1 1 j 1 3 > i 1 1 1 IH Received Marine Biological Laboratory !:ov. 14, 1941 Accession No. Dr. Gregory'' Shwartznian ^'"^^ ^^ I't, .3iiial IIobplLal, New York ClL y Place, PHENOMENON OF LOCAL TISSUE REACTIVITY and its IMMUNOLOGICAL, PATHOLOGICAL AND CLINICAL SIGNIFICANCE I'm NoMi N().\ oi l.o( \i Skin Ri \(ii\ii'i io I> aci i kiai liLTRAirs. len preparatory units intraderinally and 25 reacting units intravenously of meningococcus filtrates. Appearance of reaction four hours after the intra- venous injection. Pliotograph of li\ing rabbit by Mr. Harold Fowler, New York. PHENOMENON OF LOCAL TISSUE REACTIVITY AJ^D ITS IMMUNOLOGICAL, PATHO- LOGICAL AND CLINICAL SIGNIFICANCE by GREGORY SHWARTZMAN, M.D. Bacteriologist, Mount Sinai Hospital, New York Foreword by JULES BORDET, M.D. Paris WITH 67 ILLUSTRATIONS AND ONE COLOR PLATE •1> PAUL B. HOEBER, I^ MEDICAL BOOK DEPARTMENT OF HARPER & BROTHERS NEW YORK MCMXXXVn COPYRIGHT, 1937, BY PAUL B. HOEBER, INC. ALL RIGHTS RESERVED. THIS BOOK OR ANY PART THEREOF MAY NOT BE REPRODUCED IN ANY FORM WITHOUT THE PERMISSION OF THE PUBLISHERS. PUBLISHED OCTOBER, 1937. PRINTED IN THE UNITED STATES OF AMERICA. PREFACE SINCE the classical and tundaiiiental ^voik ot Roiix, Versin, Behring, Kitasato, Bordet, P^liiiich, Theobald Smith, Ro- senaii, and others, it has been generally acknowledged that the discovery ot antigenic and solidjle exotoxins ^vas one ot the most triiittul and imjiortant contriijutions to innnnnology and medicine. The ideas stimidated jjy this ^vork brought about eluci- dation ot inninnerable problems in j^athology and clinical medi- cine and ga\e important residts in specitic treatment and prexen- tion ot certain intectious and contagious diseases. Ever since 1892, numerous in\estigators attempted to develop antitoxic sera tor almost exery known bacterial infectious disease. It xvas, hoxvever, soon learned that the possibility of producing antitoxins was limited only to a small gTOup of diseases because the method of demonstration of toxins failed xvith a large percentage of micro- organisms tested. The investigations xvere then side-tracked by the theories of endotoxins. These theories apparently handicapped the original and finidamental progress for many years to come. Ever since the demonstration ot the phenomenon of bacteriolysis by PfeifTer and his formulation of the endotoxin theory, investi- gators used amazing ingenuity in the development of methods for cell destruction. However, the split products and the laboratory artefacts thus obtained coidd not be reasonably considered as responsible agents in e\olution of infectious diseases because of complete lack of specific pathological effect upon experimental animals. The opposing viexvs have been steadily gaining groinid that some unknoxvn agents as yet may be responsible for systemic intoxications in intectious diseases. For some years I employed various experiments in order to determine whether meningococcus, B. tyjjhosus and other micro- organisms produced sid)stances similar to soluble exotoxins. As a working hypothesis, I assumed that the microorganisms in \vhich the demonstration of exotoxins Avas hitherto impossible may pro- duce injuries in animal tissues through a mechanism altogether different from the direct injurious effect exerted by diphtheria, botulinus, tetanus, and other true exotoxins. \Vith this idea in VIU PREFACE luiiid, \aii()us exjx'i iniciital altcinj)ts were made which hardly need any detailed description in this monograph. They were in- tended to ()\'erc()me a hypothetical natural state ot resistance as- sumed to exist in the annual cell. In some experiments, rabbits received intraocular injections of B. typhosus culture filtrates with subsequent intracranial injections of the same filtrates; in others, leucocytes cidttned in vitro, according to Carrel's tech- nique, \vere repeatedly exposed to bacterial culttne filtrates at various intervals of time. The residts were suggestive Ijut rather too indehnite for publication. Following these preliminary experi- ments, the phenomenon of local skin reactivity to bacterial filtrates was oliserved by me in September, 1927 (1928 a, b) . Re- capitidation of extensive experimental work on the phenomenon thus far accimiulated; together with a review of certain previously knoAvn facts directly belonging to related fields, are embodied in this monograph. The scope of this book, intended for presenta- tion of original observations directly pertaining to the phenom- enon of local skin reactivity to bacterial filtrates made by various in\estigators and myself, does not allow a thorough sinvey of the related literattue. Attempts w^ere made to divide the material into chapters each consistently dealing ^vith a separate topic. Repetitions Avere tui- a\oidable because of close inter-relationships between certain ob- servations and because of their \arious implications. The monograph includes data published by various investiga- tors and myself uj) to the beginning of 1936 and also personal unpublished obser\'ations made by me dining the past se\'en years up to date. Inasmuch as the ^vork thus far done has demonstrated certain definite relations of this phenomenon to various immunological processes; to pathology in general; and to problems of biological therapy, the presentation of the monograph to the medical pro- fession seems to be justified.^ ^ Small print is used for technical descriptions, protocols and special facts. Readers desiring to obtain only general information concerning tlic phenomenon may omit the reading of the small print. Gregor-^' Shwartzman Nexu York, N. Y. August, 1937. ACKNOWLEDGMENTS It gives me great pleasure to ackno\vledge my indebtedness and gratitude: To Mr. George Bliuiienthal, President of the Board of Trustees of The Motuit Sinai Hospital, Mr. George Bernheim, Chairman of the Laboratory Committee, their Colleagues of the Board, and Dr. Josepli Turner, Director of this Hospital, whose constant interest and fidl support made this Avork possible. To Mrs. Kiut Simon, ^vhose help allowed the preparation of antitoxic sera described in this monograph. To Mr. Eugene Meyer (Manhattan Research Foundation) , Alt- man Foundation, Friedsam Foinidation, Mr. Louis Strasser, and the Committee on Scientific Research of the American Medical Association, Dr. Ludvig Hektoen, Chairman, for their generous grants during the past years. To the Josiah Macy, Jr. Foundation, Dr. Ludwig Kast, Presi- dent, AV'hose generous grants and continued interest permitted the preparation of this monograph and the work on the pathogenesis of the phenomenon under progress. To Mr. Eli Lilly for the support of the work on the chemical aspects of the phenomenon under progress and Dr. G. H. A. Clowes, Mr. W. A. Jamieson, and Dr. H. M. Po^vell of the Eli Lilly Research Laboratories for help and interest. To the Blanche and Frank Wolf Foinidation for the support of the ^vork mider progress on the relation of the phenomenon to neo- plastic growth. To Dr. Alexis Carrel, Dr. f^manuel Libman, Members of the clinical and laboratory staff of The Mount Sinai Hospital, in par- ticidar, to Dr. George Baehr, Dr. A. A. Berg, Dr. L Friesner, Dr. Robert T. Frank, Dr. Louis Gross, Dr. Paul Klemperer, Dr. B. S. Oppenheimer, Dr. B. Schick and Dr. Harry Sobotka, for encoinage- ment, help and cooperation. To Dr. G. W. McCoy, Director of the National Institute of Health of Washington, D. C, Dr. Rigney D'Aunoy, professor of pathology and bacteriology, of the Louisiana Medical School, and W. Y. Hollingsworth, Smgeon, U. S. Public Health Service, whose X ACKNOWLEDGMENTS efforts gave the opportunity to test the typhoid antitoxic serum in the C-haritv Hospital and l^ S. Marine Hospital ol New Orleans, I, a. To my collaborators, Dr. I. E. Gerber, Dr. Joseph L. (ioldman, the late Dr. Herbert Klein, Dr. Le\v'is H. Koplik, Dr. Sam Morell and Dr. S. S. Schneierson tor understanding and valuable co- operation. To Miss Gladys Cameron, Miss Edith Mandel, Miss Esther Dob- kin, and Miss Mildred Shapiro lor accinate and capable laboratory assistance. To Miss Helen DiVeglio for very capable secretarial help in preparation of this monograph and index, and Miss Ethelbert McKennon for most valuable work on the bibliography. I also take pleasine in acknowledging cordial cooperation of the publisher, Mr. Paid Hoeber, and Miss F. Fuller of his staff. G. S. FOREWORD Le reniaiquable jjhenoniene doiit Tetiide a etc conduite par M. Shwaitzman avec tant d'ingeniosite et de perseverance a suscite, parmi les bacteriologistes et imnuinologistes. le plus vif interet. Sa decouverte hit, a juste titre, consideree d'emblee comme tui evene- luent scientitK|ue important. II laut dire qu'on ne s'y attendait guere, car s'il est tnie etude dont on a miniuieusement envisage tons les aspects et que Ton a consciencieusement louille, c'est bien celle des reactions de lOrganisnie vis-a-vis des produits microbiens. On n'eut pas devine quelle n'avait pas livre tons ses secrets et notis reser\ait encore des sinprises. II pent paraitie singulier que le phenomene auquel le nom de M. Sh^vartzman reste attache ne se soit pas revele plus tot, puisqtie, se denotant par des accidents de caractere hemorrhagique, il donne lieu a des manifestations particulierement visibles et frappantes, et puisque les conditions necessaires a sa production sont selon toute vraisemblance, assez souvent realisees soit experimentale- ment soit au cours des maladies spontanees. II suffit en effet, potn^ le faire apparaitre, que les produits elabores par certains microbes penetrant dans la circulation apres avoir agi stn^ une region deter- minee et notamment sm^ inr point du revetement cutane ou mu- queux. Et parmi les microbes capables de developper ces effets, il en est, tel le bacille coli, qui sont tres repandtis et auxquels notre organisme est fort habitue. Or, ce bacille coli precisement avait servi aux experiences deja assez anciennes de M. Sanarelli, dont les restdtats fort ciuieux purent s'interpreter beaucoup plus aisement a la lumiere des donnees que M. Sinvartzman ajjporta dans la suite. M. Sanarelli a\ ait montre qu'une injection intra\ einetise, bien toleree par Tani- mal temoin, de cidttue filtree de bacille coli, proxoque au niveau de I'intestin luie congestion hemorrhagique mortelle lorsqu'on opere stu" des lapins qui }3eu de temps aupara\ant ont re^u dans la veine une dose inframortelle de vibrion cholerique. Ce phe- nomene de Sanarelli se rattache au phenomene de Shwartzman par tme parente etroite qtie signalerent, en 1931, des recherches (Paid Bordet) consistant dans rinoculation des produits du bacille coli XU FOREWORD a (Ics cohaves dont Ics ^anf4lic)iis rciilci niaient des inicr()l)es vivants (bacilles tiil)erculcu\ attenues B.Cl.Ci.) , puis les experiences de Gratia et Linz qui operaient sur d'autres intections, notamment la vaccine. II est extremement probable que ces phenomenes inter- N'iennent dans le determinisme de certaines complications au cours des infections. Mais je ni'arrete, car il n'est pas oj>portini que je penetre dans I'intimite du probleme; en ettet, M. Shwartznian le traite de la ia^on le plus complete. Assemblant une imposante documenta- tion, il expose ses propres experiences et celles que de nombreux cherchems, seduits par la nouveaute et I'attrait du sujet, se sont empresses d'instituer. Malgre les efforts depenses, le probleme n'est pas encore completement elucide, la genese des symptomes ob- serves n'a pas perdu tout son mystere. On doit savoir gre a M. Sh^vartzman d'avoir decouvert le pbenomene decrit dans le present ouvrage, on doit lui savoir gre aussi d'avoir procede, pom- ce qui concerne cette importante question, a une mise au point tres precise et que lui seid pouvait effectuer. CrxdjU- TRANSLATION OF THE FOREWORD The remarkable plienonienon which Dr. ShAvart/nian has studied w itli such ingenuity and perserverance has aroused the most vivid interest among bacteriologists and immunologists. From the very beginning his discovery has been justly recognized as an important scientific development. The discovery of the phenomenon was totally unexpected, it nuist be acknowledged, for if there is any field which has been minutely investigated from every point of view and conscientiously searched it is that of the reactions of the organism to bacterial products. One could not have suspected that in this field secrets still remained unraveled and surprises were in store for us. It may seem strange that the phenomenon to which the name of Dr. Shwartzman remains attached was not discovered earlier, since denoting itself l^y accidents, hemorrhagic in nattire, it gives rise to striking and especially conspicuous manifestations, and since the conditions necessary for its reproduction are in alf probability fre- cpiently encountered experimentally and in the course of spon- taneous infections. In order to reproduce the phenomenon it is sufficient indeed that the products elaborated l)y certain micro- organisms penetrate into the circulation after having acted at a given site, notably skin or mucosa. Among the microorganisms capable of developing these effects there are those like B. coli, which are Avidespread and to Avhich the human organism is well adapted. B. coli ^vas employed some time ago in the experiments of Dr. Sanarelli ^vhose \ery curious results could be interpreted more easily in the light of the more recent findings of Dr. Sh-wartzman. Dr. Sanarelli had shown that an intravenous injection of B. coli culture filtrate, althotigh easily tolerated by the control animals, elicits a fatal hemorrhagic congestion in the intestinal tract of rabbits ^vhich had previously received a stiblethal dose of cholera vibrio. The phenomenon of Sanarelli is closely related to the ShAvartzman phenomenon, the fact pointed out in 1931 by investi- gations (Paul Bordet) consisting in inoculating guinea pigs, Avhose lymph nodes contained li\'ing microorganisms (B.C.G. attenuated XIV TRANSLATION OF THE FOREWORD l)aclllus liihcK iilosis) with products ol li. (oli, and later l)\ tlu" experiments ol Gratia and Lin/., who worked witli other infections, namely, vaccinia virus. It is hiojily jjrobable that these phenomena determine the pathogenesis ol ( ertain complications in the course ol infections. But I shall say no more, tor it is not litting ior me to penetrate into the intimacy of the prol)lem. In fact, Dr. Shwartzman treats it in a most complete fashion. In an impressive documentation he exposes his own experiments and those which nimierous other in- vestigators, attracted by the novelty and interest of the subject, have eagerly tnidertaken. In spite of the efforts spent, tlie problem has not yet been completely elucidated since the genesis of the symptoms observed is still some^vhat of a mystery. A debt of grati- tude is due to Dr. Shwartzman for having discovered the phenom- enon descril)ed in this book; a debt of gratitude is also due him for having proceeded Avith his researches on this imjDortant prob- lem with the great precision ^vhich he alone could achieve. cttcLl^- CONTENTS Preface "^ii Acknowledgments ix FoRE^voRD xi Translation of the Foreword xiii Llst of Illustrations xxv Chapter I INTRODUCTION Basic Experiments i Phenomenon of Local Skin Reactivity to Bacterial Filtrates; Early Observations on Routes of Injection; Early Observations on Specificity of the Phenomenon of Local Skin Reactivity to Bacterial Filtrates; Early Observations on Time Recjuired for Elicitation of State of Reacti\ity and on its Dination. Terminology 8 General Considerations Concerning the Phenomenon 9 The QuANTiTATnE Estimations of Potency of Active Prin- ciples 23 Method i: Reciprocal Titrations of Various Skin-Preparatory Doses against Various Reacting Doses; Method 11: Titrations of Skin-Preparatory Factors against Constant Intravenous Doses; Method in: Titration of Reacting Factors against Con- stant Skin-Preparatory Doses. Recapitulation 28 Chajjier II PHENOMENON OF LOCAL SKIN REACTIVITY TO VARIOUS MICROORGANISMS Active Principles of Meningococcus 31 Meningococcus "Agar Washings'" Culture Filtrates. Active Principles of Gonococcus 34 Active Principles of B. Typhosus 35 B. typhosus "Agar W^ashings" Filtrates. Active Principles of B. PARArvpHosus 37 XV 54121 XVI CONTENTS Active Principi.ks oi H. Coi.i 37 Active Princiimfs of B. Friedi.ander 38 Active Princii'ees of B. Dysenteric. 38 AcnvF Principles of B. Prodkwosus 39 Active Principles of B. Melitensis and B. Abortus 40 Active Principles of Various Members of Hemorrhagic Sep- ticemia Group 40 Active Principles of Various Members of Hemogi.obinophii.ic Group 42 Active Principles of Vibrio CHoi.ERyE 44 Active Principles of Staphylococcus 45 Active Principles of Streptococcus 45 Reactions with Live Cultures of Styel^tococcus Hcniolxiicus Scarlotinae: Reactions with Toluene-killed Cultures: Super- natant Fluid ot Centrituged Cultures and Filtrates ot Cidtures of SirclJlococcus Hemolyti( us S( (i)l(itin(ie; Reactions with Mis- cellaneotis Strains of Styel)t()C()C( us Hcmolyticus; Reactions with Indifferent and Green-Producing Streptococci; Reactions with Streptococcus S( holl )n iiller. Active Principles of Pneumcjcoccus 50 Active Principles of B. Tuberculosis 52 Reactions with Products of B. tulx'K ulosis in Combination with Heterologous Filtrates of Ascertained Phenomenon- producing Potency. Active Principles of Anaerobic Gram-Negative Bacilli Iso- lated from Chronic Lung Abscesses 56 Active Principles of Yeasts 57 Active Principles of Ascaris Lumbricoidis 57 Active Principles of Ricin 58 Active Principles of Diphtheria Toxin 58 Recapitulation 60 Chapter III PHYSICO CHEMICAL PROPERTIES OF THE ACTIVE PRINCIPLES OF THE PHENOMENON OF LOCAL SKIN REACTIVITY The Effec:t of Heat Upon the Active Principles 61 The Effect of Different Hydrogen-ion Concentrations Upon Active Principles 63 CONTENTS XVll Effect of Ultraviolet Radlvtion Upon Active Principles .... 65 Effect of Dissolution and Arioi.vsis Upon the Active Prin- ciples 65 Meningococcus Preparations; B. Typhosus Preparations. Fluctuations in Potency of Active Principles 70 Filtration and Dialysis of Active Principles 71 Formalin Detoxification of Active Principles 75 Use of Methods of Ecker and Rimington for Concentration and Purification of the Active Principles 75 Ammonium Sulphate Precipitation of B. Typhosiis Active Principles 76 Purified and Preserved Streptococcal Reacting Factors .... 78 Purification of B. Coi i Preparations 79 Chemical Studies on Active Principles of Vibrio Choler.e .... 81 Chemical Studies on Active Principles of B. Pestis 82 Nucleoprotein of B. pestis; Capsular Material of B. pestis. Chemical Studies on Active Principles of B. Dysenterle 83 Attempts at Fractional Precipitation of Potent Bacterial Filtrates 84 Recapitulation 87 Chapter IV IMMUNOLOGICAL BEHAMOR OF ACTR E PRINCIPLES OF THE PHENOMENON OF LOCAL SKIN REACTIVITY Effect of Immune and Normal Sera Upon Skin-Prepar.\tc^ry Factors 8g Effect of Normal and Immune Sera Upon Reacting Factors 93 Effect of Bacterial Variation Upon the Specificity of Active Principles 101 Description of Strains and \'ariants; Serum Neutralizability of Reacting Factors of Variants of B. typhosus and B. coli: Morphological Appearance of Colonies and the Neutraliza- bility of Reacting Factors; Serimi Agglutinabilitv of Various Stock Strains and \'ariants of B. typhosus and the Neutraliza- bility of Reacting Factors; Serum Precipitability of Culture Filtrates of Various Stock Strains and Variants of B. typhosus and Neutralizability of Reacting Factors; Virulence of B. typhosus \'ariants and Neutralizability of B. typhosus React- ing Factors. XVIU CONTENTS CONN'ERSION OF ACTIVE PRINCIPLES i O ToXOID Ill Reacung Potency of PREc:iiMrArE.s anb Supernatani Ki rins Derived from Mixtures of Ac:hve Bacterial Filiraies wiiii Immune Sera 113 Neutralization of Active Principles in "Multiple Propor- tions" 117 An Antibody Auxiliary 10 Serum Neuiralization of Menin(;o- coccus Reacting Factors 1 20 Reactivating Effect of Blood Sera Upon Completely Neu- tralized Toxic Filirates 123 Recapitulation 127 (]]}(( l)ier V IMMUNITY TO THE PHENOMENON OF LOCAL SKIN REACTIVITY Natural Immunity of Rabbits 130 Spontaneous and Acquired Active Immunity as Determined by THE Use of Graded Amounts of Active Principles 133 Spontaneous Inmiunity; Active Acquired Immunity. Phenomenon of Local Skin Reacitvity to Bacterial Fil- trates IN Animals Other Than the Rabbit 138 Passive Immunity to Active Principles 140 Recaphulahon 144 dial) fey 17 ROLE OF INFLAMMATION IN THE PHENOMENON OF LOCAL SKIN REACTIVITY— OPTIMUM CONDITIONS FOR REPRODUCTION OF THE PHENOMENON Introduction 146 Effects of Intradermal Injection of Bacterial Produc;is in Their Relation to the Phenomenon 147 Skin-preparatory Effect of Non-Bacterial Substances 151 Skin-preparatory Potency of Sterile Tryptic Digest Broth; Skin- preparatory Potency of Turpentine; Skin-preparatory Potency of Gum Tragacanth, Crystalline Egg Albumin, Blood Sera, Sodium Arsenate, Spirits of Nitroglycerine, India Ink; Skin- preparatory Potency of Mixtures of Animal Protein Antigens CONTENTS XIX with Antibodies; Unpublished Observations on the Skin- preparatory Effect of Various Non-bacterial Substances: Spleen, liver, adrenals, kidney and brain extracts; sera treated with agar; calcium gluconate; xylol; eserine salicylate; cantharidin plaster; hypertonic NaCl solution; silver nitrate; histamine; acetylcholine; allylamine; Croton oil; urethane ethyl; calcium chloride; antispleen sera; Sid^stances Studied by Others. Effect of Non-bacterial Substances Upon the State of Re- activity TO Bacterial Filtrates 160 Ascorbic Acid; Alypin; Cocaine; Acetylcholine; Pvridine; Anti- platelet Serum; Prolan; General Anesthesia; India Ink. Studies on the Behavior of Capillaries in Sites Prepared WITH Active Principles 162 Morphological Studies on the Phenomenon 164 Route of Injection of Reacting Factors and iNCUBArioN Period Necessary for Elicitaiion of the State of Reac:- TiviTY 175 Fate of Active Principles Injected Into the Blood Stream . . 182 Recapiiulation 184 Chapter VII METHODS OF ELICITATION OF THE PHENOMENON OF LOCAL TISSUE AND ORGAN REACTIVITY Perivascular Reactivity 186 Local Vascular Reactivity 193 Elicitation of Renal Vascular Reactivity; Elicitation of Vas- cular Reactivity in the Rabbit's Ear: Preparatory intravenous injections of active principles accompanied by chilling; pre- paratory intravenous injections of active principles in xylol- treated ears; preparatory intravenous injections of active jMinciples combined with ethyl urethane, acetylcholine, pilo- carpine hydrochloride, atropine, calcium gluconate and guinea pig liver extract; preparatory intravenous injections of active principles in combination with histamine; preparatory in- travenous injections of active principles combined with adrenalin; preparatory intravenous injections of active princi- ples in combination with pituitrin; preparatory intravenous injections of active principles accompanied by application of heat; preparatory intravenous injections of active principles combined with tcsticidar extract. XX CONTENTS Elic:itaii()n of Reachvhy by Way oi Genkrai. Circuiation 200 EHett ol Single lmra\cii()iis Injedioiis ol Badcrial Filtrates; Ellecl ol Repeated liilraxeiious liijec lions ol Bacterial Filtrates. Tissue REAcrivrtY 10 Live BActERiA and Viruses Induced in Internal Or(;ans by Intrava.scui.ar Injeciton or Live Bac- lERiA 210 Course of Induced Infecfions in Aniaiafs Previously Pre- pared WITH Active Principles 212 The Effect of Active Principles Upon Infecied Skin Sites .... 213 REClAPiniLAIION 2ig Cliapler VIII REACTIVITY OF MALIGNANT NEOPLASMS TO ACTIVE PRINCIPLES OF THE PHENOMENON OF LOCAL SKIN REACTIVITY Introduction 222 Effect of Active Principles of the Phenomenon Upon Tumors 225 Positively Reacting Tmnors; Negatively Reacting Tumors. Effect of Spontaneous Infections on the Development of Mouse Sarcoma 180 242 Effect of Induced Infechions on Development of Mouse Sar- coma 180 245 Relation of Mortality to Hemorrhagic Necrosis and Regres- sion of Tumors Elicited by Active Principles of the Phe- nomenon 249 Recapitulation 256 Chapley IX REACTING POTENCY OF NON BACTERIAL SUBSTANCES Reacting Potency of Mixtures of Animal Proteins avifh Homologous Antisera 258 Basic Experiments. Reacting Potency of Complexes of Rabbit Seriim vvirn Anti- Rabbit Horse and Guinea-Pig Sera 259 Reacting Potency of Agglutinated Red Blood Cells 260 CONTENTS XXI riiRAiioN oi Reacting Poiency of Precipitates 260 Reacting Potency of Precipitates and Supernatant Fluids Derived from Completely and Partially Precipitated Mix- tures 2()1 Zone Reactions in Antigen-|-Antibody Complexes in Relaiion TO Their Reacting Potency 262 Combinations of Himian Sennn \viih Horse Ami Human Serum: Mixtures ol Horse Serum with Rabbit Anti-Horse Serum. Origin of Reacting Potency of Serum pRECiPirATES 263 Reacting Potency of Purified Horse Sera and Antisera .... 264 Reacting Potency of Mixtures of Inactive Bacterial Prepa- rations WITH Specific Immune Sera 264 Reacting Potency of Mixtures of Detoxified Bacterial Preparations with Specific Immune Sera 265 Reacting Potency of Miscellaneous Antigens with Homol- ogous Antisera 266 Skin-preparatory Pofency of Precipitates and Supernatant Fluids Derived from Antigen-|-Antibody Mixtures 266 Re.\cting Potency of Non-bacterial Substances 267 In Vivo Formation of Reacting Factors in Rabbits Actively Sensitized to Proteins 270 Formation of Reacting Factors in Rabbits Passively Sensi- tized TO Animal Proteins 276 A New Method for Demonstration of Antigen-^Antibody In- teraction 278 Effect of Reacting Faciors Formed In Vivo Upon Skin Sites Prepared by Non-bacterial Factors 279 Effect of Intradermal Reinjection of Antigen-)-Antibody Mixture Into Sites Prepared with Bacterial Filtrates .... 279 Formation of Reacting Factors in Rabbits Actively Sensi- tized TO Bacterial Substances 280 Formation of Reacting Factors in Rabbits Passively Sensi- tized TO Bacterial Filtrates 281 B. luhcvi ulo.sis: Pneinnococcus Tyj)e in. Formation of Reacting Factors Through Interaction of Antigen or Antibody Injected Into the Prepared Site with Antibody or Antigen Injected Intravenously 281 Effect of Bacterial Filtrates Injected Intravasculari.y Upon Reactions to ANTiGEN-f Antibody Complexes 283 XXll CONTENTS Ei I I'd oi In iR.WKN'oi s iNjKcrioN OK An ri(;KN-|-AN IIIU)1)^ Cioxi- PLF.XKS Upon Skin Sn i.s I isikh vvnii Horsk Sk.rum in Rahui is Sensitized to Horse Serum 2H8 Effect OF Bacterial Fieiraiks Injecied In iravasculari.^ LJi'on Skin Sites Tesied wmi Horse Serim in RAniiirs Senshi/ed WITH Horse Serum 289 Effect of In Vivo Interachon of Anikjens wifh Antibodies Upon Skin Sites Infected with B. C. G 289 Recapitui.ation 290 Chaj)ley X NATURE OF THE ACTIVE PRINCIPLES OF THE PHENOMENON OF LOCAL TISSUE REACTIVITY Summary of Methods Used in the Production of the Phenom- enon 294 Relation of Active Principles of the Phenomenon to Exo- toxins 294 Relation of the Active Principles of the Phenomenon to Endotoxins ^^10 Relation of the Phenomenon-producing Potency to Lethal Effect of Culture Filtrates 311 Relation of the Active Principles of the Phenomenon to Anaphylatoxins 319 Relation of the Active Principles of the Phenomenon to FORSSMAN ANTIGEN-)-ANril50DY COMBINATION AND NoRMAL Toxicirv OF Serum 324 Recapitulation 327 Chapter XI IMMUNOLOGICAL RELATIONSHIPS OF THE PHENOMENON OF LOCAL TISSUE REACTIVITY The Relation of the Phenomenon to Anaphylaxis 329 The Relation of Toxin Hvpersusceptibility to the Phenom- enon 347 The Relation of Spontaneous Bacterial Hypersensitiveness TO THE Phenomenon 348 Considerafions on the Mec:hanism and Nafure of the Phe- NC^MENON 34() Recapitulaiion 352 CONTENTS XXIU Chajjlcr XII IMMUNOLOGICAL, PATHOLOGICAL, AND CLINICAL SIGNIFICANCE OF THE PHENOMENON OF LOCAL TISSUE REACTR ITY The Relation of the Phenomenon to Bacterial Allergy .... 354 Focal Reactions of Bacterial Allergy in Relation to the Phe- nomenon; Local Reactions of Bacterial Allergy in Relation to the Phenomenon. The Possible Relation of the Active Principles of the Phe- nomenon TO the Arthus Phenomenon in Man 366 Toxic Manifestations in the Course of Typhoid Fever in Their Relation to the Phenomenon 360 Neutralizing Potency of Sera from Meningococcus Infec- tions 371 Neutralizing Potency of Sera from Gonococcus Infections 372 Neutralizing Potency of Sera from Typhoid Infections 375 Relation of the Phenomenon to Hemorrhagic Diathesis .... 376 Effect of the Phenomenon Upon Evolution of Local Infec- tions 387 The Relation of the Phenomenon to Mixed and Secondary Infections 388 Recapitulation 390 Chapter XIII APPLICATION OF THE PHENOMENON OF LOCAL SKIN REACTIVITY TO TREATMENT OF CERTAIN HUMAN CONDITIONS Introduction 393 Studies on the Phenomenon-Neutralizing Antibodies of Anti- Meningococcus Immune Sera and Their Use in Cases of Meningococcus Meningitis 395 Studies on Phenomenon-Neutralizing Anti-Typhoid Horse Sera and Their Use in Typhoid Fever Patients 405 Previous Observations on Therapeutic Effect of Anti-typhoid Sera; Prejiaration of Phenomenon-neutralizing Anti-t\ j)hoid Horse Sera; Clinical Observations with Phenomenon-neiuraliz- ing Anti-typhoid Horse Serum: Effect upon toxemia; effect upon duration of disease; temperature effect; serum sickness; XXIV CONTENTS cllcc I ol SCI iim oil aL;,L;liil iiiiiis lor I>. Ixjtho.sii.s: cllc(l on l);u- ic'ic'iiiia: niorialiiN : (oiiiiols. SrUDlKS ON PUKNOMI N()N-NKlirRAI.I/.lN(; AnTICOLI SeRA AND Their Use in C^Iases oe Non-specieic Ulcerative Coi.ins .... 418 Clinical Ob.servations wiili IMunomcnon-neutrali/int; anii-coii Sera: Material; outcome. Moccasin Venom Immunization in Ceriain Forms oe HEM(m- RHAGic: Diathesis and Recurreni Epistaxis }2i Recapitulation 424 Bibliography 426 Index oe Personal Names 441 Index of Subjects 447 LIST OF ILLUSTRATIONS FIGURE PAGE Phenomenon of local skin reacti\ity to bacterial filtrates (Color) F)())iti.sj)iece 1, 2. Apjjearance of skin reaction five hours after intravenous injection of B. typJiosus filtrate, at skin sites previoush in- jected with same filtrate 3, 4 3. Section of skin showing microscopic appearance of hemor- rhagic reaction five hours after intravenous injection of B. typhosus filtrate 5 4. Section of skin showing microscopic appearance of reaction produced by repeated skin injections of B. typhosus filtrate 6 5. I, Section of hypogastric lymph node six hours after in- travenous injection filtrate from culture of B. typhosus fol- lowing preparatory injection of same filtrate into afferent lymphatic. 2, Section ot arteriole in hypogastric node, show- ing swelling and hyaline degeneration of wall and necrosis of endothelial cells, j. Sections of pectoral nodes lollo\\-ing intravenotis injection of filtrate from culture of B. typhosus twent)-four hoius after jjreparator) injection of same fil- trate into skin of abdomen. _/, Section of vein of common iliac node which received intra\enous injection of filtrate from culture of B. typhosus twenty-four hours after prepar- atory injection of filtrate into the skin of medial surface of thigh. 5, Section of regional inguinal lymph node removed twenty-four hours after intradermal injection of filtrate of culture of B. typhosus into abdomen 17 6. Skin of back of rabbit following intradermal injections of meningococcus active principles and, twenty-four hours later, intravenous injection of same principles 18 7. Phenomenon of local skin reactivity to meningococcus fil- trate in rabbit ear 20 8. Section from ear following same treatment as in Figure 7 21 9. Section from area 3 cm. away from site of reaction 21 10. Neutralization of meningococcus active principles by anti- meningococcus neutralizing serum 99 XXV XXVI LIST OF ILLUSTRATIONS 11. Section ol ^kiii iiiulcr low and liii^li p()vver magnification showing fully developed phenomenon three and one-halt houis after intravenous injection of reacting factors iGS 12. A, Settic:)n of skin site after prej^aration with B. lypJiosus filtrate almost completely neutralized by homologous anti- toxic horse serum, followed by intravenous injection of same filtrate. B, Section of skin site after preparation with non-neutralized B. iyphosiis filtrate followed by intravenous injection of same filtrate 1 70 13. A, Section of skin site after intradermal injection of typhoid bacillus filtrate. B, Histologic appearance of typical phe- nomenon after skin preparation with typhoid bacillus filtrate followed by intravenous injection of same filtrate 171 14. Low and high power views of large hepatic vein with many parietal thrombi, high power view showing structure of thrombus 201 15. Photomicrographs of parietal venous thrombi in lung and spleen, and filling siurounding splenic sinuses 203 16. Gross appearance of kidneys, showing focal and diffuse corti- cal necrosis 206 17. Area of focal tiiliular necrosis, with capillary thrombosis of adjacent glomeruli and hemorrhagic necrosis of renal cortex 208 18. Sarcoma 180 in control series of mice 226 19. Histologic appearance of normal twelve day old sarcoma 180 227 20. Ttimor reactions in mice bearing twelve day old sarcoma 180 which had received B. lypJiosus "agar washings" filtrate intravenously 229 21. Histologic appearance of sarcoma 180 in mouse which received B. typhosus "agar washings" filtrate intravenously, four and twenty-four hours after injection 231 22. Histologic appearance of sarcoma 180 in mouse which received B. typJiosiis "agar washings" filtrate intravenously, forty- eight hours and eight clays after injection 232 23. Sections of same tumor nineteen days after injection 233 24. Section of twelve day old tiny necrotic sarcoma 180 in mouse with spontaneous infection of B. enteritidis 245 25. Histologic appearance of twelve day old sarcoma 180 twenty- iouv hours after intravenous injection of live culture of rough B. enteritidis 247 LIST OF ILLUSTRATIONS XXVll 26. Seclion of twelve day old sarcoma 180 tweniy-iour hours after intravenous injection of a B. ciilciilidis filtiate . 248 27. Appearance of reaction at jjrepared skin site oi^e-half hour after intra\enous injection of horse serimi following sensi- tization by single injection of horse serum six days pre- viously, followed by intradermal injection of luidiluted B. typhosus active jjrinciples 270 28. Appearance of reaction at j^repared skin site one hour after intravenous injection of horse serum 271 29. Appearance of reaction at |irepared skin site four hoins after intravenous injection of horse serum 272 30. Appearance of reaction at prepared skin site twentv-four hours after intra\enous injection of horse serum 273 31. Appearance of reaction at prepared skin site ninety-six hours after intravenous injection of horse serum 274 32. Histologic ajjjjearance of reaction at prepared skin site forty- five minutes after intraxenous injection of horse serum, showing perivascidar indanmiation and thrombosis 275 33. Histologic apj^earance of reaction at prepared skin site four hours after intraxenous injection of horse serimi 276 34. Appearance of lesion loin days after intradermal inocidation of rabbit with bovine B. tuberculosis 290 35. Severe circumscribed hemorrhagic reaction at site of B.C.G. inoculation in rabbit sensitized by single injection of horse servmi six days previously, followed by intradermal injec- tion of B.C.G. and four days later intraxenous injection of horse serum 290 36. Hemorrhagic reaction at site of intradermal injection of bovine B. tuberculosis followed four days later by intra- venous injection of B. typhosus active principles 291 37. Hemorrhagic reaction at site of intradermal injection of bovine B. tuberculosis followed three days later by intra- venous injection of B. typhosus active principles 291 38. /, Typical leaction of skin: petechial hemorrhages following application of suction to normal portion of skin distant from site of typical reaction. 2, Histologic appearance of typical reaction. 5. Histologic appearance of petechial hemorrhages 380, 381 39. Photomicrograph from autopsy stud\ in case of acute thrombocytopenic purjjura: platelet throndnis xvithin ar- teriole of kidnev 383 XXVlll LIST OF 1 I.I.I SIRATIONS |(). SiHlion lioiii swiiic wiili liot^ (liolcKi; iii;in\ t^loiiiciuli wilh (;i|)ill;ii\ llnoinhi ^84 |i. IMioioinic i()L;ia|)li Irom ;nii()|)s\ sIikIn in i;encrali/ed miliary I iibfi ( iilosis; j^loiiui iiliis wiili L;iaiuilai thrombi within dilated ( a])illarifs 385 .J2. Section horn hoisc imnunii/cd lor two years by weekly sub- cutaneous and intraxcnoiis injeciions ot meningococcus filtrates and vaccines; glomerulus with capillary thrombi 386 13-47. temperature charts in five cases following injection oi seium 412-415, 417 DIAGRAMS 1, 2. Results ol neutralization experiments with various amounts ol reacting factors of different toxic filtrates mixed vviih constant amounts of certain immune sera 104. 105 3. Effect of spontaneous epidemics on growth of mouse sar- coma 180 244 4. Effects of B. I\j)li().su.s toxin and mixtures of such toxin with anti-typhoid horse sera on mouse saicoma 180 251 TISSUE REACTIVITY Chapter I INTRODUCTION BASIC EXPERIMENTS Phenomenon of local skin reactivity to bacterial filtrates: Tlie strains ol B. typhosus employed for the early experiments were Tl and T040. They were obtained from the stool of con- valescent typhoid fever patients one to four years prior to the ^vork, respectively. Both strains appeared smooth and Avere agglu- tinated by B. typhosus antisera to a high titer. The floccules Avere large. The strain T040, ho"\vever, ^vas inagghuinable during the first month of cultivation. The toxic substances ^vere prepared as follows: Two-hundred cubic centimeters of tryptic digest broth of initial pH 7.8 were added to 2000 c.c. Erlenmeyer flasks to give a large sinface area. The entire growth of one twenty-four hoiu' old agar slant culture of B. typhosus was suspended in 10 c.c. of 0.85 per cent NaCl solution. Ten cubic centimeters of the suspension were added to each Erlenmeyer flask. The period of incubation ^vas six days. The cultures were then filtered through paper and cotton and finally through Berkefeld "V" candles. The filtrates were tested for sterility, stored in the refrigerator and used for a period of approximately t^vo to three ^veeks following their preparation. No preservative was added. The entire skin of the abdomen of rabbits was epilated with barium sidphide. The animals ^vere injected not less than tAventy- four hours after shaving ^vhen no signs of irritation "were seen. The local injections described belo^v were made intradermally. Four sites, i.e., the upper and lower right and left quadrants of the skin of the abdominal Avail, about 2 inches apart from each other, Avere used for the injections. The amount of fluid injected into each area Avas usually 0.25 c.c. I 2 LOCAL ILSSl'E RKAClIVrrY Protocol i: I'pixi and lower rij^lil and Iclt areas ol tlic skin oi llic alxlonun ol Rabhiis i| i and i.|-2 were injecled each with 0.5 ex. ol /.'. l\l>li()SHs ir\])ii( dincsi hioih cnhurc filtrate (strain I]). Twenty-lour houis hiter Raijbil i.]-! showed no skin reactions. Rabbit \,\-2 sliglitly ])ink reactions (2-)-) in the injected areas. At this time single intravenous injec- tions ol ;} (.c. per kilo ol body weight ol the filtrate were given to both labbiis. Ihe skin ol the animals was examined every hour alterwards. Approximately two hoins alter the intravenous injections, blue discolora- tion appeared at the site of previous skin injections in Rabbit 14-1. The skin ol Rabbit 14-2 remained imchanged. The discoloration observed in Rabbit i.pi rapidly increased until the reactions became extremely pro- nounced in about loin- hours alter the intravenous injection. All four areas were then dark blue in the center with a deep red zone at the periphery. The skin over the hematomas was glossy and swollen. The size ol these areas was considerable, the reaction in the upper right corner measuring 2x2 cm., the lower right 3x2 cm., the upper left 4x2 cm., and in the lower left corner 3 x 2.5 cm. The reaction described resembled a severe bruise. The animal was killed 5 hours alter the intravenous injection and sections ol the skin were made. Rabbit 14-2 showed no reaction in the course of twenty- four hours folloAving the intravenous injection. Protocol 2 : Four areas of the skin of the abdomen of Rabbit 7-6 were injected each with 0.25 c.c. of B. typhosus tryptic digest broth culture fil- trate (strain T240) . Twenty hoins later there were moderate erythemas at the site of skin injections. The rabbit was then injected intravenously with the filtrate in a dose of 1 c.c. per kilo ol body weight. The previously in- jected skin areas showed discoloration in about two hours after the intra- venous injection. The reactions again, progressively, increased and about four and one-half hours alter the intra\enous injection they were extremely severe. The areas were hemorrhagic. They appeared dark blue, glossy and swollen. The upper right corner reaction measured 3.5 x 3 cm., the lower right 2 X 2.5 cm., the upper left 2x2 cm., and the lower left 1.5 x 3 cm. Rabbit 7-6 ^vas re-examined twenty-four hours later. The sites were then black with a dark red zone at the periphery. The size remained unchanged. It appeared that the reaction reached its maximum size in about fi\e hours after the intravenous injection. The healing of the hemorrhagic areas was slow. Sloughs formed in about lorty-eight hours after the intravenous in- jection were followed by scab formation, gradual separation ol scales and scarring. The complete process ol healing took about eight days. Histological examinations of the hemorrhagic areas were made. Sections of these areas from Rabbit 14-1 obtained five hours alter the intravenous injection can be described briefly, as follows: The general impression was that ol the severest type of hemorrhage and necrosis. The skin was edematous in places. Some of the blood vessels were ruptured. The subcutaneous tissue was engorged with l^lood. There was also an extensive migration of polymorphonuclear neutrophile leucocytes. There \vas obserxed pronounced necrobiosis of these cells located inside and outside the blood vessels. Some of the blood vessels contained small parietal thrombi. While it was clear that the process afTected the veins, it remained questionable whether there was any primary injury to the arteries. Some of INTRODUCTION 3 the arteries Avere found normal. Others were ahiiost entirely occluded. The obliteration, however, was probably secondary to the hemorrhagic infiltra- tion outside the arteries. Occasionally, hyalinization was seen in the blood I'K.. 1. I'i\e sites of skin were injected each with 0.3 c.c. of T. D. l\ filtrate; 24 hours later rabbit recei\ed intra\enoiis injection of same filtrate. Dose was 2.5 c.c. per kilo of body weight. Appearance of reaction at site of prior skin injections 5 hours after intravenous injection. (Shwartzman, 19286.) \'essels. The hemorrhage and necrosis extended to the corium of the skin, which -was thin and broken in places. The sections of the hemorrhagic areas of Rabbit 7-6 made twenty-four hours after the intravenous injection were almost identical with the sections of Rabbit 14-1. Early observations on routes of injection : It became obvious from the ttnidamental experiments that the use of the intravenous route for the second injection ^vas an es- sential recjuirement of the phenomenon (Shu'artzman, 1928/?, iggofl) . Repeated skin Injections of the fikrate within tuenty- four h()in~ interxal between injections did not resuk in the hemorriiagic and necrotic type of reaction of the phenomenon. 4 LOCAL ILSSLE REACTIVITY I hey iiicicK •^ 4 « i •• ' *^ ^ f *f * • *#^«. A -' * ■• ' ••'1 ' '' I'lG. 4. Hemaluxylin-cosin. > 270. Section ot skin trom Rabbit 16-1. Mioostupit appearance of reaction produced by repeated skin injections of T. D. I'l- No necrobiosis of \vhite l)lood cells. (Shwart^man. 1928/).) Protocol 4: Twelve rabbits received each three simultaneous intradermal injections of filtrates of certain microorganisins. The upper right areas ot the skin of the abdomen of each of these rabbits were injected with 0.25 c.c. of meningococcus culture filtrate ("A^q") , the lower right areas with 0.25 c.c. of B. typhosus culture filtrate ("Ajj^") , and the upper left areas with 0.25 c.c. of B. coll cuhure fihrate ("587") . Twenty-four hours later the rab- bits were divided into 3 groups of 4 each. Each group received intravenously a different filtrate. The rabbits of the first group received, per kilo of body weight. 1.5 c.c. of "Aj,j" previously diluted 1:5, the second group received 1.5 c.c. of "Aj/' previously diluted 1:5, and the third group received 1.5 c.c. of "587" previously diluted 1:5. Four hours later there were found 3 dead rabbits in the first group. The () surviving rabbits showed severe hemorrhagic necrosis ranging in size from 2 x 2.5 cm. to 4 x 4 cm. in all tlie three pre- pared areas. INTRO DICTION 7 Early observations on time required for elicitation of state of reactivity and on its duration: The following experiments were made in order to determine the optimum time interval between the injections necessary to elicit the reactivity (Shwartzman, 1928/;) : Protocol 5 : Four areas of the skin of 6 rabbits (Nos. 48-2 to 48-6 were injected with 0.25 c.c. of B. typhosus (TjJ tryptic digest culture fihrate. Rabbits 48-2, 48-3, and 48-4 received intravenous injections of the filtrate two hours after the skin injections. The dose was 3 c.c. per kilo of body weight. No reactions were seen in these rabbits in the course of the following forty- eight hours. Rabbits 48-5, 48-6, and 48-7 were injected intravenously with 3 c.c. of the filtrate, per kilo of body weight, forty-eight hours after the skin injections. The skin remained unchanged for forty-eight hours. Protocol 6 : The upper right areas of the skin of the abdomen of Rab- bits 1-7, 1-8, 2-0, 2-6, 2-7, and 43-9 were injected with 0.25 c.c. of tryptic digest Tl filtrate. Twenty-four, forty-eight, and fifty-five hours later, 0.25 c.c. of the same filtrate was injected into the lower right, upper left, and lower left areas of the skin of the abdomen, respectively. Fifty-six hours after the first skin injection, tryptic digest Tj^ filtrate was injected intravenously into these rabbits. The dose was 3 c.c, per kilo of body weight. The morning following the intravenous injections, namely, fifteen hours after the intra- venous injections, the following results were obtained: Rabbit 1-7 was found dead and no reading was possible. Rabbits 1-8, 2-0, 2-6, 2-7. and 43-9 showed no reactions in the upper right and lower left areas. Both lower right and upper left areas showed very pronovuiced hemorrhagic reactions which varied in size from 1 x 1 cm. to 2 x 2 cm. From the above experiments the follo^ving coidd be concluded: For the reproduction of the described phenomenon a definite interval of time was required between the preparatory skin in- jection and the intravenous injection of the filtrate. Althotigh no exact information ^vas a\'ailable on the exact ninn- ber of hotns necessary for tlie preparation of the skin, neverthe- less, it became obvious that two hours were insufficient and that the state of reactivity did not last longer than thirty-t^vo hours. The phenomenon could be invariably reproduced if an interval of twenty-foin- hours \vas allowed bet^veen the skin and intra- venous injections. It may be also noted from Protocol 6 that the skin reactivity appeared as soon as eight hours after the skin-preparatory injec- tion. In these animals, however, the reading of the skin reaction was made fifteen hotus follo\ving the intraxenous injections, in- 8 LOCAL TLSSUE REACTIVITY stead ol alter (lie ( u.sU)niai y (iNe lioins. The si^iiiluauee of tliis obsei xalioii will he e(jnsi(leie(l suhscquenll). TERMINOLOGY Soon alier the ahoxe deseiil)e(I basic: experiments ^vere made, it became necessary to select an appropriate terminology. Inas- much as the exact natme of the oloservations was unknown, it seemed advisal)le to choose a new terminology which woidd avoid a preconceived association ^vith phenomena of anaphylaxis 'with ^vhich there aj^peared to be a superficial resemblance. The phenomenon descriljed here is termed as the "])henome- non of local skin reactivity to bacterial filtrates," ^vhen the pre- j)aratory injection is made into the skin; or as the "phenomenon of local tissue reacti\ ity to bacterial filtrates," ^vhen the prepara- tory injection is made into other tissues or organs. Elicitation of the state of reactivity by means of a local injec- tion into the skin or other tissues and organs of a potent bacterial filtrate is termed as "preparation" and the factors capable of eliciting this state are respectively called "preparatory" factors. By the term "reactivity" is meant to imply the ability of a given tissue to respond ^vith a typical lesion follo^viug the intravenous injection of a potent preparation and is not designated neces- sarily to associate this state with processes of anaphylaxis. The terms "vulnerability" and "preparedness" are used alternatively Avith the term "reacti\ity." Apitz (193-^^'') substituted the term "reactivity" by "Ueberstimmiuig" ^vhich seems to be expressive and for which I was inial)le to find an equivalent word in the English language. As will be seen from the ^vork described in this monogiaph, the essential prerequisite for elicitation of the reaction of the phe- nomenon under discussion is that the second injection be given via the blood stream. The factors invohed in production of local hemorrhagic reactions following the intravenous injection ^vere called "reacting factors." It may be ^vell then, to emphasize that the term "reacting factors," employed here, is strictly limited to substances capable of eliciting hemorrhagic reaction in the pre- pared tissue upon their introduction into the general circulation. Karsner and Moritz (19^34) introduced the term "provocative factor" and Gratia and Linz (1932c) the term "facteius dechai- nants." I believe, however, that it is safer to avoid the latter term INTRODUCTION 9 inasmuch as it implies a connection with strictly anaphylactic processes. Alechinsky (1935, 193^) proposed the term "angiorrhexis" (vascular rupture) for the lesion of the phenomenon. This term seems to be hardly advisable inasmuch as actual rupture of the blood vessels is only an incidental and rather irregular feature, as 'vvill be noticed from the description of the morphological changes of the phenomenon. When both skin-pre}:)aratory and reacting factors are discussed collectively, they are referred to as the "active principles" of the phenomenon. The reaction at the prepared skin sites is usually described as hemorrhagic necrosis because of its appearance in the gross. His- tologically, the reaction is predominantly hemorrhagic. As will be seen later, it is possible that the necrosis appears subsequently to a profuse hemorrhage. GENERAL COXSmERATIONS CONCERNING THE PHENOMENON Skin-preparatory injections of bacterial filtrates employed in this ^vork produce variable but mild primary effects depending on the microorganisms and methods employed for preparation of the filtrates. It is beyond doid^t that the primary reactions are histologically and in the gross conspicuously different from the typical reaction of the phenomenon of local skin reactivity to bacterial filtrates (Chapter vi) . Twenty-four hoin^s after the intradermal injection, the pre- pared sites may sho^v no reaction, or local erythema of varying intensity accompanied by s\velling, or s^velling alone ^vithout ery- thema. A primary hemorrhagic reaction is only rarely observed with a limited nimiber of preparations of meningococcus and was never seen ^vith preparations deri\'ed from other microorganisms employed thus far. The primary effect of a single intradermal injection of 0.25 c.c. of bacteria filtrates was studied on many occasions (Shwartz- man, 1928(7, /;, 1929/;, e) . The usual type of primary reaction in the site of the local injection of filtrates of six day old tryptic digest broth cultures of B. f\j)Jiosus was an erythema. The size of the reaction varied from 1 X 1 cm. to 2 X 2 cm. in diameter. When four areas of the skin of the abdomen were injected, occasionally diffuse redness ap- peared o\er the entire skin of the a])domen. About 50 per cent 10 LOCAI, TISSUE REACTIVITY ol all the iioinial animals tested showed a j)ositive erythematous reaction to the injedions ol these filtrates. The reactions differed considerably in intensity. Only a small percentage of the posi- tively reacting animals gave stiong erythematous reactions (12 per cent) , the majority of positive rabbits, ho^vever, reacting weakly. Different areas of the skin of the abdomen also presented considerable \ariations in jjrimary reactions to B. typJiosiis toxic filtrates. These residts ^vere substantiated l)y protocols of several hundred rabbits. A report on these animals woidd unnecessarily lengthen this monograph. "Agar washings" hltrates of the typhoid-paratyphoid-dysentery groups of bacteria give still less significant primary reactions, as illustrated l)y the following typical experiment: Protocol 7 : Forty rabbits each received six simultaneous injections of B. lypliosus "agar washings" filtrate in dilutions of 1:2, 1:8, 1:16, 1:24, 1:36, and 1:48. The readings of the primary effect of these injections were made twenty-three to twenty-four hours later. The dilution 1:2 produced well- marked erythema (4-(-) in one rabbit, moderate erythema (2-f-) hi 3 rab- bits, slight reddening (i-|-) in 7 rabbits, and doubtful reddening or no erythema in the remaining 29 rabbits. With the dilution 1:8, only one rab- bit showed a moderate erythema (2-)-) , and the remaining rabbits developed very slight erythema, or no reactions whatsoever. None of the higher dilu- tions had any visible primary effect upon the skin. A single intradermal injection of the undihited B. typhosus "agar washings" filtrate gives primary reactions of similar mild- ness and variability. Although the reactions in thotisands of rab- bits Asere recorded, it appears to be sufficient to report the restilts in r,'^7 animals, as follo^vs: Fotu" htmdred and seven rabjjits sh(^Aved no detectable reac- tions; 22 rabbits had a doubtftil erythema; 73 erythemas were recorded as 2-|-. and 12 labbits developed marked erythema clas- sified as 4-j-i 19 rabbits showed a distinct swelling at the pre- pared skin sites. There appeared to be no relationshi}) betAveen the intensity of the erythema and swelling. Four rabbits showed small hemorrhagic lesions which could be confused with the reac- tions of the phenomenon. The primary effect of meningococcus "agar washings" filtrates upon the skin of normal rabbits was similarly sttidied. Protocol 8: Thirty-six rabbits received each five simultaneous intra- cutaneous injections of 0.25 c.c. of undiluted meningococcus "agar washings" filtrates and of the same filtrate diluted with 0.9 per cent NaCl 1:10, 1:20, INTRODUCTION 11 1:40, and 1:80. The filtrate was of a somewhat lower potency than those described later, shice from 15 to 20 c.c. was used for washing each Kolle flask instead of the customary 3 to 4 c.c. The readings of the primary effect of these injections were made twenty-three to twenty-four hours later. The undiluted filtrate produced moderate erythema (2-j-) in three rabbits, slight reddening (i-[-) i'^ ^ rabbits, and doubtful reddening, or no erythema in the remaining 25 rabbits. With the dilution 1:10, 2 rabbits showed a moder- ate erythema (2-f-) , and the remaining rabbits very slight erythema or no reactions whatsoever. The dilutions 1:20 and 1:40 produced 2-|- erythema in 3 rabbits. None of the remaining injected areas showed any visible primary effect of the filtrate. In a stdjsequent experiment, 24 rabbits received each six simultaneous in- tradermal injections of 0.25 c.c. of the above undiluted filtrate and of the filtrate diluted with 0.9 per cent NaCl solution 1:2, 1:4, 1:6, and 1:10. The primary effect of these injections, as read twenty-four hours later, was in- significant. Only 3 rabbits showed moderate erythema and the remaining rabbits had either "i-|-" reddening or no erythema whatsoever. Occasionally, there were swellings. The primary effect of a single injection of meningococcus "agar washings" filtrates of concentration higher than in the fil- trates studied above (i.e., each Kolle flask cultine washed Avith 3 to 4 c.c. of saline containing 0.4 per cent phenol) , was as follows, in experiments on 1002 rabbits. Forty-tAvo and four-tenths per cent of rabbits showed no pri- mary reactions in the gross, t^venty-four hoius after the injection; 39.2 per cent gave doubtful primary reactions; whilst in 7.7 per cent erythemas were elicited which could be classified as 4-]-; the rjemaining rabbits sho\ved erythemas of an intensity varying from i-|- to 3-|-; 17.3 per cent showed distinct swellings at the prepared sites Avith and without erythema. Here again, the swell- ings appeared to bear no relationship to the erythematous pri- mary reactions. About 2.5 per cent of the rabbits showed strong primary hemorrhagic reactions and about 1.7 per cent shoAved pimctiform hemorrhagic reacticjns. Thus, in 4.2 per cent there appeared reactions -which could be confused Avith the reaction of the phenomenon of local skin reactivity to bacterial filtrates. These reactions were obtained Avith but a iew preparations and did not appear if th^^in was injected Avith a dilution of 1:2 or 1:3 thereof. J^ As may be seen, tWe above described primary reactions obtained with bacterial filtrates employed in this work are sharply con- trasted by the hemorrhagic reaction folloAving the intravenous in- jection of potent filtrates. It may be important to emphasize again 12 LOCAL TISSUE REACTIVITY that ihc essential leatiire ^vhich caniK^t possibly escape observa- tion, histologically and in the gross, is the hemorrhage following the intravenous injection, riie readings ot the reactions recorded in this monograph are based on this essential ieatine and are only made in rabbits surviving lor at least tour hours folloAving the intravenous injection ol reacting factors. If the animals die ^vithin the first four hours following the in- jection of a bacterial filtrate, the reading of the reaction is im- possible in most instances. The reaction may diffuse through the skin of the a])dominal wall, and assume a livid appearance which is very diflicult to interpret. Only occasionally it is possible to observe well dehned hemorrhagic reactions in dead rabbits. It seems advisable, therefore, to disregard the readings of rabbits which die 'within the fh\st four horns follo"\ving the injection. As a rule, the reactions of siu\iNing rabbits become visible in the gross as early as one-half to one hotu' after the intravenous injection. In typical strong reactions the early appearance is that of a crop of petechiae ^^■hich continuously increase in size until there results an extensive confluent hemorrhaoe forming" a sac filled ^vith blood. The color rapidly changes from blue to violet and almost black. There is an angry red zone at the periphery. The entire process may be so rapid that the petechial stage is indistinguishable in the gross. With highly diliued preparations, Aveak reactions may be de- layed in appearance, i.e., ten to t^vehe hours instead of the custo- mary foiu' to five hours. Some ^veak reactions may appear promptly but disappear the folloAving day. The reactions of this type may be best described as punctiform. They appear as a crop of petechiae "^vhich do not become confluent and have a tendency to fade promptly. Sometimes follo^ving the intra\enous injection one may see an accentuation of preexisting swelling and erythema. In imminiological studies concerning this phenomenon, Avhen the record of the results is based in most instances on the appearance of the reaction in the gross, weak reactions should be considered as doubtfid and not much, stress laid upon them, since they may only present a mere accentuation of a primary inflammatory re- action and not be an expression of the phenomenon of local skin reactivity to bacterial filtrates according to the present conception. Inasmuch as the skin preparation is elicited by means of an intradermal injection, the reaction is well defined in the super- ficial layers of the skin and penetrates, in many instances, into the INTRODUCTION I3 entire thickness ol the abdominal wall. The readino of the in- tensity and the record of the size of the reaction is made thereby exceedingly easy. In some instances, Avhich remain ^vithout ex- planation as yet, and also ^vhen the preparatory injection is made subcntaneously, reactions may be obtained which do not extend to the superficial layers of the skin and, therefore, appear as poorly defined subciitaneons hematomas. In order to differentiate these from brtiises, it is necessary to excise the suspected portion of the abdominal w^all. In the gross, typical snbcutaneoiis reactions are tniiform hemorrhages in nearly all layers of prepared sections of the abdominal ^vall well demarcated from adjacent normal tissue. The healing of the strongly hemorrhagic reaction is sIoav. Sloughs which may form in about forty-eight hours after the in- travenous injection are follo^ved by gradual separation and scar- ring. The complete process of healing takes about ten days. This coinse of events ^vas similarly described by Ecker and Welch (1930), Gratia and Linz (1932c), Stolyhwo (1935), Apitz (1933^) , Gerber (1936(7) , Kielanowski and Selzer (1934/;) . and others. Bock (1932) attempted the percutaneous method of prepara- tion. The ear of an albino rabbit ^\ as defatted whh soap and ether. Cotton immersed in B. coli broth culture filtrate was applied to the surface of the ear for one hour. Follo^vino the intravenous in- jection of the same filtrate into another ear, he obtained, in many instances, reactions of various intensity ranging from single l3luish petechiae to diffuse dark blue infiltrations and necrosis of the ear. Control defatted ears {i.e., not j^repared percutaneously ^vith the filtrate) ga\e only hyperemia folloAving the intra\enous injection of B. coli culture filtrate. Similarly, in a few experiments I at- tempted to prepare the skin of the abdominal "\vall by vigorous rubbing ^vith a sponge soaked in bacterial filtrate. Follo^ving the intravenous injection of reacting factors there appeared infre- quently scattered petechial lesions resembling purpura of men- ingococcemia. The results ^vere rather irregular and apparently depended upon variations in skin permeability. W'hen several sites of the skin of the abdominal ^vall are simul- taneously injected, the subsequent intravenous injection of a sufficient dose of bacterial filtrates elicits reactions of uniform severity in the prepared sites. This is illustrated by the following typical experiment (Shwartzman, 1928/;) : 14 LOCAL TLSSUE REACTIVITY Protocol 9: A group of 24 rabbits was employed. The skin-preparatory injection of a B. lyj)liosus tryptic digest broth culture filtrate was made sinuiltaneously into the upper right, lower right, upper left and lower left (juadrants of the skin of the abdominal wall. The doses employed for the intradermal injection were either 0.25 c.c. or 0.5 c.c. Positively reacting ani- mals showed iniiiormly extremely severe hemorrhagic reactions in prepared skin sites following intravenous injections of B. typhosus culture filtrates in various doses. The only variation observed was in the size of the reaction. The smallest reaction was 1 x 1 cm., and the largest 4x4 cm. The size of the reaction did not clearly depend on the amoimt of filtrate used for the skin prejjaration. For the same animals variations in the size in different areas lay between 1 cm. and 3 cm. in diameter. Ordinarily, the difference was 1 cm. to 2 cm. Occasionally, confluent reactions were obtained. No well marked relationship between the intravenous dose and the size of the re- action of the skin was observed. The reaction following the intravenous injection is not strictly localized to the site of the preparatory injection. Some of the sites may have a tendency to spread more than others, i.e., the central portion of the skin of the abdominal wall. The skin of the upper left quadrant invariably shows reactions smaller than those of the right cpiadrant. The question of spreading of the reaction was recently investigated by Bier (1932, 1933) , and Diiran-Reynals (1933) in an attempt to determine the effect of the Reynals factors iq)on the phenomenon. In 1928 I^iiran-Reynals (1929, 1933) reported that extracts of normal animal testicles enhance to a marked degree the activ- ity of nenro and dermal strains of vaccinia virus. The author as- sumed that the effect of the tissue extract was apparently on the cells of the host rather than on the virus. This was indicated by the fact that virus injected intravenously localized most readily in an area of skin previously injected with testicle extract. Enhanced lesions also resulted when a virus was injected into an area as long as three days prior to the injection of the testicle extract into the same area. Kidney and probably skin, brain, and liver extracts were reported by him to possess the enhancing properties but to a much lesser degree. McClean (1930) brought additional e\i- dence to the fact that Reynals factors produce the enhancing effect not through modification of the virus biu through some change in the receptivity of the host. He showed that the activity of the extract is associated \vith an increase of the dermal perme- ability ^vhich persists for at least t^venty-four hours. Accordingly, this increase of dermal permeability allo\vs the spreading of tiie virus over a large skin area. Pijoan (1931) foimd tiiat the in- INTRODUCTION I5 fectious activity ot cultures ot tAventy different bacteria may be increased I)y Reynals factors (B. I\l)h(>sits, B. paratyplwsus A and B, B. dysenteriae, Vibrio cholerae, enterococctis, B. prodigiosus, M. TetragouLs, Streptococcus hemolyticus, B. coli, pneimiocctis Type III, etc.). It is also of interest that in Favilli's (iQ.'^i) ex- periments the testictilar extract notably increased the fragility of red blood cells. In Bier's experiments the toxic material \vas the stipernatant fltiid of a centrifuged agar culture of B. typJwsiis trittnated in NaCl. Six rabbits were injected in one ear with 0.2 c.c. of a mix- tine of toxin plus testicular extract. The other ear was inocidated with 0.2 c.c. of mixture of toxin plus physiological saline. Fifteen horns later, the rabbits received a provocative injection of 1 c.c. of the toxin intravenously. In 3 rabbits the testictdar extract pro- duced no effect. In the remaining rabbits there was an increase in the zone of reaction in sites prepared ^vith mixtmes of active principles and testictdar extract. The residts obtained by Bier were reinvestigated by Dinan- Reynals. Rabbits were prepared by the intracutaneous injection of 0.25 c.c. of filtrate from a six day old broth cidtine of B. coli mixed with testictdar extract, and control animals received the same amoinit of filtrate diluted with Ringer's solution. The areas of lesion produced by the fdtrate and testictdar extract were tm- dotd3tedly larger than in the controls but what the lesions had gained in extent, they had lost in intensity. Instead of the Avell defined and severe edematous and hemorrhagic character, the lesions were flat, diffused and mild. In the animals in which the control lesion was mild, the one resulting from the filtrate plus testictdar extract mixttne was extremely benign or in some cases completely suppressed. In an additional experiment by Duran- Reynals, the procedtne was varied by injecting the testictdar extract into the prepared area at the same time as the intravenous filtrate injection was given. This provided a similar spreading out of the reaction with a corresponding lessening of the severity or even suppression of the reaction. The author concluded from his experiments that the testicular extract gave no true enhance- ment of lesions produced by bacterial toxins used and had no influence tipon the subsequent development of the phenomenon. The findings of Duran-Reynals were confirmed by Cassiito (1933//) . The latter author also gained the impression that in- l() LOCAL ILSSl'E REACTIVITY jcttioii ol siil).staiuc.s hjweiiiig the capillary pernical)ility (uie- tliaiic. j)laiii hroiii, and Besredka antivirus) into the skin sites prepared with bacterial fdtrates enhanced the severity of lesions provoked by the intraxenous injection of the bacterial filtrate. In my studies on the titration of skin-j^reparatory factors sidj- sec[uently described, it was essential to determine xvhether va- rious sites of the skin of the abdominal xvall responded similarly to diluted fdtrates. Incidental observations xvhich xvere never car- ried out systematically, suggested that high dilutions (i:()4) of the filtrates elicited xveaker reactions in the upper and lo\\'er left cpiadrants o^ the skin of the al)dominal xvall than in the upjier and Ic3wer right (juadrants. Recent investigations of Koplik (19^57) demonstrate that the prej)aratory factors are rapidly spread by means of the lymphatics inasmuch as lym])h nodes adjacent to the prepared skin sites shoxv distinct hemorrhagic reactions as \ve\\ illustrated by his figmes. It is possible, therefore, that the spreading of a reaction may depend on the richness of the lym- phatic network in a given area. The skin of \arious parts of the body of the rabbit seem to differ to a certain extent in susceptibilit) to the phenomenon. In my recent unpublished experiments, rabbits received four 1. Section of hypogastric lymph node, six hours after intravenous injection of filtrate from cuUme of B. txplwsits given following preparatory injection of same filtrate into afferent lvm|3hatic. Dilated veins are partially thrombosed. Macrophages containing red blood cells are in lymph sinuses. Hematoxylin and eosin. X 160. 2. Arteriole in hypogastric node, showing swelling and hyaline degeneration ol wall and necrosis of endothelial cells. There is perivascidar edema but no cellular infiltration. Animal killed six hours after intravenous injection of filtrate of menin- gococcus cultine administered twentv-four hours subsequent to injection of this filtrate into circumflex abdominal lymphatic. Hematoxylin and eosin. X 600. 9. Sections of right (R) and left (L) pectoral nodes. Animal was given in- travenous injection of filtrate from a culture of B. typhosus t\venty-foin- horns after l^reparatory injection of same filtrate into skin of right side of abdomen. Right pectoral node is swollen, congested and contains a large area of hemorrhage. Left pectoral node is normal in appearance. 7. Section of vein in left common iliac node, which received intravenous injection of filtrate from a culture of B. typhosus twenty-four hours after preparatory in- jection of filtrate into skin of medial surface of left thigh. Vein contains a thrombus adherent to endothelial cells which are swollen and take nuclear stain lightlv. Note there is no perivascular cellular infiltration. Reaction was negative at injected skin site. Hematoxylin and eosin. X 650. 5. Regional inguinal lymph node removed twenty-four hours after an intra- dermal injection of filtrate of cidture of B. typlwsus into right lower quadrant of abdomen. Section shows edema of node. Blood vessels are empty and not dilated or thrombosed. Compare with /. Hematoxylin and eosin. X 160. (Koplik, 1937.) i8 I.OCAI. TISSUE REACTIVITY simultaneous injections ol xarious dilutions oi meningococcus "agar washings" filtiates iiUo liie (iciniis ol the i)ack. These dilu- tions Aveie 1:100, 1 :()()(), 1:1000, and 1:1 ]oo. l'\)ll()\ving the in- TiG. (). skill ol liack of rabbit iccci\etl eight intiatlennal injections ot nieningococcus active principles in cliliitions ranging trom 1:2 to 1:300. Twenty-four fiours later, 100 tmits of meningococcus acti\e principles were injected intravenously. Note severe hemorrhagic reactions at prepared sites. travenous injection ot an appropriate dose, reactions were ob- tained in areas prepared with dilutions of filtrate 1:400, 1:600, and 1:1000 and no reactions in the sites prepared with the dilu- tion 1:1400. In another group of six rabbits, the sites of the back \vere prepared with dilutions 1:600, 1:800, 1:1000, and 1:1200. Follo^ving" the intravenous injection of the same dose as employed in the previous experiment, no reactions appeared in the prepared skin sites. Thus, the paradoxical observation Avas made that sites prepared INTRODUCTION I9 \\ ith dilutions as high as 1 : 1000 reacted, pro\'ided the same animals ^vere also prepared by injection of lower dilutions. On the other hand, if higher dilutions alone were employed for preparation, no reactions ^vere obtained. The only possible explanation is that the more concentrated material may spread to the sites receiving a lesser concentration and thus induce a state of reactivity in these sites. On the other hand, no reactions result if low concen- trations are employed for preparation of all the sites. Here again, the richness of the lymphatic network may constitute an im- portant factor in this diffusion phenomenon. The observations of Joukow-Werejnikow and Lipatova (1933- 34) suggest that the richness of the venous supply to a given site may also influence the severity and size of the reactions fol- loAving the intravenous injection. Thus, very severe reactions were obtained mostly ^vhen the preparatory injections ^vere made in jDroximity to a large subcutaneous \ein. (iratia and Linz (1932c) produced the phenomenon in the skin of the ear. They injected 0.5 c.c. of the B. coli culture filtrate intradermally and followed the pre])aration by an intravenous injection of the same filtrate t^senty-foiu" hoins later. The reac- tions obtained were se\ere and diffuse. I also foinid that with appropriate doses of the toxic material for the preparatory and provocative injections, strong reactions can be obtained in the ear (Shwartzman, 1935c) . The reactions may be well circumscribed and intensely hemorrhagic. In some instances, ho^ve\er, there is a tendency towards the generaliza- tion of the reaction through the entire ear. Frequently there is also obser\ed blue discoloration in j^arts of the ear not in\ol\ed by the reaction. Histologically, the reaction is accompanied by hemorrhage, pronounced capillary thrombosis, subsequent severe necrosis and inflammation. The discolored parts showing no re- action may have extensive thrombosis in the venules and no hemorrhage. Comparative titrations in the skin of the ear and in the abdominal Avail Avith the same material brought out the fact that the skin of the ear is considerably more resistant to the phenomenon than the skin of the abdomen. Ten times the provoc- ative dose is recjuired if it is given into the marginal vein of the ear prepared by the intradermal injection of bacterial filtrate and thirty times the dose if it is given by way of the vein of the non-prepared ear. The ol)\ ious interpretation of the latter is that 20 LOCAL TISSUE REACTIVITY a txTtain anioiml ol actixe })rinciples injected into the vein of the n()n-j:)iej)aie(l ear is lost in the general circulation before reachiiiL!, liic skin of the j)rej)are(l ear. Im(.. 7. I'hfiionienon of local skin rcacti\it\ to nieniiigococciis filtrate in rabbit car. (Mcnin.nococcus, Cironp III (44B) "agar washings" filtrate (T.1968) .) Twenty-five hundredths cubic centimeter diluted 1:2 injected intradermally and 100 reacting units injected intraveirously twenty-four hours later, (a) Typical lesion four hours after intravenous injection. (/;) Control ear which received no preparatory injection. (Sh\\'art/nian, 1935c.) In connection Avith the spreading of reactions beyond the pre- pared sites, it is of interest to consider the following experiment of Gratia and Linz: These authors gave an injection of a large preparatory dose of B. coli filtrate into the dermis of the rabbit's ear followed 24 Fig. 8. Treatment same as in Ear a of Figure 7. Note severe hemorrhage, edema, inflammatory reaction and vascular thrombosis. X 50. (Shwartzman, 1935c.) Fit;. 9. Treatment same as in Ear a of Figure 7. Section taken of area 3 cm. a\vay from site of reaction and showing in gross cyanosis and edema. Note thrombi in venules. X 400. (Shwartzman, 1935c.) Fig. 9. 22 LOCAL riSSUE REACTIVITY hours later by the iiitia\ eiioiis injection ol the reacting factors. In some instances they obtained reactions in the non-prepared ear. These reactions coirsisled ol petechial hemorrhages along a large vein. In another group ol experiments they introduced a fragment of sponge impregnated with 0.5 c.c. of B. coli ciilttire lilt rate into a subcutaneous pocket of the abdominal ^vall of a normal rabbit. The pocket was closed by an application of a ligature. Twenty-foiu" hours later the sponge was removed and macerated in saline. The fluid obtained Avas used for a prepara- tory injection in the rabbit's ear. Animals thus prepared received B. coli cidture filtrates. A few days later in the non-prepared and prepared ears of positive raljbits, as Avell as in the non-prepared and prepared ears of negative rabbits, there were noticed pur- pmic lesions appearing along the capillary network of the margin of the ear. This ptnpma ascended and finally involved the entire ear. The ears became edematotis. The observations of Gratia and Linz Avere corroborated l)y Pabst and Branham. In a few experiments of my own, the above observations with B. coli and B. typhosus culture filtrates could not be reproduced at Avill. The spreading seemed to occin- rather infrequently and oirly in rabbits in \\hich the prepared ear reacted Avith unusual severity. The spreading of the reactions to the non-prepared ear deserves, however, special consideration inasmuch as it suggests the ten- dency towards symmetrical distribution of hemorrhagic lesions observed in the Schoenlein-Henoch syndrome of htmian hemor- rhagic diathesis. It may be also in place to mention here Avidespread petechiae extending from the local lesions observed sometimes with pneu- mococcus cidture filtrates. In these experiments the skin sites were prepared either by heterologous filtrates or pneimiococcus preparations (Shwartzman, 1932/) . Intravenously, the rabbits received large doses of autolyzed pneumococcus glucose broth cid tines. In some rabbits there ap- peared typical reactions which, however, frequently Avere very large. If several sites Avere prepared, there Avas a decided tendency for the reactions to become confluent. Petechial hemorrhages Avere sometimes observed Avell beyond the prepared sites. It seems to be of still more interest that in other instances it Avas possible to obtain petechial lesions scattered over the greater portion of the skin of the abdominal Avail, and of equal severity in prepared and non-prepared sites. Thus far, the observations Avere made INTRODUCTION 23 only ^vith })neumococcus preparations and could not be repro- duced at will. The individual susceptibility of rabbits apparently ^vas an important factor. Ihiforttniately, the preparations were not tested for primary pneumococcus purpurogenic substances of Julianelle and Reimann (1926, 1927) . It is quite possible that pneinnococcus material employed in my experiments contained the purpurogenic factors in addition to the active principles of the phenomenon. This assiunptic:)n is supported by the fact that lesions of the type described were not obtained by me thus far with any other bacterial preparations, julianelle and Reimann found that extracts of strains of Staphylococcus aureus, Strepto- coccus viridaus, Streptococcus Jieiuolyticus, B. coli, and the men- ingococctis do not contain the pinpurogenic factors. The experiments ^vith pneumococcus material described here are instructi\e for pinposes of differentiation of the lesions of the phenomenon from possible superadded primary lesions elicited by certain purpurogenic factors. Lesions at sites prepared with 0.25 c.c. which exceed 5 x 5 cm. and also lesions spreading over a large area and appearing of equal intensity at the prepared site, as well as beyond it, shoidd be considered as those compli- cated by some additional purpm'ogenic factors. The work of Dman-Reynals (19-^2, 1933) has recently demon- strated that invasive strains of staphylococcus and streptococcus contain a soluble factor which markedly increases tissue perme- ability and enhances the infection produced by these organisms, by other bacteria and by vaccine virus as well. The non-invasive strains of the same species of staphylococci and streptococci do not contain this factor. Reynal's spreading bacterial factors are, ho'wever, apparently absent from the meningococcus, according to this author, and from B. coli, B. typJwsus ctdture fdtrates, according to my o^vn observatic:)ns and, therefore, j^resent no con- cern in the studies of the phenomenon of local skin reactivity to bacterial fdtrates. Their role in the phenomenon of local skin reactivity to streptococcus remains to be determined. THE QUANTITATIVE ESTIMATK^NS OF POTENCY OF ACTIVE PRINCIPLES It was realized by me early in the ^vork that quantitative esti- mations of potency of active principles are essential for studies on various aspects of the phenomenon. The methods described below% as illustrated by typical experiments, were employed in these studies. 24 LOCAL ILSSUE REACTIVITY Method I : Reciprocal titrations of various skin-preparatory doses against various reacting doses : In these titrations several sites oi the skin Avere prej^ared simul- taneously with various dilutions ol tlie material tested and twenty- loin- hoin\s later, rabbits thus prepared were divided into groups of 4-6, and injected intravenously. Each group of rabbits received intravenously a different dilution of the same material (Shwartz- man, 1929/;, e) . Meningococcus "agar washings" culture filtrates Avere used: Protocol 10: Thirty-six ral)bits recci\cd each five simultaneous intra- dermal iujec tions ot 0.25 t.c. ol inuliluted filtrate and of the filtrate diluted with o.g per cent NaCl solution 1:10, i.-^o, 1:10. and 1:80. Twenty-four hours after the intradermal injections, these i\6 rabbits were divided into 6 equal groups. Each group received a single intravenous injection ot the filtrate. The first group received 2.5 c.c. per kilo of body weight. The second group received 1.5 c.c, the third group 1.0 c.c, the fourth group 0.5 c.c, the fifth group 0.25 c.c, and the sixth group 0.1 c.c, respectively, per kilo of body weight. The intravenous injection proved very toxic. Ten rabbits of the first 2 groups (2.5 c.c, and 1.5 c.c), and 2 rabbits ot each of the following 4 groups died within one to five hours following the intravenous injection. The skin reactions of the surviving 18 rabbits were read five hours after the intra- venous injections. All of these rabbits showed severe hemorrhagic necrosis varying from 3x2 cm. to 3 x 4 cm. in size in areas previously prepared with luidiluted filtrate. Only one rabbit in each of the six groups showed hemor- rhagic necrosis in areas prepared with the dilution 1:10. All the areas pre- pared with higher dilutions remained negative. Protocol II : Twenty-four rabbits received each six simultaneous intra- dermal injections of 0.25 c.c of undiluted filtrate, and of the filtrate diluted with 0.9 per cent NaCl solution 1:2. 1:4, 1:6. 1:8, and 1:10. Twenty-four hovus after the intradermal injections, the rabbits were divided into 4 groups of 6 each; each group received a single intravenous injection of the filtrate. The first group received 0.4 c.c. per kilo of body weight. The second group received 0.3 c.c, the third group 0.2 c.c, and the fourth group 0.1 c.c, re- spectively, per kilo of body weight. Three rabbits of the first group, one rabbit of the second, and one rabbit of the third died within one to five hours following the intravenous injection. All of the remaining iq rabbits showed severe hemorrhagic necrosis in areas previously prepared ^vith mi- diluted and diluted 1:2 filtrate. There were fluctuations in the skin-prepara- tory eftect of higher dilutions. Nevertheless, reactions were obtained even in sites prepared with dilution 1:10, provided the intravenous dose was not less than 0.3 c.c per kilo of body weight. This is corroborated by the experi- ment recorded in the following protocol. Protocol 12: Fourteen rabbits received each six simultaneous intra- dermal injections of the diluted filtrate. The upper and lower right and left areas of the abdominal wall were injected with the dilution 1:2, and the middle right and left areas with the dilution 1:3 of the filtrate. Twenty- INTRODUCTION 25 lour hours alter the intradermal injections, the rabbits received a single in- travenous injection of 0.3 c.c. of the filtrate, per kilo of body weight. Four rabbits died shortly afterwards. Nine of the surviving rabbits showed uni- formly severe hemorrhagic necrosis in each of the 6 prepared areas. One rabbit showed no reactions. Analogous results were obtained Avith other batches of identi- cally prepared filtrates. As is seen from the abo\e experiments, it is possible to deter- mine the potency of a given preparation by a reciprocal titration. In order to avoid confluent reactions enough distance (approxi- mately 1 to 11/2 inches) should be allowed between the sites pre- pared. It is inadvisable to give more than six injections into the skin of the abdominal wall, and the amount of material injected should be limited to 0.25 c.c, or preferably 0.2 c.c. From the analysis of the reciprocal titrations it becomes at once evident that there exists a distinct reciprocal relationship between the amoinits of filtrate necessary to be injected intra- dermally and intravenously in order to elicit the phenomenon. Thus, even the highest dilution employed is able to prepare the skin of some rabbits for severe hemorrhagic necrosis pr(j\'ided large amoiuits of the filtrate are injected intravenotisly. As the intravenotis dose is decreased, ^vell-pronoimced reactions are elicited only in areas prepared ^vith louver diliuions. Moreover, it is of interest to point oiU that insufficient intra\'enous doses only very rarely produce ^veak reactions in all the sites prepared. In most instances, the reactions appear of greatest intensity in sites receiving sufficiently large preparatory doses and are totally ab- sent in sites prepared ^vith lower doses. If the preparations employed contain a high phenomenon-pro- ducing potency, the diltuion necessary for elicitation of uniform reactions in all sites prepared may be beyond its lethal effect. In weak preparations the lethal titer may lie very closely to the phenomenon-producing titer and thus, if the necessary concen- tration of the filtrate is used, a high percentage of rabbits may die before the reading of the reaction becomes possible. Pabst and Branham (1933) apparently attempted the above reciprocal titra- tions in rabbits prepared with several injections of the menin- gococcus toxic material into the skin of the abdominal wall as ^vell as into the skin of both ears of the rabbit. As discussed on p. 19 the skin of the ears is considerably more resistant to the phe- nomenon than the abdominal skin. If the intravenotis injection 26 LOCAL TISSUE REACTIVITY of a moderate ainouiil ol reacting fa( tors is made through the marginal vein oi one ol tlie |)rej)ared ears, it may be reason- al)l\ c\j)c'(lc(l lliat most ol iliat material is fixed in the prepared ear and only an excess readies the other sites prepared. This pos- sibility was a])parently resjionsil)le for the highly (onfusing re- sults of titrations rej)orted in the first part of Palost and Branhams' paper. Method II: Titrations of skin-preparatory factors against con- stant intravenous doses: Titrations of skin-preparatory factcMS against constant intra- venous doses consisted of a single intradermal injection of a hltrate undiluted and in various dilutions. Groups of 3 to 5 rab- bits were ordinarily used for each preparatory dose. Twenty-four hours later the rabbits of all the groups received the same intra- venous dose. This may be illustrated as follo^vs: (Shwartzman, 1930^^. i935«) • Protocol 13 : B. typhosus "agar washings" filtrate was used for a single intradermal injection. The following dilutions were prepared: 1:5, 1:10, 1:30. i:()o, 1:120, 1:200, and 1:300. Three rabbits were prepared with each dilution. Twenty-four hours later, 1 c.c. of a dilution 1:10 of the same fil- trate, per kilo of body weight, was injected intravenously. Following the in- travenous injection, 2 to 3 rabbits of each of the groups receiving dilutions up to 1:60 reacted with severe hemorrhagic necrosis. Only 1 to 2 rabbits of the groups prepared with dilutions 1:120 and 1:200 gave positive reactions. Dilution 1:300 gave no reactions. On the ])asis of the above typical experiment, it may be stated that the titers of the skin-preparatory factors can be fairly ac- curately determined. If the residts of this type of titration are compared ^vith the reciprocal titration, it becomes obvious that a considerably smal- ler amount of reacting factors is required in older to elicit reac- tions in one site than Avhen several sites are prepared (approxi- mately ten times less) . This again is taken as further proof as to the reciprocal relationships f^etween the intradermal and intra- venous doses necessary for the elicitation of the phenomenon. Method III: Titration of reacting factors against constant skin-preparatory doses : The quantitative measurement of the reacting factors is con- veniently carried out as follows: INTRODUCTION 2'] The rabbits used for titrations each receix e a single intradermal injection of 0.25 c.c. of the undiluted filtrate or filtrate diluted 1:2 and di\ided into groups of 3. Twenty-four horns later, a single intra\enous injection of the filtrate diluted in 0.85 per cent NaCl solution is gi\'en to each rabbit. The dose is 1 c.c. per kilo of body ^veight. Each group of 3 rabbits receives intrave- nously a different dilution of the filtrate. The local reactions are read foiu" to fi\e hoins after the intravenous injections. The titra- tions are carried imtil the lowest dilution is found which gives no reactions in the 3 rabbits tested as ^vell as the highest dihuion \vhich gives reactions in one or more rabbits of the group. The minimal dose of reacting factors (reacting unit) is then consid- ered as lying between these two figures. If a given filtrate is em- ployed for any length of time, repeated control titrations are done. In these control tests the dilutions employed are both the highest dilution capable of eliciting reactions and the lowest dilu- tion giving no reaction. It may be of interest to point out here that provided one area of the skin is prepared, the amoimt of reacting factors necessary to elicit reactions in the skin may be amazingly small and ^vell beyond its lethal potency. The percentage of positive rabbits with high dilutions depends largely on the batch of filtrate employed and especially on the microorganism used for its preparation. In such titrations, B. typhosus "agar washings" filtrates usually yield 25 to 50 per cent and meningococcus "agar washings" filtrates 25 to 75 per cent of positive reactions. It may frequently happen that after a certain lo\v concentration is reached, a small percentage of rabbits con- tinues to react with minute amounts of toxic substances of a rather wide range. The comparison of percentages of positive results obtained with high dilutions in small groups of rabbits may, therefore, serve as a sotnce of error. It is essential to con- tinue titrations until a dilution is determined which elicits no reactions in all the rabbits of the group tested. Once the "end-point" (one reacting unit) is foimd, several multiples thereof give consistent and regular results in a high percentage of rabbits tested. This is best shown in Table i. Pabst and Branham (1933) entirely confhnied my findings that titrations to the end-point with single prepared sites give reac- tions of high degree of regularity. They also state that 6 reacting luiits gi\e at least 75 per cent of severe reactions. 28 LOCAL TISSUE REACTIVITY Table i TITRATION OF REACHINCi FAC;TOR.S AGAINST CONSTANT SKINT^REPARATORY DOSES Fil- trate A36 989 852 971 1078 A3 4 A38 1139 A32 Aio Strain used Jar preparation of filtrate Ty Tl stock Ty Tl 245 ser Ty 157 stock Ty 159 44D ser.23 (group i) . . 44D mouscie (group i) 44B (group III) 44D ser. 33 (group i) . . , 44D (group i) 44D (group i) Highest dilution oj filtrate eliciting reactions 600 200 1000 1 100 1200 1000 800 1000 300 500 Per cent positive rabbits with highest dilution eliciting reactions 50 50 25 50 25 50 75 50 25 25 Lowest dilution of filtrate showing 800 400 1200 1400 1400 1200 1000 1200 400 700 No. of reacting 2inits per c.c. of filtrate 700 300 1 100 1250 1300 I 100 900 1 100 350 600 Per cent of posi- tive rabbits with various numbers oJ reacting units 6 re- 10 re- acting acting units units 80 80 90 100 70 100 80 80 75 100 90 100 80 80 It may be concluded that the titrations of the reacting factors against a constant skin-preparatory dose (single site prepared) to the end-point afford the method of choice. Dikited materials in- jected intravenously prodtice no lethal effect. Moreover, inasmuch as in all netitralization experiments, to be subsecpiently described, large amounts of the reacting factors are titrated against a con- stant amoimt of serum, the natinal variations in stisceptibility of animals is completely overcome and, therefore, acciuate deter- minations of the neiUralizing potency are made possible throtigh the tise of this method. RECAPITULATION Microorganisms in which hitherto no true exotoxins have been demonstrated were selected for the \vork about to be described in this monograj)h. It \vas assumed that those microorganisms may produce injuries in animal tissue through a mechanism alto- gether different from the direct injurious effect exerted by diph- theria, botidinus, tetanus and other exotoxins. With this assump- tion in mind, the following basic experiments were carried out. Rabbits received a single intradermal injection of B. typhosus culture filtrate followed twenty-foiu' hours later by an intravenous INTRODUCTION 29 injection ot the same filtrate. Four hours after the intravenous injection there appeared severe hemorrhagic necrosis at the pre- pared skin site. In the gross it was dark blue, swollen, ^vith an angry red periphery and histologically it showed disruption of the venules, extensive hemorrhage, thrombosis and necrobiosis of all the cells. The reaction extended from the superficial layers of the skin through the entire thickness of the abdominal wall to the peri tone inn. For the reproduction of this phenomenon it is essential to gi\'e the second injection into the blood stream. Repeated skin injec- tions of the filtrate with twenty-four hour interval between injec- tions does not result in the hemorrhagic and necrotic type of reaction of the phenomenon. The phenomenon can be also produced ^vhen the skin prepara- tion ^vith a given bacterial filtrate is followed by the intravenous injection of filtrates derived from biologically and serologically unrelated microorganisms. A definite interval of time is required bet^veen the preparatory skin injection and the intravenous injection of the filtrate. A t^vo hoiu' interval is insufficient. The state of reactixity appears as early as eight hoins after the skin-preparatory injection and with the filtrates employed in the basic experiments, does not last longer than thirty-t^vo hours. The optimimi incubation period is twenty-foiu' hoins. Primary reactions obtained upon intradermal injections of bac- terial filtrates employed in this work are sharply contrasted by the hemorrhagic reaction following the intravenous injection of potent filtrates. No relation exists between the size and intensity of the primary reaction and the lesion produced upon intravenous injection of reacting factors. There are also discussed in this chapter the susceptibility of xarious sites of the rabbit's skin; the possible spreading of the local lesion to adjacent non-prepared skin sites; involvement of regional lymph nodes; the effect of Reynal's spreading factors upon intensity and size of lesions of the phenomenon; and the possible superadded effect of Julianelle and Reimann's pmpuro- genic substances. The quantitative estimations of potency of active principles of the phenomenon may be made by the use of three methods: Reciprocal titrations of various skin-preparatory doses against \arious reacting doses (Method I) ; titrations of skin-preparatory 30 LOCAL TLSSUE REACTIVITY factors against c:onstant doses (Method II) ; and titration of react- ing factors against constant skin-preparatory doses (Method III) . Method III affords accurate determinations of the plienomenon- producing potency of the material and may be advantageously used for determinations of the neutralizing potency of sera. There exists a strict quantitative reciprocity Ijetween the skin- preparatory and reacting factors. If advantage be taken of the fact and several multij)les of the minimal dose of reacting factors be used, a high incidence of positive restdts may be obtained Avhich may reach loo j^er cent ^vith filtrates of some microor- yanisms. Chapter II PHENOMENON OF LOCAL SKIN REACTIVITY TO VARIOUS MICROORGANISMS FOLLOWING the basic experiments just described, attempts were made, ^vith \arying degrees ot success, to obtain the actixe principles of the phenomenon from a large number of microorganisms. Whilst some bacteria yielded, under simple conditions of cultivation, principles endowed with high skin- preparatory and reacting potencies, others prodticed the prin- ciples in \o\v and variable concentrations. In order to demon- strate the active principles in the latter group, it x\as necessary to take advantage of the non-specificity of the phenomenon and combine the intradermal injection of preparations of prede- termined high potency xvith intravenous injections of prepara- tions of bacteria imder investigation. Some groups of organisms failed to elicit the phenomenon in spite of repeated attempts. As will ])e seen from a later chapter, the knowledge concerning the piiysico-chemical properties of the active principles still remains meager. It may be achisable to postpone the experiments with these microorganisms for the time when more information is ac- ciunulated concerning the natine of the acti\e principles. ACTIVE PRINCIPLES OF MENINGOCOCCI S In attempts to reproduce the phenomenon \vith meningococcus culture filtrates a great variety of fluid cultines were first used. Media consisted of various combinations of tryptic digest broth with glucose, whole blood, blood sertim and ascitic fluid and also of combinations of plain broth of various pH xvith glucose, whole blood, blood serum and ascitic fluid. The period of inctiba- tion varied from one to seven days. The cultures were subjected to alternating freezings and tha^vings rej^eated at frequent inter- vals of time. Materials obtained in this manner were titrated for skin-preparatory potency. It xvas found that the filtrates were caj)able of eliciting the state of reacti\ity proxided a dilution not higher than 1:6 or 1:8 was used for skin preparation and as 31 '^,2 LOCAL ILSSLE REACIIVLIY inucli as 3 c;.c. pei kilo ol body wcighl, ol llic same material was injected intravenously. When the preparations ^vere re-titrated after short intervals ot time, in many instances there was observed either a partial or comjDlete disappearance of the potency. The irregularity of the results with the above descriljed methods of ]:)rcparation, whilst demonstrating the possil:)ility of eliciting the phenomenon, necessitated the development of a more reliable jjrocedure. With this purpose in mind, the "agar washings" ( ul- ture filtrates of meningococc us described below ^vere used. Meningococcus "agar washings" culture filtrates: Meningococcus is inoculated into one per cent rabbit blood broth of pH '7.2-7.4. After t^venty to twenty-two hours' incuba- tion, the supernatant broth cidtme, free of red blood cells, serves as the inoculum. Three to 4 c.c. of the inoculum are pomed on the surface of 0.7 per cent glucose veal infusion agar in Kolle flasks. After twenty to twenty-two hoins' incid)ation, the growth of each flask is washed off 'vvith 2-4 c.c. of 0.9 per cent NaCl solu- tion containing 0.4 per cent phenol. The washings are then pooled, centrifugated within the following one to two hours and the clear supernatant fluid is filtered through Berkefeld "V" can- dles, shortly after centrifugalization (Shwartzman, 1929/;) . In vie^v of the possibility of contamination during the process of washing of the agar it seems advisable to pool the washings of not more than 5 flasks, make culture controls on blood and glucose agar plates and filter the collected ^vashings of each 5 K(jlle flasks through a separate candle. The following day the filtrates of washings proxen as imcontaminated are pooled. These precautions are especially warranted in the prej^aration of men- ingococcus toxic substances because of more frequent contamina- tions observed with this microorganism. The reasons for the occurrence of the contamination is probably due to the use of rabbit blood in the inocidum. The sterility of the filtrates shoidd be controlled by subcultures on enriched media and incubated for at least forty-eight hours, inasmuch as meningococcus may take this period of time for its appearance. If the xvork is done cm masse and many candles are used for filtration, whereby it may pro\e diflicult to check the permeability of the filters, it is advis- able to heat the fUtrates at 56° c. for fifteen miniues. thus de- stroying the meningococcus cells that may pass through the fdters. This temperature does not appreciably reduce the toxic titer. REACTIVITY TO VARIOUS MICROORGANISMS 33 In tlie Foregoing maiinci, highly potent prep:nations were ob- tained Irom meningococcus ot xarioiis serological groups. Several straiirs ol each group Avere tested. All the materials pre}:)ared since i();o were titrated, as a routine, lor reacting potency against a constant skin-preparatory dose (Method III) . In this test a single site was prepared ordinarily by an injection ot 0.25 c.c. ot filtrate diluted 1:2. The tollowing is the range of the reacting potency of numerous preparations derived from \arious serological groups of meningococcus: Group 1, strains 44D and 44F — 500 to 1500 reacting units; Group 1, strains 397, 130, and 401 — 400 to 700 reacting units; Group 2, strain 383 — 400 to 1350 reacting units; Group 3, strain 44B — 600 to 4000 reacting units; Group 3, strains 39() and 101 — ■ 500 to 1000 reacting tuiits. No extensive titrations were carried out on the skin-prepara- tory potency of meningococcus "agar ^vashings" filtrate. The im- pression \vas, howexer, gained that dilutions not higher than 1:25 should be employed for rabbits receiving 25 reacting tmits intra- venously in order to obtain not less than 75 per cent of positive results. Most of the work in the past years xvas carried out with the strain 44B and on frequent occasions, titers as high as up to 4000 were obtained. From the analysis of the residts just recorded, it may be safely stated that Group 3 strains yield toxic materials of the highest potencies xvhilst strains of Groups 1, 2, and 4 fluctuate within the limits of a lower range. As will be discussed again later, the selection of the strain is a highly important consideration if one wishes to obtain potent materials. In addition to the above enumerated strains a number of cultures recently isolated from sj)inal fluid of meningococcus meningitis cases xvere tested for their ability to produce the active principles. In all, about 15 various strains were employed. No serological identification xvas done. The average titers obtained ranged between 200 and 2400 reacting luiits, per 1 c.c. of "agar xvashings" filtrate. Pabst and Branham (1933) also found that with some "agar washings" filtrates of meningococcus pronoiniced hemorrhagic necrosis xvas obtained in dilutions as high as 1:800 xvhilst xvith others the activity was much less. In their experiments there seemed to be no relation between the activity of the filtrates and serological grotiping though some indi\'idual strains regularly produced more active filtrates than others. Length of time during 34 I-OCAI, TISSliE REACTIVITY which the stock cultures weic niaintained in the hihoiatoiy had no ic'lalioii to the potency ol the active ])riii(i|)les jMochiced. Variations in {potency Aseic lound ainc)n<> (Hrterent strains of the same group and in various lots ot filtrates. The incidental observa- tion which seems worthy of mention is that maintenance of men- ingococcus strains on egg medimn may bring about occasional loss of ability to produce active principles. Whilst the efficiency of meningococcus "agar washings" filtrates was repeatedly corroborated by a large group of ^vorkers (Sickles, 193^5; Powell and Jamieson, 1931; Pabst and Branham, 1933; Apitz, 1935/^; Witebsky and Neter, 1936) , others (Burnet, 1931; Malcolm and White, 1932; and Riley and Wilson, 1932) studied the potency of materials prepared t^y a variety of methods, i.e., autolyzed cultures of Besredka (1906) ; defatted meningococ- cus antigen; heated antigens; antigens treated with alkaline and acids; washed suspensions of meningococcus cells, etc. Although no quantitative determinations of the active principles were made with the latter grotip of materials, it is suggestive that all these pre])arations. \vith the exception of washed cells, possess a certain amotmt of actixe j^rinciples. ACTIVE PRINCIPIF.S OF CiONOCOCCUS In my early experiments I succeeded in reproducing the phe- nomenon of local skin reactivity to gonococcus with filtered saline washings of stock strains of gonococcus grown on North medimn, liver hormone agar pH 6.8, and ascitic fluid 1 per cent glucose agar pH 7.2. The suspensions were centrifuged at high speed and filtered through Berkefeld "V" candles. Rabliits prepared by a sin- gle intradermal injection of 0.25 c.c. of the undiluted material developed severe reactions following the intravenous injection of 1 c.c. of filtrates diluted as high as 1:25. ^^ fmther \vork was done in order to select the method of choice. The preparations were stable and were acctnately titrated to the end-point by Klein (1933) in my laboratories, for his neutralization studies, described in a later chapter. It is interesting that several batches of gonococcus \accines (New York City Board of Health) , Avhich were washed cell sus- pensions killed by heat, failed to elicit the phenomenon Avhen used undiluted for intradermal and intravenous injections, i.e.. in quantities at least twenty-five times larger than those neces- REACTIVITY TO VARIOUS MICROORGANISMS 35 saiy for elicitatioii ot the phenomenon with the above described "aoar ^\'ashinos" filtrates. Cohn (1931) failed to obtain any reactions \vith filtrates of 10 per cent serimi broth cultures incubated for five to seven days. In contrast, filtered washings of young blood agar cultures gave strong reactions in 20 per cent of a small group of rabbits tested. In recent investigations Morimoto (1935) washed w'ith saline twenty-four hour old cultures of gonococcus on blood agar. The washings were heated at 60° c. for thirty minutes, frozen at 70° c. for twenty-four hours and incubated at 37° c. for t^venty-four hoins. The procedine of freezing and incubation was repeated foiu' times. The results were positive. Hormone broth cultines also yielded reactions. Older ctilttnes seemed to be more active than young ones but no cpiantitative titrations ^vere reported. This author filtered his material through Seitz discs. It should be pointed out here that the discs have been sho^vn by Binnet (1931) and others to reduce the potency of the principles. ACTIVE PRINCIPLES OF B. TYPHOSUS In the early experiments described in the introductory chapter, active principles were deri\'ed from tryptic digest broth ctiltures of a suitable strain of B. typlwsus. The skin-preparatory potency of these materials titrated as high as 1:64 per 0.25 c.c. provided at least 1 c.c. of the same undiluted material Avas employed for the provocati\e injection. The reacting potency of filtrates of tryptic digest broth cultines titrated against a single site prepared with 0.25 c.c. of the imdiluted filtrate fluctuated between 50 to 200 units, per 1 c.c. (Method III) . Plain broth cidture filtrates yield poor preparations. In recent work several batches of filtrates of three day old plain l^roth cid- tures contained only 10 reacting iniits per 1 c.c. (Sh^vartzman, i935«) • Inasmuch as the abo\e materials, in addition to low potency, also contained a large amount of extraneous substances derived from the bacteria and the medium, the "agar ^vashings" filtrates of B. typhosus were unquestionably the preparations of choice. B. typhosus "agar washings" filtrates: The "agar washings" filtrates are prepared essentially in the same manner as those of meningococcus (Shwartzman, 1929^) . Kolle flasks containing plain \cal infusion agar of pH 7.4 are 36 LOCAL TISSUE REACTIVITY seeded each \vith 3-4 c.c. oi t\vcnty liour old plain broth cultures of B. l\l)lio.sii.s diluted i:.} with o.t) }:)er cent NaCH solution. The dilution is made innnediately heiore use. After twenty to twenty- two horns of incid)ation the growth of each flask is Avashed off \vith 2-4 c.c. of 0.9 per cent NaCl solution containing 0.4 per cent j)henol. The washings are then pooled, centrifuged within the following one to t\vo hotns and the clear supernatant fluid filtered through Berkefeld "V" candles shortly after centrifuging. Numerous batches of "agar washings" filtrates of B. typhosiis W'Cre titrated for reacting potency in rabbits prepared by a single intradermal injection of 0.25 c.c. of midiluted filtrate, dining the past six years. Most of the work Avas done ^vith a strain Tl iso- lated some years ago from the stool of a typhoid fever patient. The strain showed colonies on plain agar which appeared dis- tinctly granular under the dry lo^v power lelis. Their margin a\ as ameboid and some^vhat ragged but this ^vas not sufficiently pro- nounced to classify this strain as rough. This intermediate type of roughness, is however, a stable characteristic since it persisted for a nimiber of years ^vithout any change either in the direction of smoothness or more pronoimced roughness. The stability Avas ascertained by 42 "platings" of this strain. No typical rough col- onies Avere obtained from old fluid cultures. Broth cultmes showed, in addition to the luiiform cloudiness, an insignificant amount of precipitate Avhich on shaking gave rise to easily broken up flakes. Suspensions in 0.85 per cent saline incid)ated in a Avater bath at 37° c. for two hours Avere only slightly agglutinated. The remaining strains of B. lypJwsus described below Avere typically smooth. The folloAving is the range of the reacting potency of various batches of filtrates derived from B. typ]wsus. Strain Tl — 100 to 1000 reacting units; strain 1684 — 50 to 100 reacting luiits; strain 1687 — 25 reacting units; strain 837 — 150 reacting units; strain 864 — 75 to 250 reacting luiits; strain 157 — 100 reacting units; strain 240 — 400 to 600 reacting units; strain 159 — 600 to 1000 re- acting units; Typhoid 0-901 (Felix) — 300 reacting imits; Ty- j)h()id H-901 (Felix) — 900 reacting units. As may be seen from the above, under ajjparently identical conditions various strains of B. typJwsus may substantially difTer in their ability to produce active principles of the phenome- non. The strain Tl Avhich Avas successfidly employed for produc- tion of actiAe principles for a period of eight years recently com- REACTIVITY TO VARIOUS MICROORGANISMS 37 pletely lost its activity. A number of repeated attempts to obtain active preparations from this strain for a jjeriod of six montlis made since the time this was first noticed gave only negative results. These attempts included passages of the strain through mice, modifications in meditmi composition hydrogen-ion con- centration, etc. Ecker and Welch (1930), Apitz (1933^), Gerber (iggOrt), Stolyhwo (1935), and others were afjle to elicit strong reactions with the above described "agar washings" filtrates of B. typhosus. Burnet (1931) apparently obtained active preparations in cul- ttnes autolyzed by the method of Besredka. ACTIVE PRINCIIM.ES OF B. PARATVl^HOSUS In early experiments the phenomenon was elicited with filtrates of ctdtines of B. paidtyphusus A, B, and B. enteriditis grown in tryptic digest broth for six days (Sh^vartzman, 1928a) . Amotnits as large as 2 c.c. per kilo of body ^veight, ^vere used for the provocative injection. Ecker and Welch (1930) obtained active preparations from B. paidtyphosus A, B. aertrycke, B. enteritidis, B. paratyphosus B, B. aertrycke (Nobele) , and B. paratyplwsus (Schottmiiller) in the synthetic meditnn of Ecker and Rimington (1927). In my recent experiments "agar \vashings" filtrates of B. enteritidis isolated from a spontaneous epidemic in mice yielded preparations containing ()00 to 800 reacting tuiits per 1 c.c. ACTIVE PRINCIPLES OF B. COEI Active principles of B. coli may be easily obtained in plain broth and washings on j^lain agar ctdttires. B. coli plain broth culture filtrates Avere used successfully by Gratia and Linz (1932c), Ecker and Welch (1930), Duran-Reynals (1933/0' P- Bordet (1936/;) , and others. I carried out comparative studies on plain broth and "agar washings" filtrates. Plain broth cultures were incubated for three days, centrifuged, and filtered through Berkefeld "V" candles. Four-tenths per cent of phenol \vas added for preservation. Sev- eral batches of broth culture filtrates contained bet^veen 40 to 60 reacting units per 1 c.c. The "agar washings" filtrates Avere made from cultures on Kolle flasks containing plain agar, pH 7.4, in- cubated for twenty hours. The growth of each flask ^vas ^vashed off with 2-4 c.c. of phenolized saline. The filtrates thus obtained ^8 LOCAL TLSSUE REACTIVITY xaiicd in poteiu) in ihc range lioni 50 to i^'^oo reading units per 1 ( .( . Certain strains which were typical B. coli isohited from urine and stool, yielded more active preparations than others. Moreover, various batches derived from the same strains gave preparations of varying potency. Apitz (iQ^g/;) employed for preparation of "agar washings" filtrates several strains of B. coli isolated from cases of cystitis. His preparations never gave intense hemorrhagic reactions. The aboxe mentioned extreme fluctuations in potency of active prin- ciples, partially depending on the strain employed, may have been responsible for the weak reactions obtained by Apitz. ACTIVE PRINCIPLES OF B. FRIEDI.ANDER For these experiments B. jriedUuider of certain serological groii})s Avere used. In 1926 Julianelle (19,^^0) classified B. jriedldnder by the im- miuiological reacticjns of agglutination, agglutinin absorption, and passive protection against infection into three specific types, A, B and C. These types comprised the great majority of strains and the remaining cultines were included in a heterogeneous group. The carfjohydrate or solidjle specific substance was shown to be chemically different for each of the three types. Julianelle re- ported the following statistical data resulting from a study of 80 strains: A — 52 per cent, mostly human infections; B — 15 per cent, animal soince; C — 9 per cent; x — 24 per cent. In my experiments plain broth cultures of B. jriedldnder, A and C \vere incubated for three days and after centrifugalization, filtered through Berkefeld "V" candles. Strain A filtrates con- tained approximately 10 reacting units per i c.c. whilst strain C yielded a filtrate containing 50 reacting" units per 1 c.c. The reactions were severe and obtained in a high percentage of rab- bits tested. Apitz (1935^-') obtained active principles of Fried- lander bacillus in "agar washings" filtrates. It appears from his description that the filtrates were of a comparati\ely low potency. ACTIVE PRINCIPLES OF B. DVSENTERIAE Filtrates of "Agar washings" and broth cultmes of variotis strains of B. dysenteriae possess active principles (Shxvartzman, 1928/^) . On one occasion the filtrate of a Sonne strain contained as many as 1500 reacting iniits, per 1 c.c. Recently Schneierson, in these laboratories, tested a large group of B. dysenteriae Shiga REACTIVITY TO VARIOUS MICROORGANISMS 39 j)reparati()ns. "Agar washings" filtrates contained 150 reacting imits per 1 c.c, and egg albumin broth culture filtrates prepared according to the method of Olitskv and Kligler (1920) , contained 150 reacting luiits per 1 c.c. Endotoxin also ])repared according to the method of Olitsky was active in dilution 1:150. Burnet (1931) elicited the phenomenon of local skin reac- tivity to B. dysenteriae with autolyzed preparations of Besredka. Thus far, it remains unknown, however, whether there exist any differences in the phenomenon-producing potency of the various strains of B. dysenteriae. AitofT, Dion and Dobkevitch (19360) reproduced the phenom- enon with active principles derived from a strain of B. dysenteriae of the new-born recently described by them. The preparations tested were as follows: (1) a suspension in distilled water of a twenty-four hour old agar culture left at room temperature for six days and then filtered through a Chamberlain L3 filtrate ("lysate") ; and (2) a filtrate of a fifteen day old culture in Ramon's medium ("toxin") . The phenomenon could be elicited when the "lysate" was used for both skin-preparatory and provoc- ative injections; \vhen the preparatory injection of "toxin" Avas followed by the provocative injection of "lysate"; ^vhen a pre- paratory injection of B. coli culture filtrate was followed by provocative injection of "lysate"; and when the preparatory in- jection of "lysate" was followed by the intravenous injection of "toxin." \V'hen "toxin" was used for both preparatory and provoc- ati\'e injections no reactions ^vere obtained. ACTIVE PRINCIPLES OF B. PRODIGIOSUS According to Bertarelli (1903) some strains of B. prodigiosus may exhibit considerable virulence. Aitoff, Dion and Dobkevitch (1936&) isolated from the sputum of a case of pneumonia B. prodigiosus pathogenic for mice, rats, guinea pigs and rabbits. This strain induced septicemia and the organism could be re- covered from the heart's blood of an injected mouse. "Lysates" and "toxins" prepared according to the method described by these authors in connection with their Av-ork on B. dysenteriae, elicited the phenomenon. The lesions obtained with the "lysates" were circumscribed whilst the "toxins" gave very prompt and diffuse reactions. In one instance petechiae appeared almost immediately after the intravenous injection was completed. 40 LOCAL TLSSUE REACTIVITY ACIINI I'RINCII'I.KS Ol I',. Ml iriENSIS AM) B. ABORITS 111 i<)2(), I attempted lo rcpiodiuc ihe j)henomenon of local skill iea(tl\iiy to B. 'nicliloisi.s without success. In these experi- ments there were emjiloyed tiyptic digest ascitic fluid broth cul- tiues incubated for six days, aerobic and anaerobic double strength broth cultures and "agar Avashings" filtrates. Kritschewski and Halperin (i9,'54) also failed to rejDroduce the phenomenon to B. aborliis. They used three day old cultures in plain l:)roth and injected 0.5 c.c. intradermally and 1 c.c. intra- venously, of this material. In the same rabbits they were able, however, to indtice the phenomenon with B. lyl)Ji()sii.s t^venty- four hour old broth cidtures, thus proving that the failure was due to lack of the active principles in B. abortus cultines and not to resistance of the rabbits used to the phenomenon. /.uwerkalow, Fischbein and JuranofT (1935) carried out a series of investigations on the active principles of B. abort us. Filtrates of six to twenty-one day old cultmes gave negative residts. Con- sidering my residts with meningococcirs ^vhereby it was shown that there may be encoiuitered diHiculties in obtaining the active j)rinciples in ffuid medium, they inidertook experiments with "agar washings" filtrates. Thirty-six to forty-eight hour old cid- tures of B. abortus of strain Schidgino on liver agar of pH 7.3 were suspended in saline (in proportion of i to 2 c.c. of saline to a culture) . The strain Avas recently isolated and failed to gro^v in the absence of a paraffin seal. Rabbits were prepared by four simultaneous intradermal injections of 0.25 c.c. and reinjected intravenously with 8-10 c.c. of the same material t^venty-four hours later. Four hours after the intravenous injection there ap- ])eared a typical reaction which somewhat extended twenty-four hours later. Healing was very slow. The phenomenon \\as elicited in about 50 per cent of animals tested with the afore-described material. After maintenance, the strain lost its ability to produce the active principles. In addition, one more strain of B. abortus yielded potent preparations whilst tAvo other strains gave entirely negative results. ACTIVE PRINCIPLES OF VARIOUS MEMBERS OF HEMORRHAGIC SEPTICEMIA GROUP In 1929, filtrates of B. suisepticiis and B. avicida cultured for six days in tryptic digest broth and on agar for t^venty-four hours REACTIVITY TO VARIOUS MICROORGANISMS 41 were found by me to yield active principles ot the phenomenon. No end-point titrations of the materials were made at the time. Siiwartzman (\c)2Sa) , and Hanger (1928/;) employed filtrates of se\enty-two iiour old broth cultures of B. lepiscpticus. These fil- trates gave a strong primary reaction, apparently dtie to a state of hypersensitixeness to B. Icpisepticiis in the rabbits employed by this atithor. Intravenous injection of the same filtrates inten- sified the primary reactions. Hanger states emphatically that he ne\er observed this phenomenon except among rabbits natu- rally allergic or rendered allergic by a previous infection. Al- thotigh his report appeared folloxving my publication on the phe- nomenon of local skin reactivity to bacterial filtrates and his experiments with this preparation xvere made according to the technicjue described in my paper, the above statement makes it questionable xvhether Hanger's material derived from B. Icpi- septicus contained the active principles of the phenomenon (pos- sibly because of the use of plain broth) . It is indeed clear that there exists no relationship between the allergic state of the ani- mal and its ability to respond to the phenomenon of local skin reactivity to bacterial filtrates. It suffices to mention here that with some preparations, i.e., meningococcus, the reactions may be obtained in 100 per cent of rabbits, as judged by experiments on many thousands of these animals. Moreover, no stich depend- ence on the allergy of the animals was observed by P. Boquet (i9'^5) in his successful attempts at reproducing the phenomenon with B. lepisepticus. This author grew the organism in Martin bouillon for six days and filtered the culture through infusorial earth and Chamberlain L^ filters. Rabbits were prepared by an intradermal injection of 0.5 c.c. of the filtrate and tested by an in- travenous injection of 2 c.c. per kilo of body weight. Yoiuig cul- tures were inactive. If cultures were heated at 50° c. for one hoiu' before filtration, the filtrates appeared to be of a higher potency. The filtrates apparently contained both preparatory and provoca- tive factors. Combined intradermal injection of the B. lepisep- ticus filtrate with the intravenotis injection of B. coli filtrate or a combined preparatory injection of B. coli filtrate w'lxh provoca- tive injection of B. lepisepticus filtrate also yielded positive results. Apitz (i935«) employed a strain of B. pasteurella cuniculicida which somexvhat differed from Hanger's strain of B. lepisepticus. The author emphasizes the fact that in his experiments intracu- 42 LOCAL TISSUE REACTIVITY taiK'oiis injcdion ol I he filtrate rei^iilarly pioduccd a severe kjcal iiillaniniation, \vlii(li in this respect, sharply differed from the insi^nilicant primary effect produced by active principles of other bacterial species. Jotikow-Werejnikow aird Lipatova (1933-34) were the first to reproduce the phenomenon of local skin reactivity to B. pestis. The cultures were grown on plain agar, pH 6.6 at 37° c. for thirty-six hours. LInder these conditions of cultivation, as shown by these authors, the organisms developed abimdant capsular ma- terial. The growth was washed off in distilled water and frozen and thawed ten times. The rabbits were prepared by intradermal inoculation of 0.25 c.c. of the material and twenty-four hours later, received an intravenous injection of the same. According to the description of these authors, the reactions were of extreme severity following the intravenous injection. If the preparatory injections ^vere made next to a large vein of the skin of the abdominal Avail, there also occtnred a swelling of the vein, in addition to the typical reaction at the prepared site. Subsequently a large portion of the vein became necrosed. ACTIVE PRINCIPLES OF VARIOUS MEMBERS OF HEMOGEOBINOPHILIC GROUP Frisch (1930) elicited the phenomenon to B. infiiieiizae with filtrates of cultures of the organism on chocolate agar. The po- tency of the filtrates was not determined. L. Gross (1930) observed the phenomenon of local skin reac- tivity to B. pertussis. The filtrates were made from washings of growth on chocolate agar Kolle flasks. The concentration of the ac- tive principles wms never of high potency and seemed to vary for unknown reasons. In the best preparations not more than 50 re- acting units per 1 c.c. were obtained. Mishulow, MoAvry and Scott (1930) modified the method of preparation of these active prin- ciples and seemed to have obtained positive reactions in a large percentage of rabbits. They used the follo\ving method: Cultures were grown on horse blood chocolate agar and trans- planted at twenty-four hour intervals for two or three generations. Blake bottles of chocolate agar were inoculated with growth from the slants (one slant for each bottle) and 4 c.c. of 1 per cent horse serum beef heart broth, j)H 7.2 to 7.4, were added to each bottle. Bottles were incubated at 37^^ c. for tAventy to t\venty-two hours. The growth was washed off with a buffer solution (47 gms. REACTIVITY TO VARIOUS MICROORGANISMS 43 crystal borax, 84 gms. boric acid, 99 gnis. NaCl dissolved in 16 liters of distilled water) containing 0.4 per cent phenol, using 5 c.c. of the solution iov each bottle. The bacterial snspension Avas centrifnged at once at high speed initil clear. The snpernatant fluid was filtered through a Berkefeld "V" filter and tested for sterility and when foiuid to be sterile, injected into rabbits. In recent experiments of Koplik (1934) in my laboratories, there ^\as employed brain medium prepared essentially according to the method of Toomey and McClelland (1933) with vari- ous peptones (neopeptone, proteose-peptone (Difco) , \\' itte's pep- tone) . The H-ion concentration ^vas adjusted either to pH 7.3 or 7.8. The media were seeded ^vith t^venty-four hour old cultures on chocolate agar slants. The strain used was subctiltured for at least three successive days preceding these inoculations. Berkefeld filtrates of four, seven, and seventeen day old brain medium cultures ^vere tested in rabbits for the elicitation of the phenomenon. Twenty-five c.c. of undiluted filtrate ^vere injected intradermally into rabbits and followed in t^venty-four hours by a single intravenous injection of the same filtrate undiluted or in varying dilutions. Comparative studies of the brain cidture fil- trate with chocolate agar culture washings Avere made. The re- stilts were described by him briefly as follows: It was possible to elicit the phenomenon under discussion in prepared rabbits by an intravenous injection of the brain filtrate in dilutions ranging from 1:40 to 1:400 and recently 1:1500. Some of the brain media culture filtrates when retitrated two to four months after preparation, showed a rise in efiiective titer of 5-20 fold. The optimum incubation period for brain medium cultures seemed to be about seventeen days, ^vhile four and seven day cultures yielded filtrates of lower potency. No differences ^vere noted with the various hydrogen-concentrations in the range em- ployed (6.8 to 7.8) . Neopeptone and proteose-peptone (Difco) appeared superior to Witte's peptone. Witebsky and Salm (1937) found that the phenomenon could be also produced by intradermal and intravenous injections of suspensions of live cultures of B. influenzae on Levinthal medium and suspensions killed by heating at 60° c. for one hour. Similarly, B. typJiosus "agar washings" filtrates elicited reactions in skin sites prepared with heat killed B. itifJuenzae whilst no re- actions were obtained in sites prepared \vith B. pertussis. Con- 44 LOCAL TISSUE REACTIVITY \erscly, B. influenzae and B. perhissis when injected intravenously, elicited reactions in rabbit skin prej)ared twenty-four hours pre- \iously with B. Ixphosus "agar washings" filtrates. ACTIVE PKINCn'LKS OF VHiRIO CHOLERAE In H).'^o, I elicited the tyj)ical reactions with filtrates cjf cultures ol cholera \^ibri() on ])lain agar, pH 7.8; 0.25 c.c. of undiluted material was used for the intracutanecjus injection and 1 c.c. of the same luidiluted material for the intravenous injection. Ciratia and Linz {icy^^^c) worked with filtrates of six day old j)lain brotii cidtines. Tiie material was employed for reproduction of tiie jjhenomenon in guinea pigs. Three-tenths c.c. of the filtrate was used fcjr the jireparatory injection and i c.c. for the intra- cardial injection, twenty-fom' horns later. Saburo Uyeda (1934) studied the effect of filtrates derived from J' i brio cJiolerae cultures groAvn on plain agar slants for 24 hours. The growth was suspended in 0.5 c.c. saline solution (5 c.c. to an agar slant) . The emulsic^ns were then centrifuged and the super- natant fluid filtered through a Berkefeld "V" candle. Six out of 8 strains yielded potent preparations. Apparently, large amounts were necessary for elicitation of the reactions. Similar prepara- tions made from cidtures of Vibrio El Tor, Vibrio melchiiikovi, Vibrio albensis and Vibrio tyrogenum were also able to produce the phenomenon. Cidtme filtrates of Vibrio a(j)ialilis proved in- active. Vassiliadis (1935) employed filtrates of six day old broth cid- tures of Vibrio cholerae and the El Tor vibrio for the elicitation of the phenomenon both in rabbits and in guinea pigs. He w^as able to elicit reactions in guinea pigs with either of the prepara- tions. Whilst cholera vibrio cidtme filtrates produced the phe- nomenon in rabbits, those of El Tor vibrios failed to do so. In a recent investigation Linton, Singh, and Seal (1935) were also successful in producing the phenomenon ^vith active principles of Vibrio cholerae. Twenty-four hour old broth cultures served lor inocidation of Roux flasks containing agar in amounts of 10 to 12 c.c. to a flask. Twenty hour old gro^vth was scraped into the condensation water, filtered through Kieselguhr and Chamber- lain L3 candles. (Seitz E. K. discs removed the active principles.) The phenomenon ^vas elicited by an intradermal injecticMi of 0.25 c.c. and by an intravenous injection of 1 to 1.5 c.c. The lesions were extremely severe, healing requiring about six weeks. REACTnilY TO VARIOUS MICROORGANISMS 45 Actixe principles were ol)tained l)y these authors from strains of Group 1 and 2 (cholera \ibrio) , Grotip 3 (xvater vibrio) , and Group 4 (El Tor \'ibrio) . There xvere obserxed, howexer, (jnan- titative differences in the amotmt of active principles obtained from xarioiis strains. The El Tor \ ibrio consistently gave small size reactions xvhen used for skin preparation. The latter filtrates employed intraxenotisly gave rise to sizable reactions only at sites prepared xvith J'ibrio cholerae filtrates of the other strains. ACTIVE PRINCIPLES OF STAPHYLOCOCCUS Bin net (1931) , Gratia and Linz (i932r) , and myself failed to elicit the phenomenon of local skin reactixity to filtrates of plain broth and "agar washings" of staphylococcus. H. Gross (1932) studied the primary necrotizing effect of sta- phylococcus filtrates previously described by Lingelsheim, Neis- ser and Wechsberg, Kraus and Pribram, and H. Gross, Parker, and Burnet (1929, 1930) . In these tests lixe cultin^es of hemolytic strains of Staphylococcus aureus are first injected into the skin of guinea pigs and rabbits for a presumptive test. The live cultures capable of producing necrotizing lesions one to txvo days later are selected for the preparation of toxins xvhich are best obtained in eight to fourteen day old broth culture filtrates. Filtrates incapable of producing the primary necrotizing lesions xvere tested by H. Gross for the phenomenon of local skin reactivity. Txventy-five c.c. of the filtrate xvere injected intraciUaneously into rabbits. These injections xvere folloxved by intravenous injections of 2 c.c. of the same filtrate, txventy-four hours later. Sexere hemorrhagic and necrotic lesions appeared at skin sites prepared xvith these filtrates. Alechinsky (1936) xvas able to elicit the phenomenon xvith a filtrate of a txventy day old culture of StapJiylococciis aureus iso- lated from X esicles of a case of herpetiform dermatitis of Duhring. The active principles xvere apparently of loxv potency since in order to obtain sexere reactions it xvas necessary to employ a B. coU filtrate either for the preparatory or the proxocatixe injec- tion. The intradermal injection of the staphylococcus filtrate of this author produced no primary necrotic lesions. ACTIVE PRINCIPLES OF STREPTOCOCCUS In order to facilitate the studies on the mode of preparation of streptococcus active principles, advantage xvas taken of the fact that the skin prepared xvith the factors from one microorganism 46 LOCAL TLSSUE REACTIVITY is al)lc to ic;ut to the intravenous injection ol these substances fioni other lunehited microorganisms. Reactions with live cultures of Streptococcus hemolyticus scarlatinae: In the followino exj:)eriments the rabbits were given intrader- mal injections of filtrates of B. typJiosus and twenty-foin^ hours later they were given intravenous injections of cultmes of Strep- tococcus hemolyticus scarlatinae prepared by various methods. Strain 4014 of Streptococcus hemolyticus scarlatinae was grown in various fluid media for twenty-four hours and as such injected intravenotisly into the prepared rabbits in a dose of 3 c.c. per kilo of body weight. The fluid media were plain broth of pH 7.6, plain broth of pH 7.6 containing various amounts of dextrose (from 0.1 per cent to 2 per cent) , tryptic digest broth of pH 7.6 and tryptic digest broth containing various amotuits of dextrose (from o. 1 per cent to 2 per cent) . The various cidtures were made aerobically {i.e., 100 c.c. of fluid in a 500 c.c. Erlenmeyer flask) , anaerobically {i.e., luider petrolatum seal) and partially anaerobically {i.e., 100 c.c. of fluid in a 125 c.c. Erlenmeyer flask) . Live broth cultmes ^vere able to elicit severe hemorrhagic necrosis in the B. typhosus prepared skin sites. One hundred rabbits were tested with these ctilttnes. Of these, 66 gave distinctly positive re- actions, one a doubtful reaction and the remaining^ rabbits nesfa- tive reactions. Examination of the individual groups of rabbits did not demonstrate any obvious relationship between the mode of cultivation and the resiUts obtained. Possibly, however, par- tially anaerobic cultures in plain broth of pH 7.6 containing from 0.5 per cent to 1 per cent dextrose and incubated at 37.5^ c. for twenty-four hours gave the best results. In an additional group of experiments there was compared the reacting potency of live cultures of various strains of Streptococ- cus Jiemolyticus grown in 1 per cent dextrose broth under partial anaerobiosis at 37.5° c for twenty-foin- hotns. Marked differences in reacting potency of strains '^vere noted. Reactions with toluene-killed cultures: supernatant fluid of centrifuged cultures and filtrates of cultures of Streptococ- cus hemolyticucss arlatinae: Experiments were made in order to determine ^vhether the presence of live streptococci ^vas essential for the reacting potency REACTIVITY TO VARIOUS MICROORGANISMS 47 of the cultures. Streptococci grown in various fluid cul tines, as described previously, ^vere killed by toluene or removed by cen- trifugation or removed by filtration through Berkefeld "V" can- dles. The Berkefeld filtrates of cidtines elicited severe reactions in 22 out of 44 rabbits tested (50 per cent) . Toluene-sterilized cidtures, as well as the supernatant fluid of centrifugated cultures, gave reactions in 26 of 53 rabbits tested (also about 50 per cent) . As may be seen from these results, the presence of live organisms in the material injected intravenously is not necessary for the production of the phenomenon of local skin reactivity to Strep- tococcus Jieniolyticus scarlatiuae. It may be also seen that the streptococcal reacting factors are filtrable bacterial products. As live cultines sho^\ed reactions in 66 per cent of the rabbits, it ^voldd appear that some loss in the potency of these factors occins when the streptococci are killed or removed. Attempts ^vere also made to increase the yield of toxic sub- stances in the filtrates by previously adjusting the hydrogen-ion concentrations of the cultme to various levels before filtration. These procediues had no apparent effect on the potency of the filtrates. A series of experiments in \\'hich the skin of rabbits w^as pre- pared by the injection of filtrates of cultures of the streptococcus and of toluene-killed cidtures and in which the same material was injected intravenously, gave entirely negative results (Shwartz- man. 1931^) • Reactions v^ith miscellaneous strains of Streptococcus hemo- lyticus: Strains of Streptococcus hemolyticus pyogenes isolated from various conditions, i.e., sinus thrombosis, mastoiditis and skin in- fections, ^vere grown in either 0.2 per cent glucose l^roth for pe- riods of t^vo to six days or brain medium of Toomey and Mc- Clelland (1933) for periods ranging from nine to twenty days. The skin of rabbits ^vas prepared by potent heterologous filtrates and the various streptococcus preparations tested for reacting po- tency. Glucose broth culture filtrates yielded potent reacting fac- tors on a number of occasions. In some instances, brain medium culture filtrates gave factors of high potency. Titrations to the end- point sho^ved, ho^ve\er, only slight differences in the potency of these two types of preparations. One of the strains of Streptococcus hemolyticus, isolated from a case of mastoiditis, was especially 48 LOCAL TLSSUE REAC^TIVITY \iiulciu loi iiiicc. At the tiiiic it was employed for the experi- ments, as fe\v as 100 organisms were al)le to kill a mouse \vithin three days Avheii iiijeeted intrajieritoiieally. Filtrates ot virulent cidtures in glucose broth and in brain medimn also gave active ]:»reparati()ns ^\■hi(h elicited reactions in ajjproximately 75 j)er (eni ol rabbits prexiously prepared by intradermal injection of a po- tent heterologous filtrate. Although no systematic studies were car- ried out, the casual impression was that the high viriUence of the strejitococcus strain employed ^vas not necessarily respcjnsible for production of actixe filtrates, inasmuch as fdtrates oi the same organism at the time when its virtdence xvas considerably reduced also yielded active princij^les. The various streptococcus preparations completely lacked skin- preparatory factors. Reactions with indifferent and green-producing streptococci : These experiments may be conveniently divided into three groups, i.e., 1. Streptococcus materials trsed both for preparatory and provocative injections. 2. Heterologous potent bacterial fil- trates used for preparatory injections and streptococcus materials for provocative injections. 3. Streptococcus preparations used for preparatory injections and ]:)otent heterologoirs filtrates for pro- vocati\'e injections. The first grouj) embodied experiments Avith six strains of en- terococcus isolated from the blood stream of rheinnatic fever patients (Lichtman and Gross, 1932) . Tryptic digest ])roth in amoimts of 100 c.c. xvas placed in 500 c.c. Erlenmeyer flasks. The medium was inocidated xvith the re- spective strains of enterococcus and incid^ated for six days. The cultures were centrifuged and filtered throtigh Berkefeld "V" candles. Rabbits w^ere prepared each Ijy foiu^ simidtaneous injec- tions of the inidiluted material into the skin of the abdominal wall. Twenty-four hotirs later, the same iUtrates were injected intravenously in a dose of 3 c.c, per kilo of body xveight. About 75 per cent of rabbits developed strong typical reactions at the sites prepared. Some rabbits gave weak reactions. The response Avas the same in all foiu^ sites prepared in strongly, as well as weakly reacting animals. Preparations obtained from plain broth four day old culttnes also gave consistently positive residts. The materials were not titrated to the end-point and, therefore, no REACTIVITY TO VARIOUS MICROORGANISMS 49 comparati\e data are a\ailal)le as to the exact potency of these preparations. More extensixe inxestigations \vere done with ciihmes and cid- ture lihrates of Streptococcus viridcnis. The experiments l)elonged to the three groups above entnnerated. The strains employed were either old stock or recently isolated from the blood stream of cases of sni)aciite bacterial endocarditis. In early experiments a recently ir.olated strain ol Streptococcus viridans (Joel) xvas cnlttn'ed aero- bically, partially anaerobically and anaerobically (inider a vase- line seal) in 2 per cent glucose broth of pH 7.2 and incubated for one day. The skin sites were j^repared xvith potent heterol- ogons filtrates. Twenty-fotn- liotns later, the abo\e described li\e cidtmes of Streptococcus virida)is were injected intravenously in a d(jse of '>, c.c. ])er kilo (^f body xveight. Strongly jDositive reactions were obtained in abotit 40 per cent of rabbits tested xs ith these prepara- tions. Niniierons attempts were also made to obtain potent Strep- tococcus viridans principles in filtrates. \^arious methods of cid- tivation were employed. The ctdtures were made in aerobic and anaerobic double strength meat infusion broth and such stigars as maltose and glucose were added daily in small amotuits (0.1 per cent) . Rabbits were prepared either xvith homologous or heterologous bacterial filtrates. All the residts proved negative. In sid)sequent experiments a strain of Streptococcus viridans was grown in brain infusion broth containing Difco proteose. The ctdtine xvas incubated for foiu' days. The filtrate thtis obtained was tested by intravaneous injection in a dose of 3 c.c. per kilo of JDody Aveight, into rabbits previously prepared by an intradermal injection of heterologous bacterial filtrates. Strong reactions oc- curred in only a small percentage of rabbits. From the above de- scribed and similar experiments, it seems that the actixe principles of Streptococcus viridans occtn^ in a loxv concentration and that processes of filtration inrpair their activity. Reactions with Streptococcus Schottm'uller : One strain of a strictly anaeroliic streptococctis, recently isolated from a case of ptierperal sepsis, Avas tested. The strain xvas an ac- tive gas producer. The ctdttnes yielded a foiU odor. A Florence flask containing cooked lixer mediimi was inocidated xvith a twenty-fom" hoin^ old culture of the organism, sealed "with vase- line and inctibated for fom- days. The material xvas then centri- 50 loc:al tissue reactivity tuged and filtered ihiough Berkeleld "\'" candles. The rabbits were prepared by intradermal injection ol meningococcus 44B "agar washings" filtrate. T\venty-i()ur hours later, the intravenous injection of the filtrate, in a dose of 2 c.c. jjer kilo of Ijody ^veight, gave strongly positive reactions in all rabbits tested. No titrations to the end-point were carried out. Apitz (h).^^^/^ 'ii^d Wadsworth and Sickles (1933) confirmed my hndings that the streptococctis cidtine hltrates contain react- ing factors which can be best demonstrated in rabbits prepared by potent heterologous filtrates. From the above studies on the active principles of streptococci, the following may be concl tided: Live cidtmes of Streptococcus viridans produce principles of low and variable concentration; the activity of these principles being considerably impaired by processes of centrifugalization and filtration. Apparently, the products of Streptococcus viridans are totally inactive as a preparatory agent. Indifferent and green-})roducing enterococci give rise to con- siderably mc:)re potent actixe principles, bcDth effective in prepara- tion to the phenomenon and its elicitation. The Streptococcus lieniolyticus grotip is by far superior in re- acting potency but apparently possesses an insignificant amoiuit of j^rejjaratory factors. Here again, the ability of producing the active principles varies with the strains employed. Active strains are frequently encoinitered among the Streptococcus Jiemolyticus scarl(iti)iae, and Streptococcus Jiemolyticus pyogenes group. No systematic studies were made xvith Streptococcus Jiemolyti- cus erysipelatis. The impression thus far gained is that the po- tency of its products is inferior to the above groups of Strepto- coccus Jiemolyticus. One strain of Streptococcus ScJwttmiiller yielded potent react- ing factors. ACTVE PRINCIPLES OF PNEl'MOCOCCUS In my preliminary experiments I was able to obtain irregular reactions with pnetimococci of Types I, II, and III. The prepara- tic:)ns were filtrates of forty-eight hour old cidtines of pneinno- coccus on blood agar (preferably fresh rabbit's blood spread over the surface of the agar) . (Shwartzman, 1928c.) Cope and Howell (1931) prepared pneumococcus active prin- ciples by dissolving the pneumococcus in bile. They failed to ob- REACTIVITY TO VARIOUS MICROORGANISMS 5I tain the reactions by the use of a pneumococciis filtrate ^vhen the cells were not disintegrated. In a limited ntmiber of experiments, it appeared to these authors that specificity of the pneumococciis could be demonstrated by means of the phenomenon. In 44.4 per cent of the experiments, prepared sites reacted only to the intra- venous injection of homologous preparations. In 13.8 per cent of the experiments, positive reactions were obtained with combined intradermal intravenous injections of heterologous types of pneu- mococci. In my investigations of 1931-1932 additional attempts were made to obtain the pneumococciis acti\e principles from aerobic and fully and j)artially anaerobic fluid cultures of \'arioiis hydro- gen-ion concentrations and containing various sugars and other enriching substances (Shwartzman, 1932/) . The experiments may be summarized as follows: The ^ariolls preparations employed ^vere totally devoid of skin- preparatory factors inasmuch as reactions were not obtained when pneumococciis preparations were used for the skin injections. The phenomenon could be elicited, however, when rabbits were prepared ^vith B. typhosus "agar washings" filtrates of high skin- preparatory potency and pneumococcus filtrates were injected in- travenously. The experiments thus demonstrated that it is pos- sible to obtain pneumococciis filtrates potent in reacting factors. Most of the work was done with pneumococciis Type III strains. Fairly consistently potent preparations were obtained under the following conditions: In the late afternoon pneumococciis Type III was seeded into 100 c.c. of meat infusion broth, pH 7.8, containing 0.2 per cent glucose and 0.3 per cent maltose. The container was an Erlen- meyer flask of 200 c.c. capacity (partial anaerobiosis) . The cul- ture was incubated at 37° c. until the following day, when the pH was adjusted to 7.8 and 0.2 per cent glucose and 0.3 per cent maltose were added again. This was repeated for three consecu- tive days. After the last readjustment of pH, the culture was in- cubated for one more day and after this centrifuged until the supernatant fluid became clear. The supernatant fluid ^vas fil- tered through a Berkefeld "\'" candle. The rabbits prepared by the intradermal injection of 0.25 c.c. of a heterologous potent bacterial filtrate reacted to the intravenous injection of 1 to 2 c.c. ])er kilo of body ^veight of the pneumococciis filtrate. Type I pneumococcus culture filtrates made in the above manner proved 52 LOCAL TISSUE REACTIVITY inactive. Inasimu h as in the experiments with (^her microorgan- isms strains l)elonging to the same serological group may differ in ability to prodtice the active principles of the phenomenon, other strains of pneinnococcus Type I should be tested. The pneumococcus reacting factors completely deteriorated in some filtrates shortly after their preparation, but seemed to have been preserved quite well in others for a consideral)le length of time. In one preparation there ^vas observed a partial deteriora- tion of reacting factors after three months storage. The variance of these residts ^vas prol)al)ly due to differences in initial concen- tration of the factors. No definite conclusions could be drawn concerning the opti- mum conditions necessary for preparation of pneumococcus re- acting factors, because various batches of filtrates prepared imder identical conditions but at various times differed in potency. It appeared, however, that the use of several sugars ^vas induc- ive to the development of potent factors. Observations on the favorable effect of multiple sugars on the development of toxins was previously made by Hagen and Heller (1931) ^vith B. dlph- tJieriae. It may also appear difficult to determine the optinnnn length of the incid^ation period. Twenty-fotu" and forty-eight hour in- cubation periods seemed sufficient to obtain potent filtrates. The use of starch medium advocated by PagoiUatos (1931) for preparation of Dick toxin proved of no special advantage. Favor- able residts were obtained with "double strength" broth. The mediiun ^vas successfully used by Parker and Pappenheimer (1928) for the preparation of pneinnococcus sid^stances of pri- mary toxicity. Inasmuch as the reacting factors could be obtained both anaero- bically, as weW as aerobically, it became possible to differentiate them from Parker and Pappenheimer's pneumococcus toxic sub- stances which are strictly anaerobic. ACTIVE PRINCIPLES OF B. TUBERCULOSIS In 1932, Nasta attempted to reproduce the phenomenon of local skin reactivity to B. luhcyculosis. Rabbits and guinea pigs were prepared by an intradermal injection of small amounts of tuberculin O. T., glycerine broth, and virulent live tubercle ba- cilli. Twenty-four hours later, the rabbits received 0.5 c.c. of tuberculin intravenously. The guinea pigs received smaller doses REACTIVITY TO VARIOUS MICROORGANISMS 53 of the same material. None of the animals ga\e any reactions. According to Bieling (u)'^!), Oelrichs was able to obtain the phenomenon \vith certain cultine filtrates of B. tuberculosis. In her experiments, the phenomenon-prodncing factors ^vere not re- lated to the sti])stances responsible lor tnbercidin hypersensitive- ness. P. Bordet (1935) obserxed reactions in guinea pigs at the site of intradermal inocnlations of B. C. G. ciiltines following the intra\enous injection of B. culi ciiltme materials. Freimd (1934^) reported that the intra\enous injection of B. hphosus cidtiue fil- trate produced se\ere reactions at the site of ttiberctilin intrader- mal tests in guinea pigs sensitized xvith the B. C. G. ciUtures. He obtained no reactions in normal guinea pigs. The work of Bordet and Freund xvill be discussed in more detail in a subsequent chapter. Mazur (1935) failed to obtain acti\e principles in fil- trates of glycerine broth cidttnes of acid fast and non-acid fast tubercle bacilli. Recently, I carried out a series of investigations on the skin- preparatory and reacting potency of tuberculin and tuberculous culture materials. As xvill be seen from the following description, heterologous filtrates of ascertained phenomenon-producing po- tency were used either for skin-preparatory or intravenous injec- tions (Shwartzman, 1935c) . Reactions with products of B. tuberculosis in combination with heterologous filtrates of ascertained phenomenon-produc- ing potency : Styai)is. The strains of B. tuherculusis employed Asere the Bo- vine Type — C3 458-559 (of Nexv York City Board of Health) ; Human Type — Hh-; and A\ian Type — (S23 (of the American Type Culture Collection) . Cultures. The strains were each seeded on the surface of 250 c.c. of 5 per cent glycerine broth j)H 7.2 and placed in "1000 c.c." Erlenmeyer flasks. The cotton plugs were sealed with paraffin and the cultures incubated at 37.5° c. The length of the incubation period varied from a iew days to a number of xveeks. Filtrates. The cultures were centrifuged at high speed. The clear supernatant fluid xvas decanted and filtered through Berke- feld "\'" candles. Sterility controls xvere made on plain agar slants and I.oewenstein media. Tuberculin O. T. Cultures were heated in the Arnold sterilizer for one and one-half hours and hltered through t)ne layer of ster- 54 LOCAL TLSSIIE REACTIVITY ile filter j)apci. The liltiates were translerred to sterile be.ikers and evaporated to i lo original Nolmiie in the water bath at 56° c. The length of time re(jiiired lor evaporation varied ironi twenty-foin- to thirty-six hoius. In preliminary experiments, the primary irritating effect ot intradermal injections of tiU)erciilin C). T., and various (idtme filtrates \vas stndied. Filtrates gave no local reactions. Lhidihited tuberculin frecjiiently produced pnstides with hemorrhagic zones at the peri])hery twenty-four hours after intradermal injections. Dilutions 1:^^ elicited small pustides with a bright red erythema at the periphery. Dilutions 1:5 and higher gave no noticeable in- flammatory response. The irritation described was apparently due to the high concentration of glycerine, since preparations of 25 and 50 per cent glycerine in plain sterile broth gave similar re- actions. In the experiments which follo^v, it seemed advisable to avoid primary skin irritation, since it w^as previously observed that in instances in which the preparatory factors produced no primary reaction, the lesions following intravenous injections of bacterial filtrates were clearly defined and stronger than when the intradermal injections by themselves elicited inflammation. An undiluted filtrate and several batches of tubercidin O. T., diluted 1:7.5 were employed for preparation of the skin. Twenty- foiu" hoiu's later, large doses of B. tuberculosis filtrates and tuber- cidin ^vere injected intravenously. No reactions 'were seen four and t\venty-foin' hoins after the intra\enous injection. In another series of experiments the skin of rabbits was pre- pared by single injections of 0.25 c.c. of various batches of luunan and bovine tuberculin O. T., a filtrate of bovine B. tuberculosis broth culture and purified tuberculin T. P. T. (kindly supplied by Dr. Florence Seibert) . T^venty-foiu' hours later, the rabbits re- ceived single intravenous injections of potent meningococcus and B. typhosus "agar Avashings" filtrates. In these experiments, batches of human and bovine tuberculin O. T., prepared in my laboratories, were capable of inducing the state of reactivity. Similar attempts with the Ne^v York Board of Health tid^ercidins and ^vith the piuified T. P. T. tid)erculin con- sistently failed. It became ob\'ious then that the skin-preparatory factors are foiuid in tul:)ercidins in low and variable concentra- tions. In the third series of experiments, meningococcus and B. ty- REACTIVITY TO VARIOUS MICROORGANISMS 55 jjliusus culture filtrates of ascertained skin-preparatory potency were employed for intradermal injections. Various batches of tid^erculin (). T., were injected intravenously into rabbits, twenty-four horns later. Here again, severe reactions resulted from five tuberculin batches (bovine tubercidins O. T., and one human tubercidin) but failed with three other preparations (iiuman, avian and bovine tuberculin) . It is of interest that one batch of bovine tuberculin ^vhich failed to elicit reactions in the experiments revealed here, proved potent ^vhen used as a skin- preparatory agent. As is seen from the experiments thus far described, tuberctilin O. T., contains in low concentration the factors necessary for the elicitation of the phenomenon. This becomes obvious if the doses employed in these experiments are compared with those of menin- gococcus, B. typhosus, and other cultme hltrates. The reactions imder discirssion cannot be obtained iniless non-tuberculotis bac- terial filtrates of high potency are used either for the skin prepara- tion or for intravenotis injection. In this manner the deficiency in potency of tuberculin is made up for by taking advantage of the strict quantitati\e reciprocity of skin-preparatory and reacting (intravenous) factors recorded previotisly. Seibert's purified tuberculin T. P. T., and the Ne^v York Board of Health tuberculin were inactive, whilst preparations made in my laboratories were capable of eliciting the phenom- enon. All the tuberculin jDreparations contained tuberculin sub- stances jDroper. This suggests that the factors necessary for the elicitation of the phenomenon inider discussion may not be re- lated to the tid^erculin sidjstances proper. As described above, the tuberctdin of my laboratories ^vas pre- pared by evaporation at 56'^ c. for t^venty-foiu' to thirty-six horns. The New York Board of Health tuberculin ^vas evaporated by boiling' over a free flame. As known already, prolonged exposure to heat is likely to inactivate the active principles of the phenom- enon. Presimiably, therefore, the lack of reacting potency in the NeAV York Board of Health preparation may be due to prolonged heating. Apparently, the pinification methods of Seibert also remo\'e the toxic suljstances described here. No data are available as yet on the period of incidjation of cul- tines optimimi for the production of B. tuberculosis acti\e prin- ciples. 5(3 LOCAL TLSSUE REAC'IIVITY ACriNl I'KINCIPLF.S OF ANAKROIilC: (.RAM-NEGATIVE BACILLI ISOLATED EKO.M CIIKOMC I.LNG ABSCESSES Cohen (i()^'52, 19,^5.^5) sLiidicd extensively the anaerobic flora of abscesses oi the lung. He isolated from the lesions gram-negative anaerobic bacilli which were classified for convenience into the chromogenic and non-chromogenic grcnips. An example of the chromogenic anaerobic group is Bacleriiini nielaninogenicum, a gram negative anaercjbic bacillus ^vhich produces a black, amor- phous pigment on the surface of the blood agar. The non-chromo- genic group of gram-negative anaerobic bacilli may be further divided into Subgroup A, including organisms which liquefy gela- tin (B. tJieoluidcs is an example of this subgroujj) ; and Subgroup B, including cjrganisms \\hich dcj not licjuefy gelatin [B. fincusiis is an example of this subgrou])) . The following method was con- sidered by Cohen as optimiun for the j)r()duction of active prin- ciples: To 4 per cent gelatin broth were added 0.1 per cent dextrose and 0.5 per cent sodiinn chloride. The reaction ^vas adjusted to pH 8.2. The medium was autoclaved for 15 minutes at 10 lbs. pressine. The final reaction ^vas about ])H 8.0. For the growth of the anaerobic organisms, the following pro- cedure was adopted. About from 10 to 15 c.c. of blood agar was poured into a 250 c.c. flask and allowed to solidify in the re- frigerator. Gelatin broth, freshly boiled and then rapidly cooled, was j3oured over the blood agar in the flask until the latter was two-thirds full. The gelatin broth was then allowed to congeal in the refrigerator. Then from 10 to 15 c.c. of blood agar was jjoured on top of the congealed gelatin. The medium thus pre- pared ccjntained gelatin broth bet\veen t^vo layers of solid blood agar. The flask was stored in the refrigerator. Before use, the medium was warmed in the incubator for about twenty minutes; it ^vas then inoculated with a suitable culture, sealed ^vith petrolatum and incubated at 37° c. for from two to three days. The fluid cultures ^vere poured into flasks and centrifugated, and the supernatant fltiid was filtered through a Berkefeld "V" candle. A drop of the sediment \vas spread on a blood agar plate and incubated aerobically, to rule out aerobic contamination. The filtrate was tested for sterility, stored in the refrigerator REACTIVITY TO VARIOUS MICROORGANISMS 57 and used for a period of approximately three to fotir weeks fol- lowing the preparation. No preservative ^vas added. Rabbits ^vere prepared by potent meningococcus and B. typho- sus "agar washings" filtrates. The above filtrates of gram-negative anaerobic bacilli injected intravenously into these rabbits elicited severe reactions in approximately 80 per cent of animals tested. The preparations were apparently devoid of skin-preparatory po- tency because the skin preparation of rabbits with potent heterol- ogous filtrates was essential. Filtrates of Bacterium melaninogeni- cum gro^vn in symbiosis ^vith Streptococcus gamma and those of leptothrix possessed both skin-preparatory and reacting factors, but of comparatively lo^v concentration. ACTIVE PRINCIPLES OF YEAST Bock (1932) reports that he was able to reproduce the phenom- enon of local skin reactivity with filtrates of four day old cultmes of yeasts in glucose broth. Reactions were obtained in rabbits re- ceiving preparatory injections of these filtrates and subsequent intravenous injections of B. coli culture filtrates; and vice versa, when preparatory injections of B. coli culture filtrates were fol- lowed by intravenous injections of the yeast filtrates. The reactions were characteristic of the phenomenon although some^si^at pro- longed in their appearance. Similar experiments with actinomyces and a trichophyton [Achorion Sclwenleini) failed. The observa- tions on the potency of yeast extracts Avere corroborated by I\'ano- vics (1934). Inasmuch as yeasts are kno^vn to li\'e in symbiosis with bac- teria, the possibility is not excluded that the active principles of the phenomenon could have been due to the presence of the symbiotic bacteria. ACTIVE PRINCIPLES OF ASCARIS LUMBRICOIDIS J. \V. Mu (1935) Studied the phenomenon to Ascaris lumhri' coidis. Before extraction, the ascaris worms were ^vashed in run ning water, immersed in 10 per cent formalin solution for 10 min- utes and then again washed in running ^\ater. The extract Avas prepared by placing 4 gms. of the whole ascaris ^vorm in mixture of one part of Coca's solution, and 3 parts of normal saline for two weeks. The extract ^vas shaken from time to time. After the extraction, the mixture was filtered through one layer of filter paper and 0.25 c.c. of concentrated phenol was added to each 100 58 LOCAL TISSUE REACTIVITY c.c. The extract was dried at room temperature and ground into a fine powder. Albino male rai)bit,s were used. The doses were 0.1 c.c. intradermally and •■] c.c. intravenously, per kilo of body weight. Clontrol rabbits received the ascaris extract intradermally and phenolized Coca's solution intravenously; ascaris intrader- mally and no intravenous inoculation; and finally, j)henolized Coca's solution intradermally and no intravenous injection. The intradermal injection ot the ascaris extract elicited moder- ately edematous papules. All the control rabbits were negative. The experimental rab- bits sho^ved extremely severe reactions in a high percentage of animals tested (8 out of 13) . Microscopically, the reactions fol- lowing the intravenous injection of ascaris showed marked exuda- tion, massi\e infiltration of polymorplionuclear leucocytes, pro- nounced vascidar necrosis, thrombosis, rupture of blood vessels and extensi\'e hemorrhage in the sul)ciUaneous tissue. The pri- mary reactions were those of mild edema, slight migration of polymorphonuclear leucocytes, moderate dilatation and conges- tion of blood vessels. There were no signs of hemorrhage and vascidar necrosis or rupture of blood vessels. This microscopic picture entirely conforms with that of the phenomenon of local skin reactivity. ACTIVE PRINCIPLES OF RICIN Gratia and Linz (1931(7) studied the active principles of ricin. They injected into the skin of the abdominal ^vall of rabbits 0.25 c.c. of a 1:1000 dilution of the toxin. After twenty-four hours the prepared sites appeared infiltrated and congested, at which time the rabbits received 1 c.c. of the same dilution of toxin intrave- nously. No reaction followed this injection. When the same rab- bits ^vere reinjected ^vith an active filtrate of B. coli, there ap- peared slight reaction at the sites prepared with ricin. The experiments, ho^vever, are too limited, as the authors themsehes point out, to draw any conclusions. ACTIVE PRINCIPLES OF DIPHTHERIA TOXIN Gratia and Linz (193W/) failed to induce the state of reactivity of the phenomenon under discussion by means of preparatory injections of dij)htheria toxin diluted 1:100 and of diphtheria anatoxin. Wadsworth and Sickles (1933) also mention their fail- REACTIVITY TO VARIOUS MICROORGANISMS 59 lire to induce the phenomenon of local skin reactivity to diph- theria toxin in rabbits. Recently Freund (1934c/) attempted the experiment in guinea pigs. The preparation consisted of an intradermal injection of 0.1 c.c. of diphtheria toxin in various dilutions ranging from 1:250 to 1:2500; 0.0016 c.c. of the toxin employed by him contained 1 M. L. D. B. tyjjJiosus "agar Avashings" filtrates in doses from 0.5 c.c. to 1 c.c. were injected intra\enoiisly, t^venty-four hours later. Some of the guinea pigs prepared by dilutions in the range of 1:250 to 1:750 showed primary hemorrhagic lesions before the intravenous injection of the B. typJiosus filtrate {i.e., twenty-four hours after the intradermal injection) . In one guinea pig the in- travenous injection of B. typliosus filtrate seemed to have en- hanced the primary hemorrhagic lesion. A few guinea pigs which shoAved no reactions to diphtheria toxin in twenty-four hours be- came hemorrhagic later incidentally to the intra\enous injection of B. typhosus filtrate. It is well known that the intradermal injections of diphtheria toxin produce primary hemorrhage and necrosis w^hich fully de- velop w^ithin forty-eight to seventy-t^vo hours follo^ving the in- jection. Freund concluded from his experiments that diphtheria toxin in the guinea pigs acts as a skin-preparatory factor. It should be noted, howexer, that the doses employed by Freund \\ere con- siderably larger than the amount necessary to elicit the primary hemorrhagic and necrotic reactions. According to Kellogg (1922), I/40 of a unit is the minimal amount producing a primary hemorrhage and necrosis in the guinea pig upon intradermal injection. As is seen from the de- scription of Freund's experiments, reactions ^vere obtained only with dilutions ranging from 1:250 to 1:750 {i.e., 1/4 M. L. D. to 1/12 M. L. D.) but no reactions Avere obtained with dilutions ranging from 1 : 1000 to 1:25,000 (notably doses consisting of 1 17 of M. L. D. to 1/425 M. L. D.) . It is obvious, therefore, that the reactions described by Freund could have been entirely due to the primary effect of the diphtheria toxin and apparently not related to the phenomenon under discussion inasmuch as the reactions appeared ^vithin the time most siiital)le for the appearance of pri- mary diphtheria toxin reactions, and ^vere produced only by doses capable of eliciting primary reactions. 6o LOCAL TISSUE REACTIVITY RFCAPl'IliLATION The actixc priiuij^les of the phenomenon were thus far ob- tained with \arying degrees of success from a large number of microorganisms. Some bacteria yield under sinijDle conditions of cultivation ("agar washings" culture filtrates and broth culture filtrates) ac- tive principles of considerable potency. Among these are meningo- coccus, B. typhosus, B. paratyphosus, B. coli, B. jriedldnder, B. dysenleriae, B. prodigiosus, members of hemorrhagic septicemia and hemoglobinophilic groups (B. Icpiseplicus, B. pestis, B. in- fluenzae , B. pertussis) , and Vibrio cholerae. The concentrations of preparatory and reacting factors do not necessarily run parallel. Thus, filtrates of certain strains of B. typhosus ordinarily contain as many as 150 to 200 jireparatory units and 300 to 500 reacting units. Conversely, meningococcus Group III yields preparations containing rarely more than 50 preparatory units with as many as 2000 to 3000 reacting imits. Many bacteria produce the principles in low and variable con- centration. Among these are pneiunococci, streptococci, B. tuber- eulosis, B. melitensis, B. abortus, staphylococci, anaerobic gram- negative bacilli isolated from chronic limg abscesses and Spiro- cheta pallida (p. 72) . The activity of the preparations of this group is best demonstrated ^vhen tested in rabbits prepared ^vith heterol- ogous fdtrates of high potency because most of them may be totally devoid of preparatory potency. Potent active principles were also obtained from Ascaris lum- bricoidis. The ability of yeast, ricin and diphtheria toxin to elicit the phenomenon remains questionable. Chapter III PHYSICO-CHEMICAL PROPERTIES OF THE ACTIVE PRINCIPLES OF THE PHENOMENON OF LOCAL SKIN REACTIVITY THE EFFECT OF HEAT UPON THE ACTIVE PRINCH'LES THE question of heat resistance of the active principles of the phenomenon of local skin reactivity has received con- siderable attention of investigators. No definite conchisions, however, can be dra^vn as yet concerning this problem since very few quantitative experiments ha\'e been carried out thtis far. In my early ^vork, tryptic digest broth ctiltine filtrates of se\- eral strains of B. typhosus ^vere exposed to 60- and 100° c. for one hoiu' and aiitoclaved at 15 lbs. pressme for forty-five minutes and one hotn\ The materials thtis treated and diluted 1:2 were used for skin preparation. Large doses of luitreated filtrates ^vere injected intravenously (3 c.c. per kilo of body weight) . The results indicated that the skin-preparatory factors of B. typhosus possess considerable heat resistance. Filtrates derived from certain strains lost these factors Avhen autoclaved for forty-five minutes. There was encotuitered, how'ever, a filtrate of one strain which resisted aiuocla\ing at 15 lbs. pressure for one hoin\ Thus, in these earlier qualitati\e experiments there was noticed a differ- ence in heat resistance of skin-preparatory factors deri\ed from various strains of the same microorganism (Shwartzman, 1928^'). The purpose of more recent experiments Avas to make qtiantita- tive studies on the effect of heat upon both skin-preparatory and reacting factors. As wall be described later (p. 181) the potency of the filtrates could be correlated Avith the dination of reactivity induced. The potency of preparations exposed to different tem- peratures Avas then graded by determination of the length of state of reactivity induced and by quantitative titrations of their react- ing potency. These experiments showed distinctly that heating of B. typJwsus "agar washings" filtrates at 60° c. for one hotu- did not shorten the dination of local skin reactivity and did not in- 61 62 LOCAL TISSUE REACTIVITY flucnce the reacting potciuA. Ilic exposure to loo c. lor twenty ininiites shortened the (huation ol reactivity from seventy-two to t\venty-foiir horns and reduced the reacting titer from 550 to 250 units per c.c. Thus, possibly loecause of a finer method of titration employed in these experiments, there was noted a par- tial but substantial reduction in the skin-preparatory and react- ing potencies at a temperature as lo^v as 100^' c. The other reason for the discrejiancy Ijetween the early and present exj^eriments coidd ha\e l)een due to the fact that "agar ^vashings" filtrates, instead of tryptic digest broth ctdtine filtrates, Avere used in the latter experiments. Inasmuch as the tryptic di- gest broth cidtine filtrates contain more autolytic material than the "agar washings" prej^arations, it seems of interest to investigate further whether the active principles oljtained through cell dis- integration {i.e., endotoxic j^reparations) differ from soluble exo- toxic materials in heat lability (Shwartzman, 1932^) . In experiments of Schneierson, in my laboratories, the reacting potency of B. dyseiiteriae Shiga "agar washings" filtrates and B. dyseiiteriae Shiga exotoxin prepared in egg medium by the method of Olitsky and Kligler (1920) , was reduced from 40 to 5 tuiits per c.c. and 50 to 5 tniits per c.c, respectively, after heating at 80^ c. for one hour. In contrast, Shiga bacillus endotoxin also prepared by the method of Olitsky and Kligler remained unal- tered after exposure to the same temperature. Burnet (1931) fotnid that boiling of cultmes autolyzed accord- ing to Besredka's method for five minutes, had no effect on their skin-preparatory activity. It appears from Uyeda's (1934) experiments that the skin- preparatcjry factors of "agar \vashings" filtrates of J'ihrio cholerae resist exposure to 90° c. No titrations were made in order to de- termine whether there Avas a partial reduction of the potency of the filtrates. According to Stolyh^vo (1935) , the phenomenon-producing and lethal potencies of filtrates of typhoid and paratyphoid cultin^es are not destroyed by boiling for thirty minutes. Heating of the filtrates in the autoclave under ii/^ atmospheric pressure in\ari- ably inactivates the preparations. In Boquet's experiments (1935) the active jjrinciples of B. lepisepticus Avere not inactivated by heating of filtrates at 100° c for ten minutes. Pabst and Branham (1933) suggest that 37° c. may inactivate PHYSICO-CHEMICAL PROPERTIES 63 the acti\e principles. These observations could not be confirmed in my laboratories. It may be concluded from the ^vork reviewed that the skin- preparatory and reacting factors possess a considerable heat re- sistance. The heat resistance may differ ^vith various microorgan- isms and even strains of the same microorganism. It may also be influenced by the mode of preparation of the material {i.e., amount of autolysis, medium ingredients, etc.) . It is obvious that the investigations just described have been undertaken in order to determine ^vhetiier the active principles of the phenom- enon are exo- or endotoxic in nature. The present generally ac- cepted point of \'ie^v, to be discussed in more detail in Chapter x, is, however, that the heat resistance of a bacterial toxic material cannot be used as a clear-cut criterion for this differentiation even among classical exotoxins. My present contention is that the active principles of the phenomenon are secreted by the bacterial cell during the phases of active growth and that remnants of these principles may be also obtained through cell disintegration. The observations re- corded here and subsequently in this chapter point to the possi- bility that the ability of a bacterial cell to give off the active principles depends on certain inherent characteristics differing in various microorganisms and their strains. It may be plausible to assimie that the active principles deri\ed from the bacterial cell through its disintegration may prove to be more heat stabile be- cause of some protective effect of the cell elements. THE EFFECT OF DIFFERENT HYDROGEN ION CONCENTRATIONS UPON ACTIVE PRINCIPLES In my studies on the effect of acids and alkali upon the skin- preparatory factors of B. typhosus, tryptic digest broth culture filtrates were adjusted under sterile precautions to pH 9.0, 8.6, 7.6, 7.0, 6.6, 5.4 and 4.0. The final dilution of the filtrates of various pH was 1:2. The filtrates of pH 9.0 and 4.0 were used on the day of adjustment; those of pH 8.6, 7.6, 7.0, 6.6 and 5.4 were injected t^venty-four hours after the adjustments. Before use, given ainounts of filtrates of various pH were heated for one hour at 60° c. in the Arnold sterilizer and in the autoclave at 15 lbs. pres- sure. Heated and non-heated filtrates of each pH ^vere tested for skin-preparatory potency. The non-adjusted unheated filtrate was 64 I.O(w\L TISSUE REACTIVITY used lor Lhc intravenous injection (^ c.c. per kilo of body weight) . The skin-preparatory factors in the hydrogen ion concentra- tions mentioned resisted heating to 60° and 100" c. for one hour but were invariably destroyed by autoclaving at 15 lbs. for forty- five minutes (Shwartzman, 1928^) . Bock (1932) studied the effect of hydrogen ion concentration upon the active principles as follows: Six quadrants of the al)dominal skin of three rabbits were {pre- pared with B. coli filtrate slightly alkaline, B. coli filtrate neutral to litmus paper, B. coli filtrate acid to litmus paper, sterile broth of pH 7.4, N/10 sodium hydroxide, and N/io hydrochloric acid solutions. One rabbit received a provocative injection of slightly alkaline B. coli filtrate; a second rabbit an injection of B. coli filtrate neutral to litmus paper and a third, a provocative injec- tion of a B. coli filtrate, acid to litmus paper. The first rabbit showed reactions of equal severity in the first three prepared sites and none in the sites prepared Avith sterile broth, N/io sodium hydroxide and N/io hydrochloric acid. The second and third rabbits also gave no lesions in the latter three sites and only weak reactions in sites prepared with B. coli filtrate at various reactions. In another group of experiments the preparatory in- jections of B. coli filtrate of pH 7.4, 5.5, and 8.5 were made. One of the rabbits received a provocative injection of B. coli filtrate of pH 5.5 and another of the filtrate of p?I 8.5. The rabbits re- ceiving provocative injection of filtrate of pH 7.4 and 8.5 gave uniformly severe reactions in the three prepared sites, whilst the rabbit receiving injection of the filtrate of pH 5.5 gave no re- actions. This author concluded from his experiments that the potency of the reacting factors is inhibited by an acid reaction. On the other hand, the hydrogen ion concentration seemed to have no effect upon the skin-preparatory potency of the filtrate. DeCour (1934) precipitated active filtrates with acetic acid of pH 4.7 and redissolved the coagulum in sodium hydroxide of pH 7.5. The redissolved coagulum became completely inactive. He does not state wdiether the coagulum ^vas tested for skin-prepara- tory or reacting potency. Stolyhwo (1935) investigated the effect of storage under various conditions iq^on the activity of the filtrates, i.e., filtrates exposed to light, stored at room temperature, and incubated at 37° c. in PHYSICO-CHEMICAL PROPERTIES 65 sealed ampules and test tubes with cotton stoppers. It appeared to this author that if the filtrates ^vere adjusted to pH 5.6 and 6.0, the actix ity persisted longer than when adjusted to pH 8.0. The skin-preparatory potency of "agar washings" ciUtine hl- trates of Vibrio cholerae prepared by Uyeda \vas preserved in the pH ranging from 4.6 to 8.5. Apitz noted that adjustment to pH 10.0 strikingly increased the activity of P fractions obtained in the coinse of his fractiona- tion of bacterial filtrates described later in this chapter. EFFECT OF UFTRAVIOLET RADIATION UPON ACTIVE PRINCIPLES Welch (1930) carried out comparative experiments on the effect of racUation with ultraviolet light upon various substances. It was absolutely necessary to use materials prepared in stich a manner that no protein or extremely small amounts of it were present. The medium of Cahn-Bonner Avas used. This atithor reported that detoxification of active filtrates of B. coli and B. iypliosus culttires may result from ultraviolet radia- tion. The skin-preparatory factors disappeared Avithin shorter pe- riods of exposure than the reacting factors. It ^vas mentioned in Chapter i that B. typhosus culttne filtrates may have a skin- preparatory potency approximating one-fourth of the reacting titer. It is apparent, therefore, that the partial destruction of the active principles may result in a proportionately lower skin-pre- paratory than reacting titer. Similar observations were made by me on partial detoxification of active principles by means of formalin. EFFECT OF DISSOLUTION AND AUTOLYSIS UPON THE ACTIVE PRINCIPLES As it became obvious from the experiments described in the preceding chapter, bacterial autolysis is not necessary for prepara- tion of the active principles. The asstmiption is first clearly demonstrated by the fact that fluid culttnes of B. typhosus in- cubated for several days usually yield active principles one hun- dred times weaker than filtrates of ^vashings of growth on solid media incubated only for twenty-four hours. Cultures of menin- gococcus in enriched media, inctdjated for periods as long as seven days, and stdijected to alternating freezings and thawings have irregular and low' potency. On the other hand, as mentioned 66 LOCAL TLSSIE REACTIVITY before, wasliings ot young cultures on agar may prochue filtrates potent in dilutions as high as 1:4000. Burnet (1931) compared the phenomenon-prodticing potencies of "agar washings" filtrates and disintegrates made according to Besredka's meth(jd and found that the latter were stronger. How- ever, his experiments cotdd not be considered conclusive since he did not make any cjuantitative titrations to the end-point for comparative determinations of potencies of the various prepara- tions. With this consideration in mind the following experiments were done by me. Meningococcus Preparations. Meningococcus Group I, 44D strain (Wadsworth) Avas inoculated into 1 per cent rabbit blood i)roth pH 7.4. After 22 hours of incubation the supernatant broth culture, free from red blood cells, served as the inoculum. On the surface of 0.7 per cent glucose veal infusion agar in Kolle flasks 4 c.c. of the inocuhnn ^vere poured. After t^venty-foiu' hoins of in- cubation the gro^vth of 10 Kolle flasks was each washed off with 4 c.c. of 0.9 per cent NaCl solution containing 0.4 per cent phenol. The suspension ^vas centrifuged three times, and resuspended each time in the same volume of 0.9 per cent NaCl solution containing 0.4 per cent phenol. Washings 1, 2 and <:] were the respective supernatant fluids obtained after each centrifugalization. The washed sediment ^vas then suspended again in 40 c.c. of 0.9 per cent NaCl solution and disintegrated by freezing in dry ice and tha^ving three conseciiti\'e times. The suspension of killed men- ingococcus ^vas the meningococcus disintegrate employed in the experiments about to he described. B. typhosus Preparatious. Three washings ^vere obtained from B. typhosus cultures grown on 10 Kolle flasks, in a manner simi- lar to the foreooinsf. The washed sediment, ho^ve\'er, was mixed with 0.4 gm. of NaCl, which had previously been sterilized in the autoclave at 250° f. for fifteen minutes and dried over cal- cium chloride. The mixture '^vas dried in a vacuum desiccator overnight, ground to a fine powder in an agate mortar and then suspended in 40 c.c. of sterile distilled water, containing 0.4 per cent phenol. This ^vas left overnight in the incubator and tiien centriftiged for one hour. The supernatant fluid was called "Bes- redka B. typJiosus disintegrate." Rabbits were prepared by single intradermal injections of 0.25 c.c. of undiluted material. Various dilutions of each of "agar PHYSICO-CHEMICAL PROPERTIES 67 \vasliings"' and disintegrates were used for intravenous injections, twenty-foiu- horns later. Each dihition ^vas tested in a group of three rabbits. The restdts are recorded in Table n. As is evident from Table n, acctuate determinations of the potencies of various preparations can be made only by titrations to the end-point, but not by comparing the percentage of positive results obtained with given dilutions in small groups of rabbits. Thus, if one should compare, for instance, meningococcus wash- ings 1 and 2 in diltuion 1:100 by the percentage of positive rabbits, both preparations ^vould appear of the same strength. Titrations to the end-point show that one preparation is ten times stronger than the other. This is due to natural resistance sho^vn by a certain percentage of rabbits to moderate doses of active principles, which, ho^vever. is not displayed with large doses. As is also seen from Table 11, the concentration of toxic fac- tors in the first ^vashing- of a oriven number of meningococcus cells was approximately 1000 times stronger than in the disintegrate of the same number of cells dissohed in the same volume of NaCl solution. With B. typhosus, the first washing ^vas about 30 times stronger than the distintegrate prepared according to Bes- redka's method. Evidently, then, the major portion of the factors was obtained in the first washing and only an insignificant amoiuit Avas found in disintegrates and the further washings. The process of washing probably caused \'ery little disintegration since in those experiments young bacterial cells on solid media were centrifuged as soon as they were suspended in NaCl solu- tion and the supernatant fluid ^vas separated from the sediment shortly after completion of centrifugal ization. It can be con- cluded, therefore, that the toxic substances necessary for the phe- nomenon are extracellular (Shwartzman, 19320) . (iratia and Linz (1932c) carried out a series of experiments witii bacterial cultures dissolved with bacteriophage and strepto- thrix. Staphylococcus ^vas inoctdated into the broth four hoins later. When the broth became turbid, 5 drops of anti-staphylg- coccus bacteriophage ^vere added to several of these tubes. The follo^ving day the bacteriophage-containing tubes were filtered and the filtrates tested in rabbits. Neither of the filtrates showed any activity. B. coil cultures were similarly treated. The filtrates of normal B. coll cultures and those dissolved ^vith bacteriophage yielded the active principles. It appears, therefore, from the ob- 68 LOCAL TLSSUE REACTIVITY Table n COMPARATIVE POTENCIES OF DISINTEGRATES AND "AGAR WASHINGS" FILTRATES Preparation Meningococcus disintegrate washings 3 " 3 " 3 " 3 " 3 2 2 " 2 " 2 2 B. typhosus Besrcdka disintegrate " washings a it, Amount injected intrave- nously per kilo of body weight c.c. undiluted material " diluted I : 2 " " I : 10 I : 25 " " I ".100 " I : 250 " " I : 600 : 25 : 100 : 250 : 600 : 25 : 100 : 250 : 600 I : 25 I : 100 I : 250 I : 1000 I : 1200 Total No. rab- bits tested undiluted diluted I 10 50 100 200 I : 10 I : 100 I : 200 I : 300 I : 500 luidilutcd diluted I undiluted diluted I 10 25 50 100 25 50 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 Posi- tive rab- bits Nega- tive rab- bits I 2 3 3 3 3 3 I 2 I 2 3 I 2 2 3 3 3 3 3 3 3 Washings i, 2, 3 — first, second and third washings of bacterial agar cultures. PHYSICO-CHEMICAL PROPERTIES bg servations that the bacteriophage does not destroy the active prin- ciples present in the bacterial cells nor does it liberate them from bacteria which do not ordinarily contain them. In another group of experiments of Gratia and Linz (1932c) agar slants were inoculated with B. coll and Vibrio cholerae both capable of producing the active principles, and also with staphylo- coccus and B. diplitJieriae, both incapable of producing the ac- tive principles. After twenty-foin- hours of incubation the cultures were emulsihed in distilled Avater and washed three times by centrifugalization. The ^vashed sediments were inoctilated with streptothrix and the dissolved cidtines filtered three days later. Mycolysis of the inactive bacteria (staphylococcus and B. diph- theriae) did not liberate the active principles, whilst the mycol- ysis of the B. coli did not destroy their activity. The active prin- ciples of I'ibrio clwlerae seemed, ho^\e\er, to l^e destroyed by mycolysis. In my own experiments anti-erysipelas bacteriophage described by me in 1927 \\as mixed ^vith ctilttnes of Streptococcus Jiemolyti- cus erysipelatis, and incubated for t\venty-foin" hours. Completely lysed ctilttnes were filtered through Berkefeld "V" candles and used in doses of 0.5 c.c. for preparatory injections and 3 c.c. per kilo of body weight, for intravenous injections. No reactions were obtained. Uyeda (1934) washed the gro^vth of agar cultures of Vibrio cholerae twice in 0.5 per cent saline solution and emulsified the residue either in saline or distilled water. The emulsions were incubated for ten days and filtered through Berkefeld "V" can- dles. Autolysates thus prepared failed to induce a state of reactiv- ity to the phenomenon follo^ving the intravenous injection of potent culture filtrates of the same strain of Vibrio cholerae. The endotoxin of Vibrio cJwlerae prepared by him according to the method of Besredka, contained potent skin-preparatory factors. No comparative quantitative studies were made on the "agar wash- ings" filtrates and the endotoxin prepared from the residue by the method of Besredka. In Plant's (1932) experiments, sediments of Spi rochet a pallida cultures washed with saline, autolyzed at 37° c. and dialyzed for forty-eight hours against running water were totally inactive. As will be mentioned subsequently, filtrates of whole young cultures dialyzed as above, contained active principles. According to Boquet (1935) , washed sediments derived from yo LOCAL TLSSUE REACTIVITY cultures ol B. l('jji.s('ljlicu.s, resuspended in NaCll solution in orig- inal xolunic lacked phenomenon-producing potency, (lultmes killed by heating and treated by the method oi Rowland em- ployed for liberation of endotoxin were also inactive. Filtrates of six day old cultures in Martin's bouillon heated 60° c. for one horn- prior to filtration, gave the best results. It was also sho\vn by Joukow-Werejnikow and Lipatoxa that the bacterial sediment derived from "agar washings" of B. jjc.sti.s and washed several times in j^hysiological saline is totally de- pri\'ed of the acti\'e j^rinciples. FLUCTUATIONS IN POTENCY OF ACTIVE PRINCIPLES The skin-preparatory and reacting factors fluctuate in potency in the same filtrate. On several occasions there was observed a gradual decrease in the j:otency as weW as complete inactivation of the filtrates stored in the refrigerator for se\eral months, as illustrated by the following example: One batch of meningococcus, 44B. Group IH, "agar washings" fihrates was tested at frequent intervals for a period of four months. The jjotency of the fihrate became somewhat lower four w^eeks after its preparation, re- mained imchanged for three months thereafter, and ^vas suddenly lost alto- gether four months after its preparation. It appears that the meningococcus culture filtrates are more likely to lose their potency than the B. typhosus filtrates although similar difficulties were occasionally experienced xvith the latter. On the other hand, some filtrates may retain their potency un- changed for a considerable length of time. A filtrate of B. dysoi- teriae Shiga, retested more than fourteen months after its prepara- tion, retained its original strength. Stolyhwo (1935) foimd that his filtrates of B. typliosus did not deteriorate when exposed to light and to room temperature. He evaporated the bacterial filtrates in a water bath. The dried mate- rial, powdered and stored in sealed ampules, preserved its activity for a period of six months. By suspending the powder in a quan- tity of fluid smaller than the "mother" fluid, it xvas possible to obtain a concentration of the active principles. Pabst and Branham (1933) also pointed out the possibility that some preparations may remain stable fc:)r a number of months whilst others may lose their potency rmexpectedly. They noted an increase in activity during storage in the refrigerator. Cholera vibrio active principles of Lintc:»n, Singh and Seal PHYSICO-CHEMH.AL PROPERTIES 71 (1935) deteriorated Avitliin three Aveeks lollowing their prepara- tion. It was frecjiiently noticed that the potency of filtrates may also substantially increase within one to two months follo\ving their j)re})aration, especially on storage in the refrigerator. These ob- ser\ations were rejieatedly and consistently made ^vith filtrates of "agar washings" of meningococcus, B. d\se)ileyiac Shiga, B. coli, and B. typhosus. It suffices to report the following examples: Meningococcus, Group I (44D) "agar washings" filtrates titrated May 5, 1930, showed 100 reacting units per 1 c.c, and May 27, 1930—700 reacting luiits. A second batch of meningococcus. Group I titrated May 27, 1930— con- tained 200 leacting units per 1 c.c, and June 10, 1930—1600 reacting units. A third batch oi' filtrate of the same microorganism possessed 1000 reacting units on May 15, 1930, and 2500 reacting units on June 30. 1930. Koplik in my laboratories made similar observations on fil- trates of B. pertussis cnltnres in brain mediinn. FILTRATION AND DIALYSIS OF THE ACTIVE PRINCIPLE.S In my early experiments B. tyljltosus and meningococcus active j)rinciples precipitated by satiiratic:)n with ammonium sulphate did not diffuse through parchment membranes (Shwaitzman, 1929c) . Gratia and Linz (1932c) utilized a methcxl of dialysis previously described by Gratia. A Petri dish was divided into t\\() sections by a very fine cello- phane membrane. Sterile broth \\as poured into the two com- partments separated by the membrane. The upper compartment A\as inoculated ^vitli B. coli and the Petri dish A\as incubated for three days at 37° c. There Avas an abundant gro^vth in the upper compartment, whilst the broth in the lo^ver compartment re- mained sterile. The fluids from bcjth compartments were cen- trifuged and filtered. Rabbits ^vere prepared ^vith each of the fluids and then injected intravenously with the respective fluids, twenty-four hours later. The phenomenon could be produced with the filtrate derived from the culture of the upper compart- ment but failed with the filtrate of the lower compartment. Similar experiments ^vere made ^vitli a collodion sac, filled ^vitll sterile broth and placed in a tube containing sterile broth. The broth of the sac was inoculated with B. coli and the apparatus incubated at 37 "^ c. for four to fi\'e days. The broth in the tube remained clear, whilst that in the inside of the sac became 72 LOCAL ILSSUE REACIIVLIY tiiil)i(l. riu' filtrate ol the (iiltinc in ilic .s;u {ontaiiicd the active j)riiuij)les, wliilst the filtrate of the broth outside of the sac was iliac ti\e. Plant (i<)'52) studied the ])iien()inenon-i)i()diicinj^ jjoteiicy of Beikefeld "V^" filtrates and sii})ernatant fluid ( iiltures ol Spiro- clieta jjallidd R-,'^() (Reiter) in media ol Plant and Kassowitz with entirely negative results. The materials were then precipitated by liim Avith ammonium sulphate and dialyzed against rinniing water for forty-eight honrs. Ral)bits were pre])ared by intrader- mal injection of 0.2 c.c. of the dialyzed filtrate and t\venty-four hours later, injected intravenously with five and ten times the dose of the same filtrate. Positive reactions were obtained in ral)- fjits thtis treated; althotigh the original filtrate (without ammo- nitim stilphate precipitation) employed in equivalent doses gave no reactions. In finther experiments Plant found that the dialysis alone (without preliminary precipitation with ammonium sul- phate) transformed inactive filtrates into active ones. It was sup- ]30sed by him that inhibitory substances are present in the filtrates which may be eliminated by dialysis. These inhibitory substances could be also precipitated by alcohol and successfully removed by dialysis of the precipitate against running w^ater. According to Plant, it is important to select a suitable membrane (Schleicher and Schiill (N579) (Diffusionshiilse) . The selection of the strain of spirochete also proved of importance. Whilst strains Ra,; and Kross K22 were active, rabbit testicular syphilomas inoculated Avith Trufli and Nichols strains and water spirochete (Spiroclieta pseu- doicteroides) gave negative results. It is interesting that in this author's experiments the most ac- tive principles ^vere obtained from twenty-four to forty-eight hour old cultures (33 rabbits, 100 per cent) . Six to eight day old cul- tures gave only 75 per cent of positive results ^vhilst fotu' week old cultures were completely inactive. Recently, quantitative studies on the dialysis of B. typhosus, meningococcus and B. coli culture filtrates were undertaken (Shwartzman, Morell, and Sobotka, 1937) . A simple and easily reproducible method of dialysis developed by Morell allo\\ed the carrying out of the procedure under strictly sterile conditions for as long as one week. When an active filtrate was titrated to the end-point, the titer computed was the mean of the dilutions at ^vhich positive and PHYSICO-CHEMICAL PROPERTIES 79, negati\e reactions \vere obtained. In calculating the percentage of reco\ery of activity, the maximum error of titration \vas used. For example, when injected intra\enously, a certain filtrate (be- fore dialysis) elicited positi\e reactions in prepared skin sites at a dilution 1:100 and negative reactions at a dilution 1:200. The reacting titer assigned to this filtrate was, therefore, 150 ± 50 units per c.c. The limits of the error thus were =*= 50/150 or =t 33 per cent. After dialysis, the filtrate gave positive results at dilu- tions 1:91 and negative results at dilutions 1:140. The titer com- puted was then 1:116 ± 25 reacting units per c.c. In this in- stance, the limits of error ^\ere then "^ 25 116 or ±22 per cent. During the dialysis of this filtrate, the volume had increased from 36 to 79 c.c. The percentage of reco\ery, therefore, was: (116) (79) (100) -7 : — i—Fi ^ 170 per cent =^ 57. The error =•= 57, ap- (150) (36) ' ^ ^' .'/I plied to the 170 per cent reco\ery, ^vas obtained by taking 33 per cent of the latter, i.e., the maximum of the titration errors before and after dialysis. It is realized that this does not represent the maximum possible error. Before dialysis, the same filtrate preparation was used for both the skin-preparatory and intra\enous injections. After dialysis, the fluid remo\'ed from the bag Avas titrated to the end-point for re- acting potency in rabbits prepared by single intradermal injec- tions of the original non-dialyzed filtrate. Since the titer of filtrates may change after long periods of time, care was taken to ascertain their potency just previous to dialysis and to titrate the dialyzed solutions as soon as possible thereafter. It ^vas observed that after dialysis the B. txpJwsus and B. coli solutions remained clear, ^\•hereas the meningococcus prepara- tions sometimes formed slight precipitates. The decrease of pH, Avhich practically ahvays occurred, suggested the presence of sub- stances whose isoelectric points were on the acid side. The biuret and Molisch tests were usually positi\e. After remo\ ing an aliquot sample for analysis the solutions ^vere adjusted ^vith saline to approximate normality (physiological) and stored in the ice chest. Even after six months they had remained clear. Inasmuch as the active principles of the phenomenon were con- sistently retained inside of "Cellophane" bags, the extent of jjurification attained by means of dialysis was significant. The pmi- iication thus eff^ected amounted, on the average, to about a three- '74 LOCAL ILSSLIE REACTIVITY fold increase in reacting potency per milligram ul dr)' \veight and ol nitrogen. B. coU gr()^vn in fluid synthetic meditim was most readily pini- fied: thus, the decrease in total dry weight and nitrogen, 9'^ and 89 per cent respectively, was greater than in any of the other fil- trates. Similarly, the increase in activity per milligram of dry weight and of nitrogen, 750 and 452 per cent respectively, was also the highest. When a similar jjreparation was concentrated in vacuo to one-half of its original volimie, approximately the same amoinit of total solids and nitrogen was removed by dialysis, but the increase in activity was only 234 and 175 per cent, respec- tively. Apparently, therefore, the concentration /// xxiciio may con- siderably reduce the extent of pmification. With B. typhosus, where a good purification was also obtained, a slightly greater removal of total solids and nitrogen was ob- tained in filtrates of cultures on synthetic medium agar than on veal infusion agar. Although the actual reacting titers of menin- gococcus ctdtiue filtrates were decidedly the highest, both before and after dialysis, the extent of ptnificaticjn attained by dialysis was lower than in any of the other filtrates. Thus, the percentage increase in activity in these filtrates was only 70 per cent per milli- gram of nitrogen and 136 per cent pei" milligram of dry weight. Possibly, the use of 1 per cent rabbit's blood in the inociUinn, carried over into the washings, accounts in piivi lor the relatively lower figures. It shoidd be noted, ho^vever, that the highest nimi- ber of reacting tmits per milligram of nitrogen and dry weight was observed in the dialyzed meningococcus filtrates. By calculat- ing the reciprocals of these values, it ^vas foimd that 0.0067 n^iHi- grams of total solids and 0.00087 milligrams of nitrogen were capable of eliciting the reaction in prepared skin sites at the end-point. Biunet (1931) made the observation that filtration of the ac- ti\e princijiles through a Seit/ filter resulted in a great decrease in the potency. This observation Avas confirmed in my laboratory on a number of occasions. Filtration tiirougji lierkefeld "\^" can- dles yields acti\'e jDrejiarations although it is suggestive that some of the activity is lost by the filtration. Boquet (1935), Gratia and Linz (i932r), P. Bordet (1935), Alechinsky (1936), and Aitoff, Dion and Dobkevitch (i936<7) obtained active preparation by filtration of the cultures through Chamberland L2 and L3 candles. PHYSICO-CHEMICAL PROPERTIES 75 FORMALIN DETOXIFICATION OF ACTIVE PRINCIPLES Burnet (1931) ^vas the first one to report on the effect of for- malin on the active principles of the phenomenon of local skin reactivity. A Besredka preparation of B. dyseyiteriae Flexner was treated with 0.2 per cent of formalin and the ttibe left sealed at 37° c. for one month. The skin ol rabbits ^vas prepared ^vith the un- treated preparation and the toxoid, and twenty-four hours later the rabbits received an intravenous injection either of 0.5 c.c. of untreated preparation or toxoid. The toxoid as a skin-preparatory agent provoked either no reaction or a reaction below the region injected. Toxoid used intravenotisly ^vas as effective as an unal- tered preparation. According to Burnet, a single intradermal injection of toxoid followed by a suitable intraxenous injection of the active principles gave rather varial^le results depending on the reactivity of the rabbit. A common form of the reaction xvas a ring of ptirpura around the central pale area. With very reactive rabbits there was more extensive purpura with only a small sharply outlined pale center, but in most the reactions were a small cres- cent below the infiltrated area or nothing at all. As pointed out, the reacting potency of a preparation may be considerably higher than its skin-preparatory potency. The inter- pretation of Burnet's exj)eriments l)ecomes easy if it is granted tiiat his treatment with h)rmalin elicited only a partial inactixa- tion of the preparation, Avitii a proportionately more considerable reduction in preparatory potency than in reacting potency. Reac- tions oi similar appearance \vere observed by me with xveak fil trates and liltrates partially inactivated by expostire to heat. Klein (1932) , in my laboratories, "was able to inactivate menin- gococcus "agar xvashings" filtrates l)y prolonged exposure tc^ formalin. Inasmuch as he carried out extensi\e studies on the antigenicity of the formalinized preparations, it seems more ap- {^ropriate to discuss his results in Chapter iv. USE OF METHODS OF ECKER AND RIMINGTON FOR CC^NCENTRATION AND PURIFICATION OF THE ACTIVE PRINCIPLES Ecker and Rimington (1917, 1927) showed that synthetic medium cultine filtrates of the Aerytrycke type of B. jjara- lyjjhosus. concentrated /// iinciio and dialyzed, ):)roduccd a raj)id 76 LOCAL TLSSUE REACTIVITY edematous and reddened area when injected into the skin ol nor- mal rabbits. After decomposition of a lead precipitate of the same filtrate material by means of (N.H.4) 1; S.()4 and dialysis, a marked reaction was obtained by a similar injection; the wheal being about 1.5 cm. in diameter within twenty-four hours. The center \vas alwut 4 mm. in diameter elevated and wide and the surround- ing area was hyperemic. In about four days the central area be- came necrotic and the hyperemia std)sided. The original filtrate (not inoculated, concentrated or dialyzed) failed to give reactions. The concentrated products ga\'e certain color reactions ^vhich are usually associated with the presence of carbohydrates. Small c[uan- tities of nitrogen were also obtained (0.3 to 0.4 per cent) . In the \vork in connection with the phenomenon of local skin reactivity, Ecker and \V^elch (1930) employed seven members of the colon typhoid group of organisms, namely, B. typJiosus, B. paratypJiosus A, B. paratyphosus of Aertrycke type, B. enteritidis, B. paraty- pliosus B of Schottmiiller type and B. colt. The filtrates ^vere concentrated as described above. Ecker and Welch compared their filtrates w^ith my "agar wash- ings" filtrates and concluded that my products had the capacity to prepare the skin to the provocative injection of their prepara- tions and vice \ersa. It may be assimied, therefore, that the active principles of tiie phenomenon are preserved in the jjrocess of con- centration and purification of Ecker and Rimington. There is an essential difference, ho^\•ever, between tliese preparations and "agar washings" filtrates, as strongly emphasized by Ecker and Welch and in a recent publication by Apitz (1935^) , i.e., the Ecker-Rimington preparations invariably produce a primary red edematous reaction following the intracutaneous injection, whilst the "agar washings" filtrates produce only an insignificant primary erythema. AMMONIUM SULPHATE PRECIPITATION OF B. TYPHOSUS ACTIVE PRINCIPLES The entire procedine ^vas carried out by me under sterile pre- cautions. The Tl "agar washings" filtrates were fully saturated with ammonium sulphate and kept at room temperature for about twenty-four hours. The mixture was then filtered through hard filtering paper, in a Buchner funnel, by means of negati\e pres- sure. The precipitate was taken off the paper and dissolved in 1 PHYSICO-CHEMICAL PROPERTIES 77 per cent NaCl solution. Roth the dissohed precipitate and the filtrate ^vere dialyzed inider toluol in parchment hags aoainst running Avater initil there was no amnioniimi sidphate detected inside of the bags. This usually took about forty-eight horns. The dialyzed solution of the precipitate was dried at 37.5° c. in a Freas vactumi o\'en connected to Cenco Hyxac Pinnp. The dried precipitate was ground into a fine po^vder and preserved in a desiccator over ])hosphorus pentoxide at room temperattne. just before use a weighed amotnit of the po^vder \vas dissolved in distilled ^vater. The powder never w^ent into perfect solution; however, if care was taken to add the distilled w^ater drop by drop and to mix it thoroughly with the water by means of a glass rod, the suspension was uniform and quite fine. The precipitate-free supernatant fluid was removed from the bags immediately after complete disappearance of ammoniimi sul- phate. The fluid was filtered and NaCl was added to the filtrate to the extent of 0.9 per cent; the filtrates were then stored inider tohiol in the refrigerator. In most instances the experiments were made shortly after the preparation ^\as completed. The abo\e materials Avere tested in rabbits by the method of reciprocal titration (Method I) . It became obvious from these experiments that the active principles of the phenomenon could be completely precipitated by saturation with ammonitmi sid- phate, the supernatant fluids showing no potency whatsoever. A yield of 9.1 gms. of dry po^vder ^vas obtained from 4000 c.c. of the "agar ^vashings" filtrates after precipitation and dialysis. When both the skin and intravenous injections of the precipitate Avere made into the same rabbits, 0.1 mgm. was sufficient to prepare the skin of some rabbits which sid^sequently received (per kilo) 8 mgms. of the precipitate intravenously. W^hen, however, the in- travenous dose Avas reduced to 2 mgms. of precipitate per kilo, it was necessary to use 1 mgm. intradermally in order to prepare for the reaction. Doses of the precipitates larger than 2 mgms. per kilo injected intravenously often exhibited general toxicity. The potency of the precipitate preserved over phosphorus pentoxide in a desic- cator at room temperattne, remained unchanged for seven months after its preparation Avas completed. The method of reciprocal titrations employed in these experi- ments makes difficidt the accinate estimation of the percentage 78 LOCAL ILSSUE REACniVLrY ol recovery. It may l)e concliidecl, however, that tlie process ol precipitation w iili animoniiim sul})hate offers the (jpj)ortiinity of ofjtaining' stabile active principles of skin-j5rej)aratory and react- ing potencies in a dry form (Sliwart/man, h)2()(:) . PURIFIED AND I'KF.SKRVl.D S 1 RF 1' I()C;( )C:CAL RKACTING FACTORS Attempts ^vere also made by me to purify the streptococcal reacting factors and to obtain them in stable form. In these ex- periments, some of the steps of the method of concentration of streptococcal scarlet fever toxin recently described by Shiiin (1930) , were employed Avith certain modifications, as follows: Ten and one-half liters of i per cent dextrose broth of pH 7.6 in Florence flasks (Y^ full) were inoculated \vith Streptococcus hemolyticus scarJatinae, strain 4014. The cidtures ^vere incubated for twenty-lour hours at 37.5° c. After incubation, the flasks were chilled in the refrigerator, the contents transferred to suitable sterile containers and diluted with 31.5 liters of 95 per cent alco- hol containing 0.03 \)gx tent glacial acetic acid. The mixtures Avere allo\ved to stand in the refrigerator for one hour and ^vere then filtered under negative pressure through a Buchner funnel, hard filter paper being used. The j^recipitate was taken off the paper and dissolved in 0.85 per cent sodiiun chloride solution previously adjusted to pH 8.0 and containing 0.5 })er cent phenol. The solution of the precipitate ^^•as dried in a Freas vacuum o\en at 37.5° c. The dried precipitate ^vas ground into a fine powder and preserved in a desiccator over phosphorus pentoxide at room temperature. The total yield in dry powder was 42 gms. Just be- fore use, a ^veighed amount of the po^vder was suspended in dis- tilled ^vater. The powder made a fairly good suspension. Inasmuch as the streptococcus yields skin-preparatory factors of lo^v and irregular potency, the rabbits ^vere jjrepared by intra- dermal injection of B. typhosus "agar washings"" filtrate and then given intravenous injections of suspensions of various amounts of the streptococcus powder. The powder was able t(^ elicit severe reactions at the prepared skin sites. The potency of the material could be easily titrated. The procedure employed for the prepara- tion of the powder apparently brought al)out some loss of the reacting potency of the cultures. In fact, the live cultures from which the material was derived, in the amount of 3 c.c. per kilo of body \\eight injected intravenously, gave about 66 per cent positi\e reactions, ^vhereas ^vith the powder it 'tvas necessary to physic:()-c:hemical properties 79 enijilo) the e(|ui\alent of 50 c.c. of tlie "niother" culture per kilo of body Aveight, in order to obtain approximately the same per- centage of positi\e results. The reacting factors in the form of po\vder were stable. A retest made six weeks after its preparation ga\'e identital results. In fmther experiments, all the steps in the j)rocedine of the concentration of scarlet lexer toxin described by Shinn xvere car- ried out as folloxvs: Fi\e hundred cubic centimeters of 1 per cent dextrose broth of pH 7.6 was inoculated with Streptococcus heniolyticus scar- la I i)uic, strain 4014. After txventy-four hours' incubation, the cul- tiue xvas cooled in the refrigerator and diluted xvith 1,500 c.c. of 85 per cent alcohol containing 0.03 per cent glacial acetic acid. The mixture xvas allowed to stand in the refrigerator for one hour and then Avas filtered inider negative pressure through a Buchner funnel, hard filter paper being used. The precipitate was taken off the paper and dissolved in 0.85 per cent sodium chlo- ride solution containing 0.5 per cent phenol and previously ad- justed to pH 8. The suspension Avas then centrifugated at high speed until it became perfectly clear. The sediment was discarded. The supernatant fluid xvas re-precipitated with foin- vohmies of acid alcohol and filtered as before; the supernatant fluid was dis- carded and the precipitates xvere dried as in pre\ ioirs experiments. The yield in dry powder was 1.76 gms. As may be seen, further purification of the material xvas ob- tained through the remo\al of bacterial bodies and txvo-fold pre- cipitation Avith acid alcohol. An amoimt of prjxvder equi\'alent to 50 c.c. of "mother" culture gave positixe reactions in approximately 50 per cent of rabbits tested. The xvork described demonstrated the possibility of some ]3ini- fication and preserxation of active principles of the phenomenon by means of precipitation xvith acid alcohol. Since accmate esti- mation of the percentage of recovery is lacking, it is difficidt to determine from these experiments xvhether the method offers any opportunity for concentration of the principles (Shxvartzman, PURIFICATION OF B. COI.I PREPARATIONS Gentile (1934) reported shortly upon experiments xvith a nucleoprotein fraction of B. coli. A 2 per cent solution of the (So loc;al tisste reactivity iiiu Icoprotein extract oi B. coli was injected iutradeniially. This was followed twenty lioms later by intravenous injection of an active Vibrio clwlerae cidtiire in a qtiantity of i c.c. per kilo of body weight. Strong reactions were elicited in the prepared skin sites. Equally severe reactions coidd be ol)tained when B. coli nucleoprotein was used for the intravenous injection. Andervont and Shear (1936) attempted to purify B. coli active filtrates by the use of methods of Felton, Kauftmann and Stahl (1935) developed by the latter authors for the precipitation of the solidole specific polysaccharide from pneimiococcus broth cid- tures, as follows: Phosphate was added to a filtrate of one w^eek old broth cultine of B. coli. The active principles ^vere carried down with a basic calciimi phosphate precipitate. In some cases in which precipita- tion failed to remove completely the active principles from the supernatairt fluid, more phosphate ^vas added and the pH read- justed to 9.4 with lime water. The precipitates w^ere dissolved in hydrochloric acid and the insoluble material filtered off and dis- carded. Reprecipitation was then obtained by the addition of sodiimi hydroxide to pH 9.5. The supernatant sohuion was inac- tive. The precipitate 'vvas again dissolved in acid and again pre- cipitated with sodiimi hydroxide. The reprecipitations were done in order to remove the protein carried down in the original precipitation. The final calcium phosphate precipitate was then treated Avith acid and alcohol in order to remove the calcium. Enough hydro- chloric acid was added to dissolve the precipitate and to bring the solution to pH 2.0. The insoluble matter ^vas filtered off and discarded. The active material was precipitated from solution by addition of ethyl alcohol; the calcimn remained in sohuion. Solu- tion in acid and precipitation Avith alcohol was repeated luitil the precipitate ^vas free from calciimi. The fraction ^vas dissohed in water, neutralized with sodium hydroxide and evaporated to dry- ness. Finally, excess sodium chloride was removed by repeated treatment Avith methyl alcohol. The original filtrate and the purified material were tested by these authors for the phenomenon-producing potency. Each of four rabbits received preparatory intradermal injections of three dilutions of the original filtrate and of three dilutions of the puri- fied fraction. Twenty-four hours later, two of the rabbits each PHYSICO-CHEMK^AL PROPERTIES 8l received 0.5 c.c. of tlie iincliluted filtrate intra\en()iisly while the other two recei\'ed 0.5 c.c. of a 1:10 dilution of the purified fiaction. Tlie ptirified fraction of Andervont and Shear appeared to be more potent than the filtrate. The intra\enoiis injection of a 1:10 dilution of the stock solution of the pmified fraction pro- duced reactions at practically all sites of intradermal injection both of the filtrate and of the purified fraction, while the intra- \'enotis injection of the iindikited filtrate elicited reactions only in those sites which had received the highest concentration of the filtrate or of the purified fraction. A positive reaction was obtained with 0.0005 iiigins. of the purified fraction. CHEMICAL STUDIES ON ACTIVE PRINCIPLES OF VIBRIO CHOLERAE Linton, Singh and Seal (1935) tested the phenomenon-pro- ducing potency of the following fractions: 1. \V'hole killed heavy suspension of Vibrio cliolerae derived from various groups. 2. A-fraction. This fraction dissolves out of vibrios at 50° c. in alcohol made 0.025 N with HCl. Accordino to Linton, Sinoh and Seal, the A-fraction has a specific rotation of about — 12.0° (in N/2 alkaline at 37° c.) and nitrogen distribution which is characterized by the possession of about twice the amotuit of amide nitrogen and one-half the amount of humin nitrogen occurring in the whole protein. "A" occurs in Ijoth the pseudo- and euglobulin fractions of the protein, increases ^vhen the ^vhole protein is kept in distilled water at 10° c. and destroyed by 0.125 N HCl extraction. It yields about 1 to 2 per cent of vibrio dried weight. 3. The B-fraction was obtained by Linton, Singh and Seal after removal of the A-fraction Ijy means of boiling the vibrios for fifteen minutes in 75° alcohol. The material was diluted with saline and made 0.05 N Avith HCl. B-fraction has a rotary powder of 46.0° and nitrogen distribution indistinguishaljle from that of the whole vibrio protein. It was found in both the globulin fraction and A-fraction. The yield is not affected by high concen- tration of HCl, nor does it increase in A'ibrio protein Avhen kept in the refrigerator in distilled water. B-fraction is about 6 per cent of dry weight of the protein. 4. Residue fraction is closely allied to B-fraction and consists (S2 LOCAI, TISSUE REACTIVITY ol ilic inatc'iial which iciuains alter the second acid s()liil)le IVac- tions ha\e heen renioxed. It is also identical with (he whole j)rotein. Reducing substance is j)iesent in A, B and in the residue but it is minute in the residue Traction in proporticjn to the amount. A- and B-fractions have prc)portic:»nately lar^e amounts of reduc- ing substance. They possibly correspond to the outer parts of the vibrio. A-fraction approaches the ^vhole xibrio in its serological activity. B- and residue fractions are less active as antigens and in immunological reactions in the test tube. If a strain is smooth, it contains more reducing sufjstance than B. If reducing substance is greater in B than A, the strain is rough-smooth. Hydrolized specific polysaccharide of the vibrio is considered by Linton, Singh and Seal as source of the reducing substance in cjuestion. These authors prepared rabbits by three simiUtaneous injec- tions of the various fractions described above. The "agar wash- ings" cultine filtrates containing the active principles were used for the intravenous provocative injection (i.e., 0.35 c.c.) . In their experiments A-fraction yielded ccjnstant and typical reactions. Whole killed culture, B- and residue fractions were totally devoid of the skin-preparatory effect. The work is of special interest be- cause it indicates the possibility of obtaining the active principles in fraction A, which constitiUes an important part of the cholera vibrio in its inmumological relationships. The identity of the active principles, of course, still remains iniknown, although the wcjrk offers interesting information consistent \vith obserx ations on their innmniological behaxior. CHEMICAL STUDIES ON ACTIVE PRINCIPLES OF B. PF.STIS After rejjroducing the j:)henomenon with B. pestis ciiltines, as described in Clhapter n, ]ouk<)\v-\V^erejnikow and Lij^atova (ic)'}'^-'^) attenijited chemical studies on the active principles. The following materials were tested for their ability to elicit the phenomenon: Nucleoprotein of B. pestis: The material \vas j)re}3ared according to the techniciue de- scribed by Lustig and (ialeotti, i.e., cultures \\'ere slightly alka- lini/ed with 0.75 to 1 per cent of NaOH and then j)recipitated with acetic acid. PHYSICO-CHEMICAL PROPERTIES 83 Capsular material of B. pestis: Cultures were grown on agar pH 6.6 at 37° c. for 36 hours. The 'washings were made in physiological NaCl solution, slightly alkalized with NaOH and shaken for one to two hours. The material was then heated at 56° c, and centrifuged. The authors contended that the supernatant fluid collected contained the cap- sular material. The assimiption ^vas based on the fact that the sediment consisted of approximately the same number of bacteria as the original suspension; that the shaken bacteria were deprived of capsule; that the supernatant fluid precipitated with wiiole B. pestis immune serum; and finally, becatise the addition of ^\'eak acetic acid gave no precipitate, thus excluding the presence of the nucleoprotein. Both fractions were tested for skin-preparatory and reacting potencies in rabbits and in spite of the fact that large amounts were employed, they pro\ed totally inactive. CHEMICAL STUDIES ON ACTIVE PRINCIPLES OF B. DVSENTERIAE Olitski and Leibowitz (1935) prepared a suspension of growth of B. dyseuteviae Sliiga on agar and obtained P], P- and C-fractions of Fiirth and Landsteiner (1928) and also fractions Q and T and the Haptene according to the method of White (1932, 1933) . In addition, after removing Q and T fractions, bacteria were boiled for two hoius in NaCl solution, both with and without addition of 10 per cent glycerol. The material was then fikered to remove the bacteria and the filtrate precipitated by addition of two vol- umes of alcohol. The fractions were called gi (glycerol) and Sa (saline) . All the materials were dried and resuspended in saline to have 1 mgm. of dry weight per 1 c.c. of saline. Pi fraction con- tained 2.3 of nitrogen; -within 1 hour it hydrolyzed incompletely with 4 per cent HC'l and completely with 10 per cent HCl. It was j^ositive for P^O.-, and reduced Fehling soluti(jn; Pi fraction con- tained 7.3 of nitrogen; it was completely hydrolyzed by 6 per cent HCl within one hour; it was slightly positive for P-C).-, and failed to redtice Fehling solution. Apparently then, these fractions rep- resented a mixture of protein and carbohydrate Avhich could be, however, further purified. These authors tested the various fractions for their ability to induce the state of reactivity and judged their potency by the size of the lesions obtained; 0.3 c.c. was used for preparation and 84 LOCAL TLSSLE REACTIVITY 1 ex. as a ])r()\()(at i\c dose. Tlic results were tabulated by these authors as lollows: Intravenous Injection P\ Crude Pi Purified Pj Carbohydrate Pi crude o cm. 3x5 cm. 4 x 9 cm. o cm. Pi purified 4x4 cm. 2x5 cm. 23 x 15 cm. 5x5 cm. P> 4x9 cm. 3x5 cm. 13x16 cm. 5x4 cm. Carbohydrate o cm. 5x5 cm. 5x4 cm. o cm. Inasmuch as no quantitative estimations were made, the tests may be considered as presiunptive and it seems that the follo^ving conclusions are justihed: Pi crude and carbohydrate fractions apparently possessed very small amounts of preparatory factors. They appeared to be able to induce the state of reactivity only provided potent materials as P] purified and P^ were used for intravenous injections. It also may be granted that Pi crude and carbohydrate fractions pos- sessed some reacting potency inasmuch as they ^vere able to elicit reactions in sites prepared by potent fractions. It becomes also obvious from these experiments that Pi ptnified and Po fractions contained most of the active principles of the phenomenon be- cause they gave strong reactions when used for skin-preparatory and provocative injections. The authors attempted to classify the potency of the materials on the basis of the size of the lesion. This, however, is some^vhat dangerous since the size of the lesion does not only depend on the potency of the material but also on its ability to diffuse in the injected site. ATTEMPTS AT FRACTIONAL PRECIPITATION OF POTENT BACTERIAL FILTRATES Recently Apitz (1935^) used the following chemical methods for fractional precipitation of bacterial filtrates potent in the elici- tation of the phenomenon: (a) Concentration of active filtrate in vacuo at 37° c. to ap- proximately i/io of the original volume and subsequent pre- cipitation of the concentrate with 95 per cent ethyl alcohol. (b) The solution of precipitate from (a) in NaOH. The in- soluble material was designated as the "N" fraction. (c) Addition of acetic acid to the solution from (b) imtil precipitation was at a maximum. The precipitate ^vas designated as the "P" fraction. (d) Precipitation of the soluti(m from (c) with approximately six volumes of 95 per cent alcohol. PHYSICO-CHEMICAL PROPERTIES 85 (e) AV'heii the "N" and "P"' tractions were completely re- moved by repetition of procedures (b) , (c) , and (d) , including several alcoholic precipitations from alkaline as well as acid solu- tions, the final alcoholic precipitate was designated as the "C" fraction. This author treated a nimiber of various bacterial preparations by means of the above methods and tested the fractions obtained for skin-preparatory potency. A series of injections into the skin of the shaved back ^vas made. Each material ^vas injected in foin^ to H\e different dilutions. "Agar ^vashings" filtrates of kno^vn potency were injected intravenously tw^enty-four hours later. Most of his work ^^•as done with fractions of "agar ^vashings" filtrates. Only weak reactions were oljtained with the lowest dilutions of the N-fractions and were interpreted as contaminations by other fractions. The freshly prepared C-fraction regularly sho^ved a high degree of activity ^vhich often approximated the action of the corresponding amoimt of whole filtrate. The Molisch tests ^vith the C-fraction ^vere ahvays strongly positi\e ^vhilst in several in- stances the material was found biuret-negative. This fraction showed a higii degree of lability and lost its activity if stored in the ice box in Avatery solution for forty-eiglit horns. In a few cases the C-fractions remained stable long enough to permit an attempt at further purification. The solutions were satiuated with annnonium sulphate and a fine precipitate formed ^\hich Avas separated by centrifugalization and dialysis against running ^vater. The dialysis proved deleterious to the active substance. The P-fraction obtained from "agar washings" of B. typJwsus varied considerably in amount. The largest P-fractions were found ^vhen the bacterial cells imderwent autolysis. The P-fraction was active in all cases and more stable than the C-fraction. It usually gave positi\e biuret reaction and a strong Molisch test. The potency of the P-fraction was conditioned by the amount obtained and this in turn by the occurrence of autolytic processes. Binnet (1931) and other investigators obtained a high degree of activity in whole bacteria and their extracts when prepared according to Besredka's method. I foinid by comparative titra- tions a superiority of "agar washings" filtrates to bacterial ex- tracts. It is possible, as Apitz points out, that the residts may vary with the ability of the strains to form active principles of ex- ogenous or endogenous origin. Apitz also obtained active C- and P-fractions from ground li\e (S6 LOCAL TLSSUE REACTIVITY B. l\jj/i()Mi.y (ioinjxn isoii ol the dry weights ol the sanic Iractions with those ol the corresponding washings showed a chHerent pro- jjortion between the C- and P-fractions. The predominant frac- tion in groiuid material was the acetic acid precipitate whilst in washings the C-Fraction predominated. Attempts to fractionate cidture filtrates in synthetic medinm proved troublesome and of- fered no advantage because of the strong primary skin reactions produced by these filtrates and also the fact that their skin-pre- paratory potency was ^veaker than in ordinary "agar washings." Apitz reached the conclusion that the active principles of the phenomenon are contained in two different fractions of the bac- terial substances, i.e., nucleoprotein and carbohydrate fractions. On accoimt of the complex character of the nucleoprotein frac- tion, ftnther work was considered necessary by this author in or- der to study the relation of the active principles to the species specific antigens. Although active principles of high activity coidd be obtained in the C-fraction, this author did not believe that they were identical with the type specific carbohydrates, since in contrast to the latter they ^vere extremely luistable and destroyed by prolonged heating. Ho\vever, in his opinion the |)ossibility ^vas not excluded that they represented a more labile sid)stance from which the type specific carbohydrates are derived. It is significant that the amoimt of nucleoprotein fraction increases in ctdtines tmdergoing autolytic changes whilst the C-fraction is mostly ob- tained in yoinig actively growing bacterial cidtmes ^vith little autolysis. Apitz also suggested that his resiUts indicating a doid)le soince (jf the active principles, namely, from nucleoproteins de- rived from l)acterial bodies and from the C-fraction derived from the metabolic activity of live bacteria, may explain some of the dis- crepancies in the 1 iterating concerning the origin of the active principles. As will be seen from the sidjsequent chapters, investigations demonstrating antigenicity of active principles of the phenome- non, suggest that they are either protein in nattne or perhaps associated with proteins. The investigations of Apitz seemed in- consistent with this point of view since he reported the isolation of a biuret-negative fraction which ^vas slowly dialyzed through "Cellophane." This necessitated the repetition of the Avork of Apitz (Shwartzman, Morell, and Sobotka (19-^7) in Avhich his procedine ^vas carried out with several batches of B. typhosus "agar ^vashings" filtrates. PHYSICO-CHEMICAL PROPERTIES 87 Instead of titrating tlie skin-preparatory potency, however, more accinate titrations ol reacting potency \vere done. The mean titers were derived in the same manner as in the work on dialysis (p. 73) . By accurately titrating the reacting factors of the phe- nomenon, it was fotmd that, in agreement with Apitz: (1) the alkaline insoluble N fractions were always inactive; (2) the nti- cleoprotein P fractions were practically always active; (3) the supernatant alcoholic sohitions, after concentration in vacuo, were inactive; and (4) the alcohol precipitated C fractions ^\ ere highly potent but frequently labile, in one instance being completely in- activated dining preparation. The isolation of a biinet-negative C fraction, however, has not been confirmed. Since cjualitative color tests as indications of chemical structine must be accepted with considerable reserve, attention is called to the following analytical data; the high nitrogen values con- sistently foimd in the various C fractions, {i.e., 8.1, 8.5 and 12.4 per cent) suggested the presence of appreciable quantity of pro- tein-like substances. Accordingly, when fraction C of one batch was refluxed for two and one-half hours in 2 N HCl, only 10.4 per cent reducing sul)stances, calculated as glucose, were formed. For comparison, it may be noted that on hydrolysis, the specific carbohydrates of Types II and III pneimiococcus, -which do not contain any nitrogen, yield aboiu 70-75 per cent reducing substances calculated as glucose; whereas that of Type I pneu- mococcus, containing 5.8 per cent nitrogen, is hydrolyzed to the extent of 28 per cent (Heidelberger, 1927) . It must be empha- sized, therefore, that in its present state of purity, fraction C can- not be designated as a carbohydrate. By titrating the reco\'ery of reacting factors, it has been fotmd that most of the activity origi- nally present in a bacterial filtrate ^sas lost ^vith the foregoing procedines of concentration. E\en ^vith the best recoveries ob- served, the extent of pinification attained was insignificant when the activity per milligram of dry -\\eight or of nitrogen ^vas measined. RECAPITULATION The acti\e principles of the phenomenon possess considerable heat resistance which may differ some\vhat ^vith various micro- organisms and even with strains of the same microorganisms. It inay also be influenced bv tlie mode of preparation of the ma- terial. (S(S LOCAL ILSSUE REACITIVITY Data oil tlic c'Hc'c I ol tlic Indro^cii ion (oiuentialioii upon the atlixc j)iin(ij)lcs are still nica^ic. Dctoxiiuation of actixc j)iincij)Ic's ol B. coli and B. lyjjiiosii.s cultures may result Iroin ultra\i()let radiation and also Ironi \)\o- lon^ed e\j)osure to tornialin. l\\c inactivation by means ol lormalin is acconijxinied by jjieservation of antigenieity. The active principles of the phenomenon are extracellular and soluble substances. Bacterial autolysis is not necessary for their j)reparation although they may also be ol)tained in autcjlyzed cid- tures. The best preparations are obtained imder conditions of iirsignificant cell autolysis, namely, in filtrates of washings of growth on solid media incubated for only twenty-four hours. The concentration of active principles in the first washing of a given niiml)er of meningoccjccus cells is approximately looo times stronger than in the disintegrate of the same number cjf cells chs- solved in the same volume of NaCl solution. With B. typJiosns the first washing is about ^^o times stronger than the respective cell disintegrate. There are fluctuations in potency of filtrates observed on steer- age. The filtrates may become inactive after several months of standing" in the refrigerator, and the potency of the filtrates may also substantially increase ^vithin one or t\vo months following" their preparation. Quantitative studies demonstrate that the active principles of the phenomenon are consistently retained inside of dialyzing bags. An opportiniity is thus aflforded to obtain considerable purification of the filtrates by means of dialysis. The pinification may amount to about a three-fold increase in reacting potency per milligram of dry ^veight and nitrogen. The active principles may be retained by Seitz filtering membranes. Ecker and Welch obtained concentration of the active princi- ples by the use of the method of Ecker and Rimington. The active principles may be precipitated by fidl saturation with anmioniinn sidphate and also by addition of acid alcohol. The precipitated and dried materials retain their potency for a considerable length of time. Various investigations c:)n concentration, jjurificaticen and frac- tionation of filtrates containing the active principles thus far car- ried out suggest that these principles are either protein in natine or perhaps associated with proteins and that they are closely re- lated to important antigenic ccjmponents of the preparations. Chapter IV IMMUNOLOGICAL BEHAVIOR OF ACTIVE PRINCIPLES OF THE PHENOMENON OF LOCAL SKIN REACTIVITY MOST of the observations just described having suggested a similarity between the active principles ot the phe- nomenon and bacterial exotoxins, a close study of their innniinological behavior A\as deemed important. The B. t\l)liosus active principles were selected for the basic work inasnnich as their neutralization and antigenicity ^vere obvious from the very beginning of the in\estigations. Later, the knowledge acquired was extended to studies on the immtniological behavior of men- ingococcus active principles. Since this part of the monograph deals w^th general studies on the natme of the active principles of the phenomenon, the special features concerning immunologi- cal properties of meningococcus active principles will be discussed in Chapter xiii dealing with certain practical aspects concerning the utilization of the phenomenon for preparation of immune therapeutic sera. EFFECT OF IMMUNE AND NORMAL SERA UPON SKIN-PREPARATORY FACTORS Most of the sera employed in these experiments ^\ ere prepared by the injection of toxic culture filtrates. Since a considerable ninnber of the rabbits died from intensive immunization with these hltrates, some animals were injected at the beginning of the immunization with autoclaved fdtrates; later WTth filtrates heated in the Arnold steam sterilizer; and with filtrates heated at 60° c. for one to two hours; and finally with potent toxic filtrates. Some animals received injections of w^hole bacteria, first heated to 60° c. for two hours, then lixe bacteria. Most of the injections were made intra\enously, Ijut in some instances the subcutaneous route ^vas employed. They were irsu- ally made every w^eek for three or four successive days for a period of six to seven weeks. The final Ijleedings were made about ten 90 LOCAL TLSSUE REACTIVITY clays alier the last iiijcclioii or caiiicr il llic animals aj)|K'aic(l sick. Hctcr()l()<>()iis sera Avere similarly prepared. B. lyj/ho.sii.s, B. jjdxilyjjlio.sti.s A and B, and B. shlo^a sera employed in some e\- j:)eriments, were prepared in the \e\\ ^'ork lioard ol Health Lab- oratories by immnni/ation oi horses. ()ther animals employed in this work 'were goats. As will l)e seen later from studies on the neutralizing anti- bodies, goats and rabbits Avere imsuitable lor inninmization, larger animals being more resistant to the toxic effects and yiekl- ing more satisfactory results. Rabbit sera employed in this work were selected from a large grouj) of animals receiving large doses of the toxic filtrates and sin"\i\ing for a period of sexen xveeks. Rabl)its injected repeatedly Avith large amounts of toxic fdtrates survixed only rarely. Since at the stage of the work described here nc^ large animals ^vere available, numerous rabbits used for repro- duction of the phenomenon tci B. lyj/ho.siis in its \arious aspects were saved and gixen repeated intraxenous injections. Only in this manner I was able to obtain a few sera which proved espe- cially potent. Tlie xvork described beloxv xvas carried out xvith fdtrates of B. ty/jJiosus tryptic digest broth cidtures. The toxic filtrates were mixed xvith sera in the desired proportions and the mixtiues were incubated in the xvater bath at ^^7^ c. for cjne hour. Immediately before injec:tion, the tubes xvere shaken. The fluid xvas injected into various areas, usually the loxver right and upper and loxver left areas of the epilated skin of the abdominal xvall. The upper right area xvas injected xvith a toxic filtrate alone, previously diluted xvith 0.9 per cent sodiimi chloride solution and kept in the xvater bath at 37° c. for one hour, this site serving as a con- trol of the susceptibility of the rabbit to the phenomenon. Some- times the order of injections xvas changed. Txventy-four hours later, the B. tyfjliosus culture fUtrate xvas injected intravenously in a dose of 2 c.c. jjer kilo of body xveight. The readings xvere made four to five hours after the intravenous injections. Since large intravenous doses of active principles xvere neces- sary in order to ol)tain uniform reactions in all sites tested, some of the rabl)its died earlier than four hours after the intravenous injections and, therefore, no clear-cut readings of the results could be made in these rabbits. The lethal effect of the prepara- tion could not be reasonably predicted and changed from batch IMMUNOLOGICAL BEHAVIOR Ql to hatdi. This tact constituted a serious handicaj) in attempts to (letcnnine the ability ot a seiiiin to neutralize the skin-prepara- tory factors. However, in carrying out the experiments on a large series of rabbits it was clearly seen that the skin-preparatory fac- tors could be specifically neutralized by homologous immiuie sera. In studies on serum neutralization of skin-preparatory factors, in iG per cent of rabbits no reactions were obtained cither in sites injected with bacterial filtrates alone or mixtures of bacterial fil- trates with immune sera. A certain number of normal rabbits receiving the injection of the toxic filtrates alone and especially of B. I\Ij/iosus tryptic digest broth filtrates, may remain resistant to the phenomenon. As ^vill be seen in a sidjsequent chapter, the above figine is the approximate percentage of spontaneously re- sistant animals and. therefore, can not be necessarily ascribed to the iLse of the immune serimi. There is a suggestive possibility, howe\'er, that Avhen a large amotnit of seriun is used in mixture Avith the skin-preparatory factors, a part of it may diffirse into the sites prepared with the toxic filtrate alone and thus, be respon- sible for negative results in control sites in rabbits normally susceptible to the phenomenon. This suggestion comes from the casual observation that in some rabbits in which the site injected with the toxic filtrate alone is too close to the site injected with a mixttne of the toxic filtrate ^vith the immune serum, there appears at the toxic filtrate site a reaction irregular in shape. The irregularity may invohe either the entire border of the reac- tion or only a part of it. The results in the remaining animals (those sho^ving reactions in the control sites) can be summarized as follows: There were two types of response to the injection of mixtures of specific immune sera ^\■ith the skin-preparatory factors. Out of a total of 316 tests with mixtures of B. typhosus filtrate and homologous sera, there were 135 neutralizations by the sera in various dilutions, and 181 failures to neutralize. This demon- strates the fact that neutralization of B. typhosus skin-preparatory factors by specific sera can be induced at will in a large propor- tion of experimental rabbits (about 42.5 per cent) . Varying the site of the injection remains without influence on the results, provided several multiples of the minimal skin-preparatory doses are used. The filtrate was diluted 1:2 in these experiments. This dilution combined Avith an intravenous injection of 2 c.c. of the filtrate per kilo of body weight, represented at least 32 skin-pre- ()2 LOCAL ILSSLE REACTIVITY paiatoiy doses. Alter c'stal)lisliiii^ dclmitcly the la(t thai ncutral- i/alioii ol skin-])iej)ar CO ^ « K -o - Tf « ^ -ft, >- ^■1^ HH j^ CJ) lO O^ lO f^ tJ- CT) I^ Cj -O ■-1 CO LO -. « ^ 2 ^ SS fe M 2 "^ 3 V 'U 0000 0000 ago . . > > (.o UD ^T) ^rs S aj j= ■ ■ <^'^ ^ '^ in LO LO LO • _ to <0 TO . . • • «j • • o c< cy CM cj ^-■^pq 6-5 10 H 1— ■ • oo 213 O O 5J 1^ C —, o o ^ K (J U =?^ 3 ^ £ ,-, CO CO 3 erum (o erum (o ngococc . Board i orse seru d d d :3 S (0.3 c, erum ( erum ( erum d O ,>,>, '0 •;2 -2 -s S S .^ .ti X ^w -5 c c c ,^ ^ nj CO « 5 3 >^ J3 XI .ii ^ S ffl 03 ti CO C .'>< ormal r ormal r herapei 148/43 ?derle a ^Coioioi^- ^ ,0 to <^ '^ <^ 'J c^ ^ iJ IJ ^'^ d i2 i2 2 s2 d -^ "-^ Z :2; h kJ KKffiffi S rS 1 ^ ^ «j' =:=:=: ^ s 3 3 s 3 « S " CO d ■ Tt- ^ Tf Tt- Tt- S? " 'v ' -a -a 13 -V ^ " (L) - ~ " ~ v V ■§ , -H Oj .:- 5j -i TD T3 3 3 Q t^ i '^ D -0 3 T3 ~ T3 3 3 -0 -a n: c n= c rt: c r3 r3 C C 1^ T3 3 T3 3 -a 3 t: -c 3 3 1^ d (J 0 lO 10 LO LO LO 01 01 cj a 0) 0) CI CJ 6 6 6 6 6 6 d d d d fc .2 tn N .2 3 CO c ■" -S c ^ T3 CO OJ (-1 V CO V c _:^ ^^ CO ^ OJ ■*-' tH ^ V CO ^ CO o (L) O X CO CO M ^ l - o o O CO 0) o o o LO O CO -^ o ^|.^ ^■5S o o CM o o CO o o lO C< CO ■* c< c^ o o 04-0) ^^S ^■11 IT) IT) m LD LO lO lO lO LO LO LO LO LO LO LO O LO LO lO Ci -^ '-1 ^ 2 - "5: -Si,.'-' o 3 S 3 3 o 3 3 3 jfl c. s ^ e ^ s ►i ■§ S r^ t^ j^ LO in lO r^vd '-D \o, \~ \tS \ bCi tJD bf) be _C ^ s .s - ^ C ^ w ^ ^ "^ ■5 " -S -5 " " " "5 " " " " u o o o o o o Oj C: 1^ U ^^ E s s 2 u u d d d ^ § -^ - O .^ 3 '^' :j .. d d ^ d ^ ^ c; '^ (N lO m LO LO LO ■^ ? S d (N 0) 0< 01 01 2 -^ =^ ^ c 3s2- 6^ S ►^ r -Ci, d d " lO d " '. t\l>li<).su.s stock stiains showed well pronounced motility. IWeiity-lour hour cultmes of various strains on plain agar were washed ill distilled water by centrifugalizalion and resuspended in 0.85 per cent and 1.4 per cent NaCl solutions. The suspensions were incubated in a water bath at 37.5° c. for two hours and then placed in the refrigerator for twenty-two hours. The C^r, stock suspension was completely agglutinated but the remain- ing strains stayed in suspension. The various strains were cultivated in series in mixtures of homologous immune sera with plain broth until rough variants were obtained. Some rough variants employed here were recovered from cultures cultivated for a number of generations in the presence of bacteriophage. Ihe various B. typhosus stock strains were also passed through mice in order to increase the virulence of the strains and possibly enhance their smoothness. The technique adopted for the purpose was as follows: A twenty-four hold old cgar slant of a given strain was emulsified in 0.85 per cent NaCl solution to a turbidity representing approximately one billion organisms per cubic centimeter. One c.c. of the emulsion was injected intra- jjeritoneally into two mice. Twenty-four hours later cultures were made on Endo agar plates from the heart's blood of the dead or killed mice. The twenty-four hour old growth on the plates was suspended to the above in- dicated tuibidity in 0.85 per cent NaCl solution. One-half of 1 c.c. of the emulsion was used for injection into a mouse. Only heart's blood cultures were employed for passages. The strains passed through mice were desig- nated "mouse" strains, with the name of the stock strain from ^\'hich they were derived and a number indicating the number of passages. None of the stock strains (including the Ty Tl strain) iniderwent any change in the appear- ance of the colonies in spite of numerous passages. The different stock strains of B. typhosus and B. coli, the vari- ants obtained by means of immune sera and bacteriophage and the strains passed through mice were designated accordingly and utilized as follows: Serum neutralizability of reacting factors of variants of typhosus and B.coli : The toxic filtrates necessary for the plienomenon of local skin reactivity -^vere "agar washings" filtrates. These filtrates were des- ignated by their nimiber and the name of the strain employed for their preparation. The "motise" strains did nc^t luidergo more than one passage through artificial media previous to their use for the preparation of toxic filtrates. The t^venty-four hour old heart's blood cultmes on Endo agar plates were seeded into plain broth. The plain brcjth properly diluted served as the inocidum for the Kolle flask agar cultures necessary for the preparation of the filtrates. IMMUNOLOGICAL BEHAVIOR IO3 The quantitative measurement of the reacting factors ^vas car- ried out by Method III (Chapter 1) . All the stock strains tested, as well as their variants, were ca- pable of producing highly potent toxic substances. The minimal dose of reacting factors produced reactions only in a small per- centage of rabbits in the majority of instances. Various multiples of the minimal doses were tested in groups of ten rabbits. Six and 10 iniits, however, gave a high percentage of reactions (i.e., "jry 100 per cent) . Diagrams 1 and 2 represent the results of neutralization experi- ments with various amounts of reacting factors of different toxic filtrates mixed with constant amounts of certain imnnuie sera carried out according to the method pre\iously described (Shwartzman. 19.^0/;) {pp. 95, 97) . As is seen from Diagrams 1 and 2, two types of neutralization were obtained, i.e., consistent and irregtdar neutralizations. The irregular neutralizations which take place Avhen the amount of reacting factors is increased beyond a certain maximum are most likely due t(j the fact that the mixtiuxs contain a small amount of free reacting factors so that only the rabbits possessing a high degree of reactivity are able to respond with reactions. It does not appear safe to make quantitative estimations of the neutraliza- ])ility of a filtrate by determining the largest amount of react- ing factors inei!;ula}ly neturalized by a serum since the exact mechanism of irregular neutralization is not known as yet. This type of neutralization offers only a rough cpiantitative comparison of the degree of neutralizability of \'arious hltrates by a given serum. Fairly accurate quantitative comparisons of the degree of neutralizability of reacting factors derived from various filtrates can be made, however, by determining the largest amoinit of reacting factors consistently neutralized by a given serum. The folloAving becomes evident from the analysis of data obtained. Various stock strains of B. iyjjhosus show striking differences in the serum neutralizability of their reacting factors. Out of 4 strains chosen for this ^vork, 2 occupied extreme positions, one shoAving a high degree of neiuralizability Avith strain homologous and strain heterologous sera (Ty Tl stock and Sera S5 and Sc) and another strain (Ty 159 stock) shoAving no consistent but only irregular neutralizations w^ith the same sera. A horse serum (Horses Bleedingiu) which was prepared by injection of mix- tiUTS of various strains including Ty 159 stock also failed to give 104 LOCAL TLSSIIE REACTIVITY il3noj 3 loo^s ■5 ^too'^s Tx Xx S'^esnopi "^x J^X 2j9s es^ tx -^I 9^3snon 69T ^X SaSaqd 0Z8 ^X S8asTiOK Of'a J^X Hoo^s isi Xx 2jas 923 Tx ^x SSasnopu ^-1 ^X 3I0O1.S tx -^X SSasnoji o^Z ^-l S^ssnofv 69T -^X jjooT.g 6ST 'fx Jioo^s ^99 /x 3(00^3 isx /x 923 3jss Tj, ^x 3j8S qfz T-i /x SSasnon T^x ^X 5100^3 Tx ^l T^x ::: I T ^x :'• i m s ^x ^x : ■^x ^x : t( II ::: sq.7tm Sirtio^ea jo jsqnmn IMMUNOLOGICAL BEHAVIOR lor. consistent neutrali/ations \vith Ty 159 stock reactin" factors, al- though fixing a liigh degree oi irregular neutralization. The re- maining 2 stock strains occtipied interniediate positions in the Serum Serufii Serua Seru.li Horse 5 Korse 5 Horse 15 Horse 15 Bl 114 Bl 125 DIAGRAM 2 Bl 115 Bl 126 = Not tested • • « • • • = Irreg'olar neutralization = Complete neutral! :ation = No neutralization degree of consistent neutralizations, one (Ty 8G4 stock) being of lower neutralizability than the other (Ty 157 stock) . The strains preserved the same differences Avhen tested with the polyvalent horse serum (Horse.-, Bleedingi^) . The present tentative expla- nation of these findings is as follows: The reacting factors of the strain showing the highest degree of neutralizability appear to be the simplest in their antigenic 106 LOCAL TLSSUE REACTIVITY Structure, as avcII as ol hi<>li autioeuicity. The reacting factors of the remaining stock strains may represent various degrees of in- creasing antigenic complexity, the Ty 159 stock strain being the most complex one. The additional complicating components, however, have to be considered of low antigenicity, as is seen from the fact that ]:)olyvalent sera, in the preparation of ^vhich there was included Ty 159 stock strain, and certain homologous monovalent rabbit sera, fail to give a high degree of neutralization Avith these factors. Finther proof of this contention is brought out by the following work ^vith "mouse" strains: The passage of B. h'ljJiosus stock strains through mice may or may not affect the neutralizability of their reacting factors. In one strain Avhich ^vas considered above as the simplest in the antigenic structure of its reacting factors (Ty Tl stock) there oc- curred a partial loss of neutralizability (Ty Tl Mouse,-,^ and Sera Sr, and S(;) . This fact may be interpreted as an acc|uirement of additional components. The factors of another strain considered as the most complex ones (Ty 159 stock strain) showed no change in the neutralizability of the reacting factors after 46 passages through mice. However, there occiuTed an increase in the antigenicity of the factors of this strain. This fact is ^vell illustrated by the experiments ^^ ith R-^^09 serinn homologous for Ty 159 Mouse^r; strain ^vhich shoAvs a high degree of consistent neutralization ^vith the reacting factors of the latter. The passage of a strain through mice is not necessarily accom- panied by an increase in the antigenicity of the factors of lo^v neutralizability. The Ty 240 stock strain ^vhich showed factors of low neutralizability was tested with anti-Ty Tl stock serum (S5) as ^vell as with an anti-mouse-passed-Ty 250-strain serum (Sr,) (previous unrecorded experiments) . Both sera showed low neu- tralizations ^vith the reacting factors of the mouse strain. As is also seen from Diagrams 1 and 2 there occurs an almost complete loss of neutralizability of reacting factors of the rough variants by the anti-stock sera. The change is illustrated by numerous experiments (Ty Tl 2^^8 ser.o and Ty Tl 245 ser.j, Ty 870 phages vs. Sera S-,, So, Horses » , and 15 ). The loss of neutralizability is associated Avith acquisition of a ne^v antigenic specificity. The "rough" reacting factors are consistently neutral- ized by homologous sera and by polyvalent sera, in the prepara- tion of which there were included the rough variants (Serum S3 IMMUNOLOGICAL BEHAVIOR 107 and coli rough; Typhoid rough strains and Sera R-30'^ and Horses 5/ 15 15/ ). Moreover, the 2 B. typhosus rough ,125, ! lie,. / Uh' - ■' _ ^ variants (Ty Tl 238 ser.o and Ty Tl 245 ser.2) , both derived from tlie same stock strain (Ty Tl) , and a third B. typhosus rough xariant (Ty 870 phages) deri\ed from another stock strain (Ty 870) varied in the degree of their neutraHzabihty l)y anti- rough sera (Horsesy^^^ and ^/nd ■ The question as to whether the "rotigh" reacting factors are able to stimuLate antibodies against normal reacting factors is left open.^ In the Avork which follo^vs, an attempt Avas made to correlate the above findings with the morphological appearance of colonies, agglutinability and virulence of the various strains and also ^vith the serum precipitability of the filtrates studied. Morphological appearance of colonies and the neutralizability of reacting factors: As is seen from above, the stock strains of B. typhosus produced reacting factors of various degrees of neutralizability. The Ty Tl stock strain, producing factors of high neutralizability and anti- genicity and apparently the simplest in antigenic structure, was morphologically rough-smooth and ^vas capable of prompt trans- formation into complete roughness. The reacting factors of the remaining stock strains which differed in neutralizability and Avere considered more complex did not display any difference in the colony morphology, all appearing normal. Neither ^vere these strains transformed into rough variants in spite of persistent at- tempts. It appears, therefore, that strains of normal colony ap- pearance differ in the neutralizability of reacting factors ^vhich they produce. The different rough variants produced by passage through immune serum and bacteriophage appeared of morphologically identical colony structtne, ^vhilst they differed in the neutraliza- bility of their reacting factors. The passage of stock strains through mice ^vhich may or may not result in a change of neutralizability ^vas not accompanied by any changes in the morphology of the colonies. '^The objection to this work which may be raised is that tlic toxoid formation (described later) may account for the resuUs. This is obviously not the case since neutralization of the reacting factors of the variants ^\as obtained wiili homologous sera where the stock sera failed to do so. 1()(S LOCAL riSSLE REACTIVITY Serum agglutinability of various stock strains and variants of B. typhosus and the neutralizability of reacting factors: These tests ^vere performed in the usual manner, except tliat 0.2 1 per cent NaCl sohuion was used for making serinn dihi- tions for the aggliuination of the rough variants. The gro^\th of rough \ariants on solid media used for the tests was suspended in distilled ^vater, ^vashed by centrifugalization and resuspended in 0.2 1 per cent NaCl solution. Tiie results of the agglutination tests ^vith various cultmes bring out the follo^ving: Two modes of comparison of agglutinability of various strains by immune sera were used: (a) the agglutination titer, namely, the hiohest flilution of the serimi ^vith which agglutination is obtained, and (b) the degree of agglutination in various dilu- tions of sera. The agglutination titer of the stock strains could not be correlated with the degree of neutralizability of reacting factors deri\'ed from the same strains. For instance, the agglutina- tion titer of Serimi S.-, was the same for Ty Tl stock and Ty 159 stock strains, ^vhilst there ^vas consistent neutralization of 26 re- acting units of the first and no consistent neutralization of the latter strain. In another example, the reacting factors of Ty 157 stock strain "were neutralized by serum Horse.-, Bleedingiu to a higher titer than those of strain Ty 864 stock, but the agglutina- tion titer ^vas higher for the second than for the first strain. Con- versely, it appeared that the degree of agglutination obtained in various dilutions of sera w^as an indication of differences in neu- tralizability of stock strains of B. typJiosus. Thus, shoidd one ar- range the strains in order of their comparati\e neiuralizability and in order of the intensity of agglutination in low dilutions of sera the same order is obtained for both. i.-Ty Tl stock, 2.-Ty 157 stock, 3.-Ty 864 stock, and 4.-Ty 159 stock would fall in this order. The above described observation is to be differentiated from previously reported results showing that the agglutinating and neutralizing antibodies do not run parallel in immune sera. The loss of agglutinability due to passage of strains through mice was not an indication of a change in the neutralizability of the reacting factors. For, on the one hand, there ^vas the exam- ple of the Ty Tl mouse strains Avhich, after passage through mice, produced reacting factors of louver neutralizability ^vithout change in agglutinability. On the other hand, in contrast to this strain. IMMUNOLOGICAL BEHAVIOR 1 Ot) Ty 159 mouse imdei^vent a considerable loss in aggliitinability \vithoiit concomitant loss of neiitralizability of its reacting factors. The results of agglutination tests with the rough variants of B. typhosus and the neutralization experiments indicated that the anti-rough agglutinating and neiUralizing antibodies did not rini parallel. In fact, Serum Horse.-, Bleedingii4 Avhich Avas not able to neutralize consistently the reacting factors of strain Ty Tl 238 ser.2 had a higher agglutination titer than Serimi R-203 which showed consistent neutralization of 43 reacting units of the same filtrate. As previously noted, various degrees of neiitralizability exist amono- roiiq-h variants derived from the same stock strain. There appeared to be a parallelism between this and the aggliitinability of the same variants by immune sera. (Horscir, vs. Ty Tl 238 ser.^ and Ty Tl 245 ser.o.) No antigen analysis (i.e., somatic and flagellar antigens of Theo- bald Smitii, etc.) of the various strains employed has been made as yet. Serum precipitability of culture filtrates of various stock strains and variants of B. typhosus and neutralizability of reacting factors: A quantitative analysis of the results could not be safely made since se\'eral dilutions of the antigens Avere not used in every in- stance. The following was, hoA\e\er, concluded: The "rough" reacting factors which ^vere not neutralized by anti-stock sera were also devoid of precipitinogens for these sera (Ty Tl 238 ser.o and Ty Tl 245 ser.^ strains and Serum S.-,) . The appearance of neutralizing antibodies for the "rough" re- acting factors was coincident ^vith formation of anti-"rough" pre- cipitins (Ty Tl 238 ser.o and Ty Tl 245 ser.o and Serum R-303) . Abundant precipitation betAveen the filtrate and serum could occur in spite of low neutralization (Ty 159 stock and Mouse^,; and Serum Sn) . Virulence of B. typhosus variants and neutralizability of B. typhosus reacting factors : The various strains tested were transplanted into plain broth pH 7.4. One c.c. of a broth culture previously diluted in sterile plain broth 1:4 was injected intraperitoneally into each mouse. It was necessary to test a large number of mice in x'lew of the IIO LOCAL TLSSUE REACTIVITY {()iisi(lcial)lc iiulix idiial xariations ol tlie la])()rat()ry animals in response to the effect ot the colcHi-typhoid gronp. It could be safely assumed from these experiments that the passage of the Ty Tl stock strain through mice did not increase its virulence. It was clear, therefore, that the change in the neutralizability of the reacting factors which occurred in the mouse strain ^vas not re- lated to its virulence. There \\as, ho^vever, unquestionable lower virulence on the part of the rough variants of B. typliosns, as compared to the virulence of the stock strains from which they were derived. This was previously observed for the colon-typhoid group by Topley and Aryton, Jordan, Orcutt, White, Goyle, Schiitze, and others quoted by Hadley (1927) . The lo^vering of virulence was coinci- dent with a change in the neutralizability of the "rough" react- ing factors by anti-stock sera. Ho^vever, there Avere differences in the neutral izal)ility of the reacting" factors derived from the t^V() "rough" variants by anti-rough sera. No differences could l)e recognized in the virulence of these rough variants. Thus, it may be concluded that the active principles of B. ty- pJiosiLS and B. coli possess a clear-cut strain and variant specificity. By the "strain" specificity is meant the fact that various strains, chosen at random, of a given species of bacteria produce reacting factors Avhich differ in their neutralizability. The differences are explainable on the basis of various degrees of antigenic com- plexity. The "variant" specificity is demonstrated l)y a luunber of facts. The passage through mice induces in certain strains a partial loss of neutralizability of reacting factors. This is probably due to variation to^vards smoothness Avith an increase in the antigenic complexity. As a matter of fact, the reacting factors of a strain . considered as the simplest in their structine imderwent a partial loss of neutralizability after passages through mice, ^vhilst the neutralizability of reacting factors of a strain found to be the most complex before the passage through mice ^vas not influenced by the latter procedure. The other type of "variant" specificity is conditioned by the change of a strain into rough. The "rough" reacting factors acquire a new specificity. The above described changes in the neutralizability of the re- acting factors taken together cannot be consistently correlated ^vith colon morphology, serum agglutinability and virulence of strains, nor with the serum precipitability of the culture filtrates. IMMUNOLOGICAL BEHAVIOR 111 Somewhat similar results ^vere obtained ^vith active principles of meningococcus (Chapter xni) . CONVERSION OF ACTIVE PRINCIPLES TO TOXOID The con\ersion of toxin into toxoid has been studied by Ehr- lich (1898) and Ramon (1923, 1924) for diphtheria toxin, by Behring and Ransom (1898) , Lowenstein (1909) , and Eisler and Lo\venstein (1912) for tetanus toxin, and by Weinberg and Goy (1924) for the toxins of the B. botulinus and of the gas gan- grene anaerobes {i.e., B. oedematiens, etc.) . Ramon (1925) has also con\'erted abrin and cobra venom into toxoids. The change ^vas obserxed to occur spontaneously (Ehrlich, Behring and Ran- som) , and also could l^e induced artificially by means of various physical and chemical agents; namely, for tetanus toxin iodine trichloride (Behring and Ransom), formalin (0.1 to 0.2 per cent) and exposure to a Nernst lamp (Lowenstein) , formalin (o.i to 0.2 per cent) and slight heat (30° c.) (Eisler and Lowen- stein) ; and for diphtheria toxin formalin (0.3 to 0.4 per cent) and heat (40-42° c.) (Ramon, 1924) . By similar procedures (for- malin and heat) abrin, cobra venom (Ramon, 1925) , and toxins of the B. botulinus and the gas gangrene anaerobes (Weinberg and Goy) have also been converted into toxoid states. These studies elaborated the concept of a toxoid first an- nounced by Ehrlich. In its completed form, the definition of a toxoid is that it is a toxin so altered that the toxicity is decreased whereas the antibody-combining capacity and the antigenicity are essentially luidiminished. It has been observed (Low^enstein, 1909; Ramon, 1925) that the antigenicity of tetanus and diph- theria toxoids varies with the potency of the mother filtrates. It can be estimated by the antitoxin-combining capacity, to "which it is directly proportional. Fluctuations in potency of filtrates employed in my studies suggested toxoid formation and necessitated the development of a standard procedure in Avhich the titer of the filtrates \\as calcu- lated from the dilution neutralized by a given amount of neu- tralizinoj serum selected as a standard. This Avas done as follow^s: An "agar Avashings" B. typhosus filtrate freshly made and highly active was titrated for reacting potency (Method III) (600 re- acting units per c.c.) . The neutralizing potency of an immune antityphoid seriun \\as then titrated against this filtrate. In these 112 LOCAL TISSUE REACTIVITY till at ions it was loiind that 0.25 c.c;. of the serum neutralized c:on- sistently 150 reacting units ot tlie filtrate. This serum was selected as the standard. In subsequent work, in addition to direct titra- tions of the reacting jjotency of the filtrates, there was also deter- mined the maximum amount of each filtrate completely neutral- ized by 0.25 c.c. of the standard serimi. This amoimt was then considered as containing 150 reacting units and the titer for one c.c. of the filtrate Avas derived by multiplying the dilution of the filtrate by 150. Thus, if a dilution 1:5.25 was the maximum amount completely neutralized by 0.25 c.c. of the standard seriuii the filtrate was considered as containing 750 reacting imits. When the direct and indirect titrations just descril)ed ^vere carried out Avith a large group of filtrates prepared under identical condi- tions and from the same strains, the following was foinid: There ajjpeared to exist no constant ratio between the direct and indirect titers, i.e., toxicity and antigen-combining titers. The ratios varied ^vithin the following range: 1:1, 1:1.5, i'^. i:'^ 1:4 and 1:6. Fresh filtrates ^vere likely to sho^v ratios of 1:3, 1:4, and 1:6. The louver the reacting potency, the greater ^vas the discrep- ancy between the reacting titers and the titers deri\'ed from in- direct neutralization titrations. After the filtrates were stored for t^vo to three ^veeks the reacting potency frequently increased. In these instances, the direct and indirect titers coincided more closely, sometimes changing from 1:5 to 1:1.5 and 1:2 ratios. On storage extending for sexeral months a sudden drop in potency occurred. In such cases the ratios became approximately tfie same as those Avhich existed soon after preparation of the filtrates, i.e., before the rise in the direct titer took place. When the indirect titers of many preparations were compared it became obvious that the antibody-combining power fluctuates within a considerably narrower range than the reacting potency and moreover, the antibody-combining power remains stationary in spite of the wide fluctuations of the reacting potency. These facts can be safely interpreted as toxoid formation ^vith retention of the antibody-combining power. This is further substantiated by the following observations of Klein made in my laboratories (1932) . Meningococcus "agar Avashings" filtrates were mixed ^vith for- malin in a final concentration of 0.4 per cent, placed in ground glass stoppered bottles, sealed with paraffin and kept in an in- cubator at 37° c. for 5 days. The skin-preparatory and reacting po- IMMUNOLOGICAL BEHAVIOR II3 tencies of these preparations were titrated by means ol combined intradermal injections of tlie formalinized meningococcns mate- rials Avith intravenous injections of active meningococcus filtrates; or by means of combined intradermal injections of active menin- gococcus filtrates with intravenous injections of formalinized fil- trates. In this manner it was found that the potency of the forma- linized materials, as measured by the titer of the reacting units, was markedly decreased. In contrast the antigenicity and the anti- body-combining capacity remained essentially imimpaired. The antibody-combining capacity of meningococcus toxoids ^vas studied as follows: Mixtines of meningocccus active principles with antimeningo- coccus horse serum ^vere titrated for reacting potency. Complete neutralization was obtained with 100 imits of meningococcus re- acting factors. "With the use of an equal amotuit of sertmi, com- plete neutralization was obtained wtih only 1.7 reacting tuiits of meningococcus toxoid. The ratio of the antibody-combining capac- ity of meningococcus toxoid to that of meningococcus active prin- ciples was, therefore, that of 10:14, ^^^^^^ showing in the toxoid a loss of only 28.7 per cent of the antibody-combining capacity asso- ciated with 98.8 per cent loss of toxicity as compared to the mother filtrate. It remained to determine whether 1.7 reacting units of menin- gococcus toxoid actually held in combination all of the neutraliz- ing antibody in the serum used for the neutralization titrations. If the mixture contained an excess, i.e., unbound antibody, then complete neutralization should still be obtained after the addi- tion of a small amount of active principles to the mixture. Accord- ingly, the following experiment was performed: Each of 4 rabbits received an intradermal injection of 0.25 c.c. of undiluted active meningococcus principles, followed t^venty-foiu' hours later by an in- travenous injection of a mixture of 1 c.c. of meningococcus toxoid diluted 1:10 (1.75 reacting luiits) plus 1 c.c. antimeningococcus immune horse serinu plus 0.5 c.c. of meningococcus active principles diluted 1:70 (10 reacting imits) per kilo of body weight. Reactions were elicited in 3 of the 4 rabbits (no neutralization) . The above experiment demonstrated conclusively that 1.7 re- acting units of meningococcus toxoid actually combined ^vith all of the antibodies in 1 c.c. of the immune antimeningococcus serum. It remained to ascertain the antigenicity of meningococcus toxoid. 114 LOCAL TLSSUE REACTIVITY Loweiistcin had noted that the antigenicity of a tetanus toxoid was directly proportional to its antitoxin-conihining capacity. In \ie\v oi the fact that the inenin,o()(()cciis toxoid described by Klein possessed 71.3 per cent ot the antibody-ccjmbinino capac- ity of the mother filtrate, it ^vas expected that the toxoid Avotild still l)e antigenic. The aggltitinating titers for Circnip I of the rabbit meningo- cc:)cciis sera produced by subcutaneotis and intravenous injections of Grotip I meningoccjcctis toxoid for a period of fcjiir to five weeks varied from 320 to 1280. Sera obtained from trial bleed- ings of the rabbits prior to immtniization ^vere free of spontane- ous agglutinins for meningococcus. It may be safely assiuned that the meningococctis toxoid studied by Klein retained its antigen- icity. The sera contained no precipitins either for meningococcus acti\'e preparations or for meningococcus toxoid. The sera Avere unable to netUralize meningococcus active principles. Ho^vever, it has been amply emphasized that rabbits do not readily form neutralizing antibodies nor precipitins for meningococcus bac- terial filtrates. It is also of interest that there may occur a similar spontaneous formation of meningococcus toxoid. Klein reported that the prep- aration of meningococcus "agar washings" toxic filtrate 1675 was completed by him on June 16, 1931. In the period, July 15 to 17, 1931, the titer of reacting units and the antibody-combining- capacity ^vere determined. Antimeningococcus immune horse serimi (with a small amoiuit of auxiliary antifjody, Shwartzman, 1931^) ^vas used for the neutralization experiments. In the period Dec. 18 to 23, 1931, these determinations were repeated. It W'as found in the latter experiments that the titer of react- ing imits dropped from 1300 to 150 tmits. Ho^vever, the same amoiuit of serum, which completely neutralized 1 c.c. of a 1:10 dilution of meningococcus toxin 1675 in July was necessary for complete neutralization of 1 c.c. of a 1:13.6 dilution of meningo- coccus toxin 1675 in December. Therefore, in December the toxin retained about 1.5 per cent of the July reacting titer. In contrast to this, in December it retained 73.5 per cent of the antibody-combining capacity present in July. Thus, Klein's experiments demonstrate clearly that meningo- coccus active principles treated ^vith formalin can be modified similarly to diphtheria toxin. The antigenicity and the antibody- combining capacity are essentially unimpaired. Furthermore, no IMMUNOLOGICAL BEHAVIOR II5 comparable loss of antibody-combining capacity is associated with the decrease in toxicity occmring on standing. It was the detoxi- fying alteration of the sort undergone by diphtheria toxin that lead Ehrlich to the description of diphtheria toxoid. Since menin- gococcus active principles can spontaneously or artificially be modified in a similar manner, it appears justifiable to consider such modified meningococcus active principles of the phenomenon under discussion as meningococcus toxoids. , In further attempts to determine the effect of immune sera upon the active principles of the phenomenon, it seemed im- portant to investigate ^vhether this ^vas in the nature of neutralization similar to the kno^vn toxin -|-antitoxin reactions. The other possibility still remained that precipitates, formed on addition of the specific antisera to the active bacterial filtrates, may remove the acti\e principles from the solutions by a process of adsorption. In the ^vork about to be described it ^vas sho^vn, ho^vever, that the inactivation of reacting factors occtns in "mul- tiple proportions"; that the precipitates formed do not inactivate the active principles; that the protection against the active prin- ciples may also be obtained in vivo; that the protection ifi vivo also occurs in multiple proportions; and, finally, that specific active acquired immunity can be demonstrated. REACTING POTENCY OF PRECIPITATES AND SUPERNATANT FLUIDS DERIVED FROM MIXTURES OF ACTIVE BACTERIAL FILTRATES WITH IMMUNE SERA As already noted, when a B. typhosus "agar washings" culture filtrate is mixed with an immune anti-typhoid horse serum in necessary proportions, there occurs complete inactivation of B. typhosus reacting factors. The mixtures usually form abundant precipitates. In order to determine whether the in- activiation of active principles was not merely due to their adsorp- tion by the precipitate formed, various numbers of B. typhosus reactino- factors were mixed Avith a constant amount of anti- typhoid horse serum. The mixtures made in the proportions in- dicated in Table v, precipitates and the supernatant fluids derived from these mixtures, Avere all tested for reacting potency. As may be seen from Table v, 1 c.c. of the immune serum com- pletely neutralized 750 B. typhosus reacting iniits. Precipitates derived from the mixtures ^vere respecti\ely dexoid of reacting potency. When an amount of B. typJiosus reacting factors exceed- ii6 LOCAL TISSUE REACTIVITY in*; Uic neutral i/in<; j)()tciuy ol the sciiiiii was einjiloycd, the precipitates proved active. Inasmuch as the jM)teiicy of the pre- cipitates appeared only with the increase in the amotint of B. l\l)/io.sus reacting factors in the niixttne, it became suggestive that these precipitates, in contrast to those derived from serum -|- antiserum mixtures (to be descriljed in Chapter ix) , possessed no indej^endent reacting potency. In partially neutralized mixtures {i.e., looo, 1500, 2000, 3000, B. typhosus reacting factors each mixed with 1 c.c. of serum) the non-neutralized active principles were associated with the precipitates. Thus, the precipitates de- rived from mixttnes of 1000, 1500, and 2000 reacting imits per 1 c.c. of serum, were jjotent, ^vhilst the supernatant fluid from mix- ttnes less than 3000 iniits ^vith 1 c.c. of serum was inacti\e. It may be considered as conchisive that the mere formation of pre- Table V REACTING POTENCY OF PRECIPITATES AND SUPERNATANT FLUIDS FROM MIXTURES OF B. TYPHOSUS FILTRATES WITH IMMUNE SERA Aiixtures of B. typhosus reacting factors and anti- typhoid horse serum HiBlin Reading potency of mixtures, per kilo of body weight Reacting potency of precipitates, per kilo oj body weight Reacting potency of supernatant fluids, per kilo of body weight Nos. of Amount reacting of units H^Bliyi C.C. 300 400 750 1000 + -f 1500 + + 2000 + + 3000 + -f + + = three rabbits tested showing severe hemorrhagic necrosis, o = no reactions obtained in a group of three rabbits. cipitates and possible adsorption of the active principles is not sufficient for their inactivation. This is also supported by numer- ous observations in which immune sera ^vhich formed precipitates ^vith the filtrates containing the active principles failed to neu- tralize them. Apparently, therefore, the addition of homologous neutralizing antibodies to B. typJiusus active filtrates results in IMMUNOLOGICAL BEHAVIOR II7 processes essentially consisting of two phases: 1. Flocculation ol the antioenic components of the bacterial hltrate by imnunie serum with adsorjjtion of the active principles by these complexes. 2. Neutralization of the active principles in these precipitates. It is difficult to exjjlain the fact that, on the one hand, as to be shown later, certain precipitates formed throuoh tlie inter- action of certain antigens with antibodies, possess reacting po- tency ^vhilst, on the other liand, precipitates residting from the interaction of B. typhosus with innmuie sera not only do not con- tribute to the reacting potency of the mixtiue. but are able to neutralize the actixe |)rinciples of the filtrate which they adsorb. The differences most likely depend on the strength of the luiion between the antibody and the active ])rincij)les. This, in turn, may be conditioned by the concentration of the antibodies in the immune serum or by the degree of interference of other antigenic components in the filtrates. The inhibitory effects of multiple an- tigens on the antigen -j-antibody interaction were demonstrated by the work of W'eil, Bayne- Jones, Linossier and Lemoine, Eisen- berg, Ascoli, von Dungern, Zinsser, Wells, Opie, and others (quoted by Zinsser, 1931). NEUTRALIZATION OF ACTIVE PRINCIPLES IN "multiple proportions' "Multij^le proportion" neutralization of active principles of the phenomenon was clearly demonstrated with B. typJiosus fil- trates inasmuch as immune anti-typhoid sera of high potency xvere available. A typical example xvhich xvas consistently repeated xvith many batches of different B. txphosiis culture filtrates is cited below: .A'o. of neutralized Approximate Amount of ant i- B. typhosus /•?- Multiples of multiples of typhoid serum acting factors serum used neutralization 0.1 c.c 75 O. 25 c.c 150 2.5 2 I c.c 550 9 7 It is of special interest that similarly to the dij)htheria toxin- antitoxin neutralization, the use of increasing midtiples of serum brings about a roughly proportionate increase in the amount of filtrate neutralized. The multiple jsroportion neutralization is also brought out by experiments on concentration and pinifica- tion of anti-typhoid horse serum (Reiner and Shwartzman, 1930) . Il8 LOCAL ILSSUE REACTIVLIY Table vi CONCENTRATION AND PURIFICATION OF ANTITYPHOID HORSE SERA Vol- N Neu- .Neutral Total Yield of Sample ume in mg. per ^"Agglu- tinins Precip- itins tral units units per .No. neutral- neutral- izing c.c. c.c. per c.c. mg. nitrogen izing units anti- bodies Original horse serum No. 200 II-3 10240 32 145 50 I 16000 I lo A .52% Antibody solution recov- 75 0.3 6400 16 200 2650 6000 ered Concentration per mg. N 23X 19X 53X Original horse serum No. 480 1 1.4 10240 32 145 50 278400 I lo A 40% Antibody solution recov- 96 0.3 20480 64 300 4000 1 15200 ered Concentration per mg. N 76X 76X 80X Original horse serum No. 450 "•5 10240 32 145 50 261000 133 45% Antibody solution recov- 45 I.I 102400 128 650 2350 1 1 7000 ered Concentration per mg. N 105X 42X 47X Original horse serum No. 350 n-5 12800 16 90 30 126000 133 44% Antibody solution recov- 35 1-3 25600 64 400 1200 56000 ered Concentration per mg. N Original horse serum No. 350 "•5 12800 16 90 30 126000 133 44% Antibody solution recov- 35 1-9 25600 64 400 850 56000 ered Concentration per mg. N 12X 24X 28X The chemical studies were done in the laboratory of Dr. H. Sobotka, as follows: The plasma or serum was diluted to ii4 times its original volume with distilled water. Instead of a saturated ammonium sulphate sohuion, a 47.5 per cent sodium-magnesitmi sulphate solution was used as the protein precipitant to facilitate the nitro- gen analysis. Twenty-eight parts of this solution ^vere added to 72 parts of the diluted serum. The precipitate which contained fibrinogen and part of the euglobidin was discarded. Then a suf- IMMUNOLOGICAL BEHAVIOR 1 1 9 ficient amount of the salt solution ^vas added to the filtrate to bring the stilphate concentration up to 19 per cent. The precipi- tate obtained was allo^\ ed to become almost dry on the filter paper and was then transferred to a dialyzing bag in running ^vater. After about six hours the white pasty mass became a clear brown solution Avhich gradually grew turbid and later showed a white precipitate. The dialysis ^vas continued for seventy-two to ninety- six hotns initil all the sidphate ion had disappeared and no more precipitate seemed to form. The globulin sohuion was centri- fuged and the precipitate ^vas made up with physiological salt soliuion to a con\enient volimie, e.g., 1/10 of the original seriun. Extraneous matter ^vas removed by centriftiging, and the super- natant fiuid filtered through a Berkefeld "V" candle. In the neiuralization experiments a constant vohnue of crtide sertim and the same xolume of concentrated sertmi Avas titrated against increasing amoinits of reacting factors (Method III) . As is seen from Table vi, the total yield in neutralizing anti])odies was from 44 to 52 per cent of that in the original serimi. The concentration, as indicated by the quotient Neutralizing Units/ mom. N, varied from 28 to 80 times. From these results it could be also safely concltided that the neutralization of reacting factors is not similar to the neutralization of endotoxins. Pro- vided a stifficient concentration of netitralizing antibodies is used, the active principles in doses of a ^vide range may be completely neiuralized. Experiments on midtiple proportion netitralization of B. coli reacting factors brotight oiu, hoAvever, a curious fact ilhistrated below. .Vo. of neutralized Approximate Amount of ant i- 'Q. coli reacting Multiples of multiples of coli serum factors serum used neutralization 0.25 c.c 50 0.5 C.C 175 2 3.5 1-5 c-c 350 6 7 As may be seen, increase in the amount of sertnn used from 0.25 c.c. to 0.5 c.c, 0.75 c.c, and 1.5 c.c brought about 3.5, 5 and 7 fold increases in neiuralization, respectively, instead of expected 2, 3, and 6 fold increases. On the other hand, if the neutralization figures obtained ^vith 0.25 c.c. are disregarded and multiples of 0.5 c.c. are considered, then 1.5 c.c. representing three midtiples of 0.5 c.c. bring aboiU a 2-fold increase in neiuralization. The I 2() LOCAL ILSSUE REACTIVITY latter coniparison (Icnioiistralcs. tlu'ictore, a 'imiliiplc jjiopor- tioii" neutiali/ation similar lo thai obtained in B. lyjjho.sii.s ex- })crinients. These cliscre|)ancics suggested the possibility tliat there may exist a factor in the serum which aids the neutraliza- tion and being present in insufficient amounts in 0.25 c.c, the neutralization is not brought out fully until a sufficiently large amount of serum is used which can supply this factor. The experi- ments described below deal with this problem.^- AN ANTnU)!)^ AUXII.IARV TO SERUM NEUTRALIZATION OE MENINGOCOCCUS REACTING FACTORS Studies on meningococcus brought out the fact that certain anti-meningococcus sera of a \o\v CN titer possessed a high IN titer. Moreover, if multiples of 0.25 c.c. ^vere used, the CN titer did not increase proportionately ^vhilst there Avas observed a roughly jjroportionate increase in the IN titer. It seemed possible that the inability of anti-meningococcus sera to neutralize com- pletely the toxic factors in midtiple proportions may be due to the lack of the hypothetical auxiliary factor assumed to exist in B. coli sera just discussed. A search for this hypothetical auxil- iary factor was done by studying the neutralization of meningo- coccus reacting factors by means of mixtures of anti-meningo- coccus sera 'with heterologous immime sera which by themselves Avere capable of complete neutralization in midtiple proportions of their homologous toxic sidjstances. It ^vas foinid, thus, that the mixtures of anti-meningococcus horse sera with anti-coli horse sera displayed a remarkable neutralizing potency. First, these mix- tures exhibited a substantial increase in the IN titer; second, they showed complete neutralization of large numbers of toxic units; and, third, they differed in their potency for filtrates of various serological groups. It was possible to determine the CN titers of the mixtines. Thus, if a given mixtine neutralized completely 100 reacting units and only irregularly 130 units, the CN titer Avas considered approximately 115 units. The CN determinations, as above men- tioned, are considerably more accurate than the IN titrations. Data were also obtained on the reciprocal quantitative relation- ships between the anti-coli and anti-meningococcus horse sera. In the examples studied the anti-coli horse serum Avas by itself capable of giving complete neutralization of meningococcus re- acting factors. The neutralizing property of this serum, hoAvever, IMMUNOLOGICAL BEHAVIOR 121 differed in sexeral respects from the specific antibody of anti- meningococciis serum: It had no atJgkitinins and precipitins for meningococcus; it neutralized the toxic factors of both serologicaf Groups I and III to the same degree; it ^vas capalole of complete neutrahzation in multiple proportions. The next step ^^■as to determine ^vhether the high complete neutralization obtained ^\ith the mixtines described 'was due to summation of neutralizing titers of sera making up the mixtures. The sera used for the work Avere anti-coli horse serum H„(i Bliss) and tAvo anti-meningococcus sera H7 Bbiso and H7 BL^o- the first having a considerably higher IN titer than the second. Inasmuch as 0.1 c.c. and 0.5 c.c. of anti-coli horse serum H,;,, Bliss contained abotit 15 and 45 CN units, respectively, it fol- lowed that, shoidd there be summation of titers, then a mixtme consisting of 0.9 c.c. of anti-meningococcus serum and 0.5 c.c. of serum must have a higher titer than a mixtme consisting of the same amount of the anti-meningococcus serum and only 0.1 c.c. of serum H,,,, Bbss- However, t^oth mixtines remained of the same titer. Similarly, in case of simimation of titers, any mixture containing 0.6 c.c. of the anti-coli horse serum should have at least 50 neutralizing luiits irresjjective c^f the titer of the anti-menin- gococcus seriun. But. a mixture consisting of 0.4 c.c. of anti-menin- gococcus serum H7 Blisuld be dra\vn : 1. The high CN titer of mixtures of anti-meningococcus and anti-coli sera is not due to mere summation of titers of both sera. 2. The addition of anti-coli serum to anti-meningococcus serum of a \o^v neiUralizing potency does not bring about any increase in potency of the mixture. Moreover, the independent neutralizing potency of the anti-coli serum seems to become abol- ished altogether Avhen mixed Avith an anti-meningococcus serum poor in neutralizing antibodies. 3. The addition of anti-coli seriun to an anti-meningococcus serum of a high IN titer results in a high CN titer of the mixture. These conclusions could be further supported by experiments described below. In these experiments there was tested the neutralizing potency 122 LOCAL TISSUE REACTIVITY of xarioiis ainoimls ol anti-mcnin^occxx us Sciuin II7 liLxi- and ot mixtures ot this serum ^vith diluted and undiluted anti-coli liorse serimi H^o Bliss. Various amounts ot anti-meningococcus serinn alone gave irregtdar neutralizations except when tested with small niunbers of toxic imits. The addition of 0.1 c.c. of the anti-coli serum previously diluted 1:^5 had no effect on neu- tralizing potency. Mixttnes of amoimts as small as o.'^ c.c. and 0.6 c.c. of anti-meningococcus serum \vith 0.1 c.c. of undiluted anti- coli serum did not result in any noticeable increase of the CN titer. In spite of the fact that the latter amotnit of the anti-coli serum coidd neutralize completely about 15 toxic tmits and 0.3 c.c. and o.fi c.c. of anti-meningococctis serum coidd neutralize com- pletely 10 and 15 units, respectively, the mixture of 0.3 c.c. of anti-meningococcus serimi ^vith 0.1 c.c. of undiluted anti-coli serum neutralized completely less than 8 units, and the mixtine of 0.6 c.c. of anti-meningococcus serum Avith the above amount of anti-coli serinn neutralized completely only 15 imits, i.e., the same nimiber of luiits xvhich could be neutralized independently by either ingredient of the mixtine. Howe\er, \vhen 0.1 c.c. of lui- diluted anti-coli serum was added to 0.9 c.c. of anti-meningococcus serum, the mixtine ga\e complete neutralization cjf 35 toxic units, although this amount of anti-meningococcus serum gave inde- pendently only irregular neutralization of 20 units. From the foregoing observations it is seen again that the anti- coli serum does not exhibit its independent neutralizing potency '^vhen mixed ^v^ith anti-meningococcus serum containing neutraliz- ing antibodies of low concentration (0.3 c.c. of H7 Blooo) ; that it produces no effect when added to a some^vhat larger amount of anti-meningococcus antibodies {i.e., 0.6 c.c. of H7 BI202) ; and that it converts (beyond its o^vn neutralizing potency) irregular neu- tralization into complete neutralization ^vhen added to a suflfi- cient concentration of anti-meningococcus antibodies capable of irregular neutralization. For these reasons it may be assumed that the anti-coli serum contributes some auxiliary factor to the spe- cific antibody-)-antigen combination and enters with it into an intricate relationship. Further studies ^vere made in order to define closer the nature of the auxiliary factor in the anti-coli horse sera. The anti-coli horse sera containing the auxiliary factor were totally devoid of complement. There Avas no apparent relation to the anti-complementary property of the sera because the same IMMUNOLOGICAL BEHAVIOR 12^ amount oi anti-complementary factors was foinid in sera devoid of the atixiliary factor as well as in sera containing large amounts of it. The auxiliary factor described seemed to be in the nattne of an antibody, the production of which is stimtdated by immuniza- tion ^vith B. colt heated cultures and cultine filtrates. The antibody appears shortly after the beginning of immiuiiza- tion. The length of time necessary, ho^vever, for its development differs in horses tnider the same treatment. Thtis, after ten ^veeks of immiuiization the antibody ^vas absent in one horse but pres- ent in another. The majority of horses developed the antibody in eight to ten Aveeks. The antibody also disappeared in t^vo horses during the process of immiuiization. After the disappearance of antibody, the horses rested intermittently for nine Aveeks. Three weeks of immiuiization folio-wing the last period of three ^veeks rest again stimulated the production of the antibody. The auxiliary antibody ^vas absent from a series of normal horse sera and \arious other heterologous sera (scarlet fever anti- toxin, anti-typhoid horse serum, tetanus antitoxin. Type I pneu- mococcus antibody, human convalescent pneumonia sera, and Type II pneumococcus antibody). (Shwartzman, 1931/-'.) REACTIVATING EFFECT OF BLOOD SERA UPON COMPLETELY NEUTRALIZED TOXIC FILTRATES Following observations on the heterologous antibody auxiliary to neutralization of meningococcus reacting factors, it seemed of interest to determine \vhether blood sera also contained factors having the opposite effect; that is to say, capable of reactivating completely neutralized toxic filtrates. In these experiments rabbits Avere injected intradermally Avith 0.5 c.c. of undiluted meningococcus Group III "agar washings" filtrate. The sera tested for the reactivating property were injected intravenously t^venty-three and one-half hours later. This Avas followed by the intravenous injection of the completely neutral- ized meningococcus preparation one-half hour later. The menin- gococcus preparation ^vas a meningococcus Group III "agar ^vasli- ings" filtrate. The completely neutralized mixtures consisted of one part of the filtrate containing in each 1 c.c. the necessary number of toxic units, 0.9 part of anti-meningococcus horse serum, and 0.1 part of auxiliary antibody. The mixture ^vas in- 124 LOCAL ILSSIL RKAC'IIVLIY (iil).iU'(l ;il ;;7 c. loi one hour and injcHlcd inl ra\ cnonsly in a (lose of 'J. ( .('. j)ci kilo ol body \\cii;iu into picj^aicd rabbits. Rabbits iiijcdcd intrax fiioiisly with normal iabl)it seriini showed leactions with (oniplctcly neuliali/cd meningococcus re- acting lactois. The etieet was cjiiite striking- since as little as lo iniits ol meningococcus filtrate mixed with 0.9 c.c. oi anti-menin- gococcus serum and 0.1 c.c. ol the auxiliary factor gave reactions in rabbits pre\iously treated Avith normal rabbit serum whilst as many as 100 iniits ot meningococcus filtrate mixed with the same sera gave no reactions in rabbits receiving no preliminary treat- ment with raljfjit serimi. Rabbits injected w'ith 0.5 per cent NaCi solution showed no reactions with the neutrahzed mixtures. Moreover, the rabbit serum by itself injected twice within the same intervai of time between injections displayed no reacting potency. Inasmuch as antigen -f-antibody interactions of Ireterologous blood sera in vivo may luring af)out formation of reacting factors (Chapter ix) , rab- l^its were eacli separately injected with rabbit and normal human sera, rabbit and normal horse sera and with rabbit and immune horse sera, respectively. No reactions ^vere elicited by these sera in prepared sites. The treatment of rabbits with the rabbit serum did not raise the susceptibility to the toxic filtrates alone. From the foregoing experiments it was concltided that the pre- liminary intravenotis injection of the rabbit serum induced reac- tivation of completely neutralized meningococcus reacting factors. The reactivating property of the serum was titrated in a twofold manner: 1. Prepared rabbits were injected intravenously with the same amoimt of the rabbit serum {i.e., 1 c.c. of undiluted rabbit serum per kilo of body ^veight) , and divided into a ninnber of groups of 3 each. One-half hotn- later rabbits of each group received a single intravenous injection of a mixture of a varying ninnber of toxic iniits mixed with a constant amount of immune serum. In these experiments untreated rabbits showed no reactions ^vith 120 neutralized reacting units, w^iilst the treated ones gave reac- tions with 25 and no reactions with 10 neutralized reacting units. Thus, the smallest amount of neutralized reacting factors showing no reactions in treated rabbits Avas apj^roximately 17 units. There was obtained, therefore, reactivation of six-sevenths of the neu- tralized reactinii factors. IMMUNOLOGICAL BEHAVIOR 1 25 2. Rabljits were also treated iiitia\enoiisly with various amounts of the rabbit serum and tested \vith a constant amount of neutralized reacting factors. In this manner, a dose as small as 1 c.c. of diliuion 1:50 of the serum ^vas shown to possess the reac- tivating property. In one experiment the reac ti\ating effect of the serum persisted for t^ventv-three and one-half hours. Longer periods of time ^vere not studied. In further experiments it ^vas also determined whether the rabbit serum added to the neutralized factors /// lutro would dis- play the above described reactivating property. One hundred meningococcus reacting units completely neu- tralized as above Avere mixed with rabbit Serum I luidiluted and diluted 1:10 and 1:50, respectively. Each mixture ^vas tested in a group of three rabbits. The ralolDit serum undiluted and diluted 1:10 reactivated the neiUralized mixture, inasmuch as all the rabbits tested sho\ved reactions. The dilution 1:50 had no effect, since no reactions were obtained in the group. By the use of the above methods there Avas also determined the frequency Avith Avhich the reactivating property occurs in sera of various animal species. The results are summarized in Table vn. Table vii TESTING OF VARIOUS SERA FOR REACTIVATING PROPERTY Tests in vivo Tests in vitro Animal species Serum dilution Total No. of sera tested No. of positive sera No. of negative sera Serum dilution Total No. of sera tested .Yo. of positive sera No. of negative sera Man Undiluted 2 I I I : 3 45 24 21 Horse " I I I : 3 2 I I Rabbit u 30 21 9 I : 10 9 7 2 Guinea pig ... " 4 3 I I : 5 I I As is seen from Table vn, the reactivating property described was found in sera of man, horse, rabbit and guinea pig. About 72 per cent of rabbit and 53 per cent of human sera shoAved this 126 LOCAL ILSSUE REACTIVITY j)i()j)c'rly. riic iiimil)cT ol liorsc and guinea piti, sera tested Avas too small to allo\v any statistical conclusions. The ^vork reported liius far has dealt with the reactivation ol neutralized reactino- factors of the meningococcus. In further experiments it was foiuid, liowever, that the reactivating property was non-specific inasmtich as the same sera ^vere capable of re- acti\ating neutralized reacting factors of B. culi as well. It has been noticed in the cotnse of tlie work that the reac- tivating property may disappear on storage. The observations sug- gested, therefore, studies on the relation of the reactivating prop- erty to the complement and also on their lieat resistance. It was found, liowever, that the reactivating factors had no apparent relationship to the complement content. Some guinea pig sera containing active complement were devoid of reactivating factors ^v^hilst rabbit sera containing no complement possessed the reac- tivating factors to a marked degree. The reactivating factors, how- ever, were found heat-labile, the exposure to 56° c. for five min- utes, 50° c. for t^venty-five miniues and ^',7° c. for four hours being sufficient to destroy them. Taking advantage of the above observations it ^vas planned to determine ^vhether immime sera contained native reactixating ]:)roperty, and if so, xvhether heating to a temperatiue destructive to the reactivating ])ro]:)erty but ineffective against neiUralizing antibodies woiUd raise the neutralizing potency of these sera. Various immime sera, unheated and heated in the xvater bath at 37° c. for four hotns (temperature sufficient for destruction of reactivating factors) Avere titrated for ireutralizing antibodies in the usual nranner. The anti-meningococcus sera xvere tested \vith and without the auxiliary antibody. The anti-typhoid horse sera Avere tested xvithout auxiliary antibody. The treatment enabled two batches of anti-meningococcus horse sera to neutralize com])letely 50 and 80 tuiits ol menin- gococcus reacting factors, respectively. In contrast, only irregular neutralization xvas obtained xvith the same tuiheated sera xvithout the aid of the auxiliary antibody, and only 20 iniits were com- j^Ietely neiUralized exen xvith the aid of the auxiliary antibody. The heated anti-typhoid horse serimi shoxved a Go per cent in- crease in the titer (tuiheated 375 units in 0.25 c.c, heated ()oo units in 0.25 c.c.) . Incidentally, in the course of some xvork on antibody absorp- tions, Klein tested in this laboratory for control pin poses the IMMUNOLOGICAL BEHAVIOR 1 27 neutralizing potency of an immune anti-typhoid serimi diluted 1:10 after it was incubated for several hours in a water bath at 37° c. He obtained a striking increase in the netitralizing titer. Thus, it is obvious that the native reactivating factors which may be present in the immune sera can be destroyed at a tem- perature {i.e., 37° c. for four hours) which has no destructive effect upon the neutralizing antibodies. The destruction of the factors is, then, coincidental with the considerable increase of the neutralizing potency of an immune serum. The therapeutic effect of sera treated in this manner remains to be determined. The immunological significance of the reactivating factors Avar- rants consideration, inasmuch as they appear to be non-specific, widely distributed among various species of animals (i.e., man, horse, rabbit and gtiinea pig) and are capable of displaying the effect in vivo. It is conceivable that they may play an an- tagonistic role if present in therapeutic immune sera, i.e., inter- fering both with the effect of passively acquired antibodies and also, possibly, restoring toxicity to toxic substances neutralized by actively acquired antibodies. The possible reactivating effect of normal human blood sera {i.e., about 53 per cent of sera tested) should also be taken into consideration in performing blood transfusions in cases of in- fectious diseases. Follo^ving the observations on the reactivating factors just described, Michelazzi (1933/-') studied the effect of normal sera upon completely neutralized diphtheria toxin -^antitoxin mix- tures. In these experiments he observed that not infrequently the iyi vitro addition of normal serum to the completely neutralized mixtures may restore the specific toxic effects of diphtheria toxin. Local and general pathological effects specific for diphtheria toxin may be obtained following the injection of completely neutralized diphtheria toxin-(-antitoxin mixtures if the guinea pigs are pre- viotisly treated by injections of normal horse serum. The effect is, ho^vexer, more pronounced if a serum containing anti-horse anti- bodies is used for the reactivation. RECAPITULATION Unquestionable proof is offered that skin-preparatory and re- acting factors derived from a large group of bacteria are antigenic substances Avhich may be specifically neutralized l)v iiimume sera. 128 l,()(w\I, I ISSUE REACTIVITY The ncii(iali/ini> jjotciuy ol sera may be (oiueniciitly titrated against iiieieasiii^ ainoiiiits of reacting factors. Tlie acti\e principles of B. typhosus, B. colt, and as will be shown in (Ihajiter xin, those ol meningococcus, possess "strain" and "variant" specificity. The passage through mice induces in certain strains a partial loss of neutralizability of reacting factors. This is probaf)ly due to \'ariations to^vards smoothness ^vith an increase in the antigenic cc:)mplexity. Reacting factors deri\ed from rough \ariants accpiire a ne^v specificity. Changes in the neutralizability of the reacting factors cannot be consistently correlated \vith colony morphology, serum agglutinability and virulence of strains, nor with the serum precipitabiHty of the culture filtrates. There is an indication that fluctuations in potency of filtrates are accompanied })y formation of "toxoids" retaining their power to comljine Avith the neutralizing antibodies. Meningococcus ac- tive principles can be also detoxified by formalin whereby the antigenicity and the antif^ody-combining capacity remain essen- tially unimpaired. It was shown that the neutralization of the active principles of the phenomenon by immune sera is not due merely to the adsorption of active principles but the precipitate formed. The addition of homologous neutralizing antibodies to B. typhosus active filtrates results in processes essentially consisting of two phases: (i) Flocculation of the antigenic components of the bacterial filtrates by immune sera accompanied by adsorption of the active principles by these complexes. (2) Neutralization of the active principles in these precipitates which may or may not follow this adsorption depending whether or not the antibodies pcissess neutralizing po\ver. The neutralization of the active princij^les of B. typhosus and B. coli occurs according to the Lnv of "multiple proportions," the reaction closely resembling the neutralization of diphtheria toxin by antitoxin. There is demonstrated the existence of an heterologous anti- body auxiliary to the neutralization of meningococcus and B. coli reacting factors. The presence of this antibody facilitates studies on neutralizing potency of anti-meningococcus sera. There also appear to exist certain reacti\ating factors in nor- mal and immune blood sera of various animal species. The effect of reacti\ating factors consists in restoration of the toxicity /// vixio IMMUNOLOGICAL BEHAVIOR 129 and /// vilro of completely neutralized nieningococciis and B. cvli reacting factors. The factors are non-specific, heat-labile and seem to bear no relationship to the complement. Heating of immune sera to a temperature destructive for the reactivating property l)ut innocuous to the neutralizing antibodies raises considerably tlieir neutralizing potency. Chapter V IMMUNITY TO THE PHENOMENON OF LOCAL SKIN REACTIVITY THE facts eml:)odied in the preceding chapter demonstrated that the active principles ot the phenomenon are antigenic bacterial factors specifically neutralized by immune anti- sera. Data concerning natural and actixely and passively acquired immunity to the active principles of the phenomenon are de- scribed in this chapter. NATURAL IMMUNITY OF RABBITS In experiments previously described it has already been noted that some of the animals remained resistant to the phenomenon of local skin reactivity to B. typhosus cultiue filtrates. Thus, out of 2 12 animals tested with the same doses of B. fyljliosus tryptic digest broth culture filtrates there were 45 negatively reacting rabbits and 167 rabbits which showed severe reactions. In the animals which Avere considered resistant to the phenomenon, there Avas no local hemorrhagic necrosis folloAving the intrave- nous injection. When, prior to the intravenous injection there was erythema at the prepared site, sometimes it became more pronounced four to five hours after the intravenous injection. TAventy-four hours later the skin appeared normal again. In this group of experiments no mild hemorrhagic reactions which Avould constitute an intermediate group bet^veen the negative and positive animals Avere obtained. In further experiments, the use of B. typhosus "agar Avashings" filtrates yielding potent active principles seemed to reduce the number of resistant animals, although a certain percentage of rabbits ahvays remained resistant to B. typJiosus materials. Frequent occurrence of spontaneous immunity to B. typJiosus in rabbits may be presumed to exist from findings of normal agglutinins in a large percentage of animals. Not infrequently B. typhosus agglutinins in a titer as high as 1:1000 Avere encountered in normal bleedings. Conversely. 130 IMMUNITY TO LOCAL SKIN REACTIVITY I3I meningococcus "agar washings" filtrates, ^vhen employed in suf- ficiently large doses, may be able to elicit the jjlienomenon in practically all rabbits tested. Thtis, in one group of experiments 74 rabbits stnvi\'ed an intravenous injection of a large dose of meningococcus "agar ^vashings" filtrates. Se^■enty-three of these animals showed severe hemorrhagic necrosis. When 6 to 10 menin- gococcus reacting units ^vere used, the percentage of positive re- sults was found to be between 90 to 100 per cent (Sh"^vart7man, i928t>, 1929^/, //) . Kielanowski (1935) and Kielano^vski and Selzer (1934/;) cor- roborated the obser\ations of \arious investigators on sponta- neous \ariations in susceptibility of rabbits to the phenomenon. The spontaneous imimniity. ho^vever, appeared to them not as an absolute btit only a relative refractoriness. Whilst they ^vere never able to obtain any rabbits ^vhich did not produce the phenomenon with large doses of the filtrate, there were some rabbits which in\ariably sho^ved a higher degree of stisceptibility (white and blue rabbits) . These authors therefore suggest that the expression "spontaneous immunity to the phenomenon" be replaced by the term "non-reactivity." Stolyh^vo (1935, 1936) also pointed out that the resistance to the phenomenon depends on the dose employed. Although Avith given moderate amotnrts of filtrates the rabbits coidd be grouped as partially susceptible and resistant, the employment of large doses may redtice the ntunber of negative rabbits to zero. Gratia and Linz (1932c) stated that about 60 to 70 per cent of rabbits were susceptible to the phenomenon, ^vhilst 10 to 20 per cent ^vere hypersensitive and died before the reaction took place and approximately 20 to 30 per cent ^vere refractory to the phenomenon. According to the same authors, no distinction coidd be made as to the resistance to the phenomenon between males and females. Bin net (1931) recorded comj^arative observations on domestic and Avild Australian rabbits. Twice ^vashed, killed bacteria were efTecti\e in eliciting the skin-preparation as ^vell as the reaction. Domestic rabbits reacted readily to stich preparations but ^vild Australian rabbits gave iew positive reactions when ^vashed bac- teria Avere used as the provocative factors. As pointed otit in Chap- ter ni, Avashed bacteria contain only a small amount of active principles. On the basis of Burnet's observation, the interpreta- tion may be made, therefore, that Avild Australian rabbits pos- 132 LOCAL ILSSIIE REACTIVITY scss a lower susceptibility to the plK-nomcnon than domestic rabbits. Apit/ (k)-}-^^') noted a lower dei>iee ot reacti\ity in the winter than during the smnniei months. Gratia and Linz (19326) lomid that yoimg ral)l)its one to two months old reacted more se\erely than old ones. In contrast to this, howe\cr, the new-born aj^jjeared refractory to the phenome- non. Thus, one rabbit ga\e birth to seven young ones; at the day ot birth the skin of one of the rabbits was prepared with 0.2 c.c. of an active B. coli filtrate. Twenty-four hours later 0.25 c.c. of the fdtrate was injected intraperitoneally. No reactions w^ere ob- tained. Another new-born animal ^vas treated the same Avay on the second day of its life Avith a negative result. Tw^o more rabbits were injected on the seventh day, two more on the fifteenth day, all of the rabl)its showing negative residts. The last new-l3orn rabbit ^vas tested on the twenty-first day. This time the reaction was strongly positive. Llyeda (1934) also claimed that yoimg rabbits are less susceptible to the jjhenomenon of local skin reac- tivity to vibrio cliolerae than adidt rabbits. Witebsky and Neter (1936) studied extensively the question of age susceptibility. Highly potent B. iypJiosus and meningococcus filtrates were used by these authors. One-tenth to 0.2 c.c. of the respective filtrates were injected. Out of 16 litters of 78 yoimg ral)bits, one and a half to five weeks old, 12 litters of 58 animals did not display any reaction. The remaining 20 yoinig ral)bits belonged to 4 litteis only. Two of these litters (three to five weeks old, resj^ectively) of 9 animals were strongly positive. The other 2 litters (two weeks and fom to five weeks old, respectively) consisted of 1 1 animals; 3 of them were negative, one sho^ved a doubtfid reaction and 7 Avere posi- tive. Thus, a high percentage of yoimg rabbits ^vere refractory to the phenomenon. The phenomenon-negative litters were again tested for periods of one to two weeks with the same or heterologous filtrates. The same amount of filtrate \vas used as in the first experiment irre- spective of the increased weight of the young rabbits. Many, but not all, of the originally negative animals were positive at the second, sometimes only at the third or fourth repetition of the test. It appeared beyond doubt from these experiments that there existed a definite correlation l^etween the aging of rabbits and their ability to react to the phenomenon. IMMUNITY TO LOCAL SKIN REACTIVITY M^^ SPONTANEOUS AND ACQUIRED ACTIVE IMMUNITY AS DETERMINED BY THE USE OF GRADED AMOUNTS OF ACTIVE PRINCIPLES Attempts to determine the specific ity of spontaneous and acti\e accjuired immunity were made l)y me by the use ol graded amounts ot active principles (Shwartzman, 1932^) . Spontaneous immunity : Protocol 14. Rabbits 8-30 and 8-01: May 26, 1931, these rabbits were injected iiuratlernuilly ^vitli 0.2.-, t.c. of B. coll S "rough" Filtrate 1042 and May 27 intravenously with 200 reacting units of the same filtrate. There were no reactions obtained. May 28, 1931, the rabbits were prepared by single intradermal injections of 0.25 c.c. of meningococcus 44 B Filtrate A-38 and twenty-four hours later injected intravenously Avith 200 reacting units of this filtrate. There appeared severe reactions. Rabbits 8-67, 8-55 and 8-73: These rabbits were tested in a similar manner with three filtrates, namely, B. coli S "rough" Filtrate 1042. meningo- coccus 44 B Filtrate A-38 and B. typhosus "rough" Filtrate 1067 during six consecutive days. The intravenous doses were 200 reacting units of each filtrate, respectively. The intravenous injections elicited no reactions in Rab- bit 8-67. Rabbits 8-73 and 8-55 showed no reactions to B. coli and meningo- coccus, but de\eloped severe reactions following the intravenous injection of B. typJwsus "rotigh" Filtrate 1067. These restilts demonstrate spontaneotis active immunity as con- cerns the phenomenon. In cme instance, the immunity was of a non-specific nature. In another instance, it appeared to be specific, inasmtich as the rabbits tested at short intervals of time ^vith various filtrates, proved stisceptible to one filtrate and resistant to another one. The idea of possible existence of nattiral specific immunity to the phenomenon was also brought out by ftirther experiments in ^vhich rabbits prepared by siinultaneous injections of B. typhosus, B. coli and meningococctis filtrates and injected intravenously, shoAved in some instances reactions in all sites; and in others, only in one or two sites. The latter rabbits showing reactions only in some sites, apparently had specific cellular im- munity of sufficient potency to prevent the occtirrence of the state of reactivity to active principles of some bacteria but had humoral immunity of insufficient potency to neutralize these active prin- ciples in the blood stream. The non-neutralized reacting factors were then capable of producing injury in areas prepared with filtrates asainst whicii these rabfjits had no cellular immunity. 134 LOCAL ILSSLE REACrrLVLFY Active Acquired Immunity : In view of the a()()\e lindiiiti^s, it was of finther interest to determine whetlier iintniniity could be also attively acquired. Tn some experiments rabbits were imnumized by one or two intradermal injections of bacterial filtrates. The intradermal vaccination with B. lyjjhosits cidtiue filtrates induced active im- miniity to the phenomenon of local skin reactivity to B. typhosus and also conferred resistance to the lethal effect of the toxic sub- stances. The existence of active acquired immimity having l)een demon- strated, it Avas necessary to determine the specificity oi this im- numity with graded amotuits of bacterial filtrates. In these experi- ments rabbits were tested with one unit of B. typJiosus reacting factors. According to their response to this injection, they could be grouped as I — sensitive animals, i.e., giving reactions with one reacting luiit; and II — refractory animals, i.e., sho^ving no reac- tion ^vith one reacting unit. The latter group could be finther sid)divided into Ila — those ^vhich had only partial immunity, inasmuch as retest ^vith 15 reacting luiits one week later produced reactions in them; and lib — those which did not sho^v any reac- tions when later retested with larger doses of reacting units. The tests were then followed by repeated attempts to reproduce the phenomenon in the tested animals with increasing amoiuits of various bacterial filtrates with the folloAving results: Group I. Rabbits highly sensitive to the phenomenon, i.e., those reacting to one intravenous iniit, acquired an active specific immunity to it by means of combined intradermal and intra- venous immimization with toxic bacterial filtrates. As might be expected the animals of this group showed indi\ idual variations in their response to the process of immunization. Some of them failed to acquire a state of immimity, ^vhile others showed various grades of it. Thus, one rabbit was sensitive to one B. typliosus reacting unit. Two weeks after the first injection there was obtained a doubtful immunity to 30 units and three and four weeks later the rabbit sho\ved complete immunity to as many as 60 and 100 iniits. The immimity established ^vas specific, inasmuch as retests ^vith meningococcus filtrate elicited severe reactions. Another rab- bit at first sensiti\'e to one reacting luiit of B. typliosus, promptly rf'. IMMUNITY TO LOCAL SKIN REACTIVITY 135 became imnume to 150 B. typlwsus reacting factors. Here again, the immunity was specific, since the rabbit remained sensitive to meningococcus reacting factors. In addition, it was possible to inmiunize it also to these factors within t^vo weeks. The rabbit, thus imminiized to reacting factors of t^vo microorganisms, re- mained, however, sensitive to a third microorganism; i.e., B. coli. A third rabbit illustrated an active acquired immunity to graded amoinits of B. typhosus reacting factors. The immunity was in- complete since retests with 300 and 350 tniits elicited strong reac- tions. On the other hand, a foiuth rabbit demonstrated total failtue to acquire active immunity, and a fifth rabbit showed a partial acqtiired immunity dining the course of immunization to 30 reacting units, with subsequent susceptibility to larger doses on a retest in a Aveek. It became evident, therefore, that in order to demonstrate the existence of this immunity, a sufficiently large group of animals and graded amounts of toxic material shoidd be employed. Group Ila. The rabbits of this group showed a spontaneous partial immunity inasmuch as they presented reactions with one reacting unit, but proved sensitive to 15 iniits of the same filtrate, ^vhen retested one ^\eek later. A few weeks later, the sensitivity Avas followed by a prompt acquired imminiity ^vhich ^vas of a non- specific natine. The animals gave no reactions when reinjected Avith larger amoiuits of the same filtrates as Avell as Avith large doses of other filtrates {i.e., B. coli, B. typhosus, B. typhosus rough and meningococcus of other serological groiq^s) . Group lib. These rabbits possessed a non-specific immunity as concerns the phenomenon inider consideration, inasmuch as in contrast to Group Ila none of the injections of large doses of heterologous filtrates elicited reactions at any time. On the basis of the above observations it then becomes neces- sary to differentiate between active specific and non-specific ac- quired immtniity to the phenomenon of local skin reactivity to bacterial filtrates. The active specific immunity can be induced by immunization of highly susceptible rabbits with the toxic material. The non-specific immunity is elicited in rabbits pri- marily showing partial spontaneous immimity. The question of active acquired immunity to the phenomenon under discussion has also been studied by other investigators. According to Burnet (1931), repeated reactions coidd be elicited in rabbits without diminishing their reactivity. In one l'^() LOCAL TLSSUE REACTIVl'IY or two ol his cxjieriinciits there seciii to he a dehiiitely increased reacti\ily when animals were retested alter a week or two. A transitory "desensiti/ation" was induced l)y him by an intrave- nous injection twenty-lom horns or less l)elore the test intrader- mal injections were made. Work on the j)r()l)Iem of desensitiza- tion will he discussed in Clhapter x. It should he jiointed out here that Bmiiet immunized rahhits for short intervals of time and employed non-measmed cjuantities of materials for testing of their imminiity. Moreover, no differentiation was made as to the sensitivity of the rahhits to the phenomenon prior to the experi- ments. As Avas shown in my experiments just described, luider these conditions the state of active acquired immunity coidd have been easily overlooked if, for example, rahhits highly sensitive to the phenomenon ^vere immunized for a short period of time and retested by large doses. On the other hand, prolonged im- munization ^vith graded amounts of materials may induce a stead- fast immunity against 200 and 300 and sometimes 1000 midtiples of the initial dose. Stolyhwo (1935 and 1936) noticed that forty- two days after the elicitation of the phenomencjn the rabbits de- veloped an immunity against moderate doses of the filtrate. In some instances, this immunity lasted as long as ninety-two days. The employment of very large doses could, ho^vever, break the immunity and elicit positive reactions. He also studied the spec- ificity of the active acquired immimity. For this purpose he se- lected rahhits in which the phenomenon had been previously elic- ited with B. typhosus and B. paratyplwsus cidtine filtrates and attempted to reproduce the phenomenon again Avith heterologous and homologous filtrates. These experiments suggested to this author that the elicitation of the first phenomenon induces a resistance of a relative specificity. Rabbits in Avhich positive reactions were obtained by Plant (1932) with Spiyoclieta pallida active principles, were retested after various intervals of time. No reactions could l)e elicited in these animals five days after elicitation of the phenomenon, only occasicjnal reactions occurred t^velve days later and five Aveeks later-, they became positive again. Kielano"\vski (1935) and Kielanowski and Selzer (1934't') note the loss of sirsceptiljility of rabbits and sometimes apjjearance of hypersensitiveness a few wrecks follo^ving the elicitation of the phenomenon. Michelazzi (i935«) reported that a single intracutaneous in- jection of B. coll filtrate induced a state of local imnuniity to IMMUNITY TO LOCAL SKIN REACTIVITY 1 37 the phenomenon. Reactions could ])e, howe\er, elicited in ad- joining skin sites. Parallel to my obseixations on acqiurement of non-specific ac- ti\ e immunity l)y certain rabbits follo^ving■ the repeated reproduc- tion of the phenomenon, Peck and Sobotka (1931) studied the production of the refractory state of the phenomenon follo^ving injections of \enom of the moccasin snake. In their preliminary Avork they assmned that animals de\eloping intense pinpina fob lo^ving the intracutaneous injection of snake \enom in dilutions of 1:1000 and 1:2000 are more likely to sho\v the phenomenon of local skin reactivity than those giving a ^veak reaction with snake \enom. After a period of fourteen days to one month elajDsed bet^veen the time of grading \vith \enom and the produc- tion of the phenomenon of local skin reacti\ ity, a large number of rabbits became refractory to the phenomenon. An incubation period of not less than fourteen days ^vas necessary for the de- velopment of this resistance. The incidence of refractory animals appeared to be within limits inversely proportional to the amount of active principles injected intravenously for the elicitation of the phenomenon. The intra\enoiis injection of moccasin \enom proved most efficacious in developing refractoriness, although the intradermai and intraperitoneal injections ^vere also successful. The refractory state Avas also stili present forty-four days after the primary injection of the moccasin \enom. The refractory ani- mals did not show any changed reaction to the local effect of moccasin venom in the concentration used. No circulating anti- bodies could be demonstrated to explain the refractory state. Passive immunization ^vith anti-venom had no effect on the course of the phenomenon of local skin reactivity. The authors con- cluded that the phenomenon of local skin reactivity probably de- pends on induced vulnerability of the capillaries (at the site of primary injection) to subsequent injection of bacterial filtrates. They assumed, therefore, that the injection of small quantities of vascular poison produces a change in the capillaries ^\•hich raises the threshold for the elicitation of the phenomenon of local skin reactivity. The refractory state induced by venom is apparently of a cellular nature, inasmuch as passi\e resistance could not be elicited by the use of serum from rabbits made refractory by means of the phenomenon or by injections of anti-\enom. In further experiments Peck (1933, 1934) found that the re- fractoriness to the phenomenon could not be produced by im- 1^]S LOCAL riSSLE REACTIVITY imuii/at ion with l)()llir()j).s and ralllcsnakc xenoms. In contrast, copj)eiheacl and moccasin venoms proved efficacious. Tlie mini- mnm effective dose of moccasin venom (of the sample used) neces- sary to produce tlie refractc^ry state was 0.04 mgms. intradermal ly. Solutions of moccasin venom made from different samples of pooled venom differed markedly, in their ability to produce a refractory state to the phenomenon of local skin reactivity. The ability of venom to induce the refractory state was titrated by this author. Although various moccasin venoms showed very little difference in their neinotoxic titer, there was a marked variation in their content of refractivity-producing factors. There was, then, no parallelism betAveen the content of hemorrhagins or neuro- toxins and these factors. His experiments also showed that com- plete inhibition of refractivity-inducing factors could be accom- panied by the production of anti-venom and that complete or partial inhii)ition could be brought about by various lots of nor- mal horse serum. It seemed to the author that the above experi- ments pointed very strongly against the assimiption that the re- fractory state induced by previous injections of venom, is based on a foreign protein reaction, in a broad sense of the term. II the refractoriness to the phenomenon were based on such a foreign protein reaction alone, there xvould not be any variation in the potency of the refractivity-inducing factors in the \arious venoms, and the addition of anti-venom and horse seriun xvould tend to increase such a foreign protein effect. Since the ability of the venoms to induce a refractory state to the phenomenon apparently signifies their efficacy in the production of some re- sistance of the capillary wall, i.e., decrease in its permeability and increase in the resistance to toxic influences, Peck decided to deter- mine clinically xvhether the venoms xvoidd be of value in treatment of conditions of hemorrhagic diathesis. His clinical results xvill be discussed in Chapter xni. PHENOMENON OF LOCAL SKIN REACTIVITY TO BACTERIAL FILTRATES IN ANIMALS OTHER THAN THE RABBIT Gratia and Linz (i932r) Avere the first authors to report on the phenomenon of local skin reactivity in guinea pigs. The abdomi- nal skin xvas prepared by an injection of 0.5 c.c. of a six day old Vibrio cholerae culture filtrate. Txventy-fotir hours later the provocative injection was given either intracardially or into the jugular vein. Some of the guinea pigs thus treated showed hemor- IMMUNITY TO LOCAL SKIN REACTIVITY I39 rhagic reactions 3 to 5 cm. in diameter in the prepared sites. This obserxation on the phenomenon in guinea pigs to Vibrio cholerae was corroborated by Llyeda (1934). DeCour (1934) noted that guinea pigs were considerably less sensitive to the phenomenon than rabbits. Stolyhwo (1935, 1936) ^^^s able to produce the phenomenon in guinea pigs by means of combined intradermal and intracardial injections of broth or "agar washings" culture hltrates of B. typhosus and B. paratyphosus. It was necessary, ho^vever, to employ considerably larger amoimts of the material for the intravenous injections than in rabbits. The doses were usually 1 to 2 c.c. of broth cidtiue filtrate or 0.25 c.c. of "agar xvashings" cidtine filtrates for guinea pigs weighing between 280 and 400 gms. The general lethal effect of the filtrates Avas con- siderably lower for guinea pigs than for rabbits. The hemorrhagic and necrotic reactions themselves x\ere less pronoiuiced than in rabbits and rarely measured over 1 x 1.5 cm. Rubin in my laboratories tested a large group of guinea pigs \\'\\\\ B. txplwsus and meningococcus "agar washings" filtrates. For a period of several months negative residts were obtained even \vith as many as 500 reacting luiits. This xvas followed by a period of sexeral -weeks dining which typical reactions were obtained xvith small doses in a high percentage of guinea pigs. Although a seasonal influence suggested itself, no correlation \vith age and sex could be clearly demonstrated. H. Gross (i93i«, b, c) , Freinid (i934fl) , Vassiliadis (1935) , and Apitz (1933^) failed to elicit the phenomenon in guinea pigs, whilst Kielanowski and Selzer (i934/:>) considered them of low susceptibility. \'arious authors consistently failed to produce the phenomenon of local skin reactivity in mice and rats (Gratia and Linz, 1932c; Shwartzman and Michailovsky, 1932; Stolyhwo, 1935, 1936; Apitz, 19336; Kielanowski, 1935; Kielanowski and Selzer, 1934/^'; and others) . Stolyhwo injected one or txvo sites of the skin of the abdominal wall of mice xvith filtrates of B. typhosus and B. para- typhosus. Twenty-four hours later the animals received either in- travenous injections of 0.1 to 0.5 c.c. of the homologous filtrate or 0.5 c.c. to 2 c.c. of the filtrate intraperitoneally. The mice died within twenty to txventy-four hours following the injections and no reactions were obtained. I made preparatory injections of potent filtrates into the skin of the abdominal xvall, back, tail and the scalp of mice. The provocative injections xvere given in- 140 LOCAL TLSSUE REACTIVITY traxcnously alter xarioiis inter\als oi time (one lo lurly-eight hours) . No lvj)i( al reactions were obtained. Sohel in my laboratories attempted without success to repro- duce the phenomenon in "white rats by means of combined intra- dermal and intravenous injections of enorm(nis doses of B. ty- phosus, B. paratypJwsus and meningococcus "agar ^vashings" filtrates of ascertained potency. In 1929, I observed the phenomenon in 2 goats with B. ty- pliosns tryptic digest broth fikrate and in one horse with menin- gococcus "agar Avashings" hltrate. PASSIVE IMMUNITY TO ACTIVE PRINCIPLES The problem of passive immunity to the phenomenon was studied with B. typhosus active principles (Sh^vartzman, 193 id). The skin of rabbits was prepared by a single injection ^vitii 0.25 c.c. of the toxic filtrate. After various intervals of time the pre- pared rabbits received an intravenous injection of serum fol- lowed by an intravenous injection of the toxic filtrate. The toxic filtrate ^vas injected t^venty-four hours after the skin injection of the hltrate. The interval of time between the intravenous injec- tion of the serum and the folloAving intra\enous injection of the ftltratt: varied, but in most experiments it was one-half hour, and in some experiments one and one-half and t^vo and one-half hours. Reactions were obtained in only i().() per cent of rabbits passi\'ely protected by the immune serimi. Ninety per cent of non- protected rabbits showed reactions ^vith a reacting dose even smaller than the one employed in the serum-protected group (dilution 1:10 in the control group and dilution 1:8 in the ex- perimental group) . Thus, it became at once e\'ident that it is possible to elicit passive immunity to B. typJwsus reacting factors. In additional experiments the protective potency of the im- mune serum against B. typJiosus reacting factors was titrated. Anti-typhoid horse serum Avas injected intravenously into pre- pared rabbits one-half hour prior to the intravenous injection of B. typliosiis filtrate. Rabbits protected ^vith various amounts of the serum were tested against various numbers of reacting units. The largest number of reacting units giving no reaction in all rafjbits tested was used to compute the /// vivo potency of a given amount of serum. Thus, if 2 c.c. of a certain seruin injected in- travenously per kilo of body weight, protected all rabbits tested IMMUNITY TO LOCAL SKIN REACTIVITY I4I Ratio between in vitro and in vivo neutralization - ^ No. of deaths OOOOOOOOOOCO" No. positive rabbits IN OC00100C40(NO(NOO-' No. negative rabbits -. COCO-hCOCO"CO"CO"COCO" Total No. rabbits tested CO co c^cococococoeococococo Order of I.V. injec- tions e s -^ CQ Sen fihr 1 Interval of time between I. V. in- jections « V. N^ N« \« N^ N., \., N^, v.. No. of re- acting units injected I. V. per kg. of body weight LO 0000000000000 " (NCO-^!^CO(Nivo neutralizations. It WAS found in this manner that, 1 c.c. of 1:4 dilution of the innnune serum capable of neutral i/im> /// x'iho 125 reacting iniits failed to give neutralization of as fe^v as 15 tuiits in vivo. The lack of neutralization coidd ])e attributed to a loss of antibodies in the general circulation. It was impossible to state whether the loss ^vas jaartial or complete since ral)bits protected with this amount of serum were not tested with a smaller nimiber of reacting units. As the amount of serum injected was increased it became possible to demonstrate /// vivo neutralization, de- cidedly less, ho^vever, than Avotdd have been obtained in direct ill vitro neutralization. Thus, a dose of 2 c.c. of serinn injected per kilo of body w^eight, neutralized 100 reacting units, i.e., it ^vas necessary to use ten times the number of neutralizing tniits as compared to i)i vitro neutralization ^vith the same serimi. The experiments also clearly demonstrated /// vivo neutraliza- tion of reacting factors in midtiple proportions. Thus, 250 units of serimi protected against 25 reacting iniits; 4 x 250 serum units protected against 4 x 25 reacting units; 8 x 250 serimi units pro- tected against 7.2 x 25 reacting imits. However, 12 x 250 serimi units protected against 24 x 25 reacting units and 18 x 250 serum units protected against 40 x 25 reacting units. As is seen from these figures, the protection took place in direct multiple pro- portions until after a certain increase in the amount of serum injected, ^vllen there occurred a two-fold and more than two-fold gain in the protective value of the serum injected. In the inter- pretation of these facts of which the mechanism may be quite complex, there should be borne in mind a possible change in the rate of retention (lower elimination) of antibodies in the general circulation follo\ving the increase in the volume of serum injected. This explanation suggests itself at the present moment as the simplest, but it remains to be proven. The next point of interest ^vas to determine the specificity of serimi protection /// vivo against B. tyjjhosus reacting factors. In these experiments, immune anti-meningococcus horse serum of high neutralizing potency for meningococcus reacting factors and a normal horse serimi, Avere used. In \'iew of the fact that IMMUNITY TO LOCAL SKIN REACTIVITY 143 B. t\j)hosus antibodies are encountered in normal and heterol- ogous imminie horse sera, the sera "were also tested for agglutinins, precipitins and neutralizing antibodies. Anti-meningococcus and normal horse sera in a dose of 4 c.c. per kilo of body weight, failed to protect against 40 reacting iniits, while the same dose of anti-typhoid horse serinn protected against 180 reacting imits. The normal and heterologous sera employed contained anti-ty- phoid antibodies, the concentration of ^vhich, however, in 4 c.c. {i.e., 320 and 240 units, respectively) was insufficient to neu- tralize i)i vivo 40 reacting iniits, since, as described above, it is necessary to have for this volume of serimi eleven times more antibodies for in vivo than for in vitro neutralization. In an- other grotip 7 c.c. of anti-meningococcus horse serum protected against 40 B. typhosus reacting iniits. It is obvious that this \)\o- tection could be easily attributed to the presence of anti-typhoid normal antibodies, since in 7 c.c. there were present 420 neu- tralizing units, or more than ten times the amount required for in vitro neutralization. From these results it can be safely con- cluded that passive immunity to B. typhosus reacting factors is specific and depends on the concentration of homologous, normal and immune neutralizing antibodies. It is also note^vorthy that there may occur a "prozone" effect of large amoinits of serimi. The "prozone" effect in neutralization experiments i)i vitro ^vas also frequently observed. Some studies ^vere also made on the dination of passi\e im- miniity to B. typhosus reacting factors. It appeared that the immimity established itself immediately after the intraxenous injection of immune serum, and that one-half hoin- after the serum injection the protection ^vas more effective. No definite conclusi(jns coidd be draxvn, however, since the "prozone" effect of seriun, just pointed out, coidd serve as a soinxe of error. In- deed, it is possible that dining the half hour interval, a part of the antibody \vas eliminated and thus the "prozone" effect ^vas avoided. Two exjjeriments were carried out in ^vhich the intcrxal oi time allowed between serum and filtrate intravenous injections was one and one-half and two and one-half hours, respectively. Each of the rabbits received 3000 neutralizing units and 100 react- ing units. The protection was complete in both groups. Longer intervals of time were not studied. It was of considerable interest to determine whether the oc- 144 LOCAL TLSSUE REACTIVITY (iiiieiuc ol the skin reaction could be j)re\ente(l by serum in- jection ii it followed the injection of the filtrate. The reactions were inhibited by an intravenous injection of serum following the injection of reacting factors, provided the serum dose was very large and provided the injection of serum ^vas made immediately after the filtrate injection. The skin reaction coidd not be pre- vented it the inmiune serimi was injected twenty minutes or longer after the injection of reacting factors. The results are most likely due to fixation of the reacting factors in the prepared tis- sue shortly after their introduction into the general circulation. As a matter of fact, this can be expected, since even one hour after the intravenous injection of the toxic filtrate one can see a blue discoloration at the prepared skin sites, although it takes fotu" to five hoins for the fidl development of hemorrhagic necrosis. RECAPITULATION Rabbits differ in the degree of susceptibility to the phenome- non of local skin reactivity. Approximately 20 per cent of rab- bits sho^v a high degree of refiactoriness to B. typhosus active principles. With sufficiently large doses of meningococctis the phenomenon may be produced in all the adtilt rabbits tested. A high percentage of new-born rabfjits are refractory to the phe- nomenon. The idea of possible existence of spontaneous specific im- munity to the phenomenon is brought otit by the experiments in Avhich rabbits prepared by simultaneotis injections of B. ty- pJiosus, B. coli, and meningococcus filtrates and injected intra- venously with the active principles sho^v. in some instances, reac- tions in all sites and in others only in one or t^vo sites. Immunization of rabbits with bacterial filtrates induces active immunity to the phenomenon. The specificity of this immunity may be tested with graded amounts of active principles. Two types of response follo^v immunization. 1. Specific immunity \vhich is obtained in rabbits highly sus- ceptible to the phenomenon. 2. Non-specific immunity ^vhich is elicited in rabbits ^vith partial spontaneous immunity. There is suggestive evidence that non-specific immunity may be obtained in a certain percentage of rabl)its through immuniza- tion \sTth certain snake venoms. IMMUNITY TO LOCAL SKIN REACTIVITY 1 45 The phenomenon of local skin reactivity may be produced ir- regularly in guinea pigs, goats, and horses. It fails completely in mice and rats. It is possible to elicit passive innniuiity to B. lyjjhosus active principles. /// vivo serum protection against these j^rinciples fol- lows the law of midtiple proportions. Passive immunity is best obtained when the immiuie serimi is injected intravenously one- half hotu' prior to the intravenous injection of the active prin- ciples. It is possible to prevent the occmrence of the local skin reaction by an intravenous injection of serimi following the in- travenous injection of the active principles provided, however, the serum dose is very large and the serum injection is made immediately after the filtrate injection. The greater the amount of antiserum injected intravenously, the more efficient is the /// vivo neutralization in a ratio distinctly greater than the (juanti- tative increase of serum. C H A P T K R V I ROLE OF INFLAMMATION IN THE PHENOM- ENON OF LOCAL SKIN REACrmTTY- OPIIMUM CONDITIONS FOR REPRO- DUCTION OF THE PHENOMENON INTRODUCTION THE opinion was voiced ])y Opie, Menkin (1931) and Kars- nei and Moritz (1934) that the inflammation following the preparatory injection brings about a local concentration of the injurious bacterial products subsequently injected intrave- nously and thus is responsible for the lesions obtained. HoAvever, an assumption of tliis sort cannot be accepted unless it is shoAvn that an inflammatory reaction per se is sufficient to obtain a local jjreparedness to the subsequent intravenous injection. As a mat- ter of fact, their hypothesis is contradicted by a mass of unques- tionable evidence described in this chapter. It ^vill be seen that there is no correlation Avhatsoever bet^veen the primary reaction to the intradermal injection of a bacterial filtrate and the ability of the prepared site to respond to the provocati\'e injection. The lack of correlation is clearly demonstrated by macroscopic and microscopic observations. A great variety of non-bacterial sid)- stances of various properties and producing various degrees of inflammation and vascular damage fail to prepare the tissue to the provocati\e injection of potent principles. Moreover, the j)henomenon described can be elicited only pro\'ided suitable routes are employed for the injections. From the facts jjresented, the ima\oidable conclusion is to be drawn, and this is borne out by the observations of most of the investigators in the field, that quite apart from the non-specific inflammation (which may also be almost completely absent luider some conditions of experimentation) , the phenomenon is con- tingent ujDon the elicitation of a certain state of reactivity by means of the bacterial active principles described. This state represents a modified response to the injurious effect of toxic 146 ROLE OF INFLAMMATION 1 47 ])iin(ijjlc's in the l)l()()cl stream and apparently (onstitutes a hitherto unknown mechanism of ])rodiiction ol se\ere injury in the animal cell. EFFECTS OF INTRADERMAL INJECTION OF BACTERIAL rRODUCTS IN THEIR RELATION TO THE PHENOMENON The effect of an intradermal injection of a bacterial product is conditioned by a number of factors, i.e., animal species, micro- organism employed, mode of preparation of bacterial products, previous sensitization by bacterial products, spontaneous and in- duced infections of the animals tested, etc. A review of the w'oik on reactions of bacterial hypersensitiveness does not belong in this chapter inasmuch as the phenomenon of local skin reactivity concerns the effect of bacterial filtrates upon the skin of normal rabbits. The reader may be referred to interesting and exhaus- tive studies on problems of bacterial hypersensitiveness in animals by Calmette, Gay, Zinsser, Swaft, Baldwin. Petroff, Julianelle. Ecker, Doerr, Coca, Pagel, and others. Certain endotoxic substances derived from bacteria, ^vith some of wdiich this monograph is concerned, ^vere described as capable of inducing primary skin reactions in normal rabbits: B. pertussis preparations (Teissier, Reilly, Rivalier and Cambessedes. 1929; Witebsky and Salm, 1937, and others) ; meningococcus (Mal- colm and W^hite, 1932) ; pneumococcus (Zinsser and Grinnell, 1927; Julianelle and Reimann, 1926, 1927; Reimann and Julian- elle, 1926; Parker and Pappenheimer, 1928, and others) . These preparations, hoAvever, ^vere lixe or dead bacterial cells prepared by the use of methods resulting in cell destruction. The bacterial preparations employed in the experimental work described in this monograph, are de\'oid of the primary toxic properties of the mate- rials of the foregoing authors. This ^vas emj)hasi/.ed in Chapter i, \vhere characteristic featmes necessary for recognition of the phe- nomenon of local skin reactivity were discussed. The erythema and swelling ^vhich constitute the direct reaction to the injection of active principles of the phenomenon of local skin reactivity, are extreinely variable in different rab])its and differ in intensity in various skin sites of the same rabbit. The erythema producing factors are not remo\ed by boiling and auto- claving of the filtrates and they are not inactivated by addition of immune sera. 148 LOCAL TLSSHE REACTIVITY 111 all my expciinicMits tlicre were always recorded the size and intensily of local reactions jnecedinj^ and following the provoca- tive injection, in order to determine whether the direct reaction to the bacterial filtrates bears any relationship to the size and severity of the hemorrhagic reaction of the phenomenon. The observations of this sort dealt ^vith filtrates derived from various microorganisms (i.e., B. typlwsiis, B. enteriditis, B. tuberculosis, Streptococcus hemolyticus, etc.) culttn^ed in different media and injected into one or several skin sites. The direct reactions following the intradermal injections of bacterial filtrates depend not only on the bacterial products pres- ent in the preparation but also on the medium ingredients. For this reason, "agar washings" filtrates in ^vhich there is present a comparatively small amount of extraneous material and bacterial autolytic products, produce definitely w^eaker primary skin reac- tions than broth and tryptic digest broth culture filtrates. Con- centrated products of Ecker and Rimington (1927) elicit greater primary inflammation than filtrates of broth and tryptic digest broth cultines. Conversely, the phenomenon-producing potency of the "agar w^ashings" filtrates is by far superior to the potency of filtrates of cultines made in fluid media and preparations of Ecker and Rimington. When single or several preparatory injections are made into the abdominal skin, it is seen that sites which do not show^ any primary reactions may respond severely to the provocative injec- tion. Also sites ^vith erythemas of various intensity and size show approximately equal response to the provocative injection. As pointed out before, bacterial filtrates differ in their skin- preparatory and reacting potency. B. typJiosus "agar washings" filtrates contain high skin-preparatory and comparatively high reacting potency, w^iilst meningococcus "agar ^vashings" filtrates contain a lower skin-preparatory potency, and higher reacting potency than the B. typJiosus "agar washings" filtrates. By taking advantage of the possibility of crossing active principles derived from various bacterial filtrates, rabbits were prepared with "agar washings" filtrates in various dilutions and subsequently injected intravenously ^vith 25 reacting iniits of meningococcus "agar washings" filtrates. In this manner it "was possible to obtain typi- cal reactions in skin sites prepared with diliuions of B. typhosus as high as 1: 1000 (provided only one site is prepared) . It may be stated here that intradermal doses of a dilution 1:100 produce ROLE OF INFLAMMATION 1 49 primary reactions not detectable with the naked eye. This appear- ance of the prepared skin site preceding the intravenous injection is indeed in very sharp contrast to the chaniatic hemorrhagic and necrotic lesion elicited a fe^v hours after the intraxenous injection. As will he seen later from the experiments described by Gerber, dilutions above 1:150 produce a primary microscopic inflamma- tory reaction, which by no means exceeds the effect of an intra- dermal injection of saline. The facts brought out are amply corroborated by the ^vork of \arious authors. Plant (i9^^2) described the use of a material deri\ed from Sjjiyocheta pallida endowed xvith phenomenon- producing potency, xvhich produced insignificant primary reac- tions. Old cultures incapable of eliciting the phenomenon gave stronger primary reactions than those endowed xvith the phenom- enon-producing potency. J. W. Mu (1935) who elicited the phe- nomenon with extracts of Ascaris luftibricoidis, also notes that the intensity and size of primary reaction is in no way related to the reaction following the intravenous injection of the potent material. According to Joukow-Werejnikow and Lipatova (1933- 34) , various chemical fractions of plague bacillus totally inactive for elicitation of the phenomenon may e\oke as much primary reaction as those xvhich possess a high phenomenon-producing potency. Bacterial filtrates differing in the concentration of their pre- paratory factors may produce eqtially severe primary reactions, and frequently, those possessing a low skin-preparatory potency may elicit stronger primary reactions than filtrates of high skin- preparatory potency. It has been my frequent observation that meningococcus "agar washings" filtrates possessing a skin-prepara- tory potency ranging fjetween 10 to 50 units per 0.25 c.c, when undiluted produce considerably stronger inflammation than ty- phoid "agar washings" filtrates containing from 50 to 300 pre- paratory units, when used undiluted. Stolyhwo (1935, 1936) ob- served that B. paratyphosus "agar xvashings" filtrates gave stronger primary reactions than B. typhosus "agar washings" filtrates. In this case also, B. paratyphosus preparatory factors are ordinarily of lower skin-preparatory potency than the B. typhosus filtrates. He is of the oj)iiiion that the intensity of the jjrimary reaction is completely independent of the hemorrhagic necrotic reaction fol- lowing the intra\enous injection. Apitz (1933/^') obtained excej^tionally strong j)rimary reactions 150 LOCAI. 1 ISSUE REAC'IIVITY with liliratc's ol B. Icjji.sejjliciis ol low plR'n()iiien()n-j)i()(liu in;^ jm)- tc-iuA, j)i()l)al)ly due to tlie spontaneous inlections ol rabbits Avitli this organism (Hanger. i<)L'hosus ])reparatory factors. No enhancement of the state of reactivity was observed. The last series of e.xperiments demonstrated that the addition ol histamine to bacterial filtrates lacking in skin-preparatory factors failed to elicit the state of reactivity to the phenomenon. These observatic^ns are in contrast to a casual observation ol Gratia and Linz (1932^) in which an enhance- ment of the state of reactivity is attributed to histamine. Acetylcholine: The skin-preparatory potency of acetylcholine was tested in dilutions 1:500, 1:1000, and 1:5000. Dilution 1:500 gave primary necrotic lesions. The addition of acetylcholine to B. typhosus "agar washings" filtrates in various dilutions used for skin-preparation did not modify the typical re- action of the phenomenon. AUylamine: The skin-preparatory potency of allylamine was tested in dilutions 1:1000, 1:2500 and 1:5000. The use of this substance as a preparatory agent appeared to be of spe- cial interest because of recent observations of Mellon, Baker and Mcllroy (1935) . These authors found that the intracutaneous injections of allylamine in rabbits in the above dilutions promptly led to a necrotizing arteriolitis. In its earliest stages, the outstanding features of the arteriolar disease are hemorrhage into the wall and loss of its structure. Perivascidar changes con- sist of mononuclear infiltration of tissue, polyblasts and formation of nodular periarteritis. The picture, according to these authors, is strikingly similar to that present in streptococcus human infections, which appear to have in their background the phenomenon of hypersensitiveness. Croton oil : Croton oil was painted on the skin of the abdomen and one ear two to four times at forty-eight-hour intervals. The first painting resulted in slight swell- ing and the second one gave pronounced irritation and scab formation. The provocative injections were given, following each of the paintings and at the time when the scabs fell off and young granulation tissue was present. In some rabbits, the healed and the irritated sites received preparatory injections of B. typhosus "agar washings" filtrates. The provocative injecticjn developed in these prepared sites was distinctly limited to the site of the injection. No extension of the reaction was noticed along non-prepared irritated sites and the non-prepared sites of granulation tissue. ROLE OF INFLAMMATION I57 Urethane ethyl : Urethane ethyl was used in dilutions 1:100, 1:200, and 1:500. The substance was also used in mixture with B. typhosus "agar washings" filtrate without effect. Calcium chloride: Dilutions 1:100, 1:200, and 1:500 of calcium chloride were tested. The substance in above dilutions was also re-injected into sites prepared with B. typhosus "agar washings" filtrates at various intervals of time with- out effect. Antispleen sera: Guinea pigs and a sheep were immunized with rabbit spleen, according to the method described by Bedson (1921, 1922). The guinea pig anti-serum showed precipitation Avith rabbit serum in dilution 1:128, and the sheep anti- serum precipitated with rabbit serum in dilution 1:64. The intradermal injec- tion of the serum produced reddening, swelling, and in some instances, small necrotic and hemorrhagic areas. The injection of 50 and 100 units of menin- gococcus, Group III, 44 B., "agar washings" filtrate at various intervals of time following the intradermal injection, failed to enhance the primary reactions. Substances studied by others : The inability of non-bacterial substances to elicit the state of reactivity to bacterial filtrates was amply corroborated by a large group of various investigators who studied a number of the substances described ab<)\e and also substances not previously tested by me, namely: Plain broth (Hanger, 1928/^; P. Bordet, 1936^; H. Gross, 193 1«, b, c\ and Stolyhwo, 1935, 1936) ; turpentine (Hanger, H. Gross, P. Bordet, and DeCour, 1934) ; India ink (Gratia and Linz, 1932c; and P. Bordet) ; peptone (P. l^ordet) ; horse serum (Gross and Bordet) ; milk (Gratia and Linz; P. Bordet) ; agar (Sickles, 1931, 1934) ; gum tragacanth (P. Bordet) ; 20 per cent sodium chloride solution (Aitoff, Dion, and Dobkevitch, 1936^/, b) , etc. According to P. Bordet, peptone and milk seemed to ser\'e as skin-preparatory factors in the following modification: (iuinea pig or rabbit serum was injected intracutaneously into rabbits. T^venty-four hours later, there Avas local sAvelling or thick- ening of the skin into Avhich 0.5 to 1 c.c. of peptone ^vater (2 per cent Chapoteaut peptone) Avas injected intraciuaneously. Five hoins after the peptone injection, 1 c.c. of B. coli filtrate injected intravenously elicited a local hemorrhagic reaction. The control sites injected with peptone alone or serum alone Avere negative. The same results could be ofjtained Avitli milk \vhen used instead 158 LO(^AL TISSUE REACTIVITY of peptone, althoiioh tlic injection o{ sodium chloride or honiol ogoirs or heterologous seriun into the skin alter the j^reliniinary intracutaneous serimi injection remained negative. According to P. Bordet. the preliminary local serum injection increases the skin sensiti\ity to the irritating properties of peptone and milk \vhich cannot be manifested ^vhen injected into the unj)rej)ared skin. The action of the previous serimi injection seems to him to rest in the s^velling or thickening of the skin which it causes. The objection \vhich may be raised against liordet's obser\a- tion is that a certain amoinit of bacterial toxic substances may be present in the peptone and milk. Chapoteaiu's peptone is made by the prolonged digestion of stomachs of pigs remoxed after death. Milk, even Avhen drawn under the strictest ase])tic precau- tions, contains \arious numbers and types of bacteria (V^ollum, 19-^9) ■ Freinid (i9^54<'/) reports that when concentrated broth or tur- pentine is introduced into the skin of tubercidous guinea pigs and later typhoid hltrate is injected into the vascidar system, no hemorrhages occtu" in the skin at the site of inflammation. Marginesu (1934) exposed the skin of rabbits to ultraviolet radiation following the local congestion. An intravenous provoca- tive injection of B. coli culture filtrate was made. No reactions were obtained at the site of the radiated skin areas. The above author also used copper acetate as a skin-prepara- tory agent xvithout success. It is interesting that copper acetate is a strong vascidar poison capable of producing in rabbits exten- sive vascular lesions of the internal organs. Freund reported that eel sertmi coidd serve as a skin-prepara- tory agent. Michelazzi (ig'^r^d) , however, prepared the skin with anti-rabbit hemolytic serum and eel serimi with negati\e results. B. coli filtrate xvas used for the provocative injection. Fiorito (19-53) prepared the right ear of a rabbit 'with 0.05 c.c. of a 10 per cent solution of turpentine and the left ear xvith an active fjacterial filtrate. Twenty-four hours later, he injected Vib- rio cholerae culture filtrate intravenously. Five rabbits shoxved i-[- reactions and 5 rabbits gave no reactions in the turpentine prepared skin sites. In the 5 rabbits xvhich were negati\e w'lXh turpentine, the ears prepared with the bacterial filtrate showed distinct reactions but of varying intensity. Sickles (1931, 1934) found that agar in the fbrm of Hitchen's medium, possessed no skin-preparatory potency. In recent experi- ROLE OF INFLAMMATION 1 59 ments she also failed to induce the state of reactivity with frac- tions of agar obtained l)y partial acid hydrolysis. I failed to induce the state of reactivity by intradermal injec- tion of 0.5 per cent of granulated and shredded agar in 0.(85 P^^^ cent salt solution. Stolyhwo (1935, lO'^G) produced traimia in the skin of rabbits by shaving" with a didl razor, application of ligattne, clamp, etc. Twenty-foiu' horns later, the rabbits injected intravenously with potent bacterial filtrates showed no reactions at the trainnatized sites. Gratia and Linz (1932c) failed to elicit the state of reactivity with distilled w^ater and a 20 per cent sodium chloride solution. Fretnid (ig'^^a) injected guinea pigs intracutaneously with 0.2 c.c. of an active B. typhosus culture filtrate and follo^ved this by intracardiac injection of the same filtrate in a dose of 1 c.c. None reacted with hemorrhage. He agrees, therefore, with H. Gross that the phenomenon of local skin reactivity cannot be elicited in guinea pigs with any degree of regtdarity, and states that condi- tions necessary for the phenomenon are different in the guinea pig and the rabbit. In spite of this statement, he proceeded w^ith a group of experiments on the phenomenon in the guinea pig and drew^ certain conclusions concerning the natiue of the phe- nomenon in rabbits on the basis of his experiments in gtiinea pigs. He reported that Avhen silver nitrate \vas injected into the skin of guinea pigs, redness, edema, and necrosis follo^ved, and in a few guinea pigs, small areas of hemorrhage cotild also be noted. Abotit one-half of the gtiinea pigs ^vhich recei\ed an intra- vasctdar injection of typhoid filtrate, reacted with hemcjrrhages in the site of the injecticjn cjf siher nitrate. As described cjn page 155, silver nitrate ^vhich produces a primary severe hemorrhagic necrosis has absolutely no skin-preparatory potency in rabbits (Gerl)er, 1936^/) . Debonera, Tzortzakis and Falchetti (1932) described the skin- preparatory effect of sterile \aseline in guinea pigs. According to these authors, the intraxcnous injection of 1 c.c. of B. coli cidtine filtrate elicits an immediate piuptnic reaction at the vaseline in- jected sites. These experiments were repeated by me in rabbits with entirely negati\e residts. Here again, inasmuch as the phe- nomenon of local skin reactivity is doid^tftd and extremely irreg- idar in guinea pigs, experiments of this nature cannot be consid- ered in studies on the mechanism of the phenomenon in rabbits. l60 LOCAL TLSSUE REACTIVITY It is interesting to note that some of the guinea pigs studied by Debonera, Tzortzakis and Falrhetti died shortly after the intra- venous injection of B. coli toxin. The postmortem examination of the animals showed pseudo-tuberculosis. It is quite possible that the preparatory effect of vaseline in spontaneously infected guinea pigs did not consist in the primary elicitation of the necessary state of reactivity, but simply in localization of the toxic factors in vaseline-treated skin sites with subsequent elicita- tion of reactivity by means of these localized factors. P. Bordet (19^56/;) observed that tiie intracutaneous injection of talcum did not prepare the skin to the ])henomenon. As is seen, a great variety of non-bacterial substances capable of producing various types and degrees of inflammation, \'ascidar damage, and other injinious local effects, consistently fail to elicit the state of local skin reactivity described. Thus far, the elicita- tion of this state remains an exclusive property of active bacterial principles of characteristics defined. EFFECT OF NONBACTERIAL SUBSTANCES UPON THE STATE OF REACTIVITY TO BACTERIAL FILTRATES In attempts to study the nature of the reactivity, experiments were also carried out in order to determine wdiether any of the j)harmacological substances described belo\v ^vould be capable of modifying the state of reactivity elicited by means of the active bacterial jninciples, notably: Ascorbic acid : The j)rej)arati()n \vas injected in a dose of 100 mg. per kilo of body ^veight. The injections were given inunediately after the intradermal jjreparatory injection of a ])acterial filtrate and fif- teen minutes, one-half hoiu', six hours, and t^venty-four hours following the preparation. T^venty-four hours after the intrader- mal injection of the bacterial filtrate the rabbits thus treated re- ceived 50 units of meningococcus Ciroup III "agar washings" filtrate. Ascor])ic acid failed to produce any effect upon the phe- nomenon. These experiments Avere done in collaboration ^vith Dr. B. Schick. Alypin: The time necessary to j)roduce anesthesia with the alypin em- ployed was tested on the rabbit's coujunctixa. Fi\e per cent solu- ROLE OF INFLAMMATION l6l tioii instilled into the eye produced a partial anesthesia of the con- jimctixal reflex ^vithin one minute, and a complete anesthesia \vithin three minutes following the application. The anesthesia disapj)eared completely within fifteen minutes. Upon intradermal injection, 5 per cent solution of alyj^in produced a local necrosis without hemorrhage; 2.5 per cent solution of the preparation pro- duced no necrosis but a slight inflammatory edema. Alypin ^vas added to the bacterial filtrate used for the preparatory injection. In other experiments, it was reinjected into the site prepared with the bacterial filtrate one-half, one liour and twenty-four horns follo^ving the preparation. The provocative injection of B. typlio- sus "agar \vashings" filtrate was made t\venty-four hoins after the preparatory intradermal injection of the bacterial filtrate. In all the experiments alypin failed consistently to produce any effect upon the phenomenon. Cocaine : Cocaine was used in 5 per cent and 2.5 per cent soliuions, sim- ilarly to alypin. There was observed no effect of cocaine upon the phenomenon luider discussion. Acetylcholine : The injection of acetylcholine in a dose of 1 mg. per kilo of body Aveight, preceding the injection of provocative factors at va- rious intervals of time, failed to produce any effect. Pyridine : Studies on the effect of pyridine ^vere of interest because of Bed- son's observations on the ability of pyridine to reduce substantially the number of blood platelets in rabbits; 0.06 gms. per kilo of body weight of pyridine produced no lethal effect. Rabbits were prepared with meningococcus "agar washings" filtrates, treated by intravenous injections of pyridine t^vo, twenty-four, and forty- eight hours follo^ving the skin preparation. Twenty-five reacting imits of meningococcus "agar Avashings" filtrates ^vere injected in- tra\'enously t^venty-foin- horns after the skin preparation. All rab- bits showed typical strong reactions. Anti-platelet serum: (iratia and Linz (1932c) employed Roskam's anti-platelet serinn in the follo^vino manner: l()2 LOCAL TISSME REACTIVITY Rabbits wvrv prejjared l)y intiadeniial injection of B. coli cul- ture (iltiale and fi\'e Iiouis later injected with the serum in a dose of 1 c .c . B. coli cuhure filtrate injected twenty-ioiu" hours alter (he skin pi ej)arati<)n. seemed to elicit unusually stroni^ re- actions in the 2 rabbits tested. Inasnuu ii as the sjieed with which the reaction ajjpears is sid)ject to indixidual variation, no con- ( lusion can be drawn from this exj)eriment. As mentioned before, in my e\j)eriments. anti-spleen serum pre- pared according to the methods of Bedsou, failed to elicit the state of reactivity of the phenomenon. Prolan : Three intravenous injections of large amounts of ]:)rolan (7 c.c. per kilo of body weight) given at tAventy-foin-hour intervals j^ro- duced no effect upon the phenomenon. Rabbits treated with prolan \S'ere also injected with subminimal doses of reacting factors. No enchanced susceptibility was detected in these e\]Deriments. General anesthesia: Several grotips of rabbits were prepared ^vith meningococctis "agar ^\•ashings" fdtrates. Twenty-four hotns later they were anes- thetized. LInder anesthesia they ^vere injected intravenously with 25 imits of meningococctis "agar washings" filtrates. Anesthesia was continued for one-half and one hour following the intravenous injection. There ^vas observed no effect upon the reactions. India ink: Gratia and Linz (i932r) injected intraperitoneally large dosei of India ink until the skin of the animals became intensely black. The prcxedure did not enhance, nor reduce the stisceptibility to the phenomenon. STUDIES ON THE BEHAVIOR OF CAPILLARIES IN SITES PREPARED WITH ACTIVE PRINCIPLES Histological investigations on the sites prepared with bacterial filtrates prior to the injection of reacting factors do not demon- strate any paralysis or unusual dilatation of capillaries of the area prepared. For these reasons, it appeared of interest to Kielanowski ROLE OF INFLAMMATION 163 (1935) and Kielanowski and Selzer (1934//) to determine the re- sponse of sites thus prepared to various physical and pharmaco- logical agents. Their findings were as follo^vs: The increase of temperature by application of cotton tampons dipped in hot water (50° to 75° c.) pro\oked hyperemia in the prepared site similar to the reaction of normal skin. The elexation of temperatme and the ensuing congestion did not cause the ap- pearance of petechiae. The prepared skin and its capillaries re- acted in a normal manner. Tiie elevation of temperature of the prepared area in rabbits injected intravenously with a potent cultiue filtrate accelerated slightly the appearance and course of the hemorrhagic reaction. The congestion of tiie prepared site produced by massage had the same effect. The injection of 0.5 c.c. of adrenalin solution (1:1000) into the prepared site caused the appearance of an ischemic zone due to the capillary constric- tion. The hemorrhagic reaction around this zone developed nor- mally but there were no petechiae in the ischemic area. (Similar observations were made by Klein, 1931-32.) If the adrenalin in- jection Avas gi\'en when the petechiae started to appear, there Avas a resultant anemia but no petechiae dexeloped in the ischemic area. Injection of histamine in dilution of 1:1000 had the same effect as adrenalin but the anemia ^vas more marked and lasted longer. Injection of physiological saline had no effect on the appearance and course of the phenomenon of local skin reactivity. The cooling of the sensitized skin by ether had no effect on the reaction. The authors concltided from their experiments that capillaries react normally in the prepared skin sites. They believe, therefore, that some specific biophysical ciiange takes place in the capillaries through the effect of the j^rej^aratory injection of bacterial fdtrate, which thus far, cannot be demonstrated by the ordinary studies on capillary permeability and plnsiological bcha\'ior of the capillaries. They also state that the hypothesis of formation of "jiurj^urogenic factors" offered by Ciratia and Linz (i932r), lacks experimental and morphological proof. In my own experiments, sites prepared with bacterial filtrates uere tested by application of suction prior to the injection of pro\'ocati\e factors and following it. This procedure did not dis- close any enhanced fragihty. The same suction aj:)plied to the skin after the injection of the provocative factors resulted in small 164 LOCAL TLSSUE REACIIVLIY petechial k'sioiis. wliic Ii (onsistcd nii( rose ()j)i{ally ol (apillaiy dila- tation and extraxasaiion oi erythrocytes. Enhanced liability ol the capillaries in sites distant from the local reaction coidd be denicjn- strated lor a pci iod ol loin- to lort)-eii>ht hours lollowin'^ the j)r()- v()cati\e injection ol the actixe j)iin(iples and depended on the sexerity ol the local lesion. Gratia and Linz (i9'^2c) noticed that alter the intravenous ])ro- \()cati\e injection of the active principles, pnlling of hair, or slight traimia in sites distant from the site of the typical reaction produced petechial hemorrhages xvhich sometimes became conlluent. The obser\ations demonstrating enhanced capillary fragility at sites distant from the local reactions deserve further investigations. Koplik (1937) reported that following the intrader- mal injection the preparatory factors may spread to adjacent lymph nodes by way of the lymphatics. It is possible, therefore, that the factors diluted in the lymph prepare distant sites for a reaction of lesser severity. The mild reaction thus obtained may then, in turn, be enhanced by additional mechanical stimidation. The other possibility is that there form at the site of local reaction some secondary diffusible toxic substances xvhich by themselves modify the capillary response to stimidation in sites distant from the local reaction. I also attempted to determine the rate of diffusion of saline in- jected prior to the provocative injection, into sites prepared by bacterial filtrates. Blebs produced by an injection of 0.5 c.c. of saline into the prepared sites and into normal sites persisted for an equal length of time, i.e., approximately thirty minutes. Bmnet (1931) injected intra\enously 1 to 2 c.c. of 1 per cent solution of trypan blue at different periods after the provocative injection Avas gi\en. There xvas no e\idence that any increased per- meability of capillaries in the prepared areas preceded the appear- ance of purpina. The slightly indurated prepared sites appeared dehnitely less blue than the surrounding skin until the commenc- ing pmpura ^vas clearly visible. Later there A\as a uniform staining of the skin. The obser\ation xvas confnmed by P. Borxlet. MORPHOLOGICAL STUDIES ON THE PHENOMENON In recent years there apj)eared a group of extensixe histological studies on the reaction of the phenomenon of local tissue reactiv- ity. Karsner and jVIoritz (1934) state that the preparatcny injection ROLE OF INFLAMMATION 1 65 ot bacterial jjioducts brings about an exudative inflammation. The type of inflammation observed, ho^vever, does not stiggest to them a state of previous sensitization to the substances in the bacterial preparations. My o^vn observations reported in a pre- \ ious chapter, also exclude the possibility of previous sensitization to bacterial substances being resj^onsible for the state of reacti\'ity (p. 41). The histological studies of Karsner and Moritz on the local reac- tion follo^ving the intravenons injection sho\ved extidative inflam- mation ^^■hich diflered from the primary reaction in increased severity and marked damage to blood \ essels. These authors believe that the increased se\erity may be due to concentration of the in- jurious agent at the site of inflammation determined by the {pre- paratory local injecti(jn. As already discussed, an asstniiption of the sort is strongly contradicted ])y the experiments on the lack of preparatory potency on the part of a large group of non-bacterial substances, and by other facts just presented. According to the au- thors themselves, the exudation catised by the prej^aratory injec- tion alone is edema, infdtration of leucocytes and large mono- cytes. The primary vascular injiuy, ho^vever, is morphologically demonstraljle only in a fe^v sections, notably those from periar- ticular structures. In contrast to this, the intravenous injection of acti\e principles causes increased cellular infiltration, necrosis of the exudate, phagocytosis of cell debris, dilatation of vascidar avails and severe hemorrhage. Healing is due to granulation and cicatrization. Karsner and Moritz employed bacterial filtrates con- centrated by the method of Ecker and Rimington. According to Ecker and Welch (1930) , the bacterial filtrates thirs prepared are responsible for severe primary inflammatory reactions, i.e., pro- noiniced erythema, large indurated s^velling, etc. These primary reactions, in part at least, may be responsible h)r the se\ere exudative inflammation follo^ving the primary reaction recorded by Karsner and Moritz. Severe and confusing primary reactions were also noticed by Apitz in some of his experiments ^vith fil- trates of B. coli cultiues in synthetic mediimi employed l^y Ecker and Welch. Apitz (1933) compared the histological manifestations follow- ing the preparatory injection alone and h)ll()\\'ing the combined preparatory and provocati\'e injections of active principles derived from various microorganisms. He found that there existed only a Cjuantitative difference between the intensity of reactions elicited iGf) LOCAL TLSSUE REACTIVITY willi various niicr()()r*>anisiii.s and that (|iialitati\cly the lesions were all the same irrespec ti\e ol the niiciooi^anisni enij^loyed. According to this author, histoloj^ically, the primary reaetion to the intradermal injection ol B. tyjjiiosus liltrates is irregular and considerably less pronoiniced than lollowing the injection of B. I('j)i.s('j)licu.s. The intensified tyj^e of primary reaction produced with B. U'j)i.s('j)licus cultine idtrates is probably due to the addi- tional complicating effect of spontaneous hypersensitiveness of the rabbits to widely spread natural infection with this group of organisms (Hanger) . Apitz draws attention to the possibility of differentiating the hemorrhagic lesions typical of the phenom- enon of local skin reactivity to bacterial filtrates {viz. B. ty- pliosus, etc.) from the superiinposed liemorrhage in sites reacting with primary allergic reaction (B. lepisepticus) . In his sttidies on the primary reaction to B. lypJiosus, B. coli, and other active principles of the phenomenon, he observed that the primary reaction gixes rise to an acute leucocytic and edematous infiam- mation which, in contrast to the primary allergic inffanmiation of B. lepisepticus, begins to disappear in twenty-four hoins following the injection. The appeaiance of this inflammation is entirely non-specific and histologically, totally unrelated to the reaction following the injection of the provocative factors. The reaction of the jjhenomenon is described by him as severe damage of all the blood vessels of the reactive area, the main symptom being profirse hemorrhage. The vascular damage is made obviotis by the migration of leucocytes through the vessel ^vall, conspicuous formation of parietal j)latelet throml)i, severe dilatation of lilood vessels accompanied loy k.aryorrhexis and fatty degeneration of the endothelial cells. This severe reaction is followed by necrosis lead- ing to demarcation and separation of the destroyed tissue. Kielano^vski (1935) and ICielanowski and Selzer (1934/;) studied morj)hological changes in rabbits which were prepared by two simultaneoirs intradermal injections of 0.5 c.c. of B. coli cul- tine filtrate. This was followed by an intravenous injection of the same filtrate twenty-two hours later. At intervals of twenty-two hotns after the first injection (just before the intravenous injec- tion) ; t^vo hoius and thirty-five minutes, four hc^ms and forty-five minutes, three hours and fifteen minutes, four hours, four hours and t^venty-five minutes, six hours and forty-five minutes, twenty- loin hours, and sixty-five and one half hours after the second injec- tion, sections were taken fcjr histological examination. Seven of ROLE OF INFLAMMATION 167 these rabbits showed a positive reaction and 2 were refractory. There was no relationship ])etween the local manifestations follow- ing the preparatory injection and the severity of the reaction after intraxenoiis injection. It is of considerable interest that according to their findings the histological pictmes of all the animals, in- cluding the refractory ones, were identical before the intraxenoiis injection. Prior to the intravenous injection there were no special changes observed in the epitheliinii. All layers of the skin tissue were defi- nitely edematous and the collagenous fibers ^vere swollen. The capillaries were either empty or filled with blood. The skin ar- terioles were normal. These authors observed, however, platelet thrombi of various sizes. The polymorphonuclear leucocytes formed several layers around the vessel. The severe primary re- actions observed by Kielanowski and Selzer may be accounted for by the fact that filtrates of seven day old broth cultures of B. coli were injected in large amounts (two sites 0.5 c.c. each) . In positive animals, following the intravenous injection, the edema and leucocytic infiltration showed no increase. Tlie venous thrombi formed after the intra\en()us injection completely filled the lumina and the necrobiosis of the red blood cells was dehnite. In the early stages (up to two hours after the second injection) in addition to the typical changes described, a strong dilatation of the capillaries was also demonstrable. This occurred in the blood vessels situated perpendictdarly or obliquely to the super- ficial layers of the skin. The endothelium ^vas flattened otit but no changes were ofjserved. The nuclei ^vere irregtilarly f)ut intensely stained with hematoxylin. Isolated petechiae were seen only in the early stages. Hemorrhage occurred at first only in the upper layers of the skin, i.e., in the papillary layer, btit later it Avas possible to observe small hemorrhages in the striated muscle. Petechiae were often circumscribed and clearly separated from the surrounding tissues and the spaces were filled with red blood cells. Kielanowski and Selzer maintain that the reaction seemed to depend on sudden rupttire of blood vessels and increased perme- ability of the capillaries. Also, the macroscopic picture of the phe- nomenon, i.e., the sudden apj^earance of hemoi rhage ^vithout pre- liminary swelling seemed to them to corroborate the hypothesis of capillary rupttire as the cause of the reaction. They think that the subsequent necrobiosis is dtie to a disturbance in the blood supply. .i & ' t>*%,' ROLE OF INFLAMMATION 169 Under experimental conditions some^vhat different from those employed by the autliors just quoted, Gerber, in my laboratories, attempted systematic studies in order to determine further ^vhether any parallelism could be disco\'ered bet\\een the skin reactivity induced by active principles of the phenomenon and the inflammation resulting from the intradermal injection of these principles. Comparisons were made by him when: Group 1 — Bacterial filtrates possessing a very high skin-prepara- tory potency were used; Group 2 — Skin-preparation was performed with a large dose of bacterial filtrate of high potency and forty- eight liours elapsed before reacting factors Avere injected; Groujj 3 — Skin-preparation was performed with bacterial filtrate almost completely neutralized by homologous antitoxic horse serinn and with non-neutralize filtrate and tiie jDhenomenon elicited in both instances; Ciroup 4 — Skin-preparation by l^acterial filtrates of lo\v skin-preparatory potency ^vas employed. (Figs. 11-13.) In this work Gerber (1936^) found that tlie degree of inflam- mation that may follow skin preparation depends to a great extent on the amount of bacterial filtrate injected. Also, filtrates of various microorganisms produce different degrees of inflammation. Like- wise, filtrates of organisms which possess little skin-preparatory potency, such as Streptococcus hei)iol\ticus, occasion moderate or marked inflammation with skin-preparation. Nevertheless, there is no parallelism bet\\'een the skin-preparatory potency of a filtrate and the inflammation resulting from its intradermal injection. This is clearly demonstrated in the results of the experiments in which identical degrees of inflammation followed skin prepara- tion Avith a filtrate almost completely neutralized by homologous antitoxic horse serum and from skin preparation ^vith non-neu- tralized filtrate. The former filtrate, hoAve\er, ^vas capal)le of only partially preparing the skin so that after intravenous injection of the reacting factors only a slight reaction was obtained. Similarly, a filtrate may possess a high degree of skin-preparatory potency and yet the inflammation seen after intradermal introduction may be slight. In this instance, the elicitation of the phenomenon re- FiG. 11. A. section of skin photographed inider low magnification, sho\\ing fnllv developed phenomenon three and one-half hours after intravenous injection of reacting factcjrs. There are marked inflammation, hemorrhage and thrombosis of \eins. li, higlier magnification of section shown in Figure 1 lA, demonstrating extreme vascular engorgement and perivascular collars of leucocytes, with infiltration of walls and extravasatcd erythrocytes between connective-tissue fibers. (Gerber, i936rt.) Fig. 12. A, section of skin site after preparation with B. typhosus filtrate almost completely neutralized by homologous antitoxic horse serum, followed twenty-four hours later by intravenous injection of 25 reacting units of same filtrate. Four hours after intravenous injection there were only slight inflammation and moderate vascular congestion. B, section of skin site after preparation with non-neutralized B. typhnsus filtrate, followed twenty-four liours later by intra\enous injection of 25 reacting units of same filtrate. Four hours later tlie skin presented typical phenomenon, with marked hemorrhage and inflammatory reaction. (Gerber, 19360. ) Ik.. 13. .J, section ot skin site six hours alter intradermal injection of typhoid Ijacillus filtrate in dilution of 1:150. There is only slight inflammation, with occasional dilatation of vessels. B, histologic appearance of typical phenomenon after skin prep- aration with typhoid bacillus filtrate in dilution of 1:150, followed by intravenous injection of 25 reacting units of same filtrate. Striking hemorrhage and vascular en- gorgement are evident. (Gerber, 19360.) 172 LOCAL TLSSIIE REACTIVITY suited in as severe a reaction as tliat seen Avitli the use ol llUrates which produced much greater degrees of inflaniniation with skin preparation. ()I)vi()iisly the determining factor in the production of the phenomenon is not in the inflammation attendant on skin jireparation ])ut rather in the a( tual preparedness of the skin. The degree of inllannnation occurring with skin {^reparation is no index of the state of preparedness of the skin. Thus, in the experiments in which forty-eight hours elapsed after the skin preparation, the inllannnation was still severe, yet the state of reactivity of the skin ^vas so low that it required fotir times the usual cpiantity of liltrate to elicit the phenomenon. Moreover, the degree of inflannnation prcxluced Avith skin jjrejjaration did not determine the extent of the inflammatory reaction seen on production of the pfie- nomenon. It is still tmknown ^vhat substances in the iUtrate are responsi- ble for the inflammation. Ho^vever, since intradermal injections of a filtrate of high skin-preparatory potency result in minimal in- flammation, no greater than that of the controls, the preparatcjry factors are in themselves not responsible for the inflammation. It is significant that in Gerber's studies in ^vhich filtrates of high skin-preparatory potency '^vere used so that only small quantities were necessary, and the amoimt of non-specific irritating sub- stances was clearly reduced, the absence of hemorrhage and thrombosis ^vith skin preparation alone ^vas a constant feature. Thus, the appearance of hemorrhage and thrombosis Avith ac- centuation of the inflammation served to differentiate the phe- nomenon from skin preparation c|ualitatively as ^vell as quantita- tively. The cause of the hemorrhages could not be ascertained by this author, morphologically. Although occasional ruptured capillaries were seen in sections of skin showing the phenomenon, he cotild not attribute the extensive hemorrhage to these alterations. Gerber was also impressed by the occurrence in the reaction of the phenomenon of widespread thrombosis affecting chiefly the veins. At a site of attachment of the thrombi no endothe- lial alterations were noted by him. The affected vessels were surrounded by thick collars of polymorphonuclear leucocytes; these often infiltrated the entire wall of the vessel. Ghanges in the elastica were absent. A])itz suggested that Dietrich's concept of endothelial hyjjer- ac ti\ ity may serve to explain the origin cjj the thrombi obserxed in ROLE OF INFLAMMATION 1 73 tlic j)lien()nien()n. The endothelial hyperactivity may he initiated hy the skin-preparatory injection. \\'hen this is followed hy a sec- ond stimulus ^vhich reaches the vessel hy the way of the blood stream (reacting factors) , and \vhen there are sinudtaneous slow- ing of the circtdation and perivasctdar inflanunation, thrombosis results. Further studies on the subject are necessary, ho^vever, because morphological proof of endothelial damage w^as lackmg in the studies of Kielanowski and Selzer, Apitz and Gerber. Gerber also pointed otit that the extent of thrombosis is not quantitatively dependent on the degree of inflammation present with skin preparation or on that occurring W'ith the phenomenon. During the first f^venty-four hom^s there may be an increase in the degree of thrombosis, possibly as a result of extension of the con- ctnrent inflammation. The degree of thrombosis varied with the type of reacting factors as ^vell as ^vith the skin-preparatory quali- ties of the factors employed. For example, experiments on the elic- itation of the phenomenon by the use of the antigen-f antibody combination as the reacting factors (Chapter ix) demonstrated a more marked degree of thrombosis than ^vhen filtrates of typhoid bacilli were injected intravenously, even though in both instances the skin preparation was productive of identical degrees of moder- ate inflammatory reaction. Inasmuch as the earliest changes seen in the phenomenon of local skin reactivity are extreme vascular dilation and engorge- ment, profuse hemorrhage and thrombosis, the reaction may be sharply differentiated from reactions of protein hypersensiti\'eness even on the morphological basis. The follo^ving recapitulation of morphological appearance of the Arthus phenomenon gi\'en by Opie (1924c/, b) may be in place: "The reaction of the Arthus phenomenon makes its appearance within teir days after the injection of a protein but does not reach maximtnn intensity until the animal has received six or eight in- jections at intervals of several days. "The Arthus phenomenon resembles \'ery closely the tuber- culin, luetin and typhoidin reactions. "T^venty-four hours after the injection of the test protein, the skin and the subcutaneous tissues become edematous and infiltra- ted with polymorphonuclear neutrophile leucocytes. The epithe- lial cells of the lo\\er layers are SAvollen. After forty-eight hours the stratum corneum separates from the stratinn lucidum by exu- date rich in leucocytes. The cells of the Malpighian layer are 174 LOCAL TISSUE REACTIVITY swollen. The dermis and subcutaneous layers are invaded by an exudate Avliidi distends the tissue and compresses the blood ves- sels. About some ol the vessels there are red blood cells in large niuubers. Connective tissue fibers look homogenous. There is a sharp line oi flemarcation bet^veen the normal and the inflamed tissue. Four or li\e days later, an iU(er is formed. The caseous material at the bottom of the ulcer contains crystals of fatty acid and necrobiosed polymorphonuclear neutrophile leucocytes. Grossly, the necrosed area can be recognized by a central pur- plish-black zone. Microscopic examination shows partial disap- pearance of nuclei with recognizable l3ut s\vollen nuclei fibro- blasts and intact endothelial cells of lolood vessels. Erythrocytes may be very ntunerous in some places. The necrotic area is sinroinided by a zone of polymorphonticlear neutrophile leuco- cytes which are very ninnerous oiuside of the necrotic area and are apparently migrating towards the periphery. The zone of the surrotmding tissue immediately in contact with the necrotic area sho^vs capillaries and \'ery small \eins ^vith hyaline throm- bosis and accumulation of the polymori^honuclcar neutrophile leucocytes abotu the arterioles and small \'eins. The leucocytes may penetrate into the walls of these blood \'essels." Opie also made observations on the histological changes follow- ing the injection of a vvashed sertun precipitate. In his experiments the precipitate Avas obtained by mixing o.i c.c. of horse serum diluted to lo c.c. with 15 c.c. of anti-horse rabbit serum. "The injection of this precipitate elicits a swelling of the skin. In stained sections the precipitate may be recognized as a granular material occupying the spaces between separated fibrous Ijimdles of a liyaline appearance. There is observed a considerable infiltra- tion of polymorphonuclear neutrophile leucocytes at the periph- ery. The zone of leiicocytic infiltration is accompanied by mod- erate edema. 'Torty-eight hoins later the leucocytic infiltration becomes more pronoiniced. Fibroblasts and mononuclear wandering cells and isophiles appear in moderate ntnnber. No necrosis has occinred in any instance of the injection of the washed precipitate." In a sid)sequent chapter experiments will be described vvhereby combinations of animal protein with homologous antisera in- jected into rabbits previously prepared by an intradermal injec- tion of a bacterial filtrate produce severe hemorrhagic lesions at the point of preparation. Opie's description of the effect of the ROLE OF INFLAMMATION 1 75 precipitate deri\ed from protein antigen-antil)ocly mixtures uj)on the skill of normal rabbits is in sharp and clear contrast to the effect of the intravenous injection of such a mixtme upon a site prepared by bacterial filtrate. The latter consists of an immediate profuse hemorrhage ^vhich appears within forty-five minutes to two hours following the provocative injection and is associated w ith prompt necrosis and extensi\e thrombosis in the small blood \essels. The entire pictine reaches it maximum development Avithin foiu' horns. The acute inflammatory reaction is apparently secondary to the changes in the local vascidar system. The prepared skin site shoAvs a reaction of uniform severity and is predominantly hemorrhagic in nature. On the contrary, while the Artlius phenomenon is essentially an inflammatory reaction, it is first quite mild and only eventually becomes so intensified after a number of \veeks of sensitization, that hemorrhagic and necrotic types of inflammation may be induced. ROUTE OF INJECTION OF RE.ACTING FACTORS AND INCUBATION PERIOD NECESSARY FOR ELICITATION OF THE STATE OF REACTIVITY The basic experiments described in the introductory chapter sho^ved clearly that the essential prerequisite for reproduction of the phenomenon of local skin reactivity is that the reacting factors be introduced into the blood stream following the preparatory in- jection into the tissue after a suitable interval of time. In this man- ner extremely se\ere reactions may be elicited with minute quanti- ties of active principles. Thus, in some instances 0.0003 c.c. of meningococcus "agar washings" filtrate were capable of eliciting typical reactions in a single prepared skin site. If one considers that this amount is diluted in at least 100 c.c. of rabbit's blood, and that a part of it may be lost in the tissue before reaching the prepared skin site, the amount actually capable of inducing the lesion may be less than 0.000003 c.c. in these instances. On the other hand, no typical reaction results from an intradermal or subcutaneous reinjectioii into the prepared skin site made after \arious intervals of time. These facts were amply corroborated by Ecker and \\'^elch (1930), Burnet (1931), Gratia and Liiiz (1932c), Bock (1932), Plant (1934), Apitz (1933/-'). Stolvhwo (1935, 1936) , P. Bordet (1936/;) , and others. Apitz, who studied the effect of repeated subcutaneous and in- tradermal injections of acti\e jDrinciples of B. coli cultures noted that the first intracutaneous injection of the filtrate elicited only 176 LOCAL TLSSHE REACTIVITY a mild inflammatory icaclion. riic second injection ])r()iiti,lit al)oiit a ])i()n()inuc'd edema. I listolo^ically, there was a dittiise and dense leiieocytic inldiialion witli Inalini/ation ol the collagenous fibers aeeomj)anied by occasional small areas ol extravasation. There ^vas no damage to the blood \esi;els. Leucocytes showed no signs ol' necrcibic:)sis. According to this aiithoi, it was note^v()rthy that leucocytes infiltrating the pre})ared skin site following the in- travenous injection underwent comjilete necrobiosis. He drew attention to the fact that the intravenous injection of the toxic substances prc:)duced a massive destruction of leucocytes in the capillaries of the liver. Tt was clear, therefore, that the leucocytes thus injured showed a very low degree of resistance ^vhen they reached the prepared skin site. In recent .experiments of Gerber (ig^^Gc/) , a rabbit received in- tradermally 0.25 c.c. of B. typhosus filtrate equivalent to 1(87 re- acting imits. Twenty-fotu- hoins later a similar dose Avas injected into the same site and the skin removed 4 hoins later. In the gross, there Avas an area of edema 2.5 cm. in diameter of a deep red color. No gross hemorrhages were seen. Microscopic examination re- vealed a severe degree of inflammation characterized by marked edema and extensive polymorphonuclear leucocytic infiltration. There was marked perivasctdar infiltration and at times the cells infiltrated the entire vessel wall. Ho^vever, no thrombi were pres- ent. Hemorrhage was likewise absent. Many of the capillaries were dilated and engorged. Small foci of necrosis were scattered in the subcutaneous tissue. As was seen from the description of the histology of the phenomenon, the absence of thrombosis and hemorrhage readily serves to distinguish this picture from that seen in the typical reaction follo^ving the intravenous injection of reacting factors. It must be emphasized that the amoiurt of filtrate employed in Gerber's experiments for repeated intrader- mal injection ^vas many times greater than that used to obtain the phenomenon. In the gross there is also a sharp difference between the appearance of the reaction following repeated intra- dermal injections into the same site and the typical phenomenon. In order to determine ^vhether some sort of a modihcation wotdd enable the second intradermal injection to elicit a typical reaction, the follo^^■ing experiments uere done by me: Each experiment ^vith a given substance was performed in 6 rab- bits. Three sites were prepared by simultaneous intradermal in- jections of 0.25 c.c. of B. typJiosus "agar washings" filtrates. ROLE OF INFLAMMATION 177 Twenty-foui hours laier. one oi the sites was reiiijectecl with the B. tyfjliosiis filtrate; another site was ireiiijected with one oi the folloAving substances: 0.5 per cent oimi tragacanth, 10 per cent gelatine, 0.2 j)er cent potassimn alinn, undiluted liver extract and histamine hydrochloride diluted 1:2000; and a third. site was injected Avith a niixttne of the B. typhosus filtrate with one of the above sul)stances. T\\enty-fotu' hotns later in some areas thus treated there Avas noticed inflanmiation of varying intensity and also edema. No typical reactions were obtained. At thisti'me, the rabl)its were injected intra\enously Avith_ 100 reacting tmits of B. typhosus "agar Avashings" fUtrate. Many sites de\eloped tyj)ical reactions fotu" horns after the intravenous injection of the react- ing factors. Exceptions ^vere noticed in areas showing extensive inflammation and edema. As may be seen, the repeated intra- dermal injections of bacterial fdtrates alone in mixttires with \arious materials, ^vhich may be expected to delay the absorp- tion of the toxic fdtrates, failed to produce reactions typical of the phenomenon wdien no provocative intravenous injection Avas made. Futhermore, in these experiments, as in many others, it Avas suggestive that extensive primary inflammation is inhibitory to the elicitation of the phenomenon. In attempts to produce the phenomenon by repeated intrader- mal injections the interval of time between the injections was also varied (t^vo, fotn, six, t^venty-four, and forty-eight hotns) . The results were consistently negative. In other experiments sites prepared by intradermal injection of B. typhosus "agar wash- ings" filtrates tAventy-four horns later were reinjected with sus- pensions of monkey poliomyelitis virus, Shope fibroma and Shope jDapilloma sus}5ensions, Streptococcus hemolyticus broth cidtine filtrates, etc. In no instance Avere there observed any reactions similar to those of the phenomenon of local skin reactivity. Frisch, H. Gross, and Gratia and Linz failed to elicit any reac- tions Avhen pro\'ocative injections Avere given sid)cutaneously, in- tracutaneously, and intramusctdarly. Frisch also fed large amoinits of B. typJiosus filtrates by m(nith Avithout success. Frisch, H. Gross, (ic)l\ia, b, c) , Gratia and Linz (1932c) , and P. Bordet (iQ.S^i/^) , however, succeeded in reproducing the j)henomenon Avhen the jjro\'ocati\'e injection Avas gi\en intraperitoneall) in large doses. In the experiments of Frisch, four hours after the intraperitoneal in- jections had been made, blue discolorations ajjj^eared at the site of the previous skin injections in 4 otit of 12 animals tested. The dis- 178 LOCAL TLSSUE REACTIVITY colorations observed rapidly increased until in about six hours the reactions became extremely pronounced, the areas being dark blue in the central portion with deep red zones in the periphery. A repetition of tlie toregoing intraperitoneal experiment, this time using 12 rabbits and allowing only eighteen hours lor a skin incubation period, gave positive residts in 6 rabbits. Another repetition of the experiment allowing a skin incubation period of twenty hours and using ■r] c.c. of the filtrate, j^er kilo of body weight, gave a positive reaction in 1 out of (S rabbits so treated. This aiuhor took adxantage of his observation in order to study local j^eritoneal immunity. Rabbits in which the intraperitoneal injection of B. lyjjho.su.s (ulture filtrate Avas able to elicit the typical phenomenon in j)re\iously prepared skin sites ^vere ren- dered negative to this ])hen()men()n by repeated injections of the cidtme fiftrates intraperitoneal ly. It was then siiown that the skin reactit^n coidd still l)e jjroduced if the reacting factors were intro- duced intravenously. Frisch concluded that the immunity pro- duced tmder these circimistances was of a distinctly local char- acter involving only the peritoneimi and that the entire organism was not yet immune. He fbiuid that betAveen three and four in- jections of B. tyl)]iosi(s ctdture filtrates Avere necessary to render the peritonetmi immune to this filtrate. The question of specificity of this inmiunity was left open. It is possible that the preliminary treatment of the peritoneum by repeated injections of the filtrate called forth a non-specific response on the part of the cellular ele- ments of the peritonetim which offer a histologic barrier to alj- sorption of the reacting factors. This type of non-specific imnumity was clearly demonstrated by the work of Gay (i92()) , and Mallory and Marble (1925) , on mobilization of monocytes in acute infections of serous cavities. I attempted to determine whether intracranial injections of the provocative factors are capal)le of eliciting the phenomenon of local skin reactivity. In preliminary experiments it was found that when meningococcus "agar washings" filtrates were injected intracerebrally, they consistently killed rabbits in dilutions as high as 1:160 whilst the B. typhosus "agar washings" filtrates seemed to be of lesser toxicity. The skin of rabbits was prepared in the tisual manner with rrieningococcus or B. fypJiosus "agar washings" filtrates and intracranial injections of the respective filtrates were made tvventy-four, forty-eight, seventy-twc^, ninety- six, and one hundred and twenty hours later. The fdtrates were ROLE OF INFLAMMATION 179 able to elicit se\ere reactions in rabbits prepared witii B. lyljliosus filtrate provided the inter\ al ol time between the intradermal and intracranial injections ^vas not longer than forty-eight hours. In- terxals of time as long as se\'enty-t\vo and ninety-six hours between the preparatory and intracranial injections yielded positive re- sults in sites prejxned with meningococctis culture filtrates. This result could well be expected because, as shown before, menin- gococcus acti\e principles give a state of reactivity lasting longer than B. typhosus actixe principles. Considering the duration of the reactivity induced and the end-point titrations, it may be concluded that the intracranial route is just as effective as the intravascular one for the introduction of reacting factors. Similar conclusions were drawn from recent experiments by Alechinsky (1935, 1936) . Inasmuch as the intracranial and intraperitoneal routes may l^e used effectively, for introduction of reacting fac- tors, while the subcutaneous, intraoral and intramuscular routes are ineffective, it is suggestive that it is essential that these factors enter into the general circulation promptly. The mechanism through xvhich the blood stream participates in the elicitation of the phenomenon of local skin reactivity re- mains unexplained as yet. The possil^le explanations are as fol- loAvs: 1. The active principles introduced into the blood stream lead to formation of new toxic principles in the blood stream capable of production of the injury described. 2. The active prin- ciples of the phenomenon are capable of eliciting the reaction only by a direct effect upon endothelial elements of the blood vessels. Gratia and Linz (1932c) attempted to determine whether an "angeotoxin" is produced as a residt of the intravenous injection of the active principles. For the demonstration of this hypotheti- cal "angeotoxin" they conceived the folloxving experiment: Three-tenths c.c. of testicular vaccine virus were injected into the rabbit's testicle. Forty-eight hours later, the inoculated tes- ticle was swollen and turgid. The rabbit was bled to death and after coagulation, the blood was centrifuged and the serum re- moved. The inoculated testicle was removed aseptically and ground to a fine J^ulp in a mortar. Two c.c. of the serum of the inoculated rabbit and 0.2 c.c. of B. coU culture filtrate were mixed. The mixture was incubated for fifteen miniues. After three hours incubation, at 37° c, the material was injected intra- cutaneously into txvo young rabbits. Three days later, there was l8o LOCAL TLSSUE REACTIVITY a distinct hemorrhagic reaction. Cultures of the fluid of the edem- atotis tissues ^\ere sterile. It is ohxious that these authors must have dealt with a xaccinia \irus lesion, the severity of which was probably enhanced by the presence of the testicidar tissue (Rey- n^ils factors) and by the possible primary injinious effect of the B. coll cultme filtrate in the mixture. The appearance of the hemorrhage three days after the intradermal injection observed in severe vaccinia lesions of the dermis make it impossible to con- sider these ol)servations as having any relationship whatsoever to the phenomenon of local skin reactivity. The possible formation of new princij)les residting from the provocative injection, of coinse, is not excluded by the failure of its demonstration. There is some other suggestive evidence in favor of this hypothesis which will be discussed in a sid^sequent chapter. The other possibility still remains, however, that the direct effect of the active prin- ciples upon the endothelial elements of the l^lood vessels made reactixe by the intradermal preparatory injection, may play an important role in the mechanism of the phenomenon. Bock (1932) made a very interesting observation xvhich needs further in\'estigation. According to this author, of 7 rabbits pre- pared with bacterial filtrates and subse([uently injected into the left heart, 2 showed a very xveak reaction and 5 were negative. Of 7 rabbits injected into the right heart, 2 showed a weak re- action and 5 showed a severe positive reaction. This author notes that the possible explanation of this obserxation lies in the fact that the injections made into the right heart pass first through the lungs. He questions Avhether the limgs play any activating role in the phenomenon of local skin reactixity to bacterial fil- trates. H. Gross (1931^, b, c) attempted a reverse order of injections, i.e., preparatory injections xvere given intraxenotisly, subcutane- otisly, intramuscularly and intraperitoneally. These injections were follox\ed twenty-four hours later by an intracutaneous injection of the active principles. His restdts xvere consistently negative. In my own experiments 25 units of B. typJiosiis reacting factors were injected intravenotisly in some groups of rabbits, and in others, the same amount xvas injected intracranially. Immediately afterwards, and six hours later, the rabbits received an intrader- mal injection of the potent material. Intravenous or intracranial injections given simultaneoidy xvith the intradermal injections yielded reactions in a small percentage of rabbits. As xvill be ROLE OF INFLAMMATION l8l described later, siimiltaneous intra\en()u,s and intradermal injec- tions may bring about inhibition ol the phenomenon (Ogata, H)']6) (p. 337) . The reactions occurred irregularly and alter a prolonged period of time, i.e., approximately twenty hours in- stead of the customary loin- hours. The explanation suggested is that the pro\ocati\e factors persisting in the blood stream pro- duce the reaction after the lapse of the necessary incidjation period for elicitation of the state of local reactixity, this being responsible for the delayed reaction. The inhil)ition of the phe- nomenon described by Ogata requires the employment of an amount of preparatory factors smaller than that employed in the experiments just described, and, therefore, the residts do not offer a discrepancy. The incubation period necessary for the elicitation of the state of reactivity and the diuation of the state of reactivity induced has not received any detailed investigation as yet. As noted in the introductory chapter, the optimum interval between the txvo injections is txventy-four hours. The state of reactivity may be induced, hoxve\er, in shorter intervals of time as well, although more irregularlv than after an interval of twenty-four hours. As a rule, inter\als less than six hours are inadecpiate, although pos- sibly, exceptions may be noted. Thus, Gratia and Linz (1932c) found that occasional reactions may also be obtained after a six hour incubation j^eriod. The duration of local skin reactivity to bacterial fdtrates de- pends a great deal on the amoiuit of the material used for the preparation and also on the potency of the preparation. The state of reactivity induced by B. tyljJwsus tryptic digest broth cul- ture and plain broth cultures as weW as filtrates partially destroyed by exposures to certain temperatiues, elicit a state of reactivity wliich ordinarily does not last more than forty-eight hours. Skin sites prepared -with B. Ixphusiis "agar washings" hltrates ordinar- ily of higher potency than broth culture hltrates, retain their reactivity for seventy-t^vo hours. The longest state of reactivity is obtained by employing meningococcus "agar washings" filtrates. It may last for a period as long as ninety-six hoins and occasion- ally for one hundred and tAventy hours. The duration of reactiv- ity may be correlated, to a fair degree of accuracy, Avith the j^henomenon-producing potency of the filtrate as determined bv titrations described in Chapter i, and therefore, studies on the l82 LOCAL TLSSUE REACTIVLFY duration ol icadixity may oiler an additional method lor (juan- titati\e estimation ol the pcjtency ol the a( live bacterial jjriiu ijjles. FAIK OK ACmVI', l'RINC:iI'I.KS INJECTKD INK) rHI', BLOOD SrKl'AM In sexeial groups ol rabbits large amounts ol Ji. /v/;//o,s(/,s ac- tive ])rin( iples (2000 to '^000 reacting iniits per kilo ol l)ody weight) were injected intraxenously. The rabbits Avere bled Irom the heart at various intervals ot time lolhnving this injection. Serum and delibrinated blood obtained Irom rabbits injected with the active princij)les t\\'o, five, txventy-foin-, and forty-eight hours prior to the bleeding were consistently negative for skin- preparatory and reacting factors. A large amount of active prin- ciples was still present, however, in the serum of rabbits bled immediately after the injection of 2000 to 3000 reacting iniits per kilo of body weight. The serum diluted as high as 1:30 was capable of eliciting typical reactions in rabbits prepared with inidiluted B. tyjjJiosus "agar washings" filtrate. The same serimi undiluted xvas capable of eliciting the state of reactivity in rab- bits tested l3y provocative injection of 50 reacting tuiits of B. ty- phosus "agar washings" filtrates. In some experiments the un- diluted serimi containing the B. typJiosus active principles xvas mixed with a neutralizing anti-typhoid horse serum (H. 196) diluted 1:10, 1:20, 1:30, and 1:50. The immiuie serum diluted as high as 1:30 was capalole of neutralizing completely the active principles of the rabbit serum but failed to do so in dilution 1:50. It coidd be concluded from this that the active principles of B. typhusiis injected into the l)lood stream inay be recovered by an immediate heart bleeding. It is also of interest that the B. typhosus active principles thus obtained are apparently identical ^vith those introduced into the blood stream inasmuch as they are specifically neutralized by an anti-typhoid imminie serimi. The prompt disappearance of the active principles of the phenomenon from the blood stream suggested a search for these factors in the various organs. After various intervals of time fol- lowing the intravenous injection of a large amount of active prin- ciples, rabbits were killed and various organs removed under sterile precautions. The organs were triturated in 0.85 per cent NaCl solution, suspended in the saline solution in the ])ro})ortion of 1 gm. to 1 c.c, and cenrifuged at a low speed. The supernatant fluid was used for the experiments on the day of its prepara- tion. Spleen and liver extracts, diluted 1:4, of rabbits injected ROLE OF INFLAMMATION 1 83 A\ itli 200 reacting units of meningococcus and removed one-half Iionr after the injection, were capable of eliciting reactions in prepared rabbits. Higher dilutions of the extracts were lacking in skin-preparatory and reacting potencies. TesticiUar and kidney extracts yielded entirely negative residts. Apparently, a small amount of the active principles injected into the blood stream may be reco\ered from some organs. Titrations of the reacting potency of the organ extracts thus obtained suggested, ho^vever, that the amoiuit present may be that of the blood in the organs remoxed and certainly do not indicate a concentration of the actixe principles in any of the organs tested. Plant failed to elicit the phenomenon xvith saline extracts of rabbits' testicles infected with Spirodieta jjdllidd and also xvith the juice of the infected testicles obtained by means of a press. Stolyhxvo (1935, 1936) performed the following experiment: Eight rabbits received intravenously 3 to 5 c.c. of three different batches of filtrates of a six day old B. typlwsus l^roth culture. Three rabbits died xvithin the following ten to twenty-fom^ horns. Inci- sions xvere made into the bladder and the mine collected. LIrine cultures were sterile. Spontaneously excreted urine of the surviv- ing rabbits was also collected txventy-four hoins after the intra- venous injection. The mine xvas sterilized by boiling, inasmuch as the active principles of the phenomenon are heat-stabile. The urine samples of the dead and sm'viving rabbits were then in- jected into 2 to 4 rabbit skin sites in amoimts from 0.2 to 0.3 c.c. Txventy-four hours later, xvhen the prepared skin sites shoxs-ed erythema of varying degrees and slight swellings, 3 to 5 c.c. of a six day old B. typJwsus cultme filtrate xvas injected intravenously into the prepared rabbits. Txvo rabbits died. Six surviving rab- bits showed severe hemorrhagic and necrotic lesions in the pre- pared skin sites. Control experiments xvith 6 samples of mine of normal ral)bits gave negative results. The work of Stolyhwo suggests that the active principles in- jected intravenotisly may be excreted by the kidneys. Apparently, hoxvever, this author had to inject enormous amomits of the mate-, rial in order to obtain the elimination in the mine. In Schneier- son's experiments conducted in my laboratories, smaller amomits of active principles of meningococcus and B. typJiosus ("agar washings" fdtrates in doses from 500 to 1000 reacting units per kilogram of body weight) failed to appear in the mine immedi- 1^4 LOCAL ILSSUK REACIIVITY alcly and allci \aiiou.s periods ol time lollouiuo the intravenous injection. Thus, the (question ol the late ol the active princ i|)les of the j)henonienon in the blood stream remains open. Ihe neutraliza- tion ol the leac ting factors by the natural antibodies of the animal injected may be safely excluded, inasnuich as meningococcus re- acting factors for whicii only insignificant amount of antibodies is lound normallv, disappear from the blood stream just as piom]nly as the B. lyl)hosus active principles. Another jjossible explanation is that the active principles may be inactivated by dilution in large amoinits of whole blood. This is, however, ex- cluded by exj)eriments in which the potency of active meningo- coccus and B. Ixphosus "agar washings" liltrates was retained when mixed in \arious pioportioirs with whole dehbiinated rabbit's blood. j)lasma and serum, and stored in the refrigerator for various lengths of time; in some experiments as long as two to three months. Work on the relation of the reticuloendothelial system to the process of elimination of active principles from the blocxl stream suggests itself. RECAPITULATION There is afforded uncjuestionable proof that there exists no re- lationship whatscjever between the inflammation which may fol- low' the intradermal injection of the preparatory factors and the lesion characteristic of the phenomenon of local skin reactivity. This is supported by observations in the gross, extensive histo- logical studies, and by the fact that a large number of non-bacterial substances iDringing about inflammation and injuries of various types are incapable of eliciting the state of reactivity under dis- cussion. Microscopically and in the gross, this phenomenon can be clearly differentiated from the Arthus phenomenon. The essential prerecpiisite of the phenomenon of local skin reactivity is that the provocative factors promptly enter into the blood stream. The phenomenon is best elicited when the provoca- tive injections are given intravenously and intracranially. Intra- peritoneal injections elicit reactions in a certain percentage of animals whilst the subcutaneous, intramuscular and oral admin- istration of the active principles consistently fail to jiroduce re- actions in the prepared skin sites. Histological studies further substantiated the observation that repeated injections of active ROLE OF INFLAMMATION 1 85 principles of the phenomenon into the same skin sites, at various intervals of time, fail to produce the characteristic lesions. Tiie dmation of local skin reactivity to bacterial filtrates de- pends a great deal on the potency of tiie preparation of bacterial filtrates used (witli some preparations the reactivity lasts as long- as one hundred and t\\enty hours) . It may be correlated to a fair degree of accmacy with the phenomenon-producing titer of the fdtrates. The acti\ e principles of the phenomenon injected intrax enously promptly disappear from the blood stream. The question of their fate remains open. Chapter VII METHODS OF ELICITATION OF THE PHE- NOMENON OF LOCAL TISSUE AND ORCxAN REACTIVITY PERIVASCULAR REACTIVITY SHORTLY after the phenomenon of local skin reactivity to I)acterial filtrates Avas described by me, attempts were made to reproduce it in organs other than the skin. In these ex- j^eriments approximately the same technique \vas followed, ^vhereby the preparatory injection of a filtrate was made into the parenchyma of an organ, followed twenty-fotir hours later, and in some instances, at different intervals of time, by an injection of active principles into the general circidation. In 1929, I reported on the phenomenon of reactivity in the lungs of rabbits. A preparatory injection of 0.5 c.c. of B. typliosus "agar 'washings" filtrate ^vas given intratracheally. Twenty-foin' hours later, 1 c.c. of the filtrate was injected into the rabbit ear vein. Large hemorrhagic and necrotic areas appeared through the entire lungs of the animals thus treated. In the gross, the appear- ance of the lesions \vas typical of the phenomenon in the skin. Intratracheal injections alone jjroduced no macroscopic reaction in the rabbit's lungs. Gratia and Linz (1932c) considered it of interest to determine Avhether the phenomenon could be elicited in neoplastic tissues. Liposarcoma in the guinea pig, which ordinarily is avascular and is associated ^vith little sjjontaneous hemorrhage, ^vas selected for the work. Young guinea pigs ^vere inoctdated with this tumor. When, after several weeks, the tumors grew to the size of an egg, an intratumoral injection of 0.3 c.c. of the B. coli active prin- ciples Avas made. T^venty-four hours later, the prepared guinea pigs received a provocative injection of 1 c.c. of the same prin- ciples into the jugular vein or intraperitoneally. Some of the guinea pigs died in a few days following the injection. Post- mortem examination demonstrated an intense iiemorrhagic re- action extending throtigh the major portion of the ttimor, sharply 186 METHODS OF ELICITATION OF PHENOMENON 1 87 demarcated by a lard-like non-affected portion of the tinnor. Some of the guinea pigs siir\i\ed the treatment for se\eral ^veeks. Tiiese animals sho\ved foci of necrosis and even licjiiefaction of the tis- sues Ijut the tumor actually proliferated at the perijjhery of the necrosed site. Bet^veen the necrosed and the normal portions there \vas also observed a zone of hemorrhagic reaction in ^vhich the histological examination demonstrated an active vascular proliferation. The authors concluded from the experiments that the phenomenon of local skin reactivity could be also produced in the sarcoma of the guinea pig. Appelmans and Vassiliadis (1931) attempted to elicit the phe- nomenon of local reactivity in the gastric mucosa and in the peritoneinn. They used filtrates of cultures of B. colt and B. typliosus. The intravenous injections were made t^venty-foiu" hours folloAving the preparatory injections into the gastric mu- cosa, and visceral and parietal peritoneum. The same rabbits were also injected into the dermis of the ear. The injections elicited reactions in the ear but failed to produce any lesions in the serous membranes and in the gastric mucosa. These authors assumed that the gastric mucosa and peritoneum Avere not sus- ceptible to the phenomenon under consideration. It Avas sho^vn previously that there exist strict quantitative reciprocal relation- ships between the skin-preparatory and provocative doses. If sev- eral areas of the skin are prepared simultaneously and an insufficient amotuit of toxic material is subsequently injected intra\enously, only some of these areas ^vill react to the provoca- tive injection, i.e., apparently the most susceptible ones. If the amount of toxin injected intravenously is very large, uniform or severe reactions may be obtained in all the prepared areas. It is quite possible that the aiuoiuit of the filtrate injected intrave- nously in the experiments of Appelmans and \^assiliadis was not sufficient for the production of reactions in all the sites prepared, I.e.. dermis of the ear, gastric mucosa, visceral and parietal peri- toneum. It seems that in experiments of this type in ^vhich the susceptibility of the prepared organ and the dosage necessary are unknown, it may be important to employ only one site for the preparation; or when several sites are employed, it is advisable to inject a filtrate of ascertained high preparatory potency. The negative findings of Appelmans and Wassiliadis are contrasted by the experiments of subsecjuent in\estigators. Karsner, Ecker and Jackson (1931-32), injected 0.2 c.c. of concentrated B. co'li cul- i88 LOCAL TISSUE reac:tivity tare filtrate (Ecker and Welch (1930) into the gastric mticosa; and in other instances, into the mnscular layer of stomach of rabbits. T^vo-tenths of the same (iltrate was injected intracuta- neoLisly into the lower abdominal region. There were i() rabbits of which 12 received 2^/, c.c. of the same filtrate intravenously twenty-foiu- hours later. Animals were killed at intervals of three and one-half, five, twelve, t\venty-foin% twenty-five, seventy-two hours, and seven days after an intravenous injection. The tissues were fixed in Zenker's fluid and stained with hematoxylin and eosin. There were typical hemorrhagic and necrotic lesions in the stomach ^vall, which ^vere generally parallel to the skin reac- tions. The peak of the reaction occurred in the stomach twenty- four hours after the intravenous injection. The local injection into the stomach jjroduced an exudative inflammation, Avhich after the intra venons injection became more severe and sho^ved hemorrhage. The hemorrhage ^vas obtained only \vith both in- jections, but in the stomach it ^vas not as severe as in the skin. No ulcers ^vere observed. Moritz and Morley (i9^^i-^^2) reproduced the phenomenon of local reactivity in the knee joints of rabbits. They nsed B. coli and B. typliosus culture filtrates (Ecker and Rimington) . The preparatory factors were injected laterally and belo^v the patella with the knee joint flexed. One-tenth c.c. to 0.4 c.c. of the mate- rial was injected. In most animals a simidtaneous skin injection was made. Twenty to thirty hoius later, the intravenons injection of 1 c.c. per kilo of body ^veight, ^vas given to the same animals. There ^vere seven controls which received only the preparatory injection. The animals were killed from one to one hnndred and ninety-two horns follo^ving the intravenous injection. Six of the 10 rabbits tested showed a positive reaction in the synovial mem- branes of the knee joints. The reaction ^vas less severe and oc- curred less frequently in the joints than in the skin. Rabbits prepared simultaneously in the knee and skin and showing severe ctitaneotis reactions did not necessarily have severe synovitis and vice versa. This observation may be explained by the fact that several sites \vere prepared simidtaneously, and that possibly here again an instifficient amoimt of provocative factors was given intravenously to elicit the reaction in all sites prepared. The authors' criteria employed for recognition of the reaction 'were based on the histological sttidies, i.e., vascular damage associated METHODS OF ELICITATION OF PHENOMENON 1 89 ^vith tlironibosis and necrosis of the vessels with exudation of polymorphonuclear leucocytes and hemorrhages. Moritz and Morley failed to produce the phenomenon of re- activity to bacterial filtrates in plenral and peritoneal cavities. I.atteri ( 19^^54) injected 0.5 ex. of B. coli or B. ty})li()sus hltrate into the appendix of rabbits exposed by median laparotomy. T\venty-four honrs later an intravenc^iis injection of one of the above filtrates was made. The controls consisted of 4 rabbits, which either recei\'ed the preparatory injection of the toxic mate- rial and no intra\'enoiis injection, or a preparatory injection of physiological saline and an intravenous injection of the bacterial filtrate. Twenty-foiu" horns after the intravenous injection, a second laparotomy was performed. Microscopically, the appendix ap- peared negative. The incision was closed and the rabbits were killed tAventy-foiu' hoins after the laparotomy, i.e., forty-eight hours after the intravenous injection. Of 14 rabbits used, 3 were definitely positive and one had a slight local hemorrhagic reaction. The appendices were s^vollen, tingid and intensely hyperemic. Histologically, there was an alteration in the mucosa, submucosa and subserous layers which consisted of intense vascular dilatation, hemorrhage, leucocytic infiltration and thrombosis of the small \essels. No similar lesions were seen in foiu" control experiments. Ka\etzky (1935) reprodticed the phenomenon in rabbit joints by the use of combined intraarticidar and intravenous injections of streptococcus culture filtrates. The changes observed were hy- peremia, edema of periarticidar tissties and exudation into the artictilar cavity. This author draws attention to certain points of similarity bet^veen rheinnatism and the phenomenon in the joints, ^vhich in his opinion, suggest that the mechanism of the johenom- enon may be similar to the pathogenesis of lunnan rheinnatism, namely: the possibility of reproduction of the phenomenon ^vith streptococcus products; vascidar involvement accompanied by comparatively insignificant changes in the soft tissues of joints and in the cartilage; transitory natine of the lesions; and finally, a short incubation period necessary for appearance of the symptoms. According to this author, the pathogenesis of hiunan rheumatism may consist of localization in the joints of bacterial active prin- ciples originating from some distant bacterial focus ^vith subse- quent elicitation of the state of reactivity. The localization may be conditioned by exposure to cold or to some other non-specific stimulation. The lesions may then be produced by the effect of the 1()0 LOCAL TLSSIIE REACH IVLIY artixc })riii( i])lcs c ii( iilal iiij^ in the blood stream u|)oii the reactive joints. Loi and ("-ardia (i9-Vl) c\])ciinicnted witli dcnciAatcd kidneys ol rabbits. The nerxes ol the hiluin were cnt, the (aj)siile stri])ped and the \cssels cnt. Ihrec-tenths ex. ol a j)()tent l)a( terial Idtrate was injected into the kidney. Twenty to tuenty-tour hours later, o.^^ c.c. ol the same liltrate was injected intra\enously. Six rabbits were used lor these ex|)erinients. T\vo died tollowin^ the inlraxenoiis injection and 4 rabbits were killed twenty-lour hours alter the second injection. Postmortem examination showed swell- ing (due to decapsulation) , dilatation and necrosis smrounded by a hemorrhagic zone. The reaction was more severe in animals which died spontaneously. Histologically, there were observed dilatation of the renal ttd)ides folloxving decapsidation, and also necrotic areas, leticocytic and hemorrhagic foci in the siuToiuid- ing parenchyma zones. In the limiina of the dilated vessels the red blood cells were agglutinated (thrombi) and an ischemic infarct was observed in one case. Controls were negative. The authors asstmied that the denervated organs show a high suscepti- bility to the phenomenon of local reactivity to bacterial filtrates. Inasmuch as during the txventy-fotir hoiu" interval betxveen the operation and the intravenotis injection there should be suffi- cient time for complete degeneration of the terminal nerve sup- ply, Loi and Cardia think that the degeneration accompanied by vasoparalysis favors the prodtiction of the phenomenon in resist- ant animals. It does not seem that the number of animals employed in the authors' experiments jtistify a comparative estimation of the sus- ceptibility of normal and operated animals. It is ob\ious, how- ever, that denervation does not interfere with the production of the lesion. Koplik (1936) reported that the state of reactivity can be elicited in the lymph nodes by means of an intralymphatic prep- aration Avith the active principles of the phenomenon. In his ex- periments amounts of 0.1 c.c. to 0.3 c.c. of B. typhosus "agar xvashings" filtrate were injected into the circimiflex abdominal lymphatic in the direction of the femoral vessels. After the filtrate Avas injected into the lymphatic, the skin incision \vas closed \vith silk. Txventy-foiu' hours after the initial injection the animals were given an intravenous dose of 5 to 25 B. typhosus reacting iniits, per kilo of body xveight. The animals were killed loin" to METHODS OF ELICITATION OF PHENOMENON I9I six hours alter the intra\enotis injection and aiitopsied. Control animals received the intralymphatic injection alone. The experi- ments of this author clearly indicated that intravenous injection ot a potent bacterial filtrate elicited strong hemorrhagic reactions typical of the plienomenon in lymph nodes of rabbits prepared by an injection of the filtrate into the afferent lymphatics. Single intralymphatic or single intravenous injections of the fil- trates ^vere ineffective in evoking such changes. The state of local reacti\'ity in the lymph nodes was limited to those nodes which recei\'ed the direct effect of the filtrates. The reactions were simi- lar to those obtained in lymph nodes adjacent to the prepared skin sites (p. 17) . In Gronchi's (1934) experiments, 20 guinea pigs averaging 265 gms. in Aveight, received a preparatory injection of B. coli ctdture filtrate into the left adrenal folloAved by an intraperi- toneal injection of 0.1 c.c. of the same filtrate tAventy-foiir hoins later. Histologically, in some guinea pigs there Avas observed a focal reaction in the adrenals, which consisted of a zone of celhdar necrosis. The red blood cells lost their staining affinities. The non-injected right adrenal shoAved hyperemia, hemorrhage and leucocytic infiltration in the cortex but there Avas no necrosis or proliferation obserxed in the right cortex. In another group of experiments by this author, 15 guinea pigs received one prepara- tory injection of B. coli ctdttne filtrate into the left adrenal and two intracardial injections of the same filtrate twenty-fotu' to forty-eight horns later. Some of the guinea pigs shoAved, macro- scopically, hemorrhages in the adrenals. Microscopically, the cortical layer Avas intensely hyperemic. There Avere numerotis ex- tensive hemorrhagic foci, leucocytic infiltration, mitotic figures in the zona fasciolata, and slight edema and hyperemia of the medtdla. Cassuto (i933«) gave the preparatory injections into the con- jinictival tissue of rabbits. Macroscopically, and histologically, lesions characteristic of the phenomenon of local skin reactivity Avere observed. Accordino- to this author the octdar tissue is more sensitive than the skin for the production of the phenomenon luider discussion. Reitano and Loi (1934) produced the phenomenon Avith B. lypJiosus filtrate in the pancreas of rabbits. The preparatory intra- parenchymal injection of 0.3 c.c. Avas folloAved by the intravenous injection tAventy-four hours Later. The animals Avere killed tAventy- 192 LOCAL TISSUE REACTIVITY lour lioms ahei the intravenous injection. I'Our ol the animals showed heinoiiha^i( neciosis ol the lobes oi the pancreas. Con- trols. \vhi(h were injected into the j^ancreas ^vith saline and re- eeived an intravenous injection ol the B. lyjjhoMi.s liltrate twenty- loin hours later, remained negative. The appearance of the hemorrhages was similar to that ol necrotic pancreatitis in man. Several attemjjts to reproduce the phenomenon of reactivity in the brain consistently failed, ((iratia and Linz, and others) . Plant (19^52) obtained no lesions in the meninges following combined intracisternal and intravenous injections of active principles of Sj)i)<)cliet.a pallida (dialyzed culttne filtrates) . Bcjck (1932) made preparatory subdtnal and intracerebral in- jections of active principles of the phenomenon and then gave provocative intravenous injections after a suitable interval of time. The subdmal preparation alc:)ne and in combination with the provocative injection gave no lesions. The combination of the intracerebral preparatory injection with the intravenous provoca- tive injection of the hltrate elicited paralysis without fail. Some inflammatory reaction follcjwed the intracerebral preparation alone. The paralyzed animals showed no hemorrhage in the gross or micrcjscopical ly . Thus, the ^vork described demonstrates the possibility of elicita- tion of a state of reactivity by means of a preparatory injection of an active bacterial filtrate in the skin and parenchyma of various tissues and organs. While it is necessary that the provocative injection be given by Avay of the blood stream, the state of reac- tivity may be accomplished by means of a local preparation. As clearly seen from Chapter vi, the conspicuous lesions of the phenomenon are thc^se of severe vascular damage. Apparently, the reactivity is elicited by means of a contact of the preparatory factors with the outside ^vall of the small blood vessels for a certain period of time. For convenience, the method of elicitation of reactivity in the skin, by means of intradermal injection; and in internal organs, by means oi intraparenchymal injection, may be termed as "local perivascular preparation." The question re- mains, however, ^vhether a state of reactivity could be elicited by Avay of the vascidar system. The experiments in this directic^n could be divided into those in ^vhich a state of reactivity is in- duced in a given organ by way of its local vascidar supply, i.e., "local vascular preparation"; and those in which a state of reac- tivity is elicited in an organ or organs through the introduction METHODS OF ELICITATION OF PHENOMENON ig^ of tlie preparatory factors into the general circulation ("prepara- tion by way of general circulation") . LOCAL VASCULAR REACTIVITY In 1929, I performed the follo\\ing experiments in collabora- tion ^vith Baehr: Elicitation of renal vascular reactivity : Kidneys Avere exposed f3y median laparotomy. The left renal vein was clamped off and 0.5 c.c. of B. typhosus "agar Avashings" filtrate Avas injected into the left renal artery. The clamp Avas released 5 minutes following the injection. The right kidney received an injection of phenolized saline under the same con- ditions. Twenty-four hours later, B. typhosus "agar washings" filtrate, in a dose of 100 reacting units per kilo of body weight, was injected into the ear vein of the rabbits. TA\enty-four hours after the provocative injection the left kidneys showed extremely severe hemorrhagic and necrotic lesions of the cortex and medulla. No gross lesions were observed in the right kidneys. No systematic histological studies Avere made at the time. It was concluded from the experiments that the preparatory factors are capable of elicit- ing a state of reacti\ ity via the vascular system of the kidnev. Inasmuch as the vascular system of the kidney possesses an ini- usually high permeability which may be altogether different from other organs, it was of interest to determine whether a similar mode of preparation could be successfully used in the rabbit's ear (Shwartzman, 1935c) . Elicitation of vascular reactivity in the rabbit's ear: In this work advantage Avas taken of the fact that the circula- tion of blood could be apparently efficiently stopped in the ear by application of a surgical clamp to the base, inasmuch as Niagara sky blue (a dye of slow diffusibility extensively used by Rous and his co-Avorkers and prepared according to the method of Rous, 1930) injected into the marginal vein of a clamped ear, failed to enter the general circulation. In the first series of experiments, preparatory injections of men- ingococcus and B. typJiosus "agar washings" filtrates were made into the marginal veins of clamped and non-clamped ears and were folloAved l)y injection of reacting factors into the marginal \ein of the prepared or non-prepared ear after various interAals ig4 LOCAL TISSUE REACTIVITY ol tinie. These preparatory injections Tailed. Iiowexer, to elicit the state of reactivity of the phenomenon under discussion. The same doses oixcn intradermally elicited the state in most rabbits tested, llnlortiniately, larger doses ccjuld not be used because of the lethal effect of these preparations. It was asstmied, then, that the j^erivasciilar preparation was an essential prerecjiiisite for elicitation of the state of reactivity. In experiments such as the above, imder conditions of normal resistance, the capillaries may not allc:>^v the diffusion of the preparatory factors into the tissues Avhen given intraxenously. In vie\v of this assimiption, it ^vas decided to accompany the preparatory intra\enous injections of the active principles by varic^us agents known to modify capillary permeability, as follo^vs: Preparatory hitravoious hijcctioiis of Active Prnicijjles Accom- panied by CJiilling. The left ears of rabbits 'were chilled by means of ice fjags applied to both sides for a period of ten minutes prior and following the preparatory intravenous injection of eighty reacting tmits of meningococcus Group III (44B) "agar washings" filtrates. The clamps ^vere kept on for two minutes following the intravenous injecticm. Distinct hyperemia followed the treatment. The injection failed, ho^vever, to elicit the state of reactivity. Preparatory Intravoious Injections of Active Principles in Xylol-Treated Ears. The left ears of rabbits ^vere clamped off at the base and rubbed with xylol until the veins became very prominent. Immediately afterwards, rabbits received 50 reacting units of meningococcus, Group III (44B) "agar washings" fil- trates into the marginal veins of treated ears. The clamps w'ere removed five minutes after the intravenous injections were com- pleted. Twenty-four hours later the ears appeared s^vollen and one was slightly hemorrhagic along the marginal vein. The pro- vocative injection of active filtrate elicited reactions in 2 out of 8 rabbits thus prepared. In one of these, the reaction Avas doidot- ful and in another it represented an enhancement of a primary hemorrhagic lesion follo^ving the preparatory treatment. The re- sults of this experiment, therefore, remain inconclusive. Preparatory Intravenous Injections of Active Principles Com- bined with Etliyl Urettiane, Acetylcholine, Pilocarpine Hydro- cliloride, Atropine, Calcium Gluconate and Guinea Pig Liver Extract. In this series of experiments preparatory injections were made intravenously in combination with the follo^ving sub- METHODS OF ELICITATION OF PHENOMENON 195 stances: 10 per cent solution of ethyl urethane; 1 per cent solu- tion ol calcium gluconate; 3 per cent solution ot atropine; 2 per cent solution of pilocarpine hydrochloride, acetylcholine di- luted 1:500; and guinea pig liver extract in an amount of 0.85 per cent NaCl solution ecjtial to the moist weight of the liver. The ears of rabbits tested were clamped off at the base and the left marginal veins injected with 1 c.c. of a mixture of equal parts of 50 reacting units of meningococctis, Group III (44B) "agar washings" filtrates with each of the above substances in dilutions indicated. The mixtures failed to elicit the state of reactivity. Prepardtory hitravenous Injections of Active Principles in Conih/nafioii luith Histdmine. Histamine dihydrochloride in dilution 1:1000 was used. In one group of experiments the sub- stance ^vas injected into the dermis of clamped and non-clamped ears in the vicinity of marginal veins. The preparatory injection of varying amounts of active principles ^vas given immediately after^vards. The clamps ^vere released fne minutes after the in- travenous injections were completed. In another group, histamine mixed ^vith acti\e principles in various proportions ^vas injected intra\enously into clamped and non-clamped ears. The clamps were also removed fi\e minutes after completion of the intrave- nous injection. Large doses of active principles mixed Avith histamine elicited some primary injury in clamped ears. The state of reactivity failed to appear in both groups of experiments. Preparatory Intravenous Injections of Active Principles Com- bined ivitJi Adrenalin. The marginal veins of clamped ears Avere injected ^vith 1 c.c. of a mixture consisting of equal parts of 50 reacting units of meningococcus. Group III (44B) "agar ^vash- ings" filtrate and adrenalin chloride in dilution 1:1000. The clamps \\ere removed five minutes after the intravenous injec- tions ^vere completed. The mixtures were capable of eliciting se\ere primary hemorrhagic lesions. The primary reactions Avere not augmented by the provocative injection of active principles. Preparatory Intravenous Injections of Active Principles in Combination with Pituitrin. Pituitrin in dilution 1:200 was used. In these experiments there were made preparatory intravenous injections of active principles in mixture with pituitrin or simul- taneously Avith intradermal injections of pituitrin in the vicinity of the marginal \ein. In some groups the ears Avere clamped be- fore the preparatory injection and the clamps removed five min- utes after the injection was completed. Intradermal or intrave- 10)6 LOCAL ILSSLE REACniVITY nous injections ol |)itniliin into ( lainj)e(l ears aic prone to elicit hemonliages. One lioui latei tliey may appear along the injected \ein. Aj)pic)\iinately twenty lour hours later there develops cya- nosis and pronomited hemoi rhai^e Avhich extend along the tribu- taries ol the injected xeins. These reacticjns become more pronounced il accompanied by intravenous injections ol menin- gococcus Cirouj) III (44B) "agar washings" filtrate. The pro- vocative injecticjn of active principles into veins oF ears thus pre})ared brings about insignificant accentuation of primary re- actions. It is obvious that the reactic:)ns may be interpreted as primary damage to blood vessels ^vhich are not related tcj the state of reactix'ity inider discussion. Possibly, similar effects were ob- tained by Marcus (1920-21) and Schmidt-Weyland (i9^^2) . These authors gave repeated injections of adrenalin into rabbits pre- x'iously treated with bacterial tcjxins and, in some instances, they obtained gangrene and thrombosis. It is of interest that in spite of the primary hemorrhagic le- sions obtained Avith some of the mixtures above described, the state of reactivity did not take place. Preparatory Intravenous Injections of Active Principles Accom- panied by Application of Heat. Sausage shaped rubber bags filled Avith Avater of the desired temperature were hrmly applied to both sides of the ear for five minutes and preparatory injections were given into the marginal \ein. In scjme experiments the ear was clamped off at the base simultaneously Avith the application of heat and the clamps Avere removed five minutes after the intra- venous injections ^vere made. The provocative injections ^vere given into the \ein of the same or of the opposite ear. Clamped and non-clamped ears exposed to 45°, 50°, 55° c. for five minutes and injected Avith \arioiis amounts of actixe bacterial filtrates intra\enously showed immediate intense hyper- emia with subsecjuent swelling and moderate hyperemia t^venty- four hours after the treatment. Pro\c)cati\e injections of bacterial filtrates in various amc^unts elicited no secondary reactions in non-clamj)ed ears. The preparatory intraxenous injections of potent preparations given into the veins of ears exposed to 45 \ 50' , 55^ c. and clamped off at the base, were capable of eliciting the state of re- activity of the phenomenon. Reactions following the pro\c)cative injections were intense. Diffuse hemorrhages throughout the en- tire ear were accompanied frecjuently by deep cyanosis. Micro- METHODS OF ELI CITATION OF PHENOMENON 1 97 scopically, there was pronounced thrombosis in the venules. There were also observed petechial hemorrhages in various parts of the ear and sometimes in portions far removed from the site of the injected vein. In the gross, these petechiae closely resem- bled pinpinic spots seen in the skin of hmnan cases of menin- gococcemia. Ears exposed to 45° c. for fi\'e minutes ga\e definite reactions following the provocative injection, provided at least 50 reacting iniits were used for the preparatory intravenous injection. When exposure to 50' c. ^vas combined \vith the preparatory injection of active principles, distinct reactions were seen with as little as 5 iniits, provided the provocative injection was given into the same vein. It is cmious that a preparatory dose of 25 reacting units in ears exposed to 50° c. failed to elicit the state of reac- tivity. This observation is difficult to explain and the experiment shoidd be repeated. Provocatixe injections of acti\e principles into the ears prepared by larger doses {i.e., 50 and 80 reacting units in ears exposed to 50° and 55° c.) gave intense and diffuse reactions. It may l)e concluded from the above tiiat j^reparatory injec- tions of active principles are capal)le of eliciting the state of reac- tivity in the skin of the ear provided they are combined with temporary stasis (clamping) and thermal hyperemia. Preparatory Intravowiis Itijections of Active Principles Com- bined ivitli Testicular Extract. The rabbit testicidar extract em- ployed Avas prepared according to the method described by Duran- Reynals (1933/;) . In one group of experiments the extract was injected into the dermis in the vicinity of marginal veins of clamped and non-clamped ears. Immediately afterwards, a prepara- tory injection of meningococcus and B. typhosus acti\e principles in various doses was made into the marginal vein of the treated ear. The clamps were released fixe minutes after the intravenous in- jection Avas completed. In another group of experiments menin- gococcus and B. typhosus active principles Avere injected in mix- ture Avith \arious amounts of testicidar extract. It Avas concluded from these experiments that mixtines of bac- terial filtrate and testicidar extract are capable of inducing the state of reacti\ity in the rabbit's ear by Avay of the vascular sys- tem. The reactivity is elicited ])ro\ided the circulation is inter- rupted for a fcAv minutes and the amount of filtrate is cjuite small. It is suggestixe that the state may disapj^ear within four hours 198 LOCAL ILSSUE REACTIVITY lollowiii^ ihc j)ic'j)arati()ii. In soiiic cxpei imciits proxocatixe in- jc'clioiis ol small doses ot a liltrate <^i\cn one-liall, one and two liours lollowing the preparatory injections xvith such a mixture, elicited severe and diffuse reactions. It is notewortln that the in- cubation period and the dination ol reactivity may be consider- ably siiorter than that loUowinj^ the intradermal preparatory injection of bacterial filtrates. In the latter, at least eight hours of incubation period are recpiired and it may last as long as ninety-six hoins. Although the possibility of inducing the state of reactivity by way of the vascidar system through a local vascular preparation in combination Avith the testicidar extract is oljvious, the exact conditions of its reproduction shoidd be considered Avith a great deal of reserve. Apparently, there exist individual lluctuations in susceptibility of rabbits vvhich may serve as a source of error. These fluctuations are illustrated by some unexpectedly negative residts. Inasmuch as the state of reactivity induced by way of the vas- cular system with the aid of testictdar extract is of short dination. larger provocative doses are necessary when the interval of time is longer. Thus, positive residts were obtained in animals when as many as 50 reacting units were used for the provocative injec- tion twenty-four hours after preparation. Rabbits yielding" nega- tive results received only 5 reacting units for the provocative injection after the same interval of time. It is difficult to interpret the observation that heat and testicu- lar extract were the only agents which allowed the preparation of the rabbit's ear by way of the vascular system whilst numerous substances employed which are capable of profoundly influencing the capillary permeability, failed to do so. According to Duran- Reynals (1928), McClean (1930), and Favilli (1931), the Reynals factors are capable of producing striking and immediate increase in dermal permeability which lasts for twenty-four hours. The well known rapid spreading of testicular extract in the in- jected site is due to an extreme dilatation of the capillaries and lymph spaces. It is obvious that when an intradermal injection of a bacterial filtrate is made for the purpose of elicitation of the state of reactivity of the phenomenon under discussion, a peri- vascular depot of the injected material is formed which comes into continuous contact with the cells and blood vessels of the site injected. The fact that an incubation period of at least eight METHODS OF ELICITATION OF PHENOMENON 1 gg hours is recjuiiecl lor tlie elicitation ol the state of reactivity may be due to rehuive impernieal)ility ol the cells which does not permit a rapid entrance ot the preparatory factors into them. The role of the testicular extract and heat accompanying the intra- vascular pre{)aration may, then, be two-fold; i.e., to allow the passage of the injected substances into the siuroinidin^ tissues through a rapid increase in the capillary permeability; and to enhance the cell permeability. A shorter incubation period is, therefore, sufficient for the preparation ^vith mixtines of the pre- paratory factors ^vith testicular extract injected intravencjusly than when the preparatory factors alone are injected intradermal ly. As ])ointed out before, in all probability the state of reactivity takes place in the elements of the vascular wall. It may be as- sumed, therefore, that the natmal relative impermeability of the cells of this wall does not attain the state of reactix ity unless the preparatory factors remain in close contact ^vith them for a cer- tain length of time. Under ordinary conditions of intravascidar preparation, the contact of the preparatory factors ^vith the \ascidar \\all may last cjnly as long as the circidation is stopped and jjrobably may cease when the circulation is re-established. If, ho^vever, a state of enhanced permeability is induced by means of the testicular extract, a diffusion of the factors is allowed from the \ascidar channel or possibly, the direct contact of even short duration of the preparatory factors ^vith the permeable cells is sufficient tC3 induce the state of reactivity (Sh^\■artzman i9^^5c) . Alechinsky (ic);^^',, 19.^^6) made intravenous injections into ears ligated at the fjase. The ligature ^vas removed 6 hours later. Hemorrhagic reacticjns were obtained in the prepared ears when a provocative injection of active principles was given sid3cuta- neously, intramuscularly, intraperitoneally or intracranially. No reactions appeared in ears ligated for the same period of time but receiving no preparatory injections. It is obvious that in the experiments of this aiuhor ligation for a period as long as 6 hours must have raised considerably the local \ascular permea- bility and thus, similarly to my experiments, there Avas possible the passage of the preparatory fact(jrs from the capillaries into the perivascular tissue. \V^orthy of further study are the observations of Alechinsky that reacting factors introduced sidocutaneously or intramuscularly are capable of eliciting reactions in ears pre- pared by way of the local vascular system. As was sho^vn before. 2()() LOCAL TISSUE REACmTTY llicsc routes arc imsiiilahlc lor cli(itation ol reactions in sites j)re|)are(l l)\ iiitiadernial injections. KI.ICrrAllOX OF RI'.ACTIVIT^ HV WA^ oi gknkrai. circmlaiion As already mentioned, attempts liave also been made to re- produce a state ol \vide-spread vascular reactivity by giving the preparator) injection of active princij^les into the general cir- culation ol the animal lollowed by a second provocative injection of the same principles after a suitable interval of time. Before disctissing this topic, it is necessary to cc:)nsider first, the effect of a single intravenous injection of the acti\e {)rincij)les of the phe- nomenon upon internal organs. Effect of Single Intravenous Injections of Bacterial Filtrates: Apitz (i9'54<^<) described certain primary toxic effects of B. coli cultine filtrates upon intravenous injection into rabbits. These may consist of necrosis of the heart muscle, li\'er and spleen, fatty degeneration, enlargement of parenchymatous or- gairs and edema of the liuigs. In Gerber's (iq'^G/^) experiments there ^vere also observed lesions in the internal organs, exclusive of the kidneys follo^ving one injection cjf a bacterial filtrate. The changes 'were as follows: The heart showed severe, cloudy swelling with waxy degenera- tion of the muscle fibers, focal necrosis ^vith or witiiout calcifica- tion and interstitial cellidar infiltration with or ^vithout focal necrosis. The interstitial infiltration ccjnsisted of large monocytes, macrophages, lymphocytes and polymorphonuclear leucocytes. Occasionally, the intima of the pulmonary artery and aorta were similarly invaded. Ho^ve\er, the hearts of ten normal control animals and those receiving horse serum also showed \arying degrees of waxy change of the muscle fibers and interstitial infil- tration. The lesions in the liver consisted of focal necrosis and venous thrombi. The necrosis of the liver cells \aried in size from small foci to large areas invohing an entire lobule. The necrotic cells ^vere often surrounded by polymorphonuclear leu- cocytes and a few macrophages. Although similar foci of necrosis were seen in animals suffering from coccidiosis, the lesions ob- served in animals treated with bacterial filtrates ^vere more ex- tensive and were frequently accc:)mpanied by a greater degree of venous thrombosis than in coccidiosis animals. The thrc^mbi ^vere found chiefly in the central sublobular or large hepatic \eins. METHODS OF ELICITATION OF PHENOMENON 201 Fig. 14. A, low power view of large hepatic vein with many parietal thrombi: hematoxylin and eosin stain. B, high power view (hematoxvlin and cosin stain) ol portion of section shown in A, illustrating structure of a tlironiljus. Wall '. [jarictal \cnoiis iiironil)i in spleen and thrombi also filling srnrounding splenic sinuses; hematox^lin and eosin stain. (Gerber, 19366.) 204 ix)c;al tissue reactivity ^vere obtained Ijy Siet^niuiul, lol lowing the intraxenous injection ol li\e c ultnies ol B. Ixjjho.siis. B. coli and stapliylococ c us. Effect of Repeated Intravenous Injections of Bacterial Fil- trates: The follo^ving authors studied the lesions resulting from re- peated intra\enous injections of acti\'e principles of the phe- nomenon: Gratia and Linz (iq'^sc) injected 0.5 c.c. of B. coli active fil- trate intra\enously into four rabbits. Two of these rabbits re- ceived 1 c.c. of the filtrate intra\'enously twenty-four hours later, xvhilst the remaining txvo rabbits received, during the folloxving day, four injections of 1 c.c. of the filtrate diluted 1:1000, 1:100, 1:10 and of the undiluted material, respectively. On the day following the completion of the experiment, one of the rabbits of the first group and txvo rabbits of the second group, died. Post- mortem examination showed blood in the peritoneal ca\'ity, hemorrhagic lesions in the intestines, multiple petechial hemor- rhages in the kidney, large circular hemorrhagic areas in the lungs, as xvell as hemorrhages in the lymph nodes, thymus, and bone marroxv. Apparently, no histological studies \vere made. Apitz (1934c/) 'was the first to make histological studies on the lesions of internal organs in the phenomenon of reactivity by way of the general circulation. He also deserves the credit for first calling attention to pronounced vascular lesions in the kid- neys. The intra\enous injections of B. coli "agar xvashings" fil- trates xvere given by him at various intervals of time, i.e., txvo injections at t^venty-four hour intervals or several injections at intervals of five to six hours. The main kidney lesions described by him were: (1) cortical necrosis, (2) focal glomerular tubular nephrosis with and without infarction, (3) focal tubular nephro- sis, and (4) glomerular nephrosis. In Gronchi and Carniellis' experiments (1934), guinea pigs received two intracardial injections of 1 c.c. of filtrate of a seventy- two hour old broth cultnre of B. coli. They w^ere killed four, six, and twenty-four hours after the second injection. These authors report histological studies only on the adrenals. The cortical sub- stances shovved intense hyperemia, numerous focal hemorrhages, and leucocytic infiltration, characterized by presence of mitotic figures in the cells. The medullar substance \vas intensely hyper- METHODS OF ELICITATION OF PHENOMENON 205 emic. The appearance of the heniorrhaoe and hy]KMj)lasia suo- oested to them a capillary iragility. In Gerber's experiments animals were divided into two series. One. consisting of 29 rabbits, received two intravenous injections of bacterial filtrate six, twenty-four, or forty-eight hours apart. The second, consisting of 35 rabbits, received an intradermal in- jection of bacterial filtrate six. twenty-four hours or immediately preceding the first intravenous injection; the second intra\enous injection was given twenty-four hotns later. Of the 29 animals of the fust series, 20 (69 per cent) showed venous thrombi in one or more organs.