions. In this con- nection it may be stated that tubercles were not found in the peritoneum of any of these animals (2), although present in the liver and spleen of a few. The sera of a number of these animals were also tested and the results observed with nine are shown in table 4. All yielded negative precipitin reactions and five yelded positive comple- ISAMU OGAWA 428 = = = a =) ee eee ae) ap = 5 a = = SRL |e ste a atest ae BO —_ i as = = aS ane hate 6°0 —— = ica = ee as 2 Pe: 60 cs = sh rs = S = = g0°0 i = = ss a Be. _ ae z'0 = a = a srs = = - 9°0 — — 25 = = Rae a ae oO = a r = = = he legit) 6rd 2 x4 = i = % = = 512070 = = = ae = = = ae 6'0 = oo = aa = a» vo a 6°0 % = a = = = > le | 3830 a = = = = = = ore ee is = — = = = Se Re £0 —- = ps ae = as = =: 20 = ce = es = = = eel 20 a = = ez = = ee ea 20 = — = Sea = = = sm £0 = = om = = AS = an 20 — aa = a ba = ea _ £0 — S = = = = 2 = e°0 99 [or1yUOD A on oo ap ud ut g ¥ TINO —_——eea hk _—_———_—_—vknkea—<«< | LUVIN NOILOVAY NOILVXIT INUNATAWOO -ILNV NOIL “OVAL -GI1dWO9} NIdtdIo “Gud VITIOVa GTO “Udon [einoeyd 4yory [eanoyd 4yysry [vanoyd 4yory [veined yysry [vouo4ylIi0g [BIpsivoli8g [einoeyd 4yorT [vaneyd yy s1yy [eanoyd 4yjorT [vouoyliog [einoyd 4yory [einoyd 4ysryy [vaneyd 4yory [veined yy sry [eane]d 4yjoryT [erne]d 4yysryy [vou04yl1og [weaned 4yjory [sanoeyd 4yysry [vouozlog [eineyd 4joT [veined 4yysryy auisaL drat ial iat OL Z el reer ee a a BO ee ed ee ee ee sbid nau fo syrinajd SNOINILIQN) [D]WIULLOd XA UL SaIMpnxa JoaUopLtod pup porpavoruad ‘ounapd yjUn $}89} UoYory puamaduos pun uyrdroa4d Co syynsaay § WIAViL 429 REACTIONS WITH TUBERCULOUS. EXUDATES a a + | ++ | 44+] 44+ + 4 ++ | +44 [++4+4/++4++4 ++ | +44 |+4+4+4/4+4+4+4 +++] +++] +++ Ede pore ae re et ce | ee ceaee lr vereraere, Gra wee ied nar ar ana + | $4t [tt4t]tt+4]t+t4]t4+t4]444+ seo) GENES Stee peso teers ater £0 £0 T0 Sighs | ectectertonieteste te tet ros 0 20 GheskeP ale ctiata te esate stl een) Be teats ¢°0 ole alia £0 oe Test €°0 = tah £0 = G0°0 O'T OT O'T O'T sack Sleds 80 = Chaat 6°0 a a €°0 = ez £0 ot ollateea 8-0 St ea Stet tea OO 10 [earned 4yory [eaneyd 4qsry eaneyd 4ysry [eaneyd yyoryT [eaneyd 4yJorT [eaneyd 4y3Iyy [eaneyd yoy emnetd yysINy [eaineyd 4yorT [emngyd yay [BouopIOg [BIprvoli9g [eineld 4yysIyy jeineyd 4ye'T [eined yy sIyy [veined yyorT [eaneyd yy sry [earned 4yorT [emoyd yusIyy [eineld 4yo'T Terngyd 4ysTy 430 ISAMU OGAWA ment fixation reactions. One of these animals showed very ex- tensive bilateral pleural, pulmonary and pericardial tuberculosis. All of the sera yielding positive reactions were from animais six- teen days or longer after inoculation. The sera of thirty control guinea pigs yielded uniformly nega- tive complement fixation reactions. The results observed with pleural, pericardial and peritoneal fluids from rabbits infected with intrapleural injections of tuber- cle bacilli, are summarized in table 5. These animals did not TABLE 4 Results of precipitin and complement fixation tests with the blood sera of guinea pigs with experimental tuberculous pleuritis DAYS PRECIPI- COMPLEMENT FIXATION REACTIONS ANIMAL SINCE IN- | TIN REAC- Cores) ano 0.1 0.05 0.025 0.0125 0.006 | Control Z 5 4 a a = is a az = Z 6 8 = = = be S at: f M6 8 a = = be = af gE Z9 15 = a - ee =e = Es G3 16 a. \\ccgmmecnilaetbarineerararasceis | oaeas = Ta 21 2 ard 4 mn a = = D7 26 = Seas -F + - _ _ D8 26 = 4s a a = = = D9 26 = Siochs ais (El inte Gia + - — - show as well marked experimental lesions as the maiority of guinea pigs (2). Fourteen days or longer after infection the majority of these fluids yielded positive complement fixation reactions and es- pecially the fluids from the right pleural cavities; only one fluid (right pleural cavity) yielded a positive precipitin reaction. The sera of all animals yielded positive complement fixation reactions, but owing to the well known property of normal rab- bit sera for yielding positive non-specific complement fixation reactions, little or no significance can be attached to these results. The reactions observed with the exudates, however, were specific insofar as could be determined. 431 REACTIONS WITH TUBERCULOUS EXUDATES *sumolyoveI UIyIdI9O ~o1d aAtyedou paplers vias [[v /SUOTJOVEL UOIYexY YUEtIETdwM0d stsopnoJEqny dArzISod papyerd syIqq va Soyy JO BIOS 9YY JO TTY x ex sa ro | | | b+i++ a + - + | +4 [44+ ]44++4]4+4++4] ¢ - + | +4 [44+ ]44+4+4]4+4+4+4] © +++] a0 | - - +++ |- €°0 = = ++ | sto = - +++] ¢°0 ~ - ++ | S10 = = tes | SEO = = 6 = = 0 — — — |+t+4]++4+4]t++4+4]t+4++4+]4+4+4+4) ¢0 | - - - — | +4 | +++ |++4+4/4+4++4] eo | - - ~ ++ |44++4]+4+4+4]+4+4+4/44+4+4] 0 | - - - - + | ++ | +++ - - - + | ++ ~ - + | +++] +++ - - + | +4 | +++ ~ ~ - ~ + - - ~ - + ~ + | + | + [+++ jou | x y : NOILOVGAY NOILVXId INAWATANWOO =r €0°0 0 0 ele le schama| = Sales Seats 9°0 = = Sh iPae 9°0 = = Se aReie 9'°0 = = alaet= 9°0 = = ate €°0 = z= ac €°0 = oS ol £0 = = Samara UE) nat “ti "99 : RTE are a) rove “warioD sarogua |TTOUTaOL xsupiinayd snojnawaqn} [veuozIog [SIpivdt1eg [eeuozIIOg [BIp{vor10 g "BI 939T [BIpalvol1e8g [BIpivo1108g [eIprvot108g [vouoylIeg [BIpazvorl1e8g [eineld 4yary [vouoqIeg [bined 4yysry [vouoyl1eg [einey{d 4joT [vane] d yysry [BouoylIog [BIprvolsg [veined 4yo'T [einetd 4qsryy [eineyd 4yysry aivaoxa ce GE T€ 0€ a NOILVTOOONI GONIS BAVA 6d 8d AIXGl rd 6d Td tad Rhett 6a 6a oad IVWINY [DpUIUILAdxa UO SaIDpnxA JDIUO}IWad PUD pDUpADIWId “Younajyd ypN $7802 UoYorY yuaumajdwoo pun uyrdrwoud fo szpnsaaz $ WIAViL 432 ISAMU OGAWA DISCUSSION These results indicate that in the exudates from tuberculosis of serous cavities, precipitins and complement fixing antibodies may be found, especially the latter when a sensitive antigen and technic are employed. In a study of the Wassermann test with exudates and transu- dates in syphilis, Klauder and Kolmer (6) observed uniformly positive reactions with inflammatory exudates when the reac- tions with blood sera were negative. Secretions from chan- cres yielded positive reactions before the syphilis ‘‘reagin” could be found in the blood sera indicating a local production of the antibody responsible for the Wassermann reaction. The results of these studies in syphilis in conjunction with those reported in this paper, indicate that specific antibodies are to be found in the exudates of localized infections at a time when they cannot be demonstrated in the blood sera probably because of their high dilution in the latter. These exudates are readily adapted for complement fixation tests if special attention is given to the following technical steps: 1. The fluid should be centrifuged to remove the cells. 2. The fluid should be heated at 55°C. for fifteen minutes to remove complement, thermolabile hemolysin, thermolabile anti- lysin, if present, and the substance responsible for the proteo- tropic reaction. For these purposes longer periods of heating are unnecessary and result in a useless destruction of antibody, as shown by the studies of Koimer and his associates in syphilis. 3. The fluid should be titrated for anticomplementary activity as they vary greatly in this property and cannot be employed in a fixed amount; similar findings were reported by Klauder and Kolmer in their study of exudates and transudates in syphilis. Since some fluids are markedly hemolytic it is advisable to use varying amounts beginning with 34 or 3 of the anticomple- mentary unit as employed in this study. The nature of the hemolytic substance sometimes found in these fluids has not yet been determined. REACTIONS WITH TUBERCULOUS EXUDATES 433 4, The antigen and complement fixation technic should be as sensitive as consistent with specificity. As previously stated, the antigen employed for tests with human fluids should be free of lipoids as prepared by Kolmer, in order to avoid cross complement fixation tests with tuberculous exudates from syphil- itic individuals. 5. It is advisable to employ a control on each amount of exu- date tested in order to avoid error with fluids which are hemolytic in large amounts but anticomplementary in smaller amounts. With these technical considerations it is believed that the com- plement fixation test will prove of value in the diagnosis of tuberculous pleuritis of fifteen days or longer duration. CONCLUSIONS 1. Of a series of twelve tuberculous pleural exudates, 42 per cent yielded weakly positive precipitin and 92 per cent well defined positive complement fixation reactions. 2. Human pleural exudates and transudates of non-tuberculous origin yielded uniformly negative precipitin and tuberculosis complement fixation reactions. In syphilis, however, positive reactions may occur due to the presence of the ‘‘reagin’”’ in the exudate unless precautions are taken to remove the lipoids from the antigen of tubercle bacilli. 3. With the pleural exudates secured fifteen days or longer after experimental tuberculous pleuritis in guinea pigs, 8 per cent yielded weakly positive precipitin reactions and 89 per cent strongly positive complement fixation reactions. 4, In experimental tuberculous pleuritis and pericarditis of guinea pigs and rabbits, precipitins and complement fixing anti- bodies are not usually found earlier than twelve days after infection. 5. These results indicate that in the exudates of tuberculous pleuritis, precipitins and especially complement fixing antibodies, are found in a large percentage and that a sensitive complement fixation test with special attention to certain technical features, may prove a valuable practical aid to diagnosis. 434 ISAMU OGAWA REFERENCES (1) Oaawa, I.: Is pleuritis ““army disease?’’ Igaku Chu-O Zosshi, 1913, 169. (2) Kotmer, J. A. anp Ocawa, I.: Experimental tuberculous pleuritis in rela- tion to chemotherapeutic studies in tuberculosis. (To be pub- lished.) (3) Kotmer, J. A.: A new complement fixation test for bacterial infections. (To be published.) (4) Perrorr, 8. A.: Glycerin extract of tubercle bacilli as an antigen in comple- ment fixation. Amer. Rev. of Tuberculosis, 1918, 2, 523. (5) Kotmer, J. A., AND YAGLE, E.: Bacterial antigens in relation to complement fixation in syphilis with special reference to antigens of tubercle bacilli. (To be published.) (6) Kuauper, J. V., AND Kotmer, J. A.: Wassermann test with secretions, transudates and exudates in syphilis. Jour. Amer. Med. Assoc., 1921, 76, 1635. ON THE ORIGIN AND NATURE OF ALEXIN (COMPLE- MENT) IN GUINEA-PIG BLOOD L. F. MORRISON From the Department of Bacteriology and Experimental Pathology, University of California, Berkeley, California Received for publication May 18, 1922 Nomenclature has always played an important réle in scientific literature. The accepted ethics of terminology recognize and uphold the precedence of priority in cases of dispute providing that the earlier investigator fully describes the substance in question. Owing to individual idiosyncrasies this procedure, in some instances, has been disregarded and serious confusion and controversy has arisen. Nuttall (1), von Fodor (2) and others recognized and experi- mentally demonstrated the presence of a bactericidal substance in fresh blood and blood serum but failed to attribute a name to the substance. Buchner (3), in 1889, gave the name of ‘‘alexin”’ to this thermolable, bactericidal, ferment-like, protecting sub- stance. Bordet (4) later verified the presence of this substance and recognized the accuracy of the word ‘‘alexin.”” In addition he discovered the thermostable factor, which also takes an important part in this reaction. The name alexin was used until Ehrlich and Morgenroth (28) working with the identical, thermolabile property of blood serum some eleven years later, disregarded the existing terminology and described the substance under the name of ‘‘complement.”’ In view of the historical background and the accepted procedure for the adoption of nomenclature it is a singular fact that the standard English texts, with one exception (5) dealing, in part, with the subject of alexin have accepted the later terminology and incorrectly discuss the subject under the heading of “complement.” We can only attribute the continuance of the use of this term to the undue precedence given it by the German school. 435 436 L. F. MORRISON Apart from the question of priority it is unfortunate that | the use of the word ‘‘complement” should continue since it indicates an acceptance of the theory of Ehrlich which is certainly relinquishing its hold on the scientific world. Owing to the fact that the historical and bibliographical back- ground on the subject of alexin has been so ably presented by other investigators only such references as are deemed essential for corroboration or correlation will be incorporated. This paper will take up the subject of the alexin content and the nature of alexin in the blood of normal male guinea-pigs. I. QUANTITATIVE DETERMINATION OF ALEXIN Time and temperature relations in clotted and defibrinated blood Technic. The technic employed in obtaining guinea-pig blood and in titrating its alexin content was in all cases as nearly identical as possible. Normal male guinea-pigs were anaesthetized with ether and the blood obtained by heart puncture. The blood was either allowed to clot, was defibrinated by agitation in a sterile flask containing glass beads or was subjected to other treatment as demanded by the various experiments. In every instance the pooled serum from at least four normal male guinea-pigs and in a few instances as many as forty guinea- pigs was used in order that individual fluctuations in alexin content should not enter into the general considerations. In titrating the alexin content of each sample two series were set up. One series starting at a dilution of 1:2, the other starting at a dilution of 1:10. Subsequent dilutions of these were made in 0.85 per cent NaCl solution, running in multiples of two, resulting in dilutions of 1/2, 1/4, 1/8, 1/10, 1/16, 1/20 and so on to a point well beyond the anticipated titer limit. Intermediate dilutions were not deemed essential as gross differences in alexin content were expected and found. The hemolytic system employed was the usual rabbit serum-anti-sheep cell combination in which three units of the immune serum in 0.5 cc. normal salt solution, 0.5 cc. of a 3 per cent suspension of washed sedimented sheep cells together with varying amounts of reactivating substance in a total volume of 0.5 ce., were employed. After the contents of the tubes were thoroughly mixed, the tubes were incubated for thirty minutes at 37°C. in a water-bath. Readings were made immediately after the tubes were removed from the water-bath and again after they had been ALEXIN IN GUINEA-PIG BLOOD 437 allowed to stand over night in the ice-box. When radical differences in the two readings were encountered the experiment was repeated until check titrations disclosed the cause. The necessary controls were set up for each experiment. Experiment 1. Normal male guinea-pigs were bled from the heart and the blood of each pig divided into two portions, one of which was defibrinated, the other allowed to clot. As soon as the blood had clotted it was freed from the walls of the container by means of an ordinary nichrome-iron planting wire. After about ten minutes sufficient serum had been squeezed from the clot by contraction of the fibrin net-work to make titration dilutions. The defibrinated blood was centrifugalized for five minutes at 2000 r.p.m. and the supernatant serum pipetted off into a sterile container. Thus we have a quantity of defibrinated blood serum removed from the fibrin clot and cells and a quantity of clotted blood serum which, in this instance, was allowed to remain on the clot until used in the experiment—only sufficient being removed each time to carry on the necessary titrations. Both sera were allowed to remain at room temperature or about 22°C. An initial titration was made one-half hour after the blood was obtained. Subsequent titrations were made at two-hour-intervals for the first twenty-four hours and at twenty-four-hour-intervals thenceforth until the alexin content of both sera had become negligible. Titration curves shown in chart 1, plotted from the results of various determinations, show that there is a gross initial difference between the alexin content of the clotted blood serum and the defibrinated blood serum—the former giving a titer of 1 to 10 and the latter a titer of 1 to 64. Following the respective curves one finds that the clotted blood serum rapidly increases in alexin content for the first twenty-four hours and then decreases, at first rapidly, and then less so as time goes on, until at the end of one hundred and forty hours it ceases to be demonstrable. The defibrinated blood serum retains its potency for the first six hours and then rapidly decreases. At the end of twenty-four hours the curve flattens out and the decrease in alexin content is gradual until at the end of one hundred and thirty hours it is negligible. Several repetitions of the above experiment were performed and in no instance were gross deviations from the incorporated curves encountered. In anumber of instances the titer of the defibrin- THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 5 438 L. F. MORRISON ated blood sera was found to be much higher—in a few instances giving a titration in a dilution of 1: 80—and often the initial titer of the clotted blood serum was as low as 1 to 2. The titration limit of the clotted blood serum never reached the initial titer of the defibrinated blood serum. Following these experiments, in which the sera were kept at room temperature, a series of experiments were set up so that temperatures from 4°C. to 37°C. could be maintained. The results obtained correspond closely to those obtained by Douglas @eecoeeeceaeooeees oeceosae@rveeeceseeceeeeeseeeaeeseeseovesescesseeeeaeeoeeeeeseeesey 3 O SF 1S 20 2530 35 40 4S SOSS & 6570 75 GO BE FO FS 100 /OSHa US 120 125 1D IFT 190 19s fime in hours, I aesesot Defibrinated blood serum, Clotted blood serum, CuHart 1. GRAPHIC REPRESENTATION OF THE COMPARATIVE TITERS OF ALBEXIN IN THE SERUM OF DEFIBRINATED AND CLoTTED BLoopD IN RELATION TO TIME and Bigger (6), Bigger (7) and Massol and Grysez (8). In their experiments they used serum which had been allowed to remain on the clot for twenty-four hours before being removed and therefore contained the maximum amount of alexin produced by this method. It was found in these experiments, as well as in the former, that the alexin content of the defibrinated blood serum was highest shortly after defibrination and decreased more rapidly at first than it did during the later period of its activity. The differences noticed between the sera kept at ice-box (4°C.) and incubator (37°C.), and those kept at room temperature were that in the first case the process of the decrease in alexin content of ‘ ALEXIN IN GUINEA-PIG BLOOD 439 the serum was greatly retarded whereas at 37°C. it was greatly increased. Attempts to find why the defibrinated blood serum should have an intial alexin titer greater than that ultimately attained by the clotted blood serum allowed to remain on the clot were fruitless. Without going into unnecessary detail some of the main experiments were conducted as follows: TABLE 1 TIME A B Cc D E k hays Pe Sena | a PANN Lenre TT ny er SUNT CR RG AR NEN: 3 1 /64 1 /64 1 /64 1 /64 1 /64 1 /64 24 1 /64 1 /64 1 /64 1/40 luBy 1/40 48 1/40 1/40 1/32 1/32 1/20 1/32 72 1/32 1 /32 1 /32 1/20 1/20 1/20 96 1/20 1 /32 1/20 1/16 1/4 1/20 104 1/10 1/20 ~ 1/8 1/2 1/10 * Serum contaminated. TABLE 2 TIME A B Cc D E F G H 3 |Negative| 1/40 1 /40 1/40 1/40 1/4 1/4 Negative 24 Negative} 1/20 1/16 1/20 1 /16 1/82 1/16 |Negative 48 |Negativel 1/20 | 1/16 | 1/16 | 1/10 | 1/20 | 1/10 |Negative 72 Negative] 1/16 1/10 1/16 1/8 1/16 1/8 |Negative Experiment 2. Guinea-pigs were bled and the blood defibrinated. The defibrinated blood was treated as follows: A. Whole defibrinated blood containing serum, red cells, white cells and fibrin clot. B. Defibrinated blood removed from the fibrin clot. C. Defibrinated blood serum and red cells. No fibrin. D. Defibrinated blood serum with white cells, a few red cells and no fibrin. E. Defibrinated blood serum and fibrin clot. No cells. F. Defibrinated blood serum alone. In order to obtain the component parts essential for setting up C, D, E, and F, the defibrinated blood was centrifugalized for ten minutes at 2000 r.p.m. and the supernatant serum divided into four 440 L. F. MORRISON portions. The white layer of leucocytes was then pipetted off and placed in one of the portions giving (D). The major part of the sedi- mented red cells was decanted to eliminate as many of the remaining white cells as possible. The remaining cells were mixed with another portion of the defibrinated blood serum giving C. E was set up by the addition of a small piece of the fibrin clot, formed during the process of defibrination, which had the cells removed by washing it in sterile salt solution, to a third portion of the defibrinated blood serum. The results of the titrations are seen in table 1. After repeating the experiment several times and obtaining similar results it was thought that the leucocytic factor could be controlled by proceeding with the problem in a different manner. Experiment 3. The leucocytes used in the following experiment were produced by the intraperitoneal injection of extract broth into guinea-pigs and removal of the exudate the following day. Guinea-pigs were bled and the blood treated as follows: F. Allowed to clot. G. Allowed to clot after the addition of leucocytes. B. Defibrinated, removed from the fibrin clot. C. Defibrinated after the addition of leucocytes and the blood removed from the fibrin clot. D. Defibrinated blood serum removed from the cells by centrifugali- zation. FE. Defibrinated blood serum removed from the cells by centrifugali- zation with additional leucocytes. A. Control. Leucocytes and salt solution subjected to a similar mechanical agitation as the blood during the process of defibrination. H. Control. Mixture of leucocytes and salt solution unagitated. In the above experiment the leucocytes were added in quantities of one-tenth volume of the clotted blood, defibrinated blood or defibrinated blood serum, as the case might be, as a suspension in the peritoneal exudate. Owing to the fact that the peritoneal exudate may have acted as the inhibitory element the same protocol was employed with the addition of a control on the possible activity of the peritoneal exudate. The peritoneal exudate containing leucocytes was divided into two equal parts. One part was untreated. The other part was centrifugalized and the supernatant serous exudate decanted. The sedimented leucocytes were washed once with sterile normal salt solution and then resuspended to the former volume in additional salt solution. One series of tubes similar to the preceding protocol were ALEXIN IN GUINEA-PIG BLOOD 441 set up with the untreated peritoneal exudate. ‘Two other series were set up, one in which the supernatant serous exudate was substituted for the untreated peritoneal fluid and the other wherein the suspension of leucocytes in normal saline was used. Titrations gave similar results for all three conditions. Therefore the results tabulated in table 2 were not influenced in any way by the presence of the serous portion of the peritoneal exudate. It is evident that the leucocytes do not contribute anything to the alexin content of clotted blood, defibrinated blood or defi- brinated blood serum, but they act as inhibitory factors on the already present alexin. One further ramification in our endeavor to locate the seat of alexin production lay in the study of the actual number and integ- rity of the leucocytes, in conjunction with the respective alexin contents, of two samples of defibrinated blood. One sample of the defibrinated blood to be subjected to ice-box temperature or about 6°C. and the other allowed to remain at room temperature. Experiment 4. CGuinea-pigs were bled, the blood divided into two portions and subjected to the above mentioned temperature conditions. Immediately after the process of defibrination was completed and leucocyte counts had been made, both samples of blood were centrifu- galized for five minutes at 2000 r.p.m. and sufficient of the supernatant serum removed for titration purposes. Leucocyte counts and alexin titrations were made almost simultaneously. Subsequent leucocyte counts and titrations were made at intervals of twenty-four hours during a period of one hundred and forty-four hours. After each centrifugalization and removal of a small quantity of serum the sedi- mented cells and supernatant serum were thoroughly mixed by means of a pipette before the leucocyte counts were made. The initial titration of the serum of the two samples of blood gave a titer limit of 1/64 and a leucocyte count of 3000 per cubic millimeter of defibrinated blood. At the end of the one hundred and forty-four hour period the blood kept at ice-box temperature gave a serum alexin content titer of 1/16 and a leucocyte count of 2,700 per cubic millimeter of blood whereas the corresponding results for the blood allowed to remain at room temperature were 1/2 and and 2850. The decrease in alexin content showed no anomalies. 442 L. F. MORRISON During the experiment several smears of each sample were made and stained by Wright’s staining method to discover, if present, any structural differences in the appearance of the leucocytes. None were apparent. In the case of the defibri- nated blood kept at room temperature the erythrocytes became very fragile during the later portion of the experiment and considerable autolysis was evidenced by the deep red hue taken on by the serum. It is evident that the differences in alexin content of the serum of defibrinated blood as compared with clotted blood when the leucocyte counts are taken into considera- tion, cannot be explained by the appearance or numbers of leucocytes. Although there is a marked difference between clotted and defibrinated blood serum at first there is no reason to conclude that this difference is due to the presence or absence of compact fibrin or to the presence, absence, or destruction of either the red or white blood corpuscles. II. COMPARISON OF THE ALEXIN CONTENTS OF CLOTTED BLOOD AND PLASMA The work of Gengou (9), later substantiated by Herman (10), apparently demonstrating the absence of alexin as a normal constituent of blood plasma, caused considerable attention to be directed toward the properties of plasma. If this assertion —that plasma does not contain any demonstrable amount of alexin—were true, it would materially assist and substantiate the the claims of the Metchnikoff school in that alexin is a leucocytic product produced in serum after the clotting of blood. However such is not the case. Numerous investigators headed by Domery (11) and running through the recent work of Addis (12) have been able to refute the contentions of Gengou and Herman on a definite and concrete experimental basis. Not only did they repeat the actual experiments of Gengou and obtain diametrically opposite results but, with additional means at their command, conclusively demonstrated the presence of alexin in blood plasma obtained in a number of ways. ALEXIN IN GUINEA-PIG BLOOD 443 For the sake of brevity the results of the titrations of the alexin content of plasma obtained in a number of ways and the necessary control tubes, run simultaneously, are incorporated in a single table which follows the technical and experimental data about to be presented. After reviewing the literature on the subject of methods for obtaining plasma, three were selected. They were: Method 1. Salted plasma (Bordet and Gengou, 13) A sufficient quantity of 20 per cent NaCl solution is added to the whole blood before it clots to give a final concentration of 5 per cent NaCl per total volume of blood and diluent. The salted blood showed no tendency to clot. This salted blood was divided into two main portions, one of which was allowed to remain untreated until needed. The other was centrifugalized and the supernatant salted plasma removed. The salted blood was divided into two portions, 4 parts of distilled water added to each, and one part allowed to clot. The other portion was defibrinated, centrifugalized for ten minutes at 2000 r.p.m. and the supernatant serum removed from the cells. The process of clotting and defibrination took about thirty minutes. The salted plasma was treated in a similar manner except that in the case of the defibrinated plasma—no cells being present—centrifugalization was unnecessary and the plasma serum was decanted from the fibrin clot which had formed around the beads. In all four instances the serum obtained was of a bright red hue showing that the cells had suffered considerably from the effects of the concentrated salt solution. One point that may prove to be of interest was noticed in the process of defibrination of the salted plasma. When defibrination was started few, if any, bubbles collected on the surface of the fluid but about twenty minutes later, when definite fibrin strands were in evidence, a heavy foam formed and persisted for many hours after the process of defibrination was completed. Titration of the four types of sera along with diluted and undiluted defibrinated and clotted blood sera were run simultaneously. In the case of the diluted, clotted and defibrinated blood sera the blood, before being defibrinated or allowed to clot, was diluted with 0.85 per cent NaCl solution to a point equal to the final dilution of the salted blood and plasma sera. 444 L. F. MORRISON Method 2. Oxalated plasma (Watanabe, 14) Sufficient 1 per cent sodium oxalate solution is added to whole blood so that a final concentration of 0.001 mgm. of the salt per cubic centi- meter of blood and diluent is obtained. The oxalated blood was divided into two portions, one of which was centrifugalized for ten minutes at 2000 r.p.m. and the supernatant oxalated plasma removed. Wata- nabe found that as much as 0.005 mgm. of the sodium salt could be added without interfering with the alexin content of the plasma or of fresh guinea-pig serum to which the salt has been added after the process of clotting has taken place. The oxalated blood was divided into two portions and sufficient quantities of 1 per cent CaCl-2 solution added to each to give a final concentration of 0.001 mgm. of CaCl-2 per cubic centimeter of oxalated blood and diluent. One part was allowed to clot and the other defi- brinated, centrifugalized for ten minutes at 2000 r.p.m. and the super- natant serum removed from the cells. In this manner a crystal clear, almost colorless serum was obtained. The fact that the serum was almost colorless denotes that the cells had suffered a minimal amount of injury. The oxalated plasma was treated in a manner similar to the oxalated blood. Clotting took place in the oxalated, restored blood and plasma within three minutes after the addition of the CaCl-2 solution. Titrations of these four sera were run simultaneously with a sample of oxalated, unrestored plasma and normal clotted and defibrinated blood sera as controls. Owing to the fact that additional substances such as sodium oxalate and calcium chloride had been incorporated into the above experiment it was essential to determine their action, singly and in combination, on guinea-pig serum of known alexin content. Three series were set up, one containing fresh guinea-pig serum with the addition of 1/10 volume of 1 per cent sodium oxalate solution, another with fresh guinea- pig serum with the addition of 1/10 volume of 1 per cent calcium chloride solution, and the third containing 1/10 volumes of both salts added to fresh guinea-pig serum. The guinea-pig serum used through- out was from the same bulk material and therefore of the same alexin content titer. In the first two instances no interference with the alexin content titer was exhibited. In the third instance, wherein a precipi- tate was formed, a slight decrease in alexin content of the serum was evidenced. This decrease was probably due to the adsorption effected by the precipitate. ALEXIN IN GUINEA-PIG BLOOD 445 Method 3. Paraffin plasma (Gengou, 9) Conical centrifuge tubes were heavily coated with paraffin, chilled and kept at 0°C. until wanted. Fresh guinea-pig blood was obtained and, before it had clotted, was placed in these tubes and centrif- ugalized at a low temperature, for five minutes at 2000r.p.m. The supernatant, fluid plasma was removed by means of a chilled, previ- ously paraffined pipette and divided into two portions, one of which was allowed to clot, the other defibrinated. The plasma serum ob- tained was crystal clear and colorless. With controls of normal clotted and defibrinated blood sera the alexin contents of the clotted and defibrinated plasma sera were determined. TABLE 3 Showing the average results obtained in the various experiments dealing with the alexin content of plasma, obtained by the previously described methods, together with the average results of the alexin content determinations of clotted and defibrin- ated blood, run simultaneously, as controls ee CLOTTED BLOOD |CLOTTED PLASMA | DEFIBRINATED | DEFIBRINATED OXALATED SERUM SERUM BLOOD SERUM | PLASMA SERUM PLASMA hours ba SEL Oe 2 1/10 1 /36.1 1 /45.3 1 /36.8 1/40 24 1/27.5 1 /24.4 1/28 1 /26.4 1/30 48 1/20.6 1/26 1/20 1/18.4 1/20 72 1/17.3 1/10 1/18.6 1/10 1/18 96 1/13 1/9 1/13 1/9 1/12 Before drawing any conclusions from table 3, one point of inter- est deserves mention. During the preceding experiments with salted blood and plasma considerable dilutions were encountered. This necessitated a corresponding dilution of normal whole blood, prior to its defibrination or clotting, as a control measure and so that the subsequent dilutions in the titrations should be com- parable to those of the salted material. In the case of the defibri- nated, diluted blood serum nothing extraordinary was noticed. The initial alexin titer of the serum squeezed from the clot of the diluted clotted blood, by the natural contraction of the fibrin strands, was much higher than that of the undiluted clotted blood serum titrated concurrently. 'This may be explained by the fact that the diluted blood clot contracted much more rapidly than did the undiluted blood clot. The importance of the contraction of 446 L. F. MORRISON the fibrin net-work will be taken up in detail under the heading of “ Discussion.” The results tabulated in table 3 bring out two main points which have been subjects of much discussion. First, they refute — the work of Gengou and Herman and corroborate the results obtained by Domery (11) and his successors by conclusively demonstrating the presence of alexin in plasma in equal amounts to that found in the corresponding blood, when allowed to clot, at its highest period of alexin content. And secondly, the demon- stration of large amounts of alexin in the plasma, eliminates the fundamental argument of the Metchnikoff school in favor of the leucocytes as the source of alexin in blood serum. In comparing the results obtained with clotted and defibrinated blood plasma and serum one readily admits the advantage of the removal of the fibrin mesh work over the usual methods of allowing it to remain intact. III. REACTIVATION OF THE ALEXIN CONTENT OF AGED GUINEA-PIG SERUM On directing our attention to the properties and nature of alexin rather than to its source, it came to our attention that several investigators, working with problems necessitating the use of alexin, had, at some time, found that old guinea-pig serum whose alexin content had ‘‘gone bad,” owing to the deleterious effects of time and incomplete refrigeration, would regain consid- erable activity by the addition of relatively small amounts of fresh guinea-pig serum. They found that this ‘“‘rejuvenated alexin”’ would fulfill the alexin requirements of their experiments as well as fresh guinea-pig serum which had been allowed to remain on the clot for twenty-four hours prior to its removal and use. This procedure has been followed empirically but not consistently by any one of the investigators. The literature on the subject of alexin makes no note of any such procedure. The nearest approach is the work (15) on the activation of the serum of hereditarily alexin-deficient guinea-pigs by the addition of small amounts of fresh normal serum from various sources or by the addition of such animal fluids as lymph and egg-white. The ALEXIN IN GUINEA-PIG BLOOD 447 subject of the regeneration of alexin in heated serum and serum subjected to the detrimental effects of radiation has been care- fully investigated by S. C. Brooks (16) and others. The suggestion that the addition of equal parts, or less than equal parts of fresh serum to guinea-pig serum whose alexin content has deteriorated will bring the titer of the mixture up to the titer of the fresh serum when titrated alone seemed worthy of investigation. Experiment 5. A quantity of defibrinated guinea-pig blood serum was obtained and set aside at room temperature until it has lost its alexin content. This process took fourteen days. Combinations of the old serum and fresh defibrinated guinea-pig serum were set up as follows: . Fresh defibrinated blood serum. . Old serum. . Old serum, 1 part and of fresh serum, 9 parts. . Equal parts of old and fresh sera. . Old serum, 9 parts and fresh serum, 1 part. . Old serum, 9.1 and fresh serum, 0.9 part. . Old serum, 9.5 parts and fresh serum, 0.5 part. . Old serum, 9.9 parts and fresh serum 0.1 part. TMorteypoawne All of the mixtures were set up in separate tubes, thoroughly mixed, and allowed to stand for five minutes before titrations on their respective alexin contents were made. TABLE 4 Showing the results obtained from the titration of old and fresh serum mixtures TIME A B C D E F G H minutes 5 1/80 |Negative} 1/80 1/80 1/64 1/40 BY 1/10 From the above data it is found that when equal parts of fresh guinea-pig serum, of high alexin content, are added to old guinea- pig serum, which has lost its alexin content by being allowed to remain at room temperature, the alexin content of the mixture is equal to that of the fresh serum when titrated alone. Assuming 448 L. F. MORRISON that the old serum acts simply as a diluent one would not expect an alexin content titration greater than 1/40 when the calculations are made on a basis of actual dilution. ‘Table 4 shows a titration of two times this calculated result. When calculated on the same basis the results obtained from the titration of a mixutre of old and fresh serum in proportions of nine to one respectively are far more striking. Results similar to those presented were obtained when old clotted blood serum was substituted in the place of the old defibrinated blood serum. The question of the reactivation of serum rapidly inactivated by being subjected to a temperature of 56°C. for one hour pre- sented itself. It was found that this serum was not reactivated by the addition of fresh serum but acted as an inhibitory factor to the alexin present in the fresh serum. Assuming that the heat inactivated serum acted only as a diluent the result of titrating a mixture containing equal parts of inactivated and fresh serum should be equal to one-half that obtained when the fresh serum was titrated. This wasnotthe case. The actual titration received was weakly hemolytic at one-fourth that of the fresh serum. This diminution of the actual alexin content of the fresh guinea- pig serum can be explained on the basis that the larger molecules caused by heating the serum offer themselves as adsorbents. Taking up the subject of the stability of the reactivated serum mixtures sufficient quantities of the various mixtures were set up in separate tubes so that there would be ample for several titrations. TABLE 5 Comparison of the stability of the alexin contents of fresh guinea-pig serum and reactivated old sera EQUAL PARTS OF OLD OLD SERUM AND TIME SERUM AND FRESH FRESH SERUM IN FRESH SERUM OLD SERUM SERUM RATIO OF 9:1 2 hours 2 1/80 1/80 1/80 Negative 24 1 /32 1/2 1/40 Negative 48 Negative Negative 1/32 Negative From the preceding data it is found that the reactivation process is a transient condition and furthermore that the amount ALEXIN IN GUINEA-PIG BLOOD 449 of reactivation is somewhat dependent upon the quantity of fresh guinea-pig serum used as the activator. The reaction depends, in stability, on the amount of fresh guinea-pig serum in- corporated with the old serum. That is, where equal parts of old and fresh serum are used the reactivation is more stable than in the case where only one part of fresh serum was added to nine parts of the old serum although the initial titer of the two be the same. One additional step in the procedure was incorporated at this point and run simultaneously with the preceding experiment. The cells from the freshly defibrinated blood were recentrifugal- ized for fifteen minutes at 2000 r.p.m. The supernatant serum and the upper portion of the sedimented cells were pipetted off and discarded, thereby removing the major portion of the serum and white cells present. The remaining cells appeared as a very viscous, dark red fluid. To this lower portion, consisting mainly of red cells, an equal part by volume of the alexin-free, fourteen- day-old serum was added. No mechanical injury was sustained TABLE 6 SERUM FROM FOAM OF SERUM FROM FLUID OF FIBRINATED AGITATED AGITATED, DEFIBRIN- BLOOD SERUM ATED BLOOD SERUM SERUM FROM DEFIBRIN- DEFIBRINATED BLOOD ATED BLOOD FOAM SERUM 1 /64 1/80 1/128 1 /64 by the red cells during the process of centrifugalization as there was little or no additional red color given to the serum when it was added to and mixed with the sedimented cells. In our previous experiments on the reactivation of old serum by the addition of fresh guinea-pig serum we found that the serum, low in alexin content, obtained from clotted blood a short time after it had clotted possessed very little, if any, reactivating property. It was thought that this ‘‘aggressive” or reactivating property might be present as an excretion of or extraction from the red cells. With this object in view titrations of the mixture were run after the mixture had been allowed to remain at room temperature for 450 L. F. MORRISON one-half hour, twenty-four and forty-eight hours. Respective titrations of 1 to 2, 1 to 2 and negative titrations were obtained. Even these titrations may have been due to the adherence of a small quantity of fresh serum to the cells. The work Zinsser (5) and others in trying to obtain alexin from leucocytes and the work of Neufeld (17) wherein he demonstrated the absence of alexin within the cell wall of the leucocytes by allowing them to phagocytize highly sensitized red blood corpus- cles and observed no intracellular hemolysis or intraphagocytic “shadow forms” is detrimental evidence for the theory of the leucocytic origin of alexin as proposed by many investigators. Considering these facts, together with the data herein presented, it is well within the limits of conservatism to eliminate the blood cells as the source of alexin, at least until evidence, other than that now available, is produced. IV. THE ENZYMATIC NATURE OF ALEXIN Solutions of true enzymes, when agitated until a definite foam is produced, show that the foam contains a greater quantity of the enzyme than does the underlying fluid. With this fact in mind it was suggested by Dr. Carl L. A. Schmidt that the foam produced by the rapid agitation of fresh guinea-pig serum might show a higher alexin content and further evidence the generally conceded ferment nature of this substance. Experiment 6. Guinea-pigs hearts blood was obtained and defi- brinated. During the process of defibrination considerable foam formed on the surface of the blood. The defibrinated blood was decanted, centrifugalized at 2000 r.p.m. for ten minutes and the supernatant serum removed. The foam produced during the process of defibrina- tion was pipetted into centrifuge tubes and centrifugalized for a similar period and its alexin content determined. The underlying serum was divided into two portions, one of which was untreated. The other was vigorously agitated for a few moments and the fluid serum decanted from the foam produced. The foam was pipetted into centrifuge tubes and centrifugalized at 2000 r.p.m. for one minute to break up the air bubbles. Titrations of the four types of sera were made one-half hour after the blood was obtained. ALEXIN IN GUINEA-PIG BLOOD 451 Samples of fresh clotted blood serum, of low alexin titer, were subjected to a treatment similar to that of the defibrinated blood serum but no differences in alexin content between the foam and the underlying fluid were recognizable by the method of titration employed. Experiment 7. The enzymatic nature of alexin was demonstrated by another method. Sand was thoroughly washed with distilled water, 94 per cent ethyl alcohol and lastly with ether. In this manner it was cleaned, dried and sterilized. Fresh defibrinated guinea-pig blood serum was obtained and divided into two portions. One was allowed to remain at room temperature. The other was added to a quantity of the sand, which had been previously chilled by being allowed to stand in the ice-box overnight, and agitated for five minutes. The chilling of the sand was deemed advisable for two reasons. First, it would retard the adsorp- tion of the alexin and, secondly, it would reduce the detrimental effects of the mechanical agitation to a minimum. Titrations were made one-half, twenty-four and forty-eight hours after the blood was obtained. TABLE 7 TIME 4 HOUR 24 HOURS | 48 HOURS Wistert bed) Berane 0). 0... Ska. wlanelseie ee ele eine ae © ake 1/80 1 /40 1/32 Serum agitated with sand...................+0-+-- 1/40 1/10 |Negative Se a ——————————————————E Under these two conditions it is evident that the alexin in guinea-pig serum reacts like a true enzyme. The first experiment is of little value other than being corrobora- ative evidence of the enzymatic nature of alexin. The later experiment (exper. 7) is of great importance, as will be shown in the hypothetical explanation of the reason for the difference between the alexin content of fresh defibrinated and clotted blood serum, in that it conclusively demonstrates the possible absorp- tion of alexin by foreign substances. DISCUSSION The advantages of the use of defibrinated blood serum over those of clotted blood serum to obtain alexin aremany. By the process 452 L. F. MORRISON of defibrination one is able to obtain a serum, ready for immediate use, with an alexin content almost double that of serum produced in any other manner. This fact offers itself as a great conserva- tion procedure in establishments using large quantities of alexin. The defibrinated blood serum deteriorates in alexin content no more rapidly than does the clotted blood serum. The disadvantages are few. Owing to the traumatic injury suffered by the cells during the process of defibrination the defibrinated blood serum contains slightly more hemoglobin than does the clotted blood serum, after being allowed to remain on the clot for twenty-four hours prior to its removal. In reviewing the historical background one finds that defibri- nated blood or defibrinated blood serum has been little employed as compared with the mass of experimental data wherein clotted blood serum plays the leading réle. Von Fodor (2) and Walker (18) worked with the bactericidal value of defibrinated blood serum and blood as compared with clotted blood serum and found that the defibrinated blood and serum were superior. Gurd (19) in working with the variation in alexin content of serum and plasma did use defibrinated blood serum but did not recognise the vast difference between it and the homologous clotted blood serum due to the fact that in defibrinating he agitated the blood for thirty-five minutes at room temperature. By this prolonged agitation at room temperature from one third to one half of the total alexin content procurable was lost. This fact has been clearly demonstrated by the work of Jacoby and Schutze (20), Zeissler (21), Ritz (22) and Noguchi and Bron- fenbrenner (23) who worked out the effects, at various tempera- tures, of mechanical agitation on the alexin content of serum. Walker (18), in explaining the difference between clotted blood serum and defibrinated blood serum in their action on cultures of B. typhosus, lays great stress on the leucocytes as being the source of the bacteriolysin and states that ‘‘while in whipped blood the whole available bacteriolysin of the blood is of course available from the onset and undergoes a steady diminution from the first; that of the serum while progressively deterioriating also, is, in the earlier hours, continually receiving fresh additions ALEXIN IN GUINEA-PIG BLOOD 453 from the clot. Accordingly the evidence supports apparently the view that the bacteriolytic ‘ferment’ is a leucocytic product, and is yielded to the serum by the gradual disintegration of the leucocytes during, and subsequent to the coagulation of the blood.” A similar explanation for the increase in alexin of clotted blood serum was sponsored by Gay and Ayer (24) and expresses the generally accepted opinions concerning the origin of alexin as stated by various investigators who have worked on the problem, with the exception of Fassin (25) and possibly Marbé (26) who lay great stress on the thyroid as the source. The contention of the Metchnikoff school that alexin is of leucocytic origin depends largely upon the findings of Gengou (9), substantiated by Herman (10), that there is no alexin in circulat- ing blood. He came to this conclusion from a series of experi- ments in which he was apparently able to demonstrate the absence of alexin in clotted plasma serum. However this is not true. Domery (11), followed by a series of investigators, was able to conclusively demonstrate results diametrically opposite to those obtained by Gengou. By repeating the actual experiments of Gengou he was able to demonstrate the presence of alexin in plasma in quantities equal to the greatest amount found in clotted blood serum which had been allowed to remain on the clot. Among the other investigators who were able to refute the contentions of Gengou and simultaneously confirm those of Domery one immediately recognizes the value of the contributions of Watanabe (14) and Addis (12) who not only worked with clotted plasma serum, but, by technical deviations, were able to determine the presence of large quantities of alexin in unclotted plasma, rapidly removed from the supposed sources of alexin with a minimal amount of cell injury. Why the leucocytic theory of the origin of alexin should hold such prestigé is difficult to explain in view of the absence of substantiating data. Neufeld (17) allowed leucocytes to phago- eytize highly sensitized red blood corpuscles and observed that no intraphagocytic hemolysis took place nor were “shadow forms” seen. The digestion of the ingested, highly sensitized red blood corpuscles progressed the same as the digestion of normal unsen- 454 L. F. MORRISON sitized cells. This conclusively demonstrates the absence of alexin, as one conceives alexin, within the cell wall of the leuco- cyte. Zinsser (15) allowed living leucocytes to remain, for a considerable period of time, in serum free from alexin and was unable to induce them to produce any demonstrable amount of alexin. All of these facts, together with that portion of this paper wherein the data show that leucocytes were given most advan- tageous conditions for the production of alexin and not only failed to do so but acted as an inhibitory factor on the already present alexin, present detrimental evidence to the theory that the leuco- cytes are the source of alexin production. Heretofore the importance of the extended fibrin network on the alexin content of clotted blood serum has not been recognised. This network presents itself as a large surface to which the alexin adheres. During the first twenty-four hours the fibrin strands contract, pack the cells together and squeeze out the imprisoned serum and alexin by reducing the amount of intracellular space. The tendency of alexin to adhere to extraneous objects has been demonstrated by various investigators, who, in trying to fraction- ate alexin, passed it through Berkfeld filters and found that the initial portion of the filtrate contained little or no alexin whereas the last portions contained alexin in quantities equal to that found in the serum prior to filtration. Furthermore, in the incor- porated experiment, wherein serum of high alexin content was mixed with clean sand, it was found that the sand removed a large portion of the alexin. This adhesion accounts for the low alexin content, during the first few hours, of clotted blood serum from blood which after clotting was carefully freed from the walls of the container and subjected to no further mechanical interfer- ence with its natural rate of contraction until sufficient serum had been excluded from the clot for titration purposes. The tube containing the blood clot was then centrifugalized at a low temperature. The serum forced out by this process was found to have an alexin content twenty times greater than that which was exuded by the natural contraction of the clot. Also it is noticed that the clot is greatly contracted. ALEXIN IN GUINEA-PIG BLOOD A455 As evidence in favor of the sponge-like action of the fibrin net- work of the clot and the lack of activity of the cells we will take into consideration the comparison of the alexin contents of clotted and defibrinated blood serum when studied from a time basis. The temperature relations of the two sera must, necessarily, be kept as a constant factor. Rather than select a single experiment to illustrate the point in question the average results of some twenty experiments, as shown in table 3, columns 1 and 3, will be used. According to the generally accepted theories the clotted blood serum, when allowed to remain on the clot, receives fresh additions from it in the way of serum and particularily extracts from the degenerating leucocytes during the first twenty-four to thirty-six hours following the clotting of the blood. When the rate of de- crease in alexin content of the defibrinated blood serum, removed from all of the supposed sources of alexin by centrifugalization, during the first twenty-four hours and the alexin contents of the two sera, after this first peroid has elapsed, are compared, it becomes evident that the increase in alexin content of the clotted blood serum is not due to leucocytic products. If products of the degenerating leucocytes are added to the serum they are of no help to the alexin content. If they were of significance the alexin content of the clotted blood serum would be, at the end of twenty- four hours, greater than that of the defibrinated blood serum. Such is not the case. The relatively low initial alexin content of the clotted blood serum is due to the adhesion of the alexin to the extended fibrin mesh-work. Furthermore, histological sections of clots of normal blood made shortly after clotting has taken place and after the clot has been allowed to stand for some time show cell differences in space relation and not structural differences. The cells in the later cases have been considerably distorted due to the pressure exerted by the contracting fibrin strands but there is no visible evidence to show that the cell membrane has been ruptured and the con- tents forced out. Nor is there sufficient difference in space relation, between the cells of normal circulating blood and those imprisoned within the clot, to lead one to believe that the cell contents have been forced through the membrane. 456 L. F. MORRISON In the experiments on the reactivation of guinea-pig serum inactivated by being heated to 56°C. for one hour and serum which had been allowed to remain at room temperature until no alexin content was demonstratable, striking differences were obtained. It was found that not only was no reactivation of the heated serum obtained but that it acted as an inhibitory sub- stance. The known alexin content of mixtures of fresh and heated serum was reduced to a point lower than could be accounted for by the actual dilutions made. During the heating process of the inactivated serum the molecular structure of the serum is changed, and aggregation of larger molecules resulting. The reduction of the alexin content of the fresh serum, added to the heated serum, past the point of actual dilution is due to the adsorption of the alexin by these larger molecules. The inactivation of the alexin content of fresh guinea-pig serum by continuous agitation, as carried out by Noguchi and Bronfen- brenner (23) and their predecessors (20, 21, 22,) may be explained by the same mechanism of molecular change. With the unheated serum, allowed to remain at room tempera- ture until it had lost its alexin content, a different state of affairs was demonstrated. By the addition of a relatively small amount of fresh serum, namely, 10 per cent, to the old serum which had no alexin content, the alexin content of the mixture was brought up to the same height as that of the fresh serum when titrated alone. This reactivated serum alexin is not as stable as that of the fresh alexin. The stability is directly dependent upon the relative amounts of fresh and old serum. The works of Brand (27) Ferrata and their contemporaries, on the dual nature of alexin should not be confused with the explana- tion about to be presented. This reactivation of guinea-pig serum, allowed to remain at room temperature until it has lost its titratable alexin, may be explained on the basis that there are two thermolabile portions of alexin. One fraction is more stable than the other and is unable, of itself, to react with the immune serum-sheep cell combination to bring about hemolysis of the sensitized sheep cells. The less stable fraction, found in large quantities in fresh defibrinated blood and clotted blood serum ALEXIN IN GUINEA-PIG BLOOD 457 which has been allowed to remain on the clot for twenty-four hours prior to its removal and use, is capable of uniting with a relatively large amount of the more stable factor, present in the old serum, and to reactivate it. If this idea of alexin being composed of two parts, one more thermolabile than the other, is not tenable, as further work on the subject may show, the reactivation of the old serum may be accounted for by the enzymatic nature of alexin in that it is, under favorable conditions and in the absence of inhibiting factors, capable of being active in concentrations less than the calculated titer limit. The enzymatic nature of alexin is demonstrated by comparing the action of guinea-pig serum, high in alexin content, with the generally accepted reactions of pure enzymes. First, by agitating the serum for a short time, producing a foam, and demonstrating that the foam has a greater alexin content than the fluid beneathit. Secondly, by the adsorption of alexin by foreign bodies, offering a large contact surface. In one instance, sand was used to demon- strate this adsorption. And thirdly, by its capability, under favorable circumstances, to act in dilutions greater than the cal- culated titer limit. CONCLUSIONS A far more powerful alexin, or complement, ready for immediate use, is obtained by the defibrination, centrifugalization and re- moval from the cells of guinea-pig blood serum than is obtained by the usual method of allowing the blood to clot and removing the serum after it has been allowed to stand on the clot for twenty- four hours. The same time and temperature relations, as have been worked out for clotted blood serum alexin, hold true for the defibrinated blood serum alexin content. The leucocytic theory of the origin of alexin is not tenable when the data detrimental to such a theory, now available, are taken into consideration. The experimental evidence presented show- ing the apparent inactivity of either the red or white blood corpuscles to produce alexin, together with the fact that blood 458 L. F. MORRISON plasma contains large amounts of alexin, offer themselves as strong arguments against such a theory. There is a gross difference between serum inactivated by heating to 56°C. for one hour and serum allowed to stand at room temperature until it has lost its alexin content, in their reactibility on the addition of fresh guinea-pig serum, high in alexin content. This difference is due to the adsorption of the alexin of the fresh serum by the larger molecules of the heated serum. The reactivation of the old serum may be explained by the enzymatic nature of alexin or by assuming that there are two thermolabile fractions of alexin, one being more susceptible to the detrimental effects of time and temperature than the other. Fresh guinea-pig serum allowed to stand at room temperatures for several days showed no further evidence of ‘‘complementoids.”’ Further data in corroboration of the enzymatic nature of alexin, were obtained. REFERENCES (1) Nurratt, G.: Zeit. f. Hyg. 1888, 4, 353. (2) von Fopor, J.: Centralbl. f. Bakt., 1890, 7, 753. (3) Bucuner, H.: Centralbl. f. Bakt., 1889, 5, 817; 6, 1. (4) Borpet, J.: Ann. de I’Inst. Past., 1899, 18, 225, 273. (5) Zinsser: Infection and Resistance. (Macmillan & Co., New York, 1914.) Chapter VII, p. 168. (6) Dovetas, J. S.C., AnD Biccrr, J. W.: Lancet, London, 1918, 2, 44. (7) Biecer, J. W.: Jour. Path. and Bact., 1919, 22, 323. (8) Masson, L., anp GrysEz, V.: Comp. rend. Soc. de Biol., 1910, 68, 825. (9) Grencou, O.: Ann. de |’Inst. Past., 1901, 15, 232. (10) Herman, M.: Bull. de l’Acad. Roy. de Med., 1904, 4 ser. 18, 137. (11) Domerry, P.: Wien. Klin. Wochenschr., 1902, 15, 1025. (142) Appts, T.: Jour. Inf. Dis,; 1912, 10, 200. (13) Borpet, J.. AnD GENGou, O.: Ann. de I’Inst. Past., 1903, 17, 822; Ann. de V’Inst. Past., 1904, 18, 98. (14) WaranasgE, S8.: Jour. Imm., 1919, 4, 77. (15) Ecker, E. E.: Jour. Inf. Dis., 1921, 29, 611. (16) Brooks, S. C.: Jour. Med. Resch., 1920, 41, 411. (17) NeuFretp, F.: Arb. a.d. Kais. Gesundheitsamt., 1908, 28, 125. (18) Wauxker, E. W. A.: Jour. Hyg., 1903, 3, 52. (19) Gurp, F. B.: Jour. Inf. Dis., 1912, 11, 225. (20) Jacopy, M., anv Scuutze, A.: Zeit. f. Imm., 1910, 4, 730. (21) Ze1ssuer, J.: Berlin Klin. Wochenschr., 1909, 46, 1968. (22) Rivz, H.: Zeit. f. Imm., 1912. Orig., 15, 145. (23) Noeucui, H., anpD BRONFENBRENNER, J.: Jour. Exp. Med., 1911, 18, 229. ALEXIN IN GUINEA-PIG BLOOD 459 (24) Gay, F. P., anp Ayer, J. B.: Jour. Med. Resch., 1907, 17, 341, 361. (25) Fassin, L.: Comp. rend. Soc. de Biol., 1907, 62, 647. (26) Marst, S.: Comp. rend. Soc. de Biol., 1908-1909, 64, 1058, 1113; 65, 612; 66, 432, 1073; 67, 54, et seq.; 68, 882, et seq. (27) Branp: Berlin Klin. Wochenschr., 1907, 34, 320. (28) Enruicu, P., anp Morcenrots, J.: Berlin Klin. Wochenschr., 1900, 37, 453, 681. en x a ay Oe ae ‘y hee ni ; wiv a, { Bel Ps TA foe! HON, hee (: i“ ; a) I is Fave a 7 Ye a “yy a ae > 1 vy } ria i u CA a ‘ Y A‘ i a > r'? i a Tee ; ' eS i . re) setae Wy ' 7 oe 1 vig, | JOGA Ny NM Wea vraath, tani ‘sree Ses sede Shel ak ahs aa i =) VERY, Pails eae ne UM ay Ie, Oh Ee ape DG ANA me ape we en ee arene Shek Apso: erie a sensei apis ini aia ant mee i, ; My aks “ny iy 7 ier oT mare i v ; P ty i 4 f * t, ft ip ‘ ) ; : * vi 7, 7 vs Ge ® 7 > | 4 { 4 a bd tin} 7 ~ 4 F c — sf ) ‘ = = ‘ : ~ i = re Paes ws 7 7 ay | os “ ue, TL ie. abe Ris ts hy an Bo ik a vey ay) OE a LA ree me ee 5 r cal » Po a Wi 7 aif ry a ara | 7 i A a) 7 iid " ar s i 42) wwe x i uy io j . int) ¥ 7 bo’ UJ inn § F . A vine ee ‘ : ty 7 4£ f a vw - é y : a ~ - at Stine ’ > oA + at hy chal oe ee ae +) baer un i A STUDY OF THE HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION H. W. CROMWELL From the Department of Immunology of the School of Hygiene and Public Health, Johns Hopkins University Received for publication May 19, 1922 I. INTRODUCTION According to Ehrlich (1), a formed antigen, such as a red cell or bacterium, has a definite number of combining groups or bonds which have an affinity for the specific antibody. When the antigen is placed in contact with its antibody, a reaction ensues, which progresses until all these bonds are satisfied or until all the anti- bodies have entered into the combination. It follows from this that when all the combining affinities of the antigen are satisfied, no more antibodies can enter into the reaction regardless of the number that still remain free in the supernatant fluid. This theory implies chemical affinities and a combination according to the respective valencies of the reacting substances. The view of Arrhenius was somewhat opposed to this (2). He regarded the reaction as a physical one and explained the phe- nomenon as a distribution of a solute between two solvents, the antibodies being the solute, and cell protoplasm and the surround- ing fluid the two solvents. He states that ‘“‘the immune bodies are probably not bound by the erythrocytes, but only absorbed by them,” and that ‘‘no proof has been given of their chemical action.” Bordet’s idea (3), although not the same as that of Arrhenius, was similar in that he attempted to explain the phenomeonon according to physical laws. He, however, held that the antibody or ‘‘sensitizing substance” is adsorbed in much the same way that a filter paper takes up a dye. He states that ‘“‘the union of the antibody with the antigen depends on what is called molec- ular adhesion or contact affinity, in other words, should be classed in the category of adsorption phenomena. ”’ 461 THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 6 462 H. W. CROMWELL It has long been known that a red cell or bacterium will com- bine with much more than the amount of antibody necessary for lysis. Arrhenius (2) found that this varies with the concentra- tion, and that a red cell will take up several thousand units, if the concentration of the antibodies in the supernatant liquid is sufficiently high. Arrhenius expressed the reaction according to the equation, 5 in which B represents the amount of antibody absorbed by the cells, C the concentration remaining in the supernatant liquid after absorption, and K and n are constants. He found the value of n to be 2, and from the equation and the calculated value of n, he interpreted the reaction as a distribution of the antibodies between the red cell protoplasm and the surrounding liquid as solvents, and that two of the antibody molecules free in the serum form three of the combined molecules. Manwarning’s work (4, 5, 6, 7, 8) on this same phenomenon failed to confirm the conclusions of Arrhenius. He found that the absorption did not follow any simple physico-chemical law, and that K and n were not constants. Manwaring obtained what he calls a ‘“‘negative absorption,” i.e., the titer of the serum dilutions, when great concentrations of antibody units were used, was often greater after contact with the corpuscles than before. He concluded, therefore, that qualitative changes take place in the amboceptor due to its contact with the cells, and “that any direct quantitative comparison between it and the untreated serum gives erroneous results.” From these and other experiments, he came to the conclusion that there is a ‘‘third component” in the serum besides the antibody and complement, which varies in quantity in different animals. This third component may be ‘‘antilytic,” or “aux- ilytic,”’ but never has independent hemolytic powers, although it may be absorbed by the red cells. In addition, he mentioned several other factors influencing the antigen-antibody combi- nation, among which are the reaction and specific gravity of the medium and the amount of inorganic salts. Amato (23), in a recent study of opsonic sensitization of bac- teria, has come to the conclusion that the union of the opsonins HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 463 with the bacteria is governed by the same law that governs the union of hemolysins with the antigen. He found the same equation applicable to the opsonins which Arrhenius applied to the hemolysins and concluded that the reaction is probably a distribution of the opsonins between two solvents in which they have different molecular weights. Coulter (9, 10) has done some very definite work on the influ- ence of the reaction of the medium on the absorption of hemolytic sensitizer by red cells, and on the dissociation of the combination. He found pH = 5.3 to be the optimum H-ion concentration for the absorption. However, he also found very little difference in absorption between the values, pH = 4.5 and pH = 6.0, in the salt-free medium, and a much wider range in medium containing salt. Kahn (11), in a very recent work on the absorption of hemo- lytic sensitizer, has made some studies on the rate of the reaction at the various temperatures. He finds that the reaction is com- pleted very quickly, in every case at the end of fifteen min- utes. He also finds ‘‘that the extraction is greater at 37° than at room temperature, which in turn is greater than at ice-box temperature.” It is seen, then, that the antigen-antibody combination is influenced by time, temperature, the reaction of the medium, and the amount of inorganic salts. In addition, there is also con- siderable variation due to uncontrolled factors, which vary with the different lots of serum used—those variations which caused the differences of opinion between Arrhenius and Manwaring. In the face of all that has been done on the absorption of hemo- lytic antibodies by red cells, it would seem almost hopeless to try to add anything new, either in fact or theory. However, it is thought that the results of the experiments recorded in this paper are significant in suggesting an explanation of the differ- ences obtained by Arrhenius and Manwaring, without the neces- sity of resorting to the sub-divided ‘‘third component,” or to any qualitative changes in the amboceptor due to its contact with the red cells. In all the discussions the terms, ‘‘amboceptor,” ‘‘antibody,” “sensitizer,” and ‘“‘immune body,” will be used interchangeably. 464 H. W. CROMWELL II. THE CONCENTRATION OF THE HEMOLYTIC ANTIBODIES AS A FACTOR INFLUENCING THEIR ABSORPTION BY RED CELLS These experiments were carried out in order to determine how many times the amount of sensitizer necessary for hemolysis the red cell will absorb, and whether or not there is a definite satura- tion point, as might be assumed from Ehrlich’s theory, above which the red cells will absorb no more, regardless of the amount still remaining free in the serum. The technic was as follows: Rabbits and guinea-pigs were immunized to the red cells of the sheep, by intraperitoneal injections of the guinea- pigs and intravenous injections of the rabbits. The serum so obtained was inactivated at 56°C. for thirty minutes and then carefully titrated against fresh sheep cells that had been collected in 1 per cent sodium citrate and washed four times in large volumes of normal salt solution. The unit of cells was arbitrarily chosen as 0.1 cc. of 1:4 suspension! (measured in terms of whole blood), and the unit of complement as 0.1 ec. of a dilution of fresh normal guinea-pig serum, pooled from several animals. The unit of antibody or sensitizer was defined as the smallest amount of the immune serum which, under the above conditions, would just suffice to hemolyze a unit of sheep corpuscles in a total volume of 1 ce., in one hour’s incubation at 37°C. Protocol 1 will illustrate. Protocol1. Type for hemolytic titrations TUBES qe Se Te aA re 0.5] 0.4] 0.3/0.25 1500) 5 iA: A BEE 0.5} 0.4; 0.3; 0 | 0 Guinea-pig complement 1:5...........| 0.1} 0.1) 0.1)0.1 | 0.1) 0.1] 0.1) 0.1) 0 Siteppece lish c 4 ore 2 Meee ere ae ee 0.1] 0.1) 0.110.1..| 0.1} 0.1) 0.1) O-t)oa8 Salissolition 5.0. '%. Vatatee cst c tee 0.3} 0.4! 0.5/0.55! 0.3) 0.4] 0.5) 0.8} 0.9 Hemolysis one hourf..................} 44] 44] 44+ 4+ 4+) 3+/2+|0 | 0 * The numbers, 1: 200 and 1: 500, represent the dilutions of the immune serum used in the tubes. For instance, tubes 1 to 4 received 0.5, 0.4, 0.3, and 0.25 ee. of a 1: 200 dilution, respectively. {4+ means complete hemolysis. The degrees of hemolysis are indicated by the signs, 3+, 2+, and +. 1 All dilutions and suspensions were made in normal salt solution. A 1:4 suspension of cells means one part of cells plus three parts of salt solution. HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 465 In the above sample titration, it is seen that the hemolytic unit of this particular serum is 0.5 cc. of a 1:500 dilution, or 0.001 ce. Then 1 ee. of this serum contains 1000 hemolytic units. The titer is expressed as 1/1000, or sometimes as 1000. In making. the absorption tests, a number of tubes were arranged, each containing two units of sheep cells and sensitizerin varying amounts. Enough salt solution was added to each tube before the sensitizer to make the final volume 2 cc. The tubes were incubated in a water bath for thirty minutes at 37°C., after which the red cells were removed by centrifugalization and the supernatant fluid titrated as in protocol 1, to determine the number of units of sensitizer lost, i.e., the number removed by the red cells. The use of all the materials concerned in the reaction in only twice the amount ordinarily used for titration overcame the necessity for the use of an undue amount of immune serum and still afforded sufficient supernatant fluid for the subsequent titration. The method is shown in protocol 2. Protocol 2. Type for absorption of hemolytic units by sheep red cells* SENSITIZER TUBE a ee oT i Sa a a ios SHEEP SALT TITER AFTER UNITS NUMBER Units per CELLs 1/4 SOLUTION | ABSORPTION| ABSORBED cubie centi-| Quantity of serum | meter ee. 1 5 1.0 (1:400) 0.2 0.8 0 5 2 10 0.2 (1:40) 0.2 0.6 0 10 3 20 0.4 (1:40) 0.2 1.4 0 20 4 50 1.0 (1:40) O:2 0.8 1/2 48 5 70 1.4 (1:40) 0.2 0.4 1/8 62 6 100 Or? (=4) 0.2 1.6 1/12 88 7 200 0.4 (1:4) 0.2 1.4 1/40 160 8 500 1.0 (1:4) 0.2 0.8 1/150 350 9 1,000 | 0.5 (undiluted) 0.2 1.3 1/400 600 10 2,000 | 1.0 (undiluted) 0.2 0.8 1/1000 1000 11+ 50 1.0 (1:40) 0 1.0 1/50 | 0 12 500 1.0 (1:4) 0 1.0 1/500 | 0 * All the sets were made in duplicate. ¢ Sets 11 and 12 are controls. Tables 1 and 2 show the absorption from immune rabbit and guinea-pig serum in concentrations ranging from 5 to 2000 units per cubic centimeter. On the first line of each table are shown the concentrations with which the cells were treated, and on the 466 H. W. CROMWELL following lines are shown the amounts of antibody taken up by the cells from each concentration in the tests made with the different lots of serum. Some of the results are also shown graphically in figures 1, 2, and 3, in which are plotted the logarithms of the TABLE 1 Absorption of antibodies from rabbit immune serum by sheep erythrocytes CONCENTRATION OF ANTIBODY UNITS PER CC. IN THE SERUM DILUTIONS REMARKS 50| 100 | 200 | 500 |1000; 2000 Test 1 |20)/49) 90/180 Test 2 |20\48) 92)175|340 Titer 1/10,000 Test 3 |20/47| 86/160|300|360 Serum of test 2, stored four months, titer 1/6000 Test 4 47| 90)167|375|500| 670 Test 5 47| 92/150|375|500 Units absorbed } Test 6 |19|47| 88|150|300|500| 750 Test 7 |18|380) 44] 77 Titer 1/1000 Test 8 |20/49| 97/180)300 Test 9 49} 95)193/450/700/1200| Titer 1/40,000 Test10| |49} 98/185|350/540| 875| Serum from same rabbit as in test 9, eight days later. Titer, 1/25,000 Test11}| |47| 87)150 TABLE 2 Absorption of hemolytic antibodies from immune guinea-pig serum* by sheep erythrocytes CONCENTRATION OF ANTIBODY UNITS PER CC. IN THE SERUM DILUTIONS TITER 19 | 411} 70] 100 1/4000 Testy se: 5 10 Test Daceeas t i 12} 20; 40| 70 1/2500 Testisia sen 4 6) 127 25] 40) 70} 235 |>t7e Units absorbed ; Test4 .....| 4 6.|- 10-| 20+]. 33-1> 40:| 100131000 Testiomenee 4 5 8 50 1/1000 LEStiG., yaes< 3 4 | 10'} 20 1/125 Mest. tere 17} 20| 33 1/2000 * From 1000 units sensitizer from guinea-pig serum the absorption was so often difficult to determine that it is not listed here. Often no difference in titer could be detected after contact with the cells. ¢ Serum of test 1, stored four months. HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 467 amounts absorbed against the logarithms of the concentrations remaining in the liquid after the absorption is completed. The plot, as is seen, tends to approach a straight line. A study of tables 1 and 2 brings out several important facts. Even from comparatively low concentrations of antibody units all are not absorbed, while, if the concentration in the liquid is sufficiently high, massive quantities are taken up. The only aT SSE REE Cer Baie os oe Mea NS a Pees Co “aha rena Ce ea ase eae eS oes Fie. 1. Locaritumic PLor oF THE DATA OF TABLE 3 The amount of sensitizer absorbed is plotted against the concentration remain- ing in the supernatant liquid. Titer of serum, 1/4,000. K = 22.18,n = 0.5 instances in which the reaction seems to be complete are when very low concentrations are used, and even here, the assumption that all the sensitizer is absorbed does not seem justified, because amounts much less than one unit per cubic centimeter could not be detected by the method of titration used in these experiments. It is evident, therefore, that the number of amboceptor units taken up by the sheep red cells varies with the concentration of 468 H. W. CROMWELL those units in the surrounding medium; the more concentrated the units are, the more are absorbed. Another fact to be noted is that the absorption is nearly always higher from the rabbit serum than from the guinea-pig serum, Zs cH SB enicleciesais: Ee Sie eae ea ae i SS 6 Ga cai a a LO i i fi; ie A PEER aha ca rr eee 1. aay a ll lf gee ee ee a ee | FP De te Mier PR a ee aE 1.0 2.0 3.0 Fig. 2. Curves A anp B, Locaritumic PLots oF THE Data oF TESTS 3 AND 4 of TABLE 2 The data for these plots are also shown in table 6. For curve A, K = 3.225, n = 0.78.4 For curve B, K = 1.97, n = 0.61. and that the absorption varies with the serum of individuals of the same species. This will be referred to later. The results here recorded agree, in part, with those obtained by Arrhenius and Morganroth (2) in their work on the absorption HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 469 of immune bodies by ox and sheep cells, in which they were able to express the reaction according to the equation, 1 SIO This equation holds true, however, only when the logarithmic plot, as shown in the figures, represents a straight line; and the farther it deviates from the straight line, the more inapplicable ie Se Se tei yy ft Smee | So J aS SS eRe e e 2.0 3.9 4. Fic. 3. Locaritamic Piots, A AND B, REPRESENTING THE Data oF TABLES 4 AND 5 The sera were obtained from rabbit 69 on the tenth and twenty-fourth days, respectively. For curve A, K = 165.9, n = 0.36. For curve B, K = 80.73, n= 0.4. becomes the equation, because of the variation in the values of the constants, K and n, between the different concentrations. The tests represented by figs. 1 and 3A show practically straight line plots, so here the equation can be applied. The experimental and calculated results compare very favorably, as is shown in tables 3 and 4, which correspond to figures 1 and 3A, respectively. 470 H. W. CROMWELL The disagreement is, in most cases, within the limits of experi- mental error. The values of the constants, K and n, should be TABLE 3 Experimental and calculated results from test 6, of table 1 B* C oBSERVED C CALCULATED K n 46 4 4.4 22.18 0.5 88 12 5, 150 50 45.7 300 200 170.3 500 500 508 .2 750 1250 1144.0 | 1000 2000 2034.0 *B =the number of units absorbed by the cells, C = the concentration remaining after absorption. TABLE 4 Experimental and calculated results from absorption tests made on the serum of rabbit 69, obtained on the tenth day B C OBSERVED C CALCULATED K n 198 2, Pe 165.9 0.36 487 13 24 875 125 124 1340 660 408 2350 1650 1950 4000 6000 8000 | TABLE 5 Results of absorption tests on the serum of rabbit 69, obtained on the twenty-fourth day B C OBSERVED C CALCULATED K n 157 3 5 80.73 0.4 380 20 15 630 170 170 800 800 309 1000 3000 540 noted. The ‘‘absorption constant,” K, is 22.18 for table 3, and for table 4, K = 169.5, which is very much higher. The absorp- tion is also very much higher from all concentrations in table 4. HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 471 The value of n differs for the two tables, and in both it is very much lower than the value, 2, which was found by Arrhenius. In figure 1, n = 0.5, and in figure 3A, n = 0.36; i.e., the curve having the steeper gradient has also the higher value for n. Thus, it is seen that the constants vary with the different lots of serum used, the value of K, determining the position of the curve, and of n, the gradient. TABLE 6 Experimental and calculated results from absorption of sensitizer from guinea-pig immune serum by sheep erythrocytes B C OBSERVED C CALCULATED K n Curve A, figure 2 6 4 2.54 3.225 0.78 12 8 7.19 25 25 21.52 40 60 43 .68 70 130 101.20 235 165 622.10 6 4 5.3 1.969 0.61 10 10 11.5 20 30 32.8 33 67 68.9 40 160 91.6 100 400 362.3 In figure 2, there are shown two plots, A and B, representing the two tests shown in table 6. Curve A deviates considerably from the straight line so the values, K = 3.225, and n = 0.78, are necessarily calculated only approximately from the averages of several determinations made at intervals along the curve. There is also in the table quite a discrepancy between the observed and calculated results. The table for curve B, which approxi- mates the straight line very closely, shows a very close agreement between the observed and the calculated results. 472 H. W. CROMWELL Ill. THE RATE OF THE REACTION BETWEEN THE ANTIGEN AND ANTIBODY, IN VITRO It would seem that the union of the antigen and antibody is extremely rapid, and practically instantaneous in vivo. This is demonstrated by the anaphalactic shock in hypersensitive animals, for when the antigen is injected into the blood-stream, the shock often occurs immediately. Bull (12) has found pneumo- cocci, when injected into the blood-stream of an immune animal, to be agglutinated within a very few seconds. In the previous absorption experiments, the cells were allowed to remain in contact with the sensitizing serum for thirty minutes, which was considered sufficient time for the reaction to reach TABLE 7 Rate of the antigen-antibody reaction CONCENTRATION OF ANTIBODY UNITS PER CC. IN THE SERUM DILUTIONS TIME 60 65 80 250 1 84 82 | 100 | 120 350 | 400 5 Writs AvsOrpeds.s...1 2236 sees 87 88 | 100 | 125 375 | 400 15 87 88 | 100 | 120 37a | 400 30 | 90 90 100 | 125 | 375 | 400 | 60 87 88 | 100 | 125 375 | 400] 120 equilibrium. In order to contro] this, however, it was necessary to establish definitely whether or not the time allowed was enough, and how much it could be varied without influencing the results. In this experiment, sheep cells and sensitizer were allowed to remain in contact for varying periods of time and the amount taken up tested by titration, as in the previousexperiments. The longest time was two hours. The shortest time, which is listed as one minute in the table, could not be kept absolutely constant. The corpuscles were added to the serum dilutions, shaken up thoroughly, and immediately centrifuged at high speed. The time during which the cells were in free contact with the serum was certainly not more than two minutes in any case. The results are given in table 7. HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 473 It is evident, from this table, that equilibrium is reached very quickly. It is practically complete within five minutes, and en- tirely so in every case at the end of fifteen minutes. A slight dissociation, at the end of two hours, is indicated in the absorp- tion from 100 units, but this is not very pronounced and does not show at all in the higher concentrations. The data given here are few, but the experiment serves its purpose as a control to the other experiments, and shows that the time allowed for the absorption may be varied within a wide range without materially influencing the results. IV. THE INFLUENCE OF TEMPERATURE ON ABSORPTION OF ANTI- BODIES BY RED CELLS A temperature of 37°C. is employed inmostserological reactions. The combination of the antigen and antibody, however, takes place at much higher and much lower temperatures. Differ- ences of opinion occur among the different investigators as tothe best temperature for the combination. Neil (13) and Kolmer (15) advise the use of room temperature as the best for the sensi- tization of cells, while Hinton (14) recommends 37°C. ‘These opinions were advanced as the result of more or less extensive work on the standardization of the Wassermann reaction. Kahn (11) states that ‘“‘the extraction is greater at 37° than at room temperature, which in turn is greater that at ice-box temperature.” In the experiments here recorded, the influence of the variation of the temperature upon the absorption of hemolytic antibodies has been observed. The absorption tests were carried out as in the foregoing experiments, using the rabbit anti-sheep hemolytic system. Duplicate sets were made for each temperature and concentration tested. The antibody dilutions were made first, and the cells added after both had been brought to the required temperature. It is needless to say that the tests at all different temperatures with one lot of serum were made on the same day with the same lot of cells and complement. The results are given in table 8. 474 H. W. CROMWELL TABLE 8 The influence of temperature upon the antigen-antibody reaction (A) CONCENTRATION OF ANTIBODY UNITS PER CC. IN THE SERUM DILUTIONS TEMPERA- Series A|Series B Series C Series D TURE Serum | Serum Serum Serum 100 600 65 | 130 | 320 400 Winey Gil FLUTE & eel 2 eke ee °C Units absorbed.............. 97 475 | 62 130 | 220} 220 37 500 40 475 45 440 50 400 55 (B) Serum of rabbit 487 CONCENTRATION OF ANTIBODY UNITS PERCC.IN THE SERUM DILUTIONS TEMPERA- TURE 100 100 200 200 500 800 “é 92 93 | 180 178 | 330] 580 0 95 95 | 187 183 | 340) 600 15 95 95 | 187 184 | 340; 600 25 90 93 | 180 178 | 330) 575 37 Units absorbed:..=......... 2s. 88 175 167 575 40 83 163 160 | 320} 550 45 75 155 150 | 310} 450 50 120 300 | 400 55 75 300 60 (C) Absorption from 100 units of sensitizer from the sera, E to I TEMPERATURE ABSORPTION FROM 0° Room is 45° Serslenits cut ticle fe bectepres 88 94 95 90 Sera ee peer ie ec oss eee 93 96 94 87 Sera (Gee eee ions e Sees 100 100 100 99 Sera Het ee eek). 40 50 55 45 HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 475 TABLE 8—Continued (D) Absorption from 200 units of sensitizer from the sera, E to I TEMPERATURE ABSORPTION FROM 0 Room 37 45 erty Bite ecie tai oh acu iehstie 145 170 185 175 (SID) NR aot aaa es es 167 183 170 155 SeraiGee ane ate os see eas 199 200 200 199 IEEE ee ees ony feito cas «it 120 160 165 135 Serapleeeer oe ceriseracisr Sys ss 80 120 100 100 In the tests with the sera A, B, C, and D the absorption was greater at 37° than at the lower temperatures. That is, the absorption at 37°>25° = or>15°>0°. In the tests with the serum of rabbit B, the absorption was greater at 40° than at any other temperature. The serum from rabbit 487, however, when tested gave a differ- ent result. Seven tests were made on this one lot of serum at different times, using the various concentrations listed in the table. In every test, the temperature of 15° to 25° proved to be the optimum, lower absorptions being obtained at 0°, and 37°, and above. ‘The relative amounts of absorption for this serum may be expressed thus: 15° = 25°> 0° = or> 37°. It is true that the differences in absorption between 15 and 37° are not great, but they consistently appear in all the concentrations tested. The results obtained with this serum suggested that the optimum temperature for the absorption might be a variable, which is different for the different sera. Therefore, other lots of serum, which had been kept stored with phenol, were tested to see if any would fall into the same class with that of rabbit 487. These are listed in the table as sera E to I, inclusive. The last two, H and I, were from guinea-pigs and the rest were from rabbits. Table 8, (C) shows the absorption from 100 units of sensitizer from each of these sera at the different temperatures. In table 8, (D), is shown the absorption from 200 units. Of these four sera, F and I proved to be in the same class as that of rabbit 487; i.e., about room temperature was the optimum for absorption. 476 H. W. CROMWELL Those investigators, therefore, who claim that either room tem- perature or 37° is the best for the sensitization of the cells, were doubtless right in their observations, but any generalization that all sera react in the same way is obviously incorrect. The best temperature for the sensitization of the cells seems, rather, to be a variable, its absolute value depending on the particular lot of serum used in the test. No explanation of this variability is here advanced. V. RELATION OF THE GLOBULIN CONTENT OF THE IMMUNE SERUM TO THE ABSORPTION OF THE ANTIBODIES It was noted in section I that the absorption of hemolytic antibodies by sheep cells varies with the sera of different animals of the same species and of different species. The question at once arises as to the cause of this variation. Manwaring (16) has shown that serum proteins may be ab- sorbed by the red cells of another species. It has long been known that the immune bodies are, in most cases, thrown out of the serum along with the globulin fractions, and Hurwitz and Meyer (17 and 18) have shown that these globulins are, as a rule, although not necessarily, increased in varying amounts during the process of immunization. The increase is, in their opinion, due to, and roughly proportional to the amount of metabolic disturbances set up in the animal. The evident relation of the globulins to the antibodies, in these respects, suggested a possible connection with the absorption of the antibodies by the cell- antigen. The experiments recorded in this section were under- taken with this idea in mind. The sera of rabbits and guinea-pigs were used in this work, and quantitative determinations of the serum proteins were made on each lot of serum before the adsorption tests were done. The protein determinations were made according to the microre- fractometric method of Robertson (19), with an Abbe refractom- eter. Serum, inactivated at 56°C., was always used immedi- ately to avoid the possibility of bacterial contamination, and special attention was given to the cleaning of the glass-ware. HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 477 The results of this work are given in table 9.2. There is no evidence, from these, that the amounts of the serum proteins present have any relation to the amount of sensitizer absorbed by the cells. Rabbits 481 and 486 were bled and tested at in- tervals for 116 days after the beginning of the immunization. TABLE 9 Absorption from immune sera with quantitative determinations of the serum proteins UNITS ABSORBED IN CONCENTRATION OF ANTIBODY UNITS PER 18 Rabbit 486 18 40 71 | 134 370 600 | 3.42 | 4.25 Animal emaciated, died soon after Cc. IN THE SERUM DILUTIONS anhnne Moet LIE Test Animal 20 50 100 200 500 per cent | per cent 1 Rabbit 276 50 99 193 8, 000 IAL) 52783 2 Rabbit 374 45 80 150 2,000 1.82 | 4.26 3 Rabbit 471 47 87 150 2,500 | BA 4 Rabbit 487 49 95 193 40,000 | 1.48 | 5.62 5 Rabbit 481 Dee |) i075 10,000 | 1.95 | 4.48 6 Rabbit 481 30 50 91 160 1,500 | 3.24 | 2.63 7 Rabbit 481 18 30 55 100 800 | 2.7 | 4.18 8 Rabbit 481 15 23 37 600 | 2.74 | 4.76 9 Rabbit 481 18 30 55 102 500 | 2.34 | 4.6 10 Rabbit 481 17 35 65 120 500) e228 4.12 11 Rabbit 481 20 42 76 140 500 | 3.92 | 4.15 Animal emaciated, died soon after 12 Rabbit 486 94 |180 | 410| 8,000 | 2.41 | 4.01 13 Rabbit 486 30 52 91 | 160 1500 porAd las 14 Rabbit 486 40 80 150 | TOOO e2a2 PAL 5 15 Rabbit 486 LQ. | 3S: lea 800 | 1.96 | 5.38 16 Rabbit 486 18 30 55 102 500 | 2.19 | 4.8 7 Rabbit 486 17 | 35 65 120 600 | 2.77 | 3.85 19 Guinea-pig 15 25 40 45 2,000 | 2.51 | 2.48 20 Guinea-pig 15 20 30 40 1,500 | 2.65 | 2.08 21 Guinea-pig ws 20 oo 2,000 | 3.03 | 2.08 The globulins and albumins fluctuated considerably during that time, but the fluctuations seem not to bear any evident relation to 2 The figures given in tables 9 to 12 do not all represent the actual experimental results. In many cases the concentrations tested did not coincide with those given in the tables, so the figures had to be interpolated from logarithmic charts, for the purpose of tabulation. 478 H. W. CROMWELL the absorption. At the time of the last bleeding both animals were emaciated and the globulins of both sera were very high. They were about the same in amount as at the second bleeding, but the absorption is much more complete. VI. THE RELATION BETWEEN THE ABSORPTION OF THE ANTI- BODIES AND THE TITER OF THE SERUM It has been noted that the constants, K and n, vary in the ab- sorption from the sera of different animals. It has also been observed, incidentally, that often the serum of the higher titer showed the higher absorption. In tests 2 and 3, of table 1, the serum was tested fresh and then tested again after four months’ storage, during which time the titer dropped from 10,000 units to 6000 units per cubic centimeter. The later test gave a some- what lower absorption than when the serum was first drawn. Another serum was tested with a titer of 40,000 units per cubic centimeter, and gave a very high absorption (test 9, table 1). The rabbit was kept and its serum drawn again eight days later (test 10, table1). The titer had dropped to 25,000 units per cubic centimeter, and from all concentrations a marked lowering of the absorption was observed. It was thought, therefore, that there might be a relation between the titer of the serum and the number of antibody units the red cells would take up. A number of fresh sera of both high and low titers were tested. Some of the animals having serum of high titer were tested sub- sequently, after the titer had dropped. The results are given in table 10. In many cases, the absorption seems to vary with the titer of the serum, but this is not constant, and the differences in the titers above 2000 units per cubic centimeter seem to have no definite relation to the absorption. Normal rabbit serum, in spite of its low titer, gives complete absorption from the undiluted serum (tests 6 and 8). Most of the other low-titered sera were obtained from rabbits that had formerly had high titers, and the absorption from these was uniformly low. Tests 17 to 21, how- ever, were made with the sera of rabbits 72 and 80, drawn during the first few days of immunization. They all show uniformly HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 479 high absorption, although the titers vary from 100 to 16,000 units per cubic centimeter. These results seem to indicate, therefore, that, although in many cases the higher the titer the higher the absorption, it is Test aor WN RE 20 21 not the titer of the serum, per se, which causes the variation. Relation of the absorption to the titer of the serum UNITS ABSORBED IN Animal Guinea-pig Guinea-pig Guinea-pig Guinea-pig Guinea-pig Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit Rabbit 72 Rabbit 72 Rabbit 72 Rabbit 80 Rabbit 80 TABLE 10 CONCENTRATION OF ANTIBODY UNITS PER CC. IN THE SERUM DILUTIONS | | | | 100 77 150 100 125 160 167 160 150 150 4 10 20 8 50 10 20 33 12 24 40 19 41 70 20 50 Normal, untreated 18 30 | 44 20 50 100 Normal, untreated 45 80 18 35 55 19 37 70 20 47 92 47 90 20 47 86 47 92 19 46 88 50 | 100 340 375 300 375 300 Animal three days immune 99 | 194| 450 | Animal six days immune | 198 | 485 | Animal ten days immune | 91 | 175 | 350 | Animal five days immune | | 96 | 188| 426 | Aminal nine days immume TITER 125 1, 000 1,000 1, 000 4, 000 50 1,000 100 2,000 800 800 10, 000 6, 000 6, 000 4,000 4, 000 100 1, 000 16, 000 800 8 000 Rather, it would indicate that the extent of absorption is de- pendent on the length of time the animal is immune, the higher absorption being obtained early in the period of immunity. 480 H. W. CROMWELL This variation may be caused by some inhibiting substance, which is present in the serum in larger quantities later in the immunity period. VII. SELECTIVE ABSORPTION In the preceding section, it was indicated that the variation in absorption may be due to some disturbing element in the serum, the lowering of the absorption depending on the amount of this substance present. It was not possible to show any influence of the fluctuations of the serum proteins, so evidently the disturb- ance is not due to the serum proteins as such. The problem now is to determine the nature of this element, if possible, and its mode of action, i.e., whether it is taken up by the cell, or acts merely by virtue of its presence in the serum. It has been noted before that, when the logarithmic plots of the absorption are made, some of the curves represent straight lines while others do not. Curve A, figure 3, for which k = 165.9, approaches the straight line very closely. Curve B, in the same figure, represents a test made on the serum of the same rabbit, drawn six days later. The titer had dropped from 20,000 units to 16,000 units per cubic centimeter with a drop in the absorption constants, k = 80.73. The plot, as is seen, deviates considerably from the straight line, the deviation being most pronounced in the higher concentrations. Figures 4 and 5 repre- sent comparatively high values for K, but in neither is it so high as in curve A, figure 3. These plots, also, deviate from the straight line in the higher concentrations of antibody units. Figures 6 to 9 all represent very low values for K, and in these is seen a drop in the curve in the intermediate concentrations. This deviation of the curve from the straight line, according to a physical interpretation, means that there is a selective absorp- tion of the antibodies, the amount of selection differing with the concentration. Selective absorption may, although it does not necessarily, denote the presence of two or more substances in the serum capable of being absorbed. It was tentatively assumed, therefore, that the variations in the absorption are caused by the presence of some inhibiting sub- Si a a ae J SESS Rae ess ieee ee aN Fig. 4. Locaritamic Piotr or THE Data or Test 5, TaBLe 9 K = 65.32, n = 0.37 AE ma 24 Saeeeae - Ese ae ne lege 5. SIMILAR ay or Trst 12, TaBLE 9 K = 35.0; 7 = 0.51 481 482 H. W. CROMWELL stance in the serum; and that this substance exerts its influence by virtue of its own absorption by the cell. The following “multiple absorption” experiments were carried out to support or disprove this hypothesis. Sheep cells, one unit, were allowed to take up a certain number of units of sensitizer from a serum that, from previous tests, had been found to have a comparatively high absorption constant. (The value of K was not determined exactly in these experiments, Pi i le i aK vse in icon eG a as id i ie as es lec) Ge hs er it cl Ee | Fins eS ag i me oh DC i We ea i i ieee ee Sel aesaeRERerors” (OES ie a IE Ze ae Sea Ee ad Pa bal Cs EO cS whic an, ea a ee SNe)” ed hae BGSRRERMREB Oe A SERS 0887 SS 875888 C4 deel a7 [ae a Fia. 6. Louantiranre Piotr oF THE Data oF Tsst 14, TABLE 9 K = 24.0, n = 0.58 RS als esa i diet eli Fic. 7. Smm1uar Piotr or Test 16, TABLE 9 K = 6.53, n = 0.56 Fic. 8. Srminsr Prot or Tsst 8, TaBLe 9 K = 8.0; n = 0:31 Fig. 9. Smmi~arR Pitot or Test 7, TABLE 9 K = 15.0, n = 0.41 HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 483 but only estimated to be high or low in proportion to the amount of absorption from the concentrations tested.) Another similar lot of cells was allowed to take up sensitizer from a serum having a lower absorption constant. Both lots of cells were then separated from the supernatant liquid by centrifuging at low speed, and each treated with an equal number of units of sensiti- zer from one serum. ‘The amount of absorption was determined as in the preceding experiments. If this inhibiting element is really taken up by the cells, those which had received a certain number of units of sensitizer from a serum with a low absorption constant should be more resistant to further sensitization than cells which had received an equal number of units from a serum with a higher absorption constant. The results of the experiments described below agree with the expectations. Experiment 1 a. First absorption. Set 1:3 One unit-of sheep cells was treated with 50 units of fresh guinea-pig immune serum, titer 1/150. Titration of the supernatant liquid showed only 5 units absorbed. The value of K, therefore, is very low. Set 2: One unit of sheep cells, treated with 50 units of guinea-pig immune serum, titer 1/5000, absorbed 20 units of sensitizer. The value for K, is therefore, much higher than for the serum of set 1, although still rather low. b. Second absorption. Each lot of cells was then treated with 50 units of the second serum, titer 1/5000. By titration of the superna- tant liquid, lot 1 was found to have taken up 5 more units, while lot 2 absorbed 10 more. The serum with the lower absorption constant was therefore, more effective in preventing further sensitization of the cells than the serum with the higher absorption constant. Experiment 2 a. First absorption. Set 1: One unit of cells, treated with 100 units sensitizer from rabbit serum, titer 1/1200, absorbed 50 units. The value for K is, therefore, comparatively low. Set 2: One unit of cells, treated with 100 units sensitizer from rabbit serum, titer 1/5000, took up 90 units. The value for K then is comparatively high. 3 All sets were made in duplicate. 484 H. W. CROMWELL b. Second absorption. Each lot of cells, after separation from the supernatant liquid, was treated with 100 units sensitizer from the second serum, titer 1/5000. Lot 1 took up so small an amount in this second absorption that it was not titratable, while lot 2 took up 25 more units. Here, 50 units taken up from the serum with the lower absorption con- stant were more effective in preventing further senitization of the cells than 90 units from the second serum. Experiment 3 a. First absorption. Set 1: One unit cells, treated with 100 units sensitizer from rabbit serum, titer 1/600, absorbed completely 100 units. The value of K is high then, in spite of the low titer. Set 2: One unit of cells, treated with 200 units sensitizer from rabbit serum, titer 1/4000, absorbed 100 units. The value for K, therefore, is much lower than for the serum of set 1. b. Second absorption. Each lot of cells was-then treated with 200 units of sensitizer from the second serum, titer 1/4000. Cells of lot 1 absorbed 75 more units, and those of lot 2 absorbed 30 more. The serum with the lower absorption constant is here, also, the more effective in making the cells resistant to further sensitization. Experiment 4 a. First absorption. Set 1: One unit of cells, treated with 200 units of sensitizer from rabbit serum, titer 1/4000, absorbed 80 units. The absorption constant is very low. Set 2: One unit of cells, treated with 120 units sensitizer from rabbit serum, titer 1/5000, took up 120. The absorption constant, therefore, must be very high. b. Second absorption. Each lot of cells, after separation from the supernatant liquid, were treated with 120 units from the second rabbit serum, titer 1/500. Lot 1 absorbed 40 units more, while lot 2 took up 110 more. In this case, also, the absorption from the serum having the lower value for K* is much more effective in making the cells resistant to further sensitization than absorption from the serum with the high value for K. In every case, in the foregoing experiments, the antibodies taken up by the cells from the serum with the lower absorption 4 Where the plot deviates from the straight line the values of K and n can be determined only approximately. ee , eEeeeEeEEOOeeEeeeEeEeEeEEOeeeeEEEEeEeEeeEeEEeEEEeEEEEeEeEeEeEeEeEEEeEeEeEEeEeEeEEeEe.- ._. i HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 485 constant had much more influence in inhibiting further sensitiza- tion of the cells than a larger number of antibodies taken up from the serum with the higher absorption constant. It is obvious, therefore, that the inhibiting element, whatever its nature, can be taken up by the cells, and probably in proportion to its com- bining affinity, inhibits the absorption of the antibodies. VIII. VARIATION OF THE ABSORPTION WITH THE LENGTH OF THE PERIOD OF IMMUNITY In the foregoing section, it was found that the disturbing ele- ment, which inhibits the absorption of the antibodies, is actually taken up by the cells along with the antibodies. In section VI, it was indicated that the amount of this substance probably varied during the time that the animal is immune. In this section, it is purposed to show how the absorption varies during the immu- nity of the animal, which variation seems to give some indication as to the nature of the disturbing factor. A careful record of rabbits 481 and 486 was kept while test- ing for the relation of the serum proteins to absorption. The animals were given three injections within the first nine days and no more during the 116 days they were under observation. ‘Table 11 shows the fluctuations in the absorption from their sera, also the titers and the number of days after the first injection when the serum was drawn. In figures 10 and 11 are plotted the observations. The first tests, made on the sixteenth day, showed a high titer and a comparatively high absorption. The second test, on the twenty-fourth day, showed a great drop in the titer and a cor- responding drop in the absorption. Following the initial drop, the titer dropped very gradually until about the sixty-second day, and from then on held fairly constant until the last obser- vation on the one-hundred and sixteenth day. Further observa- tions were not possible on account of the death of the animals. In the case of rabbit 481, where the titer dropped gradually, or held almost constant, the absorption, although fluctuating some- what, also held fairly constant, and toward the end of the period 486 H. W. CROMWELL of observation, showed a distinct rise. The fluctuations of the absorption from the serum of rabbit 486 were more pronounced, showing a marked rise after the initial drop, with a not so marked rise at the end. It is highly probable, however, judging from other observations described below, that the rise at the end TABLE 11 Absorption tests on the sera of rabbits 481 and 486, drawn at intervals during a period of immunity of 116 days* | CONCENTRATION OF ANTIBODY UNITS PER CC. IN THE SERUM DILUTIONS ceed p Pere 50 100 200 500 1000 Units absorbed from serum of rabbit 481: Testule: o -aek 8 oan | 94 | 175 | 400]! 600 | 10,000 16 Mesto ee SON SOL Slt oe 1, 500 24 Tost S. git pias, HR SOA iis5 os tee 800 62 Weak tice hae Se <. 93) SZ) BE 600 80 Pest. ls sy. week le SSO) eae aes 500 91 TestiG: sok e ae I) SO oOo NE ELO 500 110 MGRG Tt Musee REE 42 | 76 | 140 | 330 500 116 Units absorbed from serum of rabbit 486: Pent i AP PRA 94 |180 | 410] 670]! 8,000 16 MestOvu alls. UIs 6lB0y t|P2524 {aL Nh Bieb 1,500 24 ee bi iccie s yee a ee ale AO |) SO unl at) 1,000 62 TSE aa alg has UI ke AN | | 00 80 Hosea eee | eo | a ee 500 91 Mesiti otek one RB BES PTE) TMI 600 110 Pentiyeiel cic bideueab ese h4O) olde LASaeaL Sew 600 116 * These rabbits received three injections of washed sheep corpuscles within the first nine days. Each injection consisted of 3 cc. of the cells, measured in terms of whole blood. would have continued higher, had it been possible to make further tests. The results from these two rabbits seemed significant, so a number of others were started. Unfortunately, all except one, rabbit 69, died so their history is not recorded here. A goat was also immunized and its history followed. Table 12 and figures 12 and 13 show the observations on these animals. SS ee ass ee HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 487 The results agree very well with those obtained with the other two rabbits. Even though the titer is not high at first, the ab- sorption is high, with a great drop within the second or third week of the period of immunity. Following this, the absorption tt 23 So A A en nes pet Kd aml a da iil eR at a ala wl (a ose i aM FS 2c) ie EEE EEEEEEE lof Pl eb ae et te Gepee eee enaee, sides eeesenee’ came ee ann Fig. 10. PLots oF THE ABSORPTION FROM THE SERUM OF RaBBiT 481 AGAINST THE TimME or Immunity. Ser TABLE 11 Curve A represents the absorption from 200 units, B, from 100 units, and curve C represents the square root of the titer. holds about constant for a time, after which it shows a more or less gradual rise. The history of these animals seems highly significant in suggest- ing an explanation as to the nature of the disturbing factor in 488 H. W. CROMWELL the absorption. In nearly every case, the sudden initial drop in the absorption is accompanied by a similar drop in titer. In the case of the goat, this is not evident, but seven days had elapsed alee eeuateey \ceuer PEN Fic. 11. Sruruar Prot ror Rassit 486, TABLE 11 Curve C represents the absorption from 50 units, and D represents the square root of the titer. between the first and second tests, during which the titer may have gone higher and dropped back to 1/2000. It is conceivable that, as the amboceptor content of the animal’s blood-stream drops, the destruction of the antibody occurs in progressive stages, the first stage being merely the loss of the _ HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 489 power to sensitize the cells to the action of complement. The ability to combine with the cell would still be retained for a time. This would be, in effect, somewhat analogous to the “agglutin- oids” of Ehrlich, which were worked out so thoroughly by Eisen- berg and Volk (22). In titrating the serum, then, these deteri- TABLE 12 Absorption tests on the sera of rabbit 69 and a goat, drawn at intervals during the period of immunity CONCENTRATION OF ANTIBODY UNITS PER cc. IN THE SERUM DILUTIONS Eee Bie lee 50 100 200 500 1000 Units absorbed from serum of rabbit 69:* Mes tiliee sw senlecckt ices 99 194 450 1, 000 6 TEENS PER ieee ee 198 485 875 | 20,000 10 PRESTR eee eas 2, 194 452 690 | 16,000 16 PRES AM ee seh 60 100 200 | 300] 2,000 27 ANS Rea Te ee Pee Se 75 120 250 400 1, 000 37 PR GRibt Oat erdarcrue aroma Metros 75 1745) 250 800 45 NEVE) 7, See Gees Ok Oi ai 75 | 130 275 | 400 600 62 LSS) tS eR Sra RN Pina. Sch 86 150 350 500 600 86 Units absorbed from serum of goat:t AND SHE eS Ree ete ee ene aero (ie 6°) 96 180 1,000 10 eStpo nee ee encsal! 620 30 70 2,000 19 WeStiore wore ee! 40 70 100 1,000 49 GSU eer er eee tet AO 70 1,000 63 estrone eee eee eee 40 70 100 1,000 65 I RTSER FLD ee erste esses neee ce WM) 75 120 1,000 92 Nesta cs see cece. AL 85 150 800 120 * Rabbit 69 received two injections with a six-day interval. { The goat received three large injections of about 8 ec. each and ten very small injections of about 1.5 cc. each. The last injection was one month before the final test. oration products of the amboceptor would not be detected, and one hemolyzing unit would contain, in addition to the antibodies, a variable amount of the deterioration products. Those which still retained their affinity for the antigen would, in the absorp- tion experiments, tend to cause a lowering of the absorption of 490 H. W. CROMWELL the antibodies. The fact that the greatest lowering of the ab- sorption is coincident with the most rapid destruction of the anti- bodies is very forcibly suggestive that the substance in the serum ney eT P| el ec CO EGET SE eee ce ee bo ges NA se he ead ler ee oe SheBGEMUS Bae SSEGRBRAME BEE ed lal fot a | Meet TL tals tenshong | | Fic. 12. Smanar Prot ror Rassit 69, TABLE 12 Curve A represents the absorption from 500 units, and B from 200 units. The titers of the sera on the various days is indicated at the top of the chart. which causes this disturbance may be defined as the product of the destruction of the antibodies. HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 49] IX. DISCUSSION It has been found that, in the absorption of hemolytic immune bodies by red cells, immense quantities, as much as 2000 units or more, may be taken up by one unit of cells, provided their con- veel Ra ll RE iH sa FSF CE a 9 GR Ws He PN PE (a SS a el Genie SESLISSSHeuEmR Fe | Za | Pa Fig. 18. Ptor Sruriar To Fiaure 11, ror THE Goat See table 12 centration in the surrounding medium is sufficiently high. On the other hand, all the antibodies are often not taken up even from concentrations as low as 50 units per cubic centimeter. This is certainly not in accordance with the theory of chemical va- lencies. It has been found, moreover, that this absorption follows, 492 H. W. CROMWELL more or less strictly, the physical law proposed by Arrhenius, which applies to the distribution of a solute between two immisci- ble solvents. The equation, B = KC", which represents this law, could, however, with equal accuracy, be applied to simple adsorption phenomena, such as the taking up of acetic acid from a benzol solution by silica gel. Arrhenius assumed that all absorption tests with the hemolytic system would follow this law within the limits of experimental error, but it was shown by Manwaring, and confirmed in this paper, that many, in fact, most of the absorption tests do not follow this law strictly, as is shown by the deviation of the logarithmic curves from the straight line. One of the most striking things about the absorption of hemo- lytic antibodies is the great differences obtained with the different sera. While with many sera the absorption is very low, incom- plete even from a concentration of antibodies as low as 20 units per cubic centimeter, with other sera the absorption is practically complete from as much as 200 units per cubic centimeter. The antigen-antibody combination takes place with great rapid- ity. A large proportion of the antibodies are absorbed almost instantly. Equilibrium is practically established within five minutes, and in every case within fifteen minutes whatever the concentration of the antibodies. The variability of the absorption from the different sera is shown also in the extraction at the various temperatures. The extraction is always less at 0° than at room temperature, and becomes progressively less again as the temperature is raised to 45° and above. The differences in the different sera become evi- dent in the absorption at room temperature and at 37°. Of the sera that were tested, the majority gave the highest extraction at 37° to 40°. Others showed the highest extraction at about room temperature, 15° to 25°. This indicates that the optimum tem- perature for the antigen-antibody combination is a variable, the absolute value of which depends on the particular lot of serum used in the test. An endeavor was made to determine the cause of the extreme variations obtained in the absorption from the different lots of ee es HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 493 serum. A large number of tests were made on the different sera, at the same time making quantitative determinations of the glob- ulins and albumins. The quantitative fluctuations of the serum proteins were not shown to bear any direct relation to the varia- tions in the absorption. They may, however, exert some more indirect influence which was not detected by this method. Although in many cases, the serum with the lower titer gave the lower absorption, this was not constant, and with titers of 2000 units per cubic centimeter or higher the differences did not seem to bear any relation to the amount of sensitizer taken up by the cells. Normal sera and sera that were drawn early in the period of immunity gave uniformly high absorption, regardless of the titer. This indicates that the lower absorption comes about as a result of some change in the serum after the animal is immunized. This lowered absorption is evidently due to some substance in the serum which is absorbed by the cells along with the sensitizer; it is probable, also, that, in proportion to the combining affinity of this substance, the absorption of the antibodies is inhibited. This is shown by the fact that cells, which take up sensitizer from a serum that gives a low absorption, are more resistant to further Sensitization than an equivalent lot of cells sensitized with a larger number of antibody units from a serum that gives a high absorption. The serum, therefore, that gives the higher absorp- tion, has the smaller percentage of this inhibiting substance, as compared to the number of antibodies. It is true that the number of animals is not large, for which the history was kept of the absorption at intervals during the period of immunity. The results, however, are consistent. The fact that, in every case, the absorption is highest early in the immunity period, and that the greatest drop in the absorption is coincident with the most rapid drop in titer, is forcibly suggestive that the inhibiting substance represents the deterioration products of the amboceptor, i.e., the amboceptors which have lost their sensitiz- ing power, but still retain their combining affinity for the antigen. This would be, in effect, somewhat analogous to the “agglutin- oids” of Ehrlich. This hypothesis would account, not only, for THE JOURNAL OF IMMUNOLOGY, VOL. VII, NO. 6 494 H. W. CROMWELL the great differences in the absorption from the different sera, but also for the fact that so many of the tests fail to conform to the equation proposed by Arrhenius. xX. SUMMARY 1. Red cells, when treated with their specific immune serum, take up amounts of sensitizer which vary with the concentration of the sensitizer in the surrounding medium. Although from very low concentrations, all is not absorbed, as much as 2000 units or more may be taken up by one unit of cells, provided the con- centration of the antibodies is sufficiently high. The logarithmic plots of some of the tests approach a straight line very closely, while many deviate considerably from the straight line. The amount of deviation of the curve is propor- tional to the amount of deviation of the tests from the physical law proposed by Arrhenius, which is expressed by the equation, B = KC". The amount of absorption varies also with the serum used. Some given practically complete extraction from as much as 200 units per cubic centimeter while with other sera, all is not absorbed from concentrations of antibodies as low as 50 units per cubic centimeter. 2. The combination of the red cell and its antibody is extremely rapid. In every case, fifteen minutes were found to be sufficient time for equilibrium to be established. 3. With the majority of sera the extraction of sensitizer by the cells was found to be greatest at a temperature of about 37°C. Some sera, however, gave more complete absorption at 15° to 25°C. than at any other temperature. 4, Neither the quantitative fluctuations of the serum proteins nor the titer of the serum was found to have any constant rela- tion to the variations in the absorption. 5. Cells, sensitized with a serum which gave a low absorption, were more resistant to further sensitization than cells, sensitized with a much larger number of units from a serum which gave a high absorption. This indicates that the substance in the serum, that inhibits the absorption, is really absorbed by the cell along with the antibodies. HEMOLYTIC ANTIBODY-ANTIGEN COMBINATION 495 6. In every case tested, the extraction of the antibodies was greatest from serum drawn early in the period of immunity, re- gardless of the titer. The greatest drop in the absorption was found to be coincident with the most rapid drop in the titer. 7. The accumulated evidence is forcibly suggestive that the great variations in absorption are due to the presence of the de- terioration products of the antibodies, i.e., the antibodies which have lost their sensitizing power but still retain their affinity for the antigen. The writer’s thanks are due Dr. R. R. Hyde, under whose direction the work has been carried forward. REFERENCES (1) Enruicu, Pauw: Studies in Immunity. 1910. Translation by Charles Bolduan. (2) ARRHENIUS, S.: Immunochemistry. MacMillan, 1907, 144-166; 257-259. (3) Borpet, J.: Studies in Immunity. Translation by Frederick P. Gay. First Ed., 1909, 440. (4) Manwarine, W. H.: The absorption of hemolytic amboceptor. Jour. Inf. Dis., 1905, 2, 485. (5) Manwarine, W. H.: The third Serum component. Jour. Inf. Dis., 1906, 3, 647. (6) Manwarine, W. H.: Analytical methods of serum pathology. Jour. Biol. Chem., 1905-1906, 1, 213. (7) Manwaring, W. H.: On the so-called physical chemistry of the hemolytic serum. Jour. Inf. Dis., 1907, 4, 219. (8) Manwarine, W. H.: Qualitative changes in the third serum component. Jour. Inf. Dis., 1908, 5, 55. (9) Courter, C. B.: The isoelectric point of red blood cells and its relation to agglutination. Jour. Gen. Physiol., 1920-1921, 3, 309. (10) Coutrrr, C. B.: The equilibrium between hemolytic sensitizer and red blood cells in relation to the H-ion concentration. Ibid., p. 513. (11) Kaun, R. L., anp Lyon, D. 8.: Studies on complement fixation. IV. On the affinity of sheep corpuscles for anti-sheep hemolysin. Jour. Inf. Dis., 1921, 29, 651. (2) Bonn, ©. G.: The agglutination of bacteria in vivo. Jour. Exp. Med., 1915, 22, 484. (13) Nem, M. H.: The complement fixation test for syphilis. Pub. Health Reports, 1918, 33, 1387. (14) Hinton, Wa. A.: A standardized method of performing the Wassermann reaction. Am. Jour. Syph., 1920, 4, 598. (15) Kotmer, J. A., Rue, A. M. anp Trist, M. E.: Studies in the standardiza- tion of the Wassermann reaction. XII. Ibid., p. 616. (16) Manwarine, W. H., ann Yosuio Kasvuma: Protein absorption by blood corpuscles. Proc. Soc. Exp. Biol. and Med., 1916, 13, 173. 496 H. W. CROMWELL (17) Hurwitz, 8S. H., anp Myer, K. F.: Studies on the blood proteins. Jour. Exp. Med., 1916, 24, 516. (18) Hurwitz, S. H., Meyer, K. F. anp Taussis, L.: Studies on the blood proteins. III. Jour. Inf. Dis., 1918, 22, 1. (19) Ropertson, T. B.: A microrefractometric method of determining the percentages of globulin and albumin in very small quantities of blood serum, Jour. Biol. Chem. 1915, 22, 235 (20) Huntoon, F. M. era. : Antibody studies. I, IJ, and III. Jour. Immun., 1921, 6, 117-200. (21) Roxuro Umemura: A serological study of cholera immunity. Jour. Immun., 1920, 5, 465. (22) E1sENBERG, P. AND VOLK, R.: Zeitschr. f. Hyg., 1902, 40, 155. (23) Amato, A. Ricerche Sperimentali sulla fagocitosi, Sperimentale, Florence, 1917, 71, 459. STUDIES ON THE TOXICITY OF HUMAN BLOOD PLASMA FOR GUINEA-PIGS I. RELATIVE TOXICITY OF FETAL AND MATERNAL PLASMA S. A. LEVINSON From the Department of Pathology and Bacteriology, and the Laboratory of Physi- ological Chemistry, College of Medicine, University of Illinois, and the Laborataries of the City of Chicago Municipal Tuberculosis Sanitarium, Chicago, Illinois Received for publication May 28, 1922 In a recent paper Sachs and Oettingen (1) call attention to certain striking differences in maternal as contrasted with fetal plasma. Their observations disclosed that the surface tension, as well as the relative stability of the plasma colloids, as deter- mined by coagulation with heat, salts, alcohol, etc., was dis- tinctly altered. They emphasize the fact that in our usual serological studies we ignore the réle of the largest and most unstable colloidal aggregate—fibrinogen—and that it is only when we work with plasma that we more closely approximate normal biological conditions. Particularly in relation to the Wassermann reaction their findings are of significance when we recall that the reaction is not infrequently nonspecifically posi- tive in the pregnant woman, while fetal serum on the contrary, is frequently negative despite positive clinical evidence of syph- ilis. This they believe may be related to the fact that the colloids of the pregnant blood flocculate more readily than do those of the fetal blood, therefore favoring nonspecific colloidal alterations. In view of the definite alterations that Sachs and Oettingen have found we thought it would be of interest to determine whether fetal and maternal plasma differed in toxicity when injected intravenously into guinea pigs. Numerous studies have been reported dealing with primary serum toxicity when 497 498 Ss. A. LEVINSON the serum of one species is injected into another; so, too, a num- ber have been reported that concern the relative toxicity of human serum (normal and from pathological cases) for the guinea-pig, but we have found no reference in the literature dealing with the particular phase of the question that we have under consideration. Insofar as we inject a colloidal mixture we must be prepared to meet with the usual mode of death that follows injections (in sufficient dosage) of any number of the more typical colloidal substances usually used in experimental work. Hanzlik and Karsner (2) have published an extensive study of the effects of the intravenous injection of various colloids which produce an “anaphylactoid”’ death, while Friedberger (8), Nathan (4), Novy (5), and others have studied in detail the phenomena associated with the injection of kaolin, starch, agar, etc. Re- lated to the phenomenon that we have under consideration must be considered the toxicity manifested by organ extracts when intravenously injected; the literature of this particular field is an extensive one. METHODS The maternal blood was collected shortly before delivery, the blood being obtained from the median basillic vein. The fetal blood was obtained from the umbilical cord. The blood was allowed to flow to the 10 cc. mark in a graduated centrifuge tube containing 2 cc. of a 2 per cent sodium citrate solution. The blood was then centrifuged and the plasma pippeted off into a perfectly clean tube and then immediately used. We are of the im- pression that the results are more accurate if the fresh plasma is used, for if the plasma is allowed to remain at room temperature or is kept in a very warm place for any length of time, colloidal changes occuring in the plasma may affect the ultimate outcome of the experiment. If a tendency of the blood plasma to clot was observed, the plasma was not used. Because of the relatively large amount of plasma needed for the experiments we collected at least two tubes of plasma whenever possible. Surface tension. In determining the surface tension of the plasma we used the Traube Stalagmometer (static method). Table 1 shows the relationship of the surface tension of fetal, pregnant, and non- pregnant female plasma. TOXICITY OF HUMAN BLOOD PLASMA FOR GUINEA-PIGS 499 Toxicity. We determined the toxicity of the plasma by slowly injecting directly into the jugular vein of guinea-pigs in the usual TABLE 1 Surface tension of blood plasma in maternal, fetal, and normal cases case xownre | arenas EA MRE” oer 1 63 59 61 57 2 62 56 62 58 3 60 55 61.5 60 4 63.5 60 62 59 5 61 Gy 63 58 6 62 58.5 59.5 i 63 55 8 62.5 59 9 63.5 58 10 60 Average..... 62.05 DIED 61.9 58.5 TABLE 1a Surface tension of blood plasma during various pertods of pregnancy CASE NUMBER DURATION OF PREGNANCY RESULT months 1 3 61 2 7 61 3 8 63 4 8 63.5 INVETARG je caches s b 62.12 TABLE i138 Surface tension of blood plasma before and after menstruation BEFORE MENSTRUATION AFTER MENSTRUATION Case number Result Case number Result 1 60 1 59 2 63 2 62 manner, beginning usually with 1 cc. of plasma per 100 grams of guinea pig weight. If immediate death resulted the dose was decreased to 0.75 cc. of plasma per 100 grams of guinea-pig, and so on. If the origi- nal dose did not kill almost immediately (one to five minutes) the dose 500 S. A. LEVINSON was then increased. Tables 2 to 7 give the amount of plasma used in the various conditions to produce immediate death in the guinea-pig. The shock resulting in these animals is not to be compared with a true anaphylactic shock, as these animals were not sensitized; it is a coagulation phenomenon that in many cases resembles an anaphylactoid shock. The phenomenon of coagulation toxicity will be described in a following paper. RESULTS The results obtained in these experiments, demonstrate the relative increase in toxicity of the maternal pregnant plasma, as compared with fetal plasma. The following is a detailed study of our investigation. 1. Toxicity of normal human plasma tmmediately before de- livery. The blood was obtained approximately one-half to one hour before the expected time for delivery. Our results show (table 2) that plasma obtained in these conditions is very toxic to the guinea pig, in a dose even as low as 0.5 ec. per 100 grams of guinea pig weight. It will also be observed that the surface tension of this plasma is high. When the heart and lungs are examined from those pigs that died immediately after the plasma was injected intravenously, a clot was found in the heart ventricles, and the lungs were markedly distended and engorged. On palpation the lungs feel like a resilient mass of tissue. In the animals dying later than 10 minutes after the plasma was injected intravenously, the heart did not contain a demonstrable blood clot, but the lungs were distended and the surfaces were frequently covered with petechial hemorrhages. As will be described in a subse- quent paper, the mechanism of shock. death resulting when such plasma is injected intravenously into a guinea-pig, is not due to a clumping of a foreign substance in the capillaries of the bronchioles, but results from a coagulation of the plasma within the pulmonary capillaries and smaller blood vessels. While blood plasma of pregnant women may contain an increased amount of fibrinogen, it is by no means to be concluded that this injected fibrinogen itself is the cause of the intravascular = oe | TOXICITY OF HUMAN BLOOD PLASMA FOR GUINEA-PIGS 501 TABLE 2 Toxicity of human plasma from normal pregnant cases immediately before delivery DOSE OF WEIGHT PLASMA PER | TOTAL DOSE ee eae ee | cee nis PIG WEIGHT grams rire iis haa, 340 0.5 ae, | Died in 24 hours 1 260 0.6 1.56 | Died in 10 minutes 280 0.7 1.96 | Died in 3 minutes ( 285 0.5 C7, | Died in 24 hours 2 280 0.7 1.96 Died in 5 minutes 240 1.0 Del Immediate death 260 0.3 0.78 Alive at end of 24 hours 3 260 0.5 Ibs Died in 10 minutes 310 0.65 Piel Immediate death 4 205 0.75 eo Died in 15 minutes 335 1.0 380 Immediate death ( 300 0.5 1153 Died in 24 hours 5 300 0.75 2.25 Immediate death 280 1.0 2.8 Immediate death 6 270 0.5 1.35 Died in 3 hours 395 _ 0.75 2.96 Immediate death 285 0.6 17 Died in 3 minutes 7 280 0.7 1.96 Died in 3 minutes 240 10 2.4 Died in 2 minutes 340 0.5 1.9 Died in 12 hours 8 260 0.6 1.56 Died in 10 minutes 280 0.7 1.96 Died in 3 minutes 280 0.5 1.4 Died in 7 hours 9 260 0.75 1.95 Died in 3 minutes 270 1.0 Dh Immediate death 310 0.5 16 Died in 6 hours 10 285 0.75 2S Died in 2 minutes 250 1.0 2.5 Immediate death 502 S. A. LEVINSON coagulation. Welker (6) has shown that serum (rattle snake) from good clotting blood produced no fatalities in the guinea- pig with doses running up to 1 cc. injected intraperitoneally. 2. Toxicity of normal human fetal plasma. Fetal plasma used in these experiments was obtained from the blood of the umbili- TABLE 3 Toxicity of normal human fetal plasma DOSE OF WEIGHT PLASMA PER | TOTAL DOSE emeen’ | OF Scene: | a nea | onreaees aris PIG WEIGHT grams cc cc 1 330 0.7 2S Alive at end of 24 hours 165 12, 1.98 Alive at end of 24 hours - 270 1.0 eT | Alive at end of 24 hours 4 210 1 ey 2.52 Alive at end of 24 hours 3 250 1e5 3.15 Alive at end of 24 hours A (| 290 1.0 2.9 Alive at end of 24 hours \ 300 12 3.6 Alive at end of 24 hours 2 ( 225 0 2.25 Alive at end of 24 hours \} 265 1.5 3.97 Alive at end of 24 hours 6 f 270 1.0 2a, Alive at end of 24 hours i 1.2 2.52 Alive at end of 24 hours Alive at end of 24 hours _ & _ eer — ives} io) = 330 1 3 : Alive at end of 24 hours : f 280 1.0 2.8 Alive at end of 24 hours | 300 1.5 4.5 Alive at end of 24 hours 9 f 240 1.0 2.4 Alive at end of 24 hours \ 290 | 1.5 4.35 Alive at end of 24 hours eal cord coming from the placenta. In comparing our results, we used only the fetal and maternal plasmas obtained from the same case. It will be observed from the table of comparative toxicities that the maternal plasma is approximately 3 times as toxic as the fetal plasma. While 0.5 cc. of maternal plasma —— TOXICITY OF HUMAN BLOOD PLASMA FOR GUINEA-PIGS 503 per 100 grams of guinea pig kills almost immediately, 1.5 ce. of fetal plasma per 100 grams of guinea-pig does not kill the ani- mal (table 3). With greater doses death may occur with autopsy findings similar to those observed with the maternal plasma. 3. Toxicity of normal human plasma during various stages of pregnancy. The blood used for investigation in this experiment was obtained from women in various stages of pregnancy, vary- ing anywhere from three months to almost term. Table 4 shows our results, and it will be observed that blood plasma from TABLE 4 Toxicity of human blood plasma during pregnancy ( 3 to 8 months) DOSE OF PLASMA TOTAL CASE DURATION | WEIGHT OF PER 100 DOSE OF NMMBER loangNancr| PIG | GRAMS OF | PLASMA, ary, PIG WEIGHT months grams cc cc 1 7 320 0.5 1.6 Died within 24 hours 200 0.75 pe5 Immediate death 340 0.5 Ib 2 Died within 24 hours 2 8.5 235 0.75 L7G Immediate death 340 1.0 3.4 Immediate death 3 8 270 0.75 2.02 Immediate death 245 0.5 ie22 Died within 24 hours 4 3 210 1.0 2AO Immediate death 245 0.75 1.83 Died within 24 hours pregnant women during the earlier months of pregnancy is as toxic as blood plasma from pregnant women at time during delivery. Just how soon the blood in pregnancy becomes toxic we have not been able to ascertain. In one case of the third month of pregnancy the toxicity of the blood was not very great. 4. Toxicity of normal human male and female plasma. For comparative purposes it was of interest to observe the relation- ship of the degree of toxicity that normal human male and female plasma may have with that of fetal and pregnant plasmas. For this purpose normal healthy medical students were used. The surface tension of this plasma was intermediate between fetal 504 S. A. LEVINSON and maternal plasmas. The toxicity of the normal plasma was less than that of the maternal plasma (table 5). The condition of the lungs in the guinea-pigs dying from the toxic dose of normal human plasma is the same as that found in the case of pregnant plasma. Inasmuch as the mode of death TABLE 5 Toxicity of human plasma from normal cases DOSE OF WEIGHT PLASMA PER | TOTAL DOSE OF GUINEA- | 100 GRAMS OF PLASMA PIG OF GUINEA- | INJECTED PIG WEIGHT CASE NUMBER RESULT grams cc. ce. 295 0.5 eA, Alive at end of 24 hours 315 On” Foe Alive at end of 24 hours 1 220 120 ae, Died in 24 hours 210 12 2.25 Died in 24 hours 350 1.5 525 Immediate death 105 0.75 0.78 | Died in 24 hours 2 125 1.0 a be Died in 24 hours 105 15 1s Immediate death 445 0.75 38) Died in 24 hours 3 365 1.0 3.65 Died in 1 hour \] 186 15 2.77 Immediate death 4 105 0.75 0.78 Died in 24 hours 150 1.5 | 2.25 Immediate death 290 1.5 4.35 Immediate death 5 275 EY 33.8) Immediate death 190 1.0 1.9 Alive at end of 24 hours 6 250 1.5 Sis Alive at end of 24 hours 205 1eT5 3.58 Immediate death in the guinea pig injected with maternal and normal human plas- mas is the same, the differentiation being one of degree, the toxic factor in the maternal plasma which causes the immediate death of the guinea pig must be present in the fetal and non- pregnant human bloods as well. 5. Toxicity of normal human female plasma before and after menstruation. Table 6 shows that the blood plasma before ——— TOXICITY OF HUMAN BLOOD PLASMA FOR GUINEA-PIGS 505 TABLE 6 Toxicity of normal human (female) plasma before and after menstruation DOSE OF eB AES Sal ieee ee eee te PIG j OF GUINEA- INJECTED PIG WEIGHT a. Before menstruation grams cc. CC. 1 195 1.2 2.34 Died in 1 minute 235 1.0 2.35 Died in 43 minutes 240 1.0 2.4 Immediate death 2 235 0.75 1.76 Immediate death 225 0.5 0.62 Alive at end of 24 hours 3 270 1.0 Pha Immediate death 210 0.75 LAG / Died in 2 hours 220 1.0 Dee, Immediate death 4 230 0.75 Lee Immediate death 205 0.5 1.02 Died in 24 hours 210 0.75 1577, Alive at end of 24 hours 5 240 1.0 2.4 Alive at end of 24 hours 200 1.25 DED Died in 3 hours 160 1.5 2.4 Immediate death b. After menstruation 225 1.0 2.25 Immediate death 1 190 0.75 1.42 Immediate death { 225 0.5 de2 Died in 24 hours 2 340 0.75 7 5) Immediate death 260 0.5 123 Alive at end of 24 hours 250 10) Det Died in 12 hours 3 265 1.25 34,84 Died in 12 hours 260 15 3.9 Immediate death 4 { 260 0.75 1.95 Immediate death ( 230 0.5 1.15 Died in 12 hours 5 225 1.0 2.25 Died in 12 hours 210 1.25 2.62 Immediate death 506 Ss. A. LEVINSON menstruation is a little less toxic than normal female plasma, while blood plasma following menstruation is almost as toxic as pregnant blood plasma. Jacobi in 1876 advanced the theory that the metabolic proc- esses In women present a distinct rhythm, gradually increasing in intensity up to the time of the menstrual flow, when they suddenly drop and reach their lowest point; after this they grad- ually rise again to attain their maximum intensity before the next menstrual period (Webster (7), Ott (8), Van der Velde (9)). These changes include in the blood—settling of blood TABLE 7 Table of average toxicities of human blood plasma in pregnancy, fetal, and normal cases CONDITION AVERAGE TOXIC DOSE Pregnancy..............| 0.5-0.75 ce. per 100 grams of guinea-pig weight Hetalearetenc nsec ie oa 1.75 cc. per 100 grams of guinea-pig weight did not kill Period. of gestation, 3-8 TMOMUNS oA Ae ae eta eae oe 0.75 cc. per 100 grams of guinea-pig weight Normal hunen {emni.| 0-12 per M00 goama of euinea oe weigh Menstruation {ter || 0.78 per 100 grams of guinea-pig weight corpuscles, leucocytosis, changes in fibrinogen content, in enzymes and antienzymes, cholesterol and lipoid balance; clini- cally the changes are manifested by temperature rise, alteration of pulse rate, altered metabolism, nitrogen excretion, etc. Numerous investigators have associated ovulation, the devel- opment of the corpus luteum, and menstruation. A similar association exists between ovulation and the development of the corpus luteum of pregnancy, and pregnancy. If fertiliza- tion takes place, the corpus luteum persists unchanged for months, and its secretion may play a prominent part in the regulation of a number of important biological alterations asso- ciated with pregnancy. If implantation does not occur the cor- 3 TOXICITY OF HUMAN BLOOD PLASMA FOR GUINEA-PIGS 507 pus luteum immediately shows degenerative changes, which in turn are supposed to lead to the onset of the menstrual flow (Ruge (10), Fraenkel (11), Halban (12), Schroeder (13)). The menstrual cycle and pregnancy having many common biological alterations, one in the toxicity such as we have observed, might seem especially logical, and seems to offer conductive evidence of the association of the two conditions. DISCUSSION AND CONCLUSIONS Parallel with the findings of Sachs and v. Oettingen that fetal and maternal plasma occupy the opposite extremes of colloidal stability (normal plasma occupying an intermediate position), we have found that of the two substances the maternal plasma is much more toxic for the guinea-pig when intravenously in- jected. The fetal plasma can be injected with impunity in rela- tively large doses. In this regard normal plasma also occupies an intermediate position. Both chemically and physically certain differences exist be- tween maternal and fetal bloods. The blood lipoids differ quantitatively and even the proteins, as Naegili points out, are dis- similar. Albrecht (14), working in this laboratory, has observed that in saturating with magnesium sulphate practically no globulins are precipitated from fetal blood. Howe (15) has recently examined the differences that exist in the blood of new born calves and finds that the euglobulin and the pseudoglobu- lin 1 fractions are absent, but appear in the blood after the ingestion of cholostrum. Fibrinogen is present in fetal blood in increasing amounts until term (16) but the amount in the ma- ternal blood is much greater (0.47 to 0.4: per cent as compared with 0.31 to 0.32 per cent normal, Gram (17). Among the physical variations, the difference in surface ten- sion is very striking, as well as the alteration in the settling time of the red blood corpuscles. Inasmuch as the death that occurs in the guinea pig is so obviously associated with clotting phenomena, the increase in the fibrinogen content of the maternal plasma might be con- sidered responsible. Welker found that in using serum from ~ 508 S. A. LEVINSON rattlesnakes the serum was toxic in a degree corresponding to the fibrinogen content remaining in the serum, as soon as the fibrinogen was quantitatively removed the serum was no longer toxic. Even the serum of the maternal organism is quite toxic for the guinea pig, but this may be, as Welker found in the rattle snake, because of the fibrinogen still remaining in the serum. Adsorption by inert agents and filtration through a Berkfeld filter largely remove the factor that causes immediate death in the pig. While we can in general state that an increase in the fibrinogen content goes hand in hand with increased toxicity, it is by no means excluded that some factor associated with the fibrinogen increase rather than the fibrinogen itself, is the fac- tor responsible. In view of the fact that fetal plasma contains no euglobulin nor pseudoglobulin I it might be assumed that these particular proteins are associated with the toxic effect. Experiments to elucidate this particular question are now being made. Until further investigation we would draw no final conclusions con- cerning the particular factors responsible for the considerable. differences in toxicity observed between maternal and fetal plasmas. SUMMARY Blood plasma obtained from pregnant women is more toxic for the guinea-pig on intravenous injection than plasma obtained from normal individuals. Human placental plasma is practi- cally nontoxic. Similar relations are found with dog plasma obtained from pregnant dogs and from full term fetuses. Plasma obtained during the menstrual cycle (before and after) is occasionally more toxic for guinea-pigs than normal female plasma. The cause of death in the guinea-pig is an intravascular clot- ting of the capillaries of the lungs, with a resulting asphyxia, and occlusion in other capillaries as well (brain, liver, etc.). Death can be prevented if the animal is hirudinized, or if the amount of citrate simultaneously injected, is increased. a ————., e TOXICITY OF HUMAN BLOOD PLASMA FOR GUINEA-PIGS 509 ’ REFERENCES (1) Sacus, H., AND von OETTINGEN, Kj.: Miinch. med. Wochenschr., 1921, 68, 351. (2) Hanzuix, P. J., anp Karsner, H. T.: Jour. Phar. and Exp. Therap., 1920, 14, 449. (8) FRIEDBERGER, E.: Zeit. f. Immunitat. Orig., 1913-14, 20, 405. (4) Natuan, E.: Zeit. f. Immunitat. Orig., 1914, 28, 204. (5) Novy, F. G.: Jour. Infec. Dis., 1917, 20, 536 and 657. (6) Wretxer, W. H., anp MarsHatt, J.: Jour. Phar. and Exp. Therap., 1915, 6, 563. (7) WessterR: Montreal Med. Jour., 1897, volume 26, WessteR, J. C.: Pelvic viscera in relation to microdrganisms in health and disease, p 91. Wesster, 2. E.: Strangulated umbilical herina, p. 359. (8) Orro, A.: Vehr. des X internat. med. Congresses, Berlin, 1891, 10. (9) VANDER VELDE: Wellenbewegung u. Menstrual. Jena, 1905. (10) Rugs, Caru: Archiv. f. Gyn., 1913, 100, 20. (11) FraEnxKEt, C.: Archiv. f. Gyn., 1910, 91, 705. (12) Haxpan, J., AND Kouuer, R.: Archiv. f. Gyn., 1914, 103, 575. (13) ScoroEpDER, C.: Archiv. f. Gyn., 1914, 101, 1. (14) Personal Communication. (15) Howsz, P. E.: Jour. Biolog. Chem., 1921, 49, 115. (16) Emmet, V. E., Levinson, S. A., Fiscu, M. E.: Jour. Exp. Med., 1920, 31, 177. (17) Gram, H. C.: Jour. Biolog. Chem., 1921, 49, 279. Pah ari, wey, Oe Fr) STUDIES ON THE TOXICITY OF HUMAN BLOOD PLASMA FOR GUINEA-PIGS II. COAGULATION TOXICITY 8S. A. LEVINSON From the Department of Pathology and Bacteriology, and the Laboratory of Physiological Chemistry, College of Medicine, University of Illinois, and the Laboratories of the City of Chicago Municipal Tuberculosis Sanitarium, Chicago, Illinois Received for publication May 28, 1922 I. INTRODUCTION In our previous report (1), we have briefly mentioned the phenomena produced by the intravenous injection of human plasma into the guinea-pig which we associated with coagulation changes. It is the object of this paper to report in some detail the results of this observation, and to show briefly the relationship it has with anaphylaxis and the associated phenomena. The theories of anaphylaxis and anaphylactic shock are so numerous that it is only possible to mention a few of them, especially those which may have any bearing on our findings. Biedl and Kraus (2), Schultz and Jordan (8) among others have shown that the peculiar and rapid fatal character of anaphylactic shock in the guinea-pig depends on the valve-like closure of the bronchioles, due partly to the special susceptibility of the bron- chiolar musculature, and partly to the thick lining of these tubes in the guinea-pig, so that when contraction of the bronchi take place, the lumen is obstructed. These findings in conjunction with other explanations that may be given as to the possible cause of anaphylactic shock, presuppose the fact that the guinea- pig has been previously sensitized against a foreign substance, and then subjected to the anaphylactic shock. A similar observa- tion was also made by Hanzlik and Karsner (4) when unspecific colloids were used in non-sensitized guinea-pigs. 511 Bio S. A. LEVINSON Vaughan (5), Schittenhelm and Weichardt (6), and numerous other investigators have shown that various products of partial protein hydrolysis produce, on injection into normal animals, symtoms in many respects resembling anaphylactic shock. Many observers have been able to produce symtoms of anaphy- lactic shock in other ways, thus Richet (7), Vaughan (5), Anderson and Frost (8), by injecting into a normal animal, serum from a sensitized animal digested with the antigen, Friedberger (9), by injecting serum from a normal animal digested with specific precipitates, while Friedemann (10) used blood corpuscles sensitized with hemolysin. Just where the specific antibody and antigen union is brought about (which is cellular or humoral) is still a moot question. Some place the formation of a non- specific poison in the circulating fluids, others in the cells partici- pating in the anaphylactic reaction. Doerr and Moldovan (11 have suggested that an immediate physical change in the blood colloids accounts for the anaphylactic reaction, a theory that is at the present under study by a number of French investigators. Dale (12), Wells (13), as well as Besredka (14) give a detailed résumé of the various theories of anaphylaxis and anaphylactic shock. These include the anaphylatoxin theory of Friedberger (9), the apotoxin theory of Richet (7), numerous ferment theories (Jobling and Petersen (15)), the physical theory of Doerr and Moldovan (11), physical nervous theory of Besredka (14), and peptone theory of Beidl and Kraus (2). When dealing with such an intoxication as that following plasma injection in the guinea-pig these theories are of courseirrelevant. Von Behring’s (16) platelet theory, or the observation of Dale and Laidlow (17) as well as Hanzlik and Karsner (4), that the clumping of platelets, or conglutination of red cells and presence of thrombi in pul- monary vessels, appears to be the most logical explanation of shock taking place following an intravenous injection of foreign substance into a non-sensitized guinea-pig. II. METHODS AND RESULTS The method used in demonstrating this phenomenon is similar to that previously described by us when plasma from pregnant TOXICITY OF HUMAN BLOOD PLASMA FOR GUINEA-PIGS 513 women was injected intravenously into guinea-pigs. We wish to emphasize the fact that in these experiments, as in the others deal- ing with the toxicity of human plasma, we used guinea-pigs that had not been used for other purposes, and were not sensitized. Mode of death. When approximately one half the lethal dose of human plasma is injected intravenously into a guinea-pig, the latter appears to be unaffected, and the animal lives for an indefinite period. When, however, a sublethal dose of plasma is injected intravenously, the following symptoms become manifest: The animal is at first restless; in a few seconds it becomes quiet and seems sick. Respirations are increased in frequency, the breathing seems labored and is more shallow than in the normal. A contraction of the muscles of the gastro-intestinal tract follows, and the animal as a rule passes feces, not infrequently a diarrhea may ensue. Convulsive seizures take place, in some cases appearing gradually, at other times very suddenly; the animal’s head may take an oscillatory motion from left to right, then suddenly the head is thrown backward and convulsive jumping movements take place. This may last for a few seconds, and as the animal becomes exhausted it lies on one side and appears as if dead. The respirations are shallow and almost invisible, then become increased in frequency and the animal struggles to get back on its feet again. Although the animal appears to be sick and gives one the impression that it may die almost any minute, the guinea-pig may live from three to twenty-four hours, and in some cases even at a longer period makes a complete recovery. When the guinea-pig receives a toxic dose of human plasma the phenomenon may be very similar to that here described, but as a rule the following sequence occurs: The animal immediately after the injection becomes very restless, runs about until almost exhausted, and then scratches its nose and appears to be very irritable. There then appears a contraction of smooth muscle fibers (there is a discharge of urine and feces) and a peculiar con- traction of the intercostal muscles, and evidently a contraction of the bronchial muscles resulting in a peculiar respiratory grunt. Pulse and respirations are quickened and the animal may lie 514 S. A. LEVINSON over on one side and get a convulsive seizure or the latter attack may effect the guinea-pig without any warning. The head is thrown backward and the back is bent, the guinea-pig appearing in opisthotonus. These convulsive attacks may be of such a severe degree as to cause the animal to be thrown off its feet. During this attack the hind legs are straightened, and as the guinea-pig makes attempts to get on its feet again there is this added difficulty of paralysis of the hind legs; at times the front legs may also become paralyzed. After a few more shallow respi- ratory efforts the guinea-pig dies. Instead of the extended and rigid condition that the animal was in a few moments ago, the entire body is relaxed, the muscles are soft and the legs flexible. Upon opening the guinea-pig and examining the heart and lungs it will be seen that the heart is distended and a large blood clot fills the ventricles. The lungs are distended, firm, and elastic and feel like a semisolid piece of rubber. Occasionally, hemorrhages may be seen on the surface of the lung. When the lung is cut open the latter still retains its solid appearance, but at times a serous exudate may be squeezed from the cut surface. Sections of the lungs of the animals that died immediately after the toxic dose of plasma was injected, were made. The tissues were fixed in Zenker’s solution for twenty-four hours, then washed in distilled water for twenty-four hours and later embedded in paraffin. Sections were later stained for fibrin by using Weigart’s staining method. Microscopic examination of these sections show fibrin strands and clumps of fibrin in practic- ally all the capillaries of the lung. Occasionally red blood cells are seen engulfed in the fibrinous mass (fig. 1). Effect of heating and filtering plasma. It was of interest to note the effect temperature changes and filtration of human plasma may have on the degree of toxicity of the latter for guinea- pigs. Sakamoto (18) has shown that lung extract, the most toxic extract of all organs, retains its toxicity for about 2 days when kept on ice and is only slightly reduced by filtration through a Berkfeld filter. This observer has also noted that when the extract is shaken with kaolin or animal charcoal the toxicity is TOXICITY OF HUMAN BLOOD PLASMA FOR GUINEA-PIGS 515 Fig. 1. Microphotograph of lung showing fibrin in the smaller vessels. There is also amarked distension of the alveoli. Zeiss X 200. 516 S. A. LEVINSON slightly reduced, but that it is not affected by a solution of egg white. When fresh normal serum is added to the extract, the toxicity is neutralized, but serum inactivated at 50°C. loses its neutralizing properties. Our findings in many respects are similar to those of Sakamoto, and we have observed no change in the degree of toxicity when sterile human plasma is kept on ice for two days. Although Sakamoto states that the degree of toxicity is slightly reduced by filtration through the Berkefeld filter we have found that the degree of toxicity was entirely reduced to that of normal. The toxic normal dose of human plasma kills the guinea-pig in a few moments, while the same dose of the same plasma (taken from normal or pathologic cases) which has been filtered through a Berkefeld filter, does not effect the guinea-pig. In fact, the animal is as active at the end of twenty-four hours as before the injection (protocol 1). When plasma is heated to 56°C. for twenty minutes the degree of toxicity is lessened. It will be noted that when twice the toxic dose is injected intravenously the animal remains unaffected. Although Sakamoto states that the toxicity of the lung extract is destroyed by heating at 100° for two hours and at 38° for ten hours we have not been able to use this factor because the blood plasma would clot. Heatingto 56°C. for twenty minutes produces the same results as filtering the plasma through the Berkefeld filter (protocol 2). Experiments with hirudin and citrate. One cubic centimeter of hirudin was injected intravenously into a guinea-pig and from three to five minutes later the toxic dose of the plasma was also injected intravenously. Instead of immediate death from the toxic dose, the animal remained unaffected, for at least five hours. When a dose of plasma was injected somewhat larger than the toxic dose after the animal received hirudin, the guinea- pig did not die immediately nor were marked convulsions pro- duced, but the animal was sick and in the course of several hours it would die. Also increasing the amount of citrate does the same thing (protocol 3). TOXICITY OF HUMAN BLOOD PLASMA FOR GUINEA-PIGS 517 PROTOCOL 1 Effect of filtering plasma on the degree of toxicity for guinea-pigs AMOUNT DOSE OF DOSE OF |OF PLASMA PLASMA PLASMA INJECTED WEIGHT INJECTED INJECTED AFTER OF PER 100 Pearce A ascent PER 100 | FiuTRa- : GUINEA- | GRAMS OF |) ecpep GRAMS OF| TION RESULT PIG GUINEA- GUINEA- | THROUGH PIG PIG A BEREKE- WEIGHT WEIGHT FELD FILTER grams cc. cc cc cc 250 1 Ate Immediate death 1 25 Alive at end of 24 hours 235, 1 2.35 | Immediate death I 2.35 | Alive at end of 24 hours PROTOCOL 2 Effect of heating plasma on the degree of toxicity for guinea-pigs DOSE OF DOSE OF | AMOUNT PLASMA PLASMA |OF PLASMA WEIGHT | INJECTED INJECTED | INJECTED OF PER 100 | “MOUNT PER 100 AFTER GUINEA- | GRAMS OF eras RESULT GRAMS OF | HEATING WOLSENEP ERY PIG GUINEA- GUINEA- | TO 56° PIG PIG FOR 20 WEIGHT WEIGHT | MINUTES grams cc. cc ce ce 260 1 2.6 Immediate death 2 Die Alive at end of 24 hours 200 1 2.0 Immediate death 2, 4.0 Alive at end of 24 hours PROTOCOL 3 Effect of previous intravenous injections of hirudin on the degree of toxicity of human plasma for guinea-pigs DOSE OF PLASMA as ped AMOUNT cats cars AMOUNT aurxea- | Gras oF [OF Pee REROED HIRUDIN | Or OGrED. eee PIG GUINEA- INJECTED PIG WEIGHT grams ce. cc. ce. ce. 240 1 2.4 Immediate death 1 2.4 Alive at end of 5 hours 275 1 2.75 | Immediate death 1 2.75 | Alive at end of 6 hours 518 S. A. LEVINSON III. DISCUSSION _ In solving this problem, the method of approach would be by the process of elimination. If previous sensitization of an animal by a foreign substance causes a liberation of certain antibodies in the blood stream, and a subsequent injection of the same or other foreign substance produces death by virtue of the antigen com- bining with the antibody in the blood vascular system, then this explanation for coagulation toxicity does not hold true, since only one foreign substance is injected and produces death. Doerr and Moldovan (11) suppose that the effect of the antigen, in contact with the antibody of the sensitized animal, alters the state of aggregation of the colloids in the blood plasma, and induces changes of surface tension, which are responsible for the effect on the plain muscle and other tissues. None of the supposed theories that have been formulated to explain anaphy- laxis or anaphylactoid shock can be used wholly or partially in explaining our phenomenon of coagulation toxicity, as long as sensitized animals are used for these experiments. Hanzlik and Karsner have demonstrated that by the intra- venous injection of various colloidal agents such as agar, congo red acacia, and other products they were able to produce inflation of the lungs very similar to that found in true anaphylactic shock. v. Behring’s platelet theory seems of considerable interest in the mechanism of death with which we are dealing, and this explana- tion comes closer to the actual state of affairs than do most of the other theories. v. Behring (16) was of the opinion that some disturbance of the platelets, endogenous or exogenous in character, results in a clumping of the platelets in the smallest blood vessels forming capillary emboli and symptoms of anaphylactic shock. Freund (19) has shown that by injecting fresh serum intrave- nously into a rabbit, the animal is killed instantly. This observer has also shown that when fresh serum is injected into cats, the temperature drops, the respiratory center becomes involved, and there is a ‘‘shock effect’? on the heart with a fall in blood pressure. Freund is of the opinion that sudden death following an intravenous injection of a foreign protein is due to broken TOXICITY OF HUMAN BLOOD PLASMA FOR GUINEA-PIGS 519 down platelets which disturb the cellular elements in the blood. The physical and chemical state of the blood platelets, may also become altered and these in time may become toxic to or occlude the smaller blood vessels. Freund and Gottlieb (20) have also noted that the breaking down of cellular elements can influence the autonomic function very markedly. In pregnancy Neu and others have demonstrated that vasoconstrictor substance affects the vascular system more so than in normal persons. This observation has also been made by Hanzlik and Karsner, who have shown that any foreign substance that may be injected intravenously into an animal, may have the power of obstructing smaller blood vessels, and thereby produce a phenomenon similar to that described by v. Behring. Many of these inorganic colloids have the property of causing a mutual precipitation when they are mixed with a solution of protein, and this probably takes place within the blood vessels when inorganic colloids are injected intravenously. It is possible that the agglutination of corpuscles and the formation of thrombi which takes place within the capil- larieson the injection of bacteria, asin pseudoanaphylaxis (Herzfeld and Klinger (21)) is the result of the bacteria] extracts from these organisms forming an active ingredient in the process of coagula- tion. It will be observed that none of the theories mentioned, or those enumerated in the various monographs and discussions on the subject deal with fibrinogen as possible factor in anaphyl- axis or associated phenomena. That the stability of the colloids in the blood stream is to be considered an important factor in the study of hypersensitiveness has recently been emphasized by Sachs (22). The fibrinogen fraction of the blood contrasted to the albumen fraction is colloid- ally very unstabile. Sachs and von Oettingen (23) have shown that blood plasma from normal human pregnant cases, from normal individuals, and from fetal cases will flocculate in the order mentioned. ‘The fibrinogen in the maternal blood plasma is very labile and therefore flocculates readily, while the fibrinogen fraction of the fetal blood plasma seems more stabile; the plasma may show a faint turbidity when subject toa variety of procedures (heat, alcohol, ete.) but no marked flocculation as is seen in the 520 S. A. LEVINSON maternal plasma. We have been able to corroborate these observations and have also noted that there was a difference in the degree of flocculation and turbidity between adult male and female plasma. There was a tendency of the latter plasma to flocculate sooner than the former, the probable reason being that the female plasma contains a higher percentage of fibrin than does themale plasma. This evidence, along with the observations of Fahraeus (24), Doerr (25), Linzenmeier,(26) Plaut and Pagniez (27) and others that the rate of sedimentation of the red blood cells in pregnancy as well as in other conditions depends upon the fibrinogen content of the blood plasma, is further proof of the importance of fibrinogen fraction. Because of the fact that the degree of toxicity for guinea-pigs is higher in those conditions in which the fibrinogen content is increased (pregnancy, acute stage of pneumonia, severe heart cases), and that we have also been able to observe fibrin strands and clumps in the smaller vessels and capillaries, we can not escape the conclusion that the phenomenon of coagulation toxicity when plasma is injected in guinea-pigs is associated with the quantitative fibrinogen content of the blood plasma injected. The mechanism causing death in the guinea pig no doubtisdue to emboliin the terminal vessels, particularly in the lungs, heart, and brain, and possibly in the liver. We have demonstrated that when an agent such as hirudin is injected intravenously into a guinea-pig with the purpose of pre- venting an intravascular coagulation, the toxic dose of the blood plasma will not kill the animal. This at once shows that the mechanism of death is due to a coagulation taking place within the blood vessels. By filtering the blood plasma through a Berkefeld filter the toxicity can be reduced, and heating the plasma to 56°C. for twenty minutes produces the same results. It is possible that the Berkefeld filter byabsorption prevents some of the colloidal fibrinogen from passing through the filter, and that heat- ing either disperses the fibrinogen or brings about absorptive aggregates with smaller colloidal particles with a lessening of potential toxicity. We have observed that when a clot forms in the plasma during centrifuging the blood, or forms when the plasma stands at room temperature for any length of time, the TOXICITY OF HUMAN BLOOD PLASMA FORGUINEA-PIGS 521 toxicity of the plasma is reduced, indicating that the fibrinogen is an active factor in determining the degree of toxicity. If we assume that ‘‘colloidoclasis’’ is one of the essential phenomena associated with anaphylaxis, a supposition that has found many adherents among French scientists (Lumiere (28), Kopaczewski (29) etc.) it is apparent that any alteration in the balance of the plasma colloids will be of profound importance, particularly if the colloidal balance is altered in the direction of a lessened stability. In such states it is obvious that anaphylactoid phenomena should be easily induced and relatively frequent. We might be war- ranted in assuming that (in view of the relative unstability of the fibrinogen fraction), toxic manifestations should become clini- cally apparent in diseases associated with an increased fibrinogen content of the plasma. Whether eclampsia and the toxemias of pregnancy are to be grouped here is not determined, but the frequency of capillary thrombosis in eclampsia might seem to favor this idea. So, too, the whole group of manifestations that French clinicians associate with the ‘rheumatoid state’ (manifes- tations in the joint membranes, and other serous cavities, in the eye, skin and intestinal tract, etc.) might be investigated profitably from this point of view. In general one might find some warrant for assuming that the fibrinogen fraction being unstable, there would be a tendency to symptoms analogous to anaphylactic shock in those patients who have a high fibrinogen content, but one must remember that in a diseased condition, especially where catabolic processes take place at a rapid rate, there is a tendency for the organism to adjust the balance or to neutralize the factor which causes this disturbance. REFERENCES (1) Levinson, 8. A.: This Journal. (2) Brepu, A., AND Kraus, R: Wiener k. Woch., 1910, 23, 385. (3) Scuuitz, W. H., anp Jorpan, H. E.: Jour. Phar. and Exp. Therap., 1911, 11, 375. (4) Hanzurk, P., anp Karsner, H. T.: Jour. Phar. and Exp. Therap., 1920 14, 463. (5) Vaueuan, Victor C.: Jour. Infect. Dis., 1907, 4, 476. (6) ScH1rreENHELM, H., AND WEIcHARDT, W.: Miinch. med. Wochen., 1912, 59, 67. 522 S. A. LEVINSON . (7) Ricnet, Cu.: L’Anaphylaxie, Paris, 1912. (8) AnpERSON, J. F., anD Frost, W. H.: Hyg. Lab. Bulletins, Washington, 1910, No. 64. (9) FriepserGer, E.: Zeit. f. Immunitat. ‘‘Orig.,’’ 1909, 2, 208. (10) FRIEDEMANN, Uuricu: Jahersb. uber d. Ergeb. d. Immunitat., orig., 1910, 6, Abt. i, 31. (11) Dorrr, R., AnD Motpovan, J.: Biochem. Zeit., 1912, 41, 27. (12) Dats, H. H.: Jour. Phar. and Exp. Therap., 1913, 4, 167. (13) Wetts, H. G.: Physiological Reviews, 1921, 1, 44. (14) BesrepKxa, A: Anaphylaxie et Antianaphylaxie, 1917, Masson et Cie, Paris. (15) Jopuine, J. W., AND PETERSEN, W.: Jour. Exp. Med., 1914, 20, 468. (16) von BeHrinea, E.: Deut. Med. Wochen., 1914, 40, 1833. (17) Daz, H. H. anp Latwtow, P. P.: Jour. Phys., 1919, 52, 355. (18) SaKamorTo, E.: Zeit. f. Immunitat., Orig., 1921, 32, 1. (19) Freunp, H.: Archiv. Exp. Path. and Phar., 1921, 91, 272. (20) Freunp, H., anp GotTtiies R.: Miinch. med. Wochenschr., 1921, 68, 383. (21) HerzreLp, E. anD KLINGER, R.: Schw. Med. Wehr. 1920, 1, 567. (22) Sacus, H.: Kolloidzeitschr. 1919, 24, 113. (23) Sacus, H., anp v. OnTTINGEN, K.: Miinch. med. Wochenschr., 1921, 68, 351. (24) Fanrazus, R.: Acta Medica Scand. 1921, 56, 3. (25) Doerr, R.: Schweiz. med. Wochen., 1921, 51, 937. (26) LinzENMEIER, G,: Med. K1., 1920, 16, 439. (27) Paantez,P.: Presse Medicale, 1921, 29, 405. (28) Lumifire A.: ‘“Antianaphylactic Shock and Colloioclasis.’’ Press med., 1921, 11, 960. Lumi&re, A.: Colloidal Shock. Paris Med., 1921, 11, 445. (29) KopaczEwskI, M. W.: Paris Med., 1921, 11, 379. ee INDEX TO VOLUME VII Acids, The toxicity of, for leucocytes, as indicated by the tropin reaction. . 271 Agglutination, Immunological studies on types of diphtheria bacilli. I., characteristics. II. Protective value of the standard monovalent anti- EAT een Say et isi onshccovsis clase tuareieuele eierae edepaze: see steve salmnenstanvenacotenen = 243 Alexander, H. L., Larsen, Nils P., and Paddock, Royce. Bronchial asthma and allied ions Giasea and immunological observations...... 81 Alexin (complement), On the origin and nature of, in guinea-pig blood...... 435 Allergens, The diagnostic cutaneous test and therapeutic injection of...... 119 Allergic reaction, An, of the tuberculous uterine horn...............+...... 47 Allergies, On the relative susceptibility of the American Indian race and the white race to the, and to serum disease...................:0.500005 201 ——, The age incidence of serum disease and of dermatitis venenata as Comipareawich, tis OL Che WAG WAL soc oye ease ae ale o 2ys =a la oe nelle eine = 193 ——, The preparation of fluid extracts and solutions for use in the diagnosis and treatment of the, with notes on the collection of pollens............ 163 Allergy), On the phenomenon of hyposensitization (the clinically lessened sensitiveness Of........... PEN MEE MAO te Re See Ma hae ene 219 —, The diagnostic cutaneous reaction in. Comparison of the intra- dermal method (Cooke) and the scratch method (Schloss).............. 97 Alpha streptococci, A serological study of, from the upper respiratory tract. 361 Antibody-antigen combination, hemolytic, A study of the................-. 461 Antiorgan sera, Relationship of various..............--2+----seeee eee eees 51 Antitoxin, Immunological studies on types of diphtheria bacilli. I. Agglu- tination characteristics. II. Protective value of the standard mono- WAIN Sas SESS. BU ola o Sones S Ce ebne Une aeme rede beadsaaddds ods 243 Asthma, Bronchial, and allied conditions. Clinical and immunological Qbsenyationseasscecc esc cra Mierateacsbeneteeeteveis acest nica See etch hal oe sea eRe 81 ——, bronchial, New etiologic factors in..............0--- eee eee eee ences 147 Bacilli, diphtheria, Immunological studies on types of, I. Agglutination characteristics. II. Protective value of the standard monovalent CNTY ITTD (eT ae eR a eR ene Ue Rm a Taree eae alec 2 Sos more 243 Bacillus diphtheriae. Immunological types; toxin-antitoxin relationship.. 69 Blood, guinea-pig, On the origin and nature of alexin (complement) in.... 435 Bronchial asthma and allied conditions. Clinical and immunological obser- CUETO (ROS oie HEUER SiN AeneO BOT ES 0. NSS REIS 2 Oe Steen ayiene Pye clt 81 —— asthma, New etiologic factors In................ eee eee eee eee ees 147 Brown, Aaron. Studies in specific hypersensitiveness. I. The diagnostic cutaneous reaction in allergy. Comparison of the intradermal method (Cooke) and the scratch method (Schloss)..........-...--+++seeeeeees 97 523 524 INDEX Buckell, George T., and Torrey, John C. A serological study of the gono- COCCUSETOUDN vtec sci seein a we canes re eRe Oe CORO CO eT en ee ee 305 Coseulatiow toxicity oe... css ose ee aed os a ee 571 Coca, Arthur F. Studies in specific hypersensitiveness. V. The prepara- tion of fluid extracts and solutions for use in the diagnosis and treatment of the allergies with notes on the collection of pollens.................. 163 ——. Studies in specific hypersensitiveness. VII. The age incidence of serum disease and of dermatitis venenata as compared with that of the PIS UT LR OTIS: cece ei atesncs ale ne ao er oe POT aT ee Diet eee 193 ——, Deibert, Olin, and Menger, Edward F. Studies in specific hyper- sensitiveness. VIII. On the relative susceptibility of the American Indian race and the white race to the allergies and to serum disease.... 201 Complement fixation reactions, precipitin and, A study of the, with tuber- culous exudates with special reference to tuberculous pleuritis........ 423 —— ,serum, On ue, pHOLOlaDMIGY. Ole. t.o5. ccs oe le i ee ee 389 Constitutional reactions, On: The dangers of the diagnostic cutaneous test and therapeutic. mjection of allergens... .... 0. si... .cteue cs. s es Jee 119 Cooke, Robert A. Studies in specific hypersensitiveness. III. On con- stitutional reactions: The dangers of the diagnostic cutaneous test and therapeutic. myection, of allergens.\.0 0.02... ..\..-.. .<. Sone os eee 119 ——. Studies in specific hypersensitiveness. IV. New etiologic factors in bronchial “asthma............:. RAs i Mga 53 SPORES ae toe 3 147 ——. Studies in specific hypersensitiveness. IX. On the phenomenon of hyposensitization (the clinically lessened sensitiveness of allergy).... 219 Cromwell, H. W. A study of the, hemolytic antibody-antigen combination. 461 Cutaneous reaction in allergy, The diagnostic. Comparison of the intra- dermal method (Cooke) and the scratch method (Schloss).............. 97 —— test, The diagnostic, and therapeutic injection of allergens............ 119 Deibert, Olin, Coca, Arthur F., and Menger, Edward F. Studies in specific hypersensitiveness. VIII. On the relative susceptibility of the Amer- ican Indian race and the white race to the allergies and to serum disease. 201 HenmAtitis Venenabay : 4s. 6. oieeee 2 «tee eee PEERED PEAS eS Se 179 —— venenata, The age incidence of serum disease and of, as compared with _ that of the natural allergies: |< oy... 2 packeee On os cee ee eee 193 Diphtheria bacilli, Immunological studies on types of, I. Agglutination characteristics. II. Protective value of the standard monovalent BAG UOOS 5 sonia sac tocar ace not ieee ee Lt Co ee eee 243 Diphtheriae, Bacillus. Immunological types; toxin-antitoxin relationship 69 Etiologic factors, New, in bronchial asthma...................-¢.--eeeee- 147 Evans, Alice C. The toxicity of acids for leucocytes, as indicated by the PLOPUT TERETION. oe 5.03. Le ee eee eee Ce eee ee 271 Exudates, tuberculous, A study of the precipitin and complement fixation reactions with, with special reference to tuberculous pleuritis.......... 423 Fetal and maternal plasma, Relative toxicity of...........-..----++++-+++ 497 Fleisher, Moyer S. Relationship of various antiorgan sera..............-. 51 ee ee eee eee — —- ——_— INDEX 525 Fluid extracts and solutions, The preparation of, for use in the diagnosis and treatment of the allergies with notes on the collection of pollens.... 163 Glycerinated virus, Prophylactic treatment for rabies by means of stand- LTO Cs ee ee as cokes een ee Pees Sya sishs Myerson Beng 409 Goldman, Agnes. Studies on acute respiratory infections. XI. A sero- logical study of alpha streptococci from the upper respiratory tract.... 361 Gonococcus group, A serological study of the....................e see eee 305 Guinea-pig blood, On the origin and nature of alexin (complement) in...... 435 Hay fever, A comparison of various pollen extracts with reference to the question of their therapeutic value in.....................2eeeeueeees 113 Heist, George D., Solis-Cohen, Solomon, and Solis-Cohen, Myer. A study of the virulence of meningococci for man and of human susceptibility to PAE INP OCU CIC IHLECULOD eat clei i- ie mtiar tee stevens = ante cee ate oiel svete 1 Hemolysin, The action of various metallic salts on...............-.-2+---- 35 Hemolytic antibody-antigen combination, A study of the.................. 461 Hodge, W. Ray, and MacLennan, M. F. The relationship of lipoids and proteins to serum reactions in tuberculosis.............-...--++++++++- 253 Horn, uterine, tuberculous, An allergic reaction of the.................... 47 Human susceptibility, A study of the virulence of meningococci for man and GE TO MENT LOCOCCLG TIME CUION «2.15.2 eyeuel sates ees a> pe «fee = eases = te 1 Hypersensitiveness, Studies in specific, 97, 113, 119, 147, 163, 179, 193, 201, 219 Hyposensitization, On she phenomenon of, (the clinically lessened sensitive- ETUC sly a) ee GES BOO OR eet COURS Ne AIC Re eM ne Oo cae cor 219 Immunological types, toxin-antitoxin relationship...................+..-- 69 Indian race, American, On the relative susceptibility of the, and the white race to the allergies and to serum disease....................+22-000- 201 Infection, meningococcic, A study of the virulence of meningococci for man angio mau suscep LibilityatO! .jhoebes Seige 3S ola che ery sical vtreie de ctalone 1 Infections, acute respiratory, Studies on...............---2s:e seers e eee 361 Larsen, Nils P., Paddock, Royce, and Alexander, H. L. Bronchial asthma and allied conditions. Clinical and immunological observations...... 81 Leucocytes, The toxicity of acids for, as indicated by the tropic reaction.... 271 Levinson, S. A. Studies on the toxicity of human blood plasma for guinea- [U1 area OBS anya ee Gre SORE CORE Gece acon ioe nT a Tur Sema tpeeC BOAT 497, 511 Lipoids, The relationship of, and proteins to serum reactions in tuberculosis. 253 Lundberg, E.G. On the photolability of serum complement.............. 389 MacLennan, M. F., and Hodge, W. Ray. The relationship of lipoids and proteins to serum reactions in tuberculosis...........-.-.-+++++e+e+ees 253 Man, A study of the virulence of meningococci for, and of human suscepti- bility to meningococcic: nfecbion.. 2.5... a.m wane ao 8 i esis ee 1 Mann, Alice G., Park, William H., and Williams, Anna W. Immunological studies on types of diphtheria bacuty, I. Agglutination characteristics. TI. Protective value of the standard monovalent antitoxin............ 243 526 INDEX Maternal plasma, fetal and, Relative toxicity of.......................4.5 497 Meningococci, A study of the virulence of, for man and of human suscepti- bility taimeningocoecic infection. . 2.62.2 +5,1.0.. +. oe eee eee 1 Meningococcie infection, A study of the virulence of meningococci for man and ofhumanisusceptibility.tov.> <-oe -ice eee eee eee eee 1 Menger, Edward F., Coca, Arthur F., and Deibert, Olin. Studies in specifie hypersensitiveness. VIII. On the relative susceptibility of the Amer- ican Indian race and the white race to the allergies and to serum disease. 201 Metallic salts, The action of various, on hemolysis......................-: 35 Morrison, L.F. On the origin and nature of alexin (complement) in guinea- pie blood ee a See ee or Re, SE en ene eee Ogawa, Isamu. A study of the precipitin and complement fixation reactions with tuberculous exudates with special reference to tuberculous pleuritis. 423 Paddock, Royce, Larsen, Nils P., and Alexander, H. L. Bronchial asthma and allied conditions. Clinical and immunological observations...... 81 Park, William H., Williams, Anna W., and Mann, Alice G. Immunological studies on types of diphtheria bacilli. I. Agglutination characteristics. II. Protective value of the standard monovalent antitoxin............ 243 Paxson, W. H., and Redowitz, Edward. Bacillus diphtheriae. Immuno- logical types; toxin-antitoxin relationship............5..0....0.2202-. 69 Phillips, James McIlvaine. Prophylactic treatment for rabies by means of standardized glycerinated virus...... SEO IRL 5. Be eee 409 Photolability of serum complement, On the.............2-...20202.00008 389 Plasma, fetal and maternal, Relative toxicity of.........................- 497 Pleuritis, tuberculous, A study of the precipitin and complement fixation reactions with tuberculous exudates with special reference to.......... 423 Pollen extracts, A comparison of various, with reference to the question of their therapeutic value in hay fever...... PN ce Cicer pease oe 113 Pollens, The preparation of fluid extracts and solutions for use in the diag- nosis and treatment of the allergies with notes on the collection of...... 163 Precipitin and complement fixation reactions, A study of the, with tuber- culous exudates with special reference to tuberculous pleuritis.......... 423 Prophylactic treatment for rabies by means of standardized glycerinated 409 Proteins, The relationship of lipoids and to serum reactions in tuberculosis... 253 Purdy, Helen A., and Walbum, L. E. The action of various metallic salts OHA MEMOLYSIS «22.246 cSe ol be woe eee eee ee ee a ee ee 35 Rabies, Prophylactic treatment for, by means of standardized glycerinated VITUS Hoe ce: Sole os bind Semen © Sic ees | ae ee a eee 409 Reaction, allergic, An, of the tuberculous uterine horn..................-- 47 Reactions, precipitin and complement fixation, A study of the, with tuber- culous exudates with special reference to tuberculous pleuritis.......... 423 Redowitz, Edward, and Paxson, W. H. Bacillus diphtheriae. Immunolog- ical types, toxin-antitoxny rélationship:. : ./2. ves. 2.2 eee ee 69 361 Respiratory infections» acuve, Studies ons. 22.20 ee eee ee coe INDEX 527 Respiratory tract, upper, A serological study of alpha streptococci from the. 361 Salts, metallic, The action of various, on hemolysis....................... 35 SOTAs ANCIOFZAN MEIAUIONSHED OF VATIONS ..0 aoc ¢ 2.0 fours cies vialt\s cv smarter 51 Serological study, A, of alpha streptococci from the upper respiratory tract. 361 Hid ys AS OL THE PONOCOECUS! PLOUP. . 2 .<.\0 oa caisdiice a neselelhe ices ee oes 305 Serum complement, On the photolability of...............5..2.0eeeceeees 389 reactions, The relationship of lipoids and proteins to, in tuberculosis. . 253 —— disease, On the relative susceptibility of the American Indian race and thenwhicemacewotne-allercieshan dion wass cee sia. veces ceri eins sere 201 —— disease, The age incidence of, and of dermatitis venenata as compared With atOmbthemathunalallercies); 5611 sake since. econ seen 193 Smith, G. H. An allergic reaction of the tuberculous uterine horn........ 47 Solis-Cohen, Myer, Heist, George D., and Solis-Cohen, Solomon. A study of the virulence of meningococci for man and of human susceptibility OMMEMINPOCOCCICMMPEC LION Ag. ey. Non aca cake von Sisie cieps eS sieve: «ae Maeotes 1 Solomon, Heist, George D., and Solis-Cohen, Myer. Astudy of the virulence of meningococci for man and of human susceptibility to meningococcic THEW HSVCLETKOY IVD pate eos od ce ce ORR ae RET oh PEI ot OLS LT mR ct CRA REE” ee CER AC toeee crt IAA 1 Spain, W. C. Studies in specific hypersensitiveness. VI. Dermatitis RASLIVETO EW Ng 8 i Sra ae See Rae ERE ania Dae ct ene fre Re re Oi micas coy OPA CPsic 179 Streptococci, alpha, A serological study of, from the upper respiratory tract. 361 Susceptibility, human, A study of the virulence of meningococci for man and Oistowneningococeicnmreetion so. 4:0 45-ccsiedad desis «+s see oe = eee 1 Torrey, John C., and Buckell, George T. A serological study of the gono- COCCUSOTOUP ert foray ct LTR MA eee nee ke ios fede ate Cee chy arena 305 BORA CTU AAO O AOL GI OT rast rozyrsve ate teh es orewe rs. yes Mae tee eee er toner a eeraeoy ht 511 ——ihelative, of fetal and maternal plasma. .:..0 225: 0..5225525 46: eee 497 , The, of acids for leucocytes, as indicated by the tropin reaction...... 271 Toxin-antitoxin relationship; Immunological types.............-..--+++++- 69 Tropin reaction, The toxicity of acids for leucocytes, as indicated by the... 271 Tuberculosis, The relationship of lipoids and proteins toserum reactions in. 253 Tuberculous exudates, A study of the precipitin and complement fixation reactions with, with special reference to tuberculous pleuritis.......... 423 —— pleuritis, A study of the precipitin and complement fixation reactions with tuberculous exudates, with special reference to............-...++-- 423 — uterine horn, An allergic reaction of the... 20.2... 2 00ce/ccn0ee concn 47 Uterine horn, tuberculous, An allergic reaction of the...................--- 47 Vander Veer, Albert, Jr. Studies in specific hypersensitiveness. II. A comparison of various pollen extracts with reference to the question of gheir therapeutic vyaluevin hay fever... 4.0: secs s- aoe: 0 nae ie oe 113 Virulence, A study of the, of meningococci for man and of human suscepti- Pali. to meningococcic, infection: «j226. 6 .4-0es605- + sani. 2 eer ee it Virus, standardized glycerinated, Prophylactic treatment for rabies by HGGTIS (Ceres Geena aon eR ten Metre clo De ete 5 ata Iona. © aT Ot 409 eis. Walbum, L. E., and Purdy, Helen A. The action of io sous ae studies on types of diphtheria bacilli. II. Protective value of the standard monovalent pa eee a Kort : a a, ie ae y it es “ ; Sey es i eee ": CR The Journal of immnolozy 180 J6 vV.7 ce P i 2. : Biological & Medical Serials PLEASE DO NOT REMOVE CARDS OR SLIPS FROM THIS POCKET UNIVERSITY OF TORONTO LIBRARY on Ptah Pear ee ee at Tektite he pre cae rent AAS ae S| 1 4 7 m4 ate) grr RT