Wyatt ae ren ra newaninn nan eA one
Reich
maton
Saeceeces
ben Univesyenye ReReSaIRS eens eee Rian Chae ee = baa ss
anicersbeena
*
/ x " : Pi ' 1% ) - un bie a & a ; ua a +a 5% SSS tak Oe ee ay \
t a
Fi ROBIN R. A, Coos.
Digitized by the Internet Archive In 2022 with funding from Kahle/Austin Foundation
https ://archive.org/details/chemicalaspectso0000hgid
THE CHEMICAL ASPECTS OF IMMUNITY
BY H. GIDEON WELLS, Pu.D., M.D.
PROFESSOR OF PATHOLOGY, UNIVERSITY OF CHICAGO, DIRECTOR OF THE OTHO S. A. SPRAGUE MEMORIAL INSTITUTE
e
American Chemical Society
Monograph Series
BO ORD EE A Rel MIN:
The CHEMICAL CATALOG COMPANY, Inc, 19 EAST 24TH STREET, NEW YORK, U. S. A, 1925
COPYRIGHT, 1925, BY The CHEMICAL CATALOG COMPANY, Izc.
All rights reserved
Printed in the United States of America by
J. J. LITTLE AND IVES COMPANY, NEW YORK
GENERAL INTRODUCTION
American Chemical Society Series of Scientific and Technologic Monographs
By arrangement with the Interallied Conference of Pure and Applied Chemistry, which met in London and Brussels in July, 1919, the American Chemical Society was to undertake the pro- duction and publication of Scientific and Technologic Mono- graphs on chemical subjects. At the same time it was agreed that the National Research Council, in codperation with the American Chemical Society and the American Physical Society, should undertake the production and publication of Critical Tables of Chemical and Physical Constants. The American Chemical Society and the National Research Council mutually agreed to care for these two fields of chemical development. The American Chemical Society named as Trustees, to make the necessary arrangements for the publication of the mono- graphs, Charles L. Parsons, Secretary of the American Chemical Society, Washington, D. C.; John E. Teeple, Treasurer of the American Chemical Society, New York City; and Professor Gellert Alleman of Swarthmore College. The Trustees have arranged for the publication of the American Chemical Society series of (a) Scientific and (b) Technologic Monographs by the Chemical Catalog Company of New York City.
The Council, acting through the Committee on National Policy of the American Chemical Society, appointed the editors, named at the close of this introduction, to have charge of securing authors, and of considering critically the manuscripts prepared. The editors of each series will endeavor to select topics which are of current interest and authors who are recognized as author- ities in their respective fields. The list of monographs thus far secured appears in the publisher’s own announcement elsewhere in this volume.
3
4 GENERAL INTRODUCTION
The development of knowledge in all branches of science, and especially in chemistry, has been so rapid during the last fifty years and the fields covered by this development have been so varied that it is difficult for any individual to keep in touch with the progress in branches of science outside his own specialty. In spite of the facilities for the examination of the literature given by Chemical Abstracts and such compendia as Beilstein’s Handbuch der Organischen Chemie, Richter’s Lexikon, Ostwald’s Lehrbuch der Allgemeinen Chemie, Abegg’s and Gmelin-Kraut’s Handbuch der Anorganischen Chemie and the English and French Dictionaries of Chemistry, it often takes a great deal of time to codrdinate the knowledge available upon a single topic. Consequently when men who have spent years in the study of important subjects are willing to codrdinate their knowledge and present it in concise, readable form, they perform a service of the highest value to their fellow chemists.
It was with a clear recognition of the usefulness of reviews of this character that a Committee of the American Chemical Society recommended the publication of the two series of mono- graphs under the auspices of the Society.
Two rather distinct purposes are to be served by these mono- graphs. The first purpose, whose fulfilment will probably render to chemists in general the most important service, is to present the knowledge available upon the chosen topic in a readable form, intelligible to those whose activities may be along a wholly different line. Many chemists fail to realize how closely their investigations may be connected with other work which on the surface appears far afield from their own. These monographs will enable such men to form closer contact with the work of chemists in other lines of research. The second purpose is to promote research in the branch of science covered by the mono- graph, by furnishing a well digested survey of the progress already made in that field and by pointing out directions in which investigation needs to be extended. To facilitate the attainment of this purpose, it is intended to include extended references to the literature, which will enable anyone interested to follow up the subject in more detail. If the literature is so voluminous that a complete bibliography is impracticable, a critical selection will be made of those papers which are most important.
GENERAL INTRODUCTION 5
The publication of these books marks a distinct departure in the policy of the American Chemical Society inasmuch as it is a serious attempt to found an American chemical literature with- out primary regard to commercial considerations. The success of the venture will depend in large part upon the measure of cooperation which can be secured in the preparation of books dealing adequately with topics of general interest; it is earnestly hoped, therefore, that every member of the various organizations in the chemical and allied industries will recognize the impor- tance of the enterprise and take sufficient interest to justify it.
AMERICAN CHEMICAL SOCIETY
BOARD OF EDITORS
Scientific Series:— Technologic Series: — Wiuuam A. Noygs, Editor, Harrison E. Hows, Editor, Gitpert N. Lewis, Water A. SCHMIDT, LAFAYETTE B. MENDEL, F. A. Lippury,
ArTHuR A. NOYES, ArtuHur D. LITTLE, JULIUS STIEGLITZ. FreD C. ZEISBERG,
JoHN JOHNSTON, R. E. WILson.
American Chemical Society
MON OG ReACE To Beha le ras PUBLISHED
Organic Compounds of Mercury. By Frank C. Whitmore. 397 pages. Price $4.50. Industrial Hydrogen. By Hugh S. Taylor. Price $3.50. The Vitamins. By H. C. Sherman and S. L. Smith. 273 pages. Price $4.00. The Chemical Effects of Alpha Particles and Electrons. By Samuel C. Lind. 180 pages. Price $3.00. Zirconium and Its Compounds. By F. P. Venable. Price $2.50. The Properties of Electrically Conducting Systems. By Charles A. Kraus. Price $4.50. The Analysis of Rubber. By John B. Tuttle. Price $2.50. The Origin of Spectra. By Paul D. Foote and F. L. Mohler. Price $4.50. Carotinoids and Related Pigments. By Leroy S, Palmer. Price $4.50. Glue and Gelatin. By Jerome Alexander. Price $3.00. The Chemistry of Leather Manufacture. By John A. Wilson. Price $5.00. Wood Distillation. By L. F. Hawley. Price $3.00. Valence, and the Structure of Atoms and Molecules. By Gilbert N. Lewis. Price $3.00. Organic Arsenical Compounds. By George W. Raiziss and Jos. L. Gavron. Price $7.00. Colloid Chemistry. 3y The Svedberg. Price $3.00. Solubility. By Joel H. Hildebrand. Price $3.00. Coal Carbonization. By Horace C. Porter. Price $6.00. The Structure of Crystals. By Ralph W. G. Wyckoff. Price $6.00. The Chemistry of Enzyme Actions (Revised Edition). By K. George Falk. Price $3.50.
American Chemical Society
MONOGRAPH SERIES IN PREPARATION Thyroxin.
By E. C. Kendall. The Properties of Silica and Silicates. By Robert B. Sosman. The Corrosion of Alloys. By C. G. Fink. Piezo-Chemistry. By L. H. Adams. Cyanamide. By Joseph M. Braham. Liquid Ammonia as a Solvent. By E. C. Franklin. Shale Oil. By Ralph H. McKee. Aluminothermic Reduction of Metals. By B. D. Saklatwalla. Absorptive Carbon. By N. K. Chaney. Refining of Petroleum. By George A. Burrell, et al. Extraction of Gasoline from Natural Gas. By George A. Burrell. The Animal as a Converter. By H. P. Armsby and C. Robert Moulton. Chemistry of Cellulose. By Harold Hibbert. The Properties of Metallic Substances. By Charles A. Kraus. Photosynthesis. By H. A. Spoehr. Physical and Chemical Properties of Glass. By Geo. W. Morey. The Chemistry of the Treatment of Water and Sewage. By A. M. Buswell. The Chemistry of Wheat Flour. By C. H. Bailey. The Rare Gases of the Atmosphere. By Richard B. Moore. The Manufacture of Sulfuric Acid. By Andrew M. Fairlie. Equilibrium in Aqueous Solutions of Soluble Salts. By Walter C. Blasdale. The Biochemistry and the Biological Réle of the Amino Acids. By H. H. Mitchell and T. S. Hamilton. Protective Metallic Coatings. By Henry S. Rawdon. Soluble Silicates in Industry. By James G. Vail. Organic Derivatives of Antimony. By Walter G. Christiansen. The Industrial Development of Searles Lake Brines with Equilibrium Data. By John E. Teeple, et al. The Chemistry of Wood. By L. F. Hawley and Louis E. Wise. Sizes, Adhesives and Cements. By S. S. Sadtler and G,. C. Lathrop.
To the memory of my friends and colleagues
Howarp Taytor RICKETTS RicHArD WEIL
Two pioneer American investigators in the field of immunology, who gave their lives, one in the service of science, the other in the service of his country. Through their early death, the progress of human knowledge and the welfare of their fellow men suffered an immeasurable loss.
PREFACEK
Originally the reactions of immunity were studied with the purpose of solving urgent problems concerning the cure, diagnosis and preven- tion of disease. After a time there came to be a growing recognition of their importance as general biological phenomena not exclusively ‘concerned with disease. For the most part their chemical significance was less appreciated, largely because they were observed as reactions to bacteria, blood serum and cells, all of which are such complex mixtures of unknown constitution that any chemical consideration of their behavior is entirely impossible. Perhaps the hypothetical presentation of the subject in the terms of the Ehrlich nomenclature, with pictorial conceptions which had no chemical significance, had some influence in satisfying many investigators that they understood the principles when they merely understood the hypothesis. As Dean has said in this connection, “Ignorance, however aptly veiled in an attractive terminology, remains ignorance.” The early recognition by Bordet, of the similarity of the reactions of immunity to the reactions of colloid chemistry, probably failed to impress the rank and file of investigators in immunology because they were, at that date, unable to appreciate the significance of the colloid-chemical viewpoint, for lack of knowledge of this new field of chemistry.
In course of time, however, it began to be appreciated more and more that these reactions of immunity are important, not merely for their application to medical practice, but as general biological phenomena and as processes of biological and colloidal chemistry; therefore there has of late been more and more consideration of immunology from these standpoints. The progress towards an understanding of the funda- mental principles has been slow, because in the study of the immu- nological processes we must have on one side of the equation a living animal and on the other the most complex of all known chemical compounds, the proteins or closely related colloidal materials. The only way in which we can simplify the equation is by using purified proteins, preferably those of as well-known composition as possible, in place of such hopelessly complex mixtures as bacteria or blood serum,
2)
10 PREFACE
and even then we have not obtained any very simple component. Undoubtedly knowledge will grow with the progress of colloidal chemistry, and such studies as those of Jacques Loeb on the behavior of protein solutions bid fair to throw more light into the knowledge of immunity than most of the direct investigations of immunological problems.
It would seem to be fitting, therefore, that a series of monographs covering modern chemistry include a discussion of immunity from the chemical standpoint, despite the evident lack of maturity of this new field of chemistry. The critical reader will appreciate that the chemistry of immunology has so far had but a fragmentary and illogical development—here, the work of an immunologist struggling with agents of unknown composition and measuring results with a yard stick of most uncertain accuracy; there, the efforts of a physical chemist apply- ing methods of great accuracy to materials of uncertain nature and to reactions modified by an infinity of unknowable variables. No one can expect that from data so derived, in a new science in which the con- tributions of tomorrow contradict many of those of yesterday, any clear picture or final statement can be presented. At most, one can consider as much of the evidence as he can digest, present as much as seems necessary to carry the thread of the argument, and hope to convey a fair and impartial impression of how the matter stands now and in. what direction the subject appears to be moving.
Since it is probable that such a presentation will have for its audience a mixed group of chemists and immunologists, their differing requirements have been kept in view. Presumably the chem- ist will wish to know to what extent immunology is a branch of chemistry. Works on immunology will be closed to him by the com- plex and vague terminology that has been developed because the lack of sufficient knowledge of the processes involved permits the use of no more exact terms. For his benefit an introductory chapter has been provided with the intention of reducing, as far as may be, the obstruction offered by this terminology. Also, each chapter is con- cluded with a recapitulation of the contents, which may serve to present the essential facts unobscured by the mass of evidence on which they are based. Probably the chemist who is seeking to familiarize himself with immunology will do best to read the recapitulations first, in order to be able to follow the line of argument in case he wishes for more details than the recapitulation furnishes.
The immunologist is entitled to expect that a monograph on the
PREFACE II
chemical aspects of immunity shall present a digest of the entire subject, so arranged as to constitute not only a summary of the situation, but also to serve as an illustrated guide to the literature dealing with these chemical aspects. He will appreciate that the bulk of the literature is so great that even were it possible to consider it all, the actual mass would render obscure the essentials of the line of development. No attempt has been made to cite all the scattered literature, but such references as are cited will be found to furnish adequate bibliographies to cover practically all contributions of im- portance.
Many colleagues are entitled to my gratitude for their assistance in ‘ the preparation of this monograph. To remove the difficulties inherent in the attempt by one man to cover so wide a field, each chapter has been read over by at least one chemist and one immunologist, and most of them by sevegal of each. It is hoped that through their help the number of actual errors of statement has been kept to a minimum and that serious omissions have not occurred. However, for such errors of commission or omission the final responsibility is mine, and by failing to give individual credit and thanks by name, I shall avoid the appearance of shirking this responsibility, without in the least decreasing my debt to my friends for their generous aid.
August, 1924. He Gaaws
>
wt t oe
en | nee 2 y
“f ese ian a Nod es oe ” i Paorabeation Sari hg ee me ag a Sat, «ia “lune OS : Hem att Be ht we nee dies ay See ee ‘Hans afte oe - fale ete ae VE whe Ag
vied Pa ae: ge yan
ot
¢ Fuca 7a 19 aan |
ats
Reroas = ey Rg ht
,
Vf he an ark z : Ce oh. Syn? | aay joke west mali ame i> Whee ne oO 2 7 =- oe ait 5 f he as! + 4 2
st - ie Hn Be A a i erict boas 7 5 : . — 4 a _ wowed Pa "Vie =a ‘id er a cy <i a) 4 . e Ca
Ss 9 - 7 bos > - a J a , « é _ - Le > 7
Spits iF ;
heh Lee oeas at ree nt saab So? _ KOE hE eve ae’ Bisrigticg gant 0 lle ae 2S ox were ote ae Ge hes ae Hints dt al pin BR aanaha seta jamal os cs i Petia tl 3 a". a Tints Ve ggesk Se Caer panshare | be eh Sere eS fies, OR FS Gar aie Anes BEricwae
ar ae moans? Te
bil
A. a ae tty
wee pee Beare tas MC an
a
2
- hes ’ -_ _ ey = we . > a u
CONTENTS
CHAPTER I INTRODUCTION . Immunological Reactions as Chemical Reactions Definition of Terms Used in Immunology .
II - ANTIGENS Nee The Character is Antigens Can Homologous Proteins be Pnceenice The Effect of Alterations in the Protein Molecule on the Antigenic Property Coagulation ee Cleavage Plasteins . Compound Proteitis as Se Annwene Artificial Compound Proteins Nucleoproteins . Hemoglobin Racemization of Proteins Bacterial Toxins : The Antigenic Capacity of Enzymes ; Tuberculin os : Mushroom Poisons . Lipoids as Antigens . Bacterial Lipoids . y LWCtaey Mie ae lee Lipoids as “Antigens” in Complement Fixation Re- actions Recapitulation
III IMMUNOLOGICAL SPECIFICITY gf comer ts cancalaaty Relation of Immunological to Biological Specificity Distinct Antigens in a Single Species Common Antigens in Unrelated Species Heterogenetic Antibodies 13
14 CONTENTS
CHAPTER
The Chemical Basis of Specificity . The Specificity of Hemoglobin . The Complexity of Proteins The Evolution of the Proteins . Immunological Specificity Is Denendent on Chemical Individuality : Evidence from Vereabie Py ora : 5 Specificity not Dependent on Entire Protein Mclecated Specificity of Blood Proteins : Chemical Differences between Serum Dioteine Bence-Jones and Noel Paton Proteins . Milk Proteins . Egg Proteins . Immunological Behavior 5; Aranclally Modiied hee teins . she ode We me he Pick’s Concepen a Sretacity, Landsteiner’s Observations . Influence of Physical Properties on Specify Non-specific Reactions . nas - Recapitulation
IV Tue Nature oF THE ANTIBODIES
Are There Different Types of Antibodies? . Evidence Favoring Unity of Antibodies Significance of Quantitative Discrepancies Bordet’s Theories . ; Sear ee Objections to the “Unitarian Hypothesis
The Nature and Properties of Antitoxins Are Antitoxins Globulins ?
Physical Properties of Antitoxins_ .
The Nature of the Amboceptors Isolation of Antibodies .
The Site of Antibody Formation .
The Coexistence of Antigen and Ratibodye in tine Blood
Recapitulation ‘
V THe NEUTRALIZATION OF TOXIN BY ANTITOXIN
Definition of Toxins , The Nature of the Toxin- Antitorin Reacuen
102 103 105
109
109 I1O
CHAPTER
VI
VII Tue Lyric Reactions (BAcTERIOLYSIS, CyroLysis, AM-
CONTENTS
Physical Chemistry of the Reaction
Relation to Enzyme Action
Ehrlich’s Theory of Toxin- Aputorin Nenttalicetion Arrhenius’s Critique of the Ehrlich Theory
The Adsorption Theory of Bordet
Recapitulation . ‘at
AGGLUTINATION AND PRECIPITATION REACTIONS .
Normal Agglutinins and Precipitins Agglutinogens Properties of emintains Principles of the Agglutinin Rcuon: The Mechanism of Agglutination . Influence of Electrolytes Resemblance to Colloidal Peacnons: Effect of H-ion Concentration . Alterations Produced by Agglutinins Ionization of Antigens Acid Agglutination
Influence of Surface Tension a Tee ea) eee
tial on Agglutination The Donnan Equilibrium Hemagglutination
THE PRECIPITIN REACTION
The Character of the Precipitate .
The Precipitin
The Mechanism of the Precipita Reraon The Zone Phenomenon Physical Chemistry
Recapitulation
BOCEPTOR-COMPLEMENT REACTIONS)
Cytolysis . asi herr Properties of Rentocenties or Saraitizers Properties of Complement or Alexin Resemblance of Complement to Enzymes . Structure of Complement
eis
15 PAGE III 112 113 114 ig
123
124 124 127, 129 130 131 131 133 134 135 136
137 139 142
146 147 148 148 150 151
157 158 159 161 164 165
16
CHAPTER
CONTENTS
Complement Fixation (Bordet-Gengou Reaction) Scope of the Complement Fixation Reaction . Relation to Other Reactions.
Physical Chemistry The Neisser-Wechsberg PHEROmenon (Complement Deviation’) :
The Abderhalden Reaction:
The Meiostagmin Reaction
The Epiphanin Reaction
Recapitulation
VIII Tue WasseRMANN REACTION AND RELATED REACTIONS
WiItH SYPHILITIC BLOOD
The Nature of the “Antigen” The Nature of the Reacting Agent (Amboceptor a the Serum i es a Be ‘ The Nature of the Resciom! Chemical Changes in the Blood in See Flocculation Reactions : Significance of the ileceulien: Recor The Active Agent in the ie Reactions Recapitulation : Be tee
IX. HyPpersENSITIVENESS—ANAPHYLAXIS—ALLERGY .
General Features : Definition of Anaphylaxis . Nature of the Antigens Nature of the Immune Body Anaphylatoxin Formation and Its erie ie ce laxis ‘ Relation of Anaptviede: to Pema The Significance of Anaphylatoxin Formation The Basis of Anaphylactic Shock Pathological Physiology Chemical Changes 5 Desensitization and Anti- Sanilac ; Anti-anaphylaxis . Anti-sensitization .
CHAPTER
CONTENTS
Concurrence of Antigens
Non-specific Transitory Reduction e Roney
Recapitulation .
X PuHacocytic IMMuNiTty . . . Chemotaxis ‘ Phapocytosis, 9. gk
Results of Phagocytosis . CPpsOUNS) ee ary eas | one Recapitulation . . .
XI Resistance to Non-ANTIGENIC Potsons
Narcotee Poisons). “« <-*. Arsenic Habituation
Defensive Mechanisms . IRECapdiaion) i ois, Gone eet
17 PAGE 218 218 219
224 224 231 232 233 235 238 238 241 242 244
THE CHEMICAL ASPECTS OF IMMUNITY
Chapter I
Introduction
Immunological reactions, the processes by which the living organism defends itself against the chemical attacks of its enemies and so is able to exist in an environment seething with such enemies, are chemical reactions. The reagents involved are substances endowed with active chemical properties, and they are the product of chemical activity of the tissues of the body. In few if any cases do we know the chemical constitution of either the poison of the parasite or the defensive agent of the host, and our knowledge is gained entirely by observing the reac- tion or the effects resulting from the reaction. Therefore the chemistry of immunity is on quite the same plane as the chemistry of the enzymes, and our lack of exact knowledge, coupled with the vitally important nature of the processes concerned, makes all the more stimulating the investigation of the unanswered problems.
Despite the lack of any definite information as to the fundamental principles and agents of the immunological reactions, these reactions have already become of practical value in the study of many problems of chemistry. They permit us to determine the presence of infinitesimal amounts of proteins concealed in mixtures of great complexity, and to ascertain many facts about these proteins which can be determined accurately by no other known chemical means. They permit us to tell from just what species of animal a minute fleck of blood or scrap of meat has come; or whether a sample of meal contains other than the sort of plant material it is supposed to be. They give the protein chemist a method of determining the purity and many other facts con- cerning the preparation with which he is working, and with the expen- diture of but trifling quantities of material. The physiologist may detect, by the methods of the immunologist, in the lymph coming from the
19
20 THE CHEMICAL ASPECTS OF IMMUNITY
thyroid gland the presence of quantities of the specific thyroglobulin that are far too small to be detected by any known analytical or physi- ological method.t These reactions indicate that the chemistry of the blood proteins is different in even closely related species of animals, but that some other proteins, such as the crystalline lens of the eye or the albumin of the egg white, may be nearly or quite identical in widely separated species, these resemblances and differences often being ex- tremely difficult of detection by any other chemical methods nowavailable.
In view of these facts, the presentation of a recapitulation of our knowledge of the chemical aspects of immunity is deemed appropriate in a library of chemical monographs. However, such a presentation meets at the outset considerable difficulty, depending on the fact that immunology, in common with other subdivisions of biological science, has built up its own nomenclature. This nomenclature at once appears as a serious obstacle to the uninitiated reader, but it cannot be evaded. As with all special nomenclatures, the new terms have been devised out of necessity to permit of the expression in a single word or phrase what otherwise would require endless wasteful and tiring circumlocution. Therefore, in a work intended at least in large part to be of use to chemists and biologists not primarily engaged in the study of im- munology, it seems necessary to present as a foreword a brief statement of the fundamental principles and descriptive definitions of the ter- minology employed in this field.?
The following paragraphs, therefore, offer definitions of the terms in most common use, with a brief statement of the fundamental prin- ciples involved.
Antigens.— As the word implies, antigens are substances which, when introduced into the body of an animal under proper conditions, stimulate this animal to produce substances which may combine spe- cifically with these antigens; that is, antigens incite the formation of antibodies. Not all foreign substances, whether poisonous or not, have this capacity to incite the production of antagonistic substances; for example, alcohol, alkaloids, mineral poisons, sugars, are not antigens. Antigens seem always to be large colloidal molecules, and in general all soluble proteins are antigenic, there being some doubt as to whether anything except protein molecules can serve as antigens.
Antibodies.—These are the substances which appear in the blood of the immunized animal and exhibit the property of reacting specifically with the antigen used in immunizing. As antibodies may be found naturally in the blood in greater or less amounts before any immuniza-
INTRODUCTION ‘21
tion has taken place, it is customary to indicate such natural antibodies as normal antibodies, in contrast to the immune antibodies engendered by immunizing. As antibodies have never been isolated in a pure con- dition we have no knowledge as to what they are, and their existence is recognized merely by the effects they produce, just as we recognize the existence and presence of enzymes. The antibodies may be recognized by their numerous different reactions, although as yet we do not know whether these several reactions are produced by one antibody or whether there are as many different sorts of antibodies as there are sorts of reactions by which they may be detected.
The most usual reactions employed in immunological work, and the ‘terms used in discussing them, are the following:
Precipitin Reaction.—I{ the antigen is soluble it will, when added to the blood serum of the immunized animal (antiserum) in proper proportions, lead to the formation of a precipitate. The antibody con- cerned in this reaction is therefore called a precipitin, the antigen is called a precipitinogen.
Agglutinin Reaction.—lf the antigen is not dissolved but is in the form of visible particles, e.g., bacteria, red corpuscles and other cells, these particles will, under the influence of the antiserum, adhere to one another to form flocculi which usually are then precipitated. If the cells are motile, e.g., typhoid bacilli, spermatozoa, they also lose their motility. In this reaction the antibody is called an agglutinin and the antigen is an agglutinogen. Evidently these two reactions (agglutinin and precipitin) are closely related, the differences depending entirely on the size of the colloidal complex serving as antigen, and there is much reason to believe that precipitins and agglutinins are identical.
Toxin-antitoxin Reaction.—lf the antigen is poisonous the antibody may be capable of neutralizing it. A poisonous antigen which engen- ders such a specific immunizing substance is called a toxin, the antibody is an antitoxin. Poisonous substances which are not antigenic, i.e., such substances as morphine or arsenic, which do not incite the forma- tion of specific antibodies, should not be called toxins. Some bacteria secrete or liberate into the fluids about them soluble toxins which are able to engender antitoxins when injected into animals, but most of the pathogenic bacteria do not do so, although when their structure is broken down poisonous materials which are not antigenic are often re- leased. These two types of poisons are distinguished by the names exotoxin, for the true soluble antigenic poison secreted by such bacteria as B. diphtherie and B. tetani; and endotoxin for the intracellular,
22 THE CHEMICAL ASPECTS OF IMMUNITY
non-antigenic poison which is liberated only On disintegration of the bacteria. There are numerous antigenic poisons besides bacterial toxins which engender specific antitoxins on immunization, such as snake venoms and the plant toxins (ricin, abrin, etc.). When the toxins are modified by physical or chemical means so that they lose their toxicity but retain their antigenic capacity to incite antibody formation, they are called toxoids.
Anaphylaxis.—If an antigen is injected into an animal, the animal may, after seven days or more of incubation, show a marked hypersen- sitivity to this antigen, so that a very minute amount may intoxicate it severely or fatally, even if the antigen, e.g., egg white, was entirely non- toxic before the animal had been sensitized by injection of the first or sensitizing .dose. This reaction was designated by Richet as anaphy- laxis, indicating that the condition is the opposite of prophylaxis. This and all other forms of altered reaction to antigens, and even to non- antigenic substances, are sometimes grouped together under the term allergy, meaning altered reactivity. The antibody invoking the ana- phylaxis reaction is usually called anaphylactin, although the term senst- bilisin has also been employed, the antigen being anaphylactogen or sensibilisinogen.
Lysis.—In case the antigen is in particulate or cellular form, the action upon it of immune serum may lead to its solution, the process being called lysis, and lysis of special forms of antigens is appropriately indicated by such terms as bacteriolysis, cytolysis, hemolysis, etc. Dis- solved antigens may also be disintegrated by immune serum (proteoly- sis). The antibody concerned is called a lysin, but it is known that in addition to this specific antibody, engendered by immunization with the specific antigen, there is required another agent to complete the lytic action. ‘This second component is present in the normal serum of unimmunized animals as well as in immune serum, and as it com- plements the action of the antibody, or lysin, it has been called comple- ment. The term alexin, meaning protective substance, has also been employed, since it is an important factor in protecting against infection. As the immune antibody or lysin involved in this type of reaction has been pictured as uniting the complement to the antigen it has been called an intermediary body; and also, because it was supposed to have an affinity or receptor for the antigen and one for the complement, Ehr- lich christened it amboceptor, Other terms have also been used, but the above are most generally employed.
Numerous other phenomena of immunity, which introduce still other
INTRODUCTION 23
terms, will be discussed in their proper place, but the foregoing state- ments cover the best known reactions and most used terms, and hence will suffice for an introduction.
REFERENCES
1 Hektoen, Carlson and Schulhof, Jour. Amer. Med. Assoc., 1923 (81), 86.
2The principles and known facts of immunity will be found adequately presented in Zinsser’s ‘Infection and Resistance,’’ Macmillan; and in Kolmer’s ‘Infection, Immunity and Specific Therapy,’ W. B. Saunders. The former is to be recommended especially for its discussion of principles, the latter for details as to methods. Those wishing to secure a working knowledge of immunology, and desiring statements more concise than are to be found in such complete treatises, are referred to Karsner and Ecker, “The Prin- ciples of Immunology,” Lippincott.
Chapter II
Antigens
By definition an antigen is any substance which, introduced into the tissues or circulating fluids of an animal, causes the appearance in these fluids, sooner or later, of substances which react specifically with the antigenic substance; i.e., specific antibodies. This definition may pos- sibly be a trifle narrow, for we may imagine specific antibodies being produced and remaining within cells, not becoming demonstrable in the circulating fluids. In artificial cultures of tissue cells, also, specific antibodies may be formed in reaction to the presence of antigens when there is no circulating fluid. But practically we cannot well determine that a given substance is antigenic unless its injection into the body of an animal in suitable quantities and under suitable conditions leads to the production of antibodies demonstrable in the blood by the reactions they exhibit in the presence of the specific antigen.t
Sometimes the converse assumption is made, viz., that a substance which reacts with a specific immune serum is an antigen, and although this assumption is usually correct there are exceptions? which make such reactions no final proof of antigenic activity; this matter will re- ceive further consideration later. (See p. 33.)
To exhibit its antigenic function the protein must penetrate beyond the epithelial surfaces of the body, and the assumption is naturally made that the development of antibodies is a defense against the pres- ence within the tissues and fluids of the body of proteins foreign to the organism, for only foreign proteins are ordinarily antigenic. The epithelium of cutaneous, alimentary, and probably also placental sur- faces is an almost perfect barrier to the penetration of foreign pro- teins, as illustrated by the fact that such toxic foreign proteins as those of snake venom are harmless when taken into the alimentary canal. This protective action of surface epithelium is not exhibited by the deeper tissue cells with which an injected foreign protein comes in contact, and the immunological reactions seem to constitute a secondary means of defense (Grosser ).°
24
ANTIGENS 25
THE CHARACTER OF ANTIGENS
For the most part antigenic substances are proteins. At the present time it has not been conclusively established that anything except pro- teins ever exhibit true antigenic activity and lead to the production of specific antibodies. On the other hand it is safe to state that nearly if not every known sort of soluble protein occurring in nature is an- tigenic, if we limit our use of the term protein to those colloidal ag- gregates of amino acids which contain the full quota of amino acids usually found in “complete” proteins.
Of course, if a protein is not soluble in the tissues of an animal it _cannot reach the sites of antibody production and therefore can exhibit no antigenic activity. Hence proteins that have been coagulated by heat commonly exhibit no antigenic activity, but proteins which are not coagulated by boiling (e.g., casein) retain their antigenic activity un- impaired after being subjected to this degree of heat, indicating that it is not the effect of alteration in the protein molecule induced by heat that destroys antigenic activity unless these alterations lead to loss of solubility.* ®
As yet we do not know to what the proteins owe their antigenic ac- tivity. Gelatin, which is not a naturally existing protein, but one de- rived by hydrolysis of the insoluble connective tissue protein, collagen, exhibits no demonstrable antigenic effect. *%*® As gelatin is char- acterized chemically by its lack of aromatic radicals, for it contains no tryptophan or tyrosine and but a small amount of phenylalanine, the in- ference seems warranted that the aromatic radicals of the protein mole- cule are of importance in determining antigenic activity. In support of this idea are the following facts:
(1). Vaughan ® showed that when toxic and non-toxic fractions are formed from proteins by cleavage, the toxic fraction contains the aro- matic radicals, and gelatin does not yield a toxic fraction when sub- jected to such cleavage. If the immunological reactions are defense reactions it is to be expected that the toxic character of the protein or its cleavage product will be a determining factor in the stimulation of antibody formation.
(2). Proteins possessing a full complement of aromatic radicals but deficient in some of the other amino acids commonly found in proteins, exhibit active antigenic properties. Among such proteins are the fol- lowing: Zein, which lacks tryptophan, lysine and glycine, but possesses
26 THE CHEMICAL ASPECTS OF IMMUNITY
a large proportion of tyrosine; gliadin, which lacks lysine ; egg albumin, which lacks glycine; casein, which lacks cystine and glycine.
(3). The protamines, which consist chiefly of complexes of diamino acids with but a small total quantity of a few of the mono-amino acids of proteins, are devoid of antigenic activity.'° '' This fact also suggests that the diamino acids are not of importance in respect to antigenic function, an assumption which is supported by the antigenic activity of hordein from barley, which protein contains no lysine and very little arginine or histidine. As no proteins are known which do not contain either histidine or arginine we cannot prove their lack of essential an- tigenic significance as we can for lysine.
(4). Obermayer and Pick?? and others have found evidence that the aromatic radicals of the proteins may be of importance in determin- ing the specific character of immunological reactions (see Specificity, Chapter iil):
Undoubtedly an important and perhaps essential factor in the anti- genic activity of proteins is their large molecular dimension with the attendant colloidal properties. Whenever the protein molecule is broken down into smaller fragments it loses its antigenic properties pari passu. Zinsser * suggests that antigens must be non-diffusible colloids, which therefore cannot enter the cells to be destroyed therein, so that it be- comes necessary for soluble, extracellular antibodies to be formed which may react with the foreign protein as a step in its destruction. When the foreign molecule is small enough to be taken into the cells by diffusion, antibody formation becomes unnecessary for its destruction, and con- sequently we have no antigens that are not colloidal. In support of this we have the fact that the cleavage products of a protein, even when injected all together, have no antigenic capacity, although when in their. original combined colloidal form they are antigenic, and when artificially reunited into colloidal molecules (called plasteins) the antigenic capac- ity of these cleavage products may be restored.
The antigenic activity, as measured by the amount or rapidity of antibody production, varies much with different proteins. Sometimes this' seems: to depend on the solubility of the antigen in the tissue fluids, as in the case of the vegetable proteins; ** but, in other cases, no such explanation can be found. For example, serum albumin seems to be ‘much less strongly antigenic than the globulins from the same serum.2®
With any given protein the amount of antibody formation does not vary directly with the amount of antigen injected, small doses often producing even larger amounts of antibody than larger doses.1° There
ANTIGENS 27,
is a great individual variation in the antibody production by different animals, even of the same species, in response to the same amount of antigen.
Falk ** found evidence that the pH of the protein solution used for immunizing has some influence on the amount of immunological re- sponse, acid solution (pH = 2.5) generally being more effective than alkaline (pH = 10), but these effects are not constant.
Can Homologous Proteins be Antigenic?
Under ordinary conditions the proteins native to an animal do not stimulate any antibody formation, but apparently under certain condi- ‘tions these proteins may become so altered as to behave as foreign proteins. For example, Doerr *® cites instances in the literature in which persons have given typical anaphylactic reactions to repeated in- jections of human blood or serum, and various authors have claimed that repeated injections of the blood, serum or tissue extracts from one animal into another of the same species has led to the production of antibodies (isoantibodies). However, usually attempts to produce isoantibodies are unsuccessful, and when successful the activity of the antisera is usually low. If antigenic activity depends entirely on par- enteral digestion it is perfectly possible for homologous proteins to. be antigenic, for we find that whenever tissues die they undergo sponta-. neous proteolysis, fibrinous exudates and extravasated blood undergo digestion and absorption, and, according to Abderhalden, all such di- gestion of homologous tissues and proteins leads to the appearance in the blood of an increased capacity to digest the same particular kind of protein, which he attributes to the presence of specific defensive pro- teolytic enzymes (Abwehrfermente). (See Chapter VII.)
When an animal’s proteins are chemically altered they may also behave as foreign proteins and lead to antibody formation. For exam- ple, if the serum of an animal is treated with formaldehyde, nitrous acid, iodin, or numerous other substances which combine with the proteins, this altered serum may serve as an antigen and give rise to the production of specific antibodies when injected into the body of an animal of the same species from which it was obtained.
THe Errect oF ALTERATIONS IN THE PROTEIN MOLECULE ON THE ANTIGENIC PROPERTY
A. Coagulation.—Complete irreversible coagulation of proteins, would, of course, render them non-antigenic, by preventing them from
28 THE CHEMICAL, ASPECTS OF IMMUNTIY
reaching the site of antibody formation. If the coagulation is reversi- ble, however, the redissolved protein is still actively antigenic if in an unchanged condition. Thus, if serum is coagulated by strong alcohol it can be redissolved in water after the alcohol has been removed, and the dissolved serum is antigenic.1® Egg albumin, however, is practi- cally irreversibly coagulated by alcohol, and the coagulated protein is in corresponding degree devoid of antigenic capacity. As a matter of fact the coagulation even of egg albumin by heat is not absolutely irreversible, and to just the extent that a suspension of coagulated egg albumin redissolves it exhibits antigenic activity. I have found that 5 cc. of a suspension of thoroughly washed heat-coagulated egg albumin, which has stood for some months in salt solution, may contain enough redissolved egg albumin to sensitize occasionally, but not always, a guinea pig so that it will give an anaphylactic reaction with egg albumin. As the minimum sensitizing dose of egg albumin is not far from 0.000,000,05 gram, it is evident that the solubility of coagulated egg albumin is extremely low. Experiments with the delicate complement fixation test establish similar values for the amount of free egg albumin dissolved in water in which coagulated egg albumin has been suspended for several weeks or months.
These figures give a striking illustration of the delicacy of the im- munological methods and their value in studying certain problems in protein chemistry. In no other way could such minute amounts of protein be detected in a solution. Furthermore, the immunological tests not only disclose the presence of this minute amount of protein in solution, but specificity tests establish also that it is the same pro- tein that was coagulated by heat which has been redissolved in its original form and not merely a product of hydrolysis of the coagu- lated protein. Chemical tests on relatively large amounts of such fluids might possibly disclose that amino compounds, perhaps even proteins or their cleavage products, are present in the solution, but they could not possibly establish the character of the dissolved molecules as intact molecules of egg albumin, a feat easy of accomplishment by the im- munological tests.
Proteins that are not coagulable by heat do not seem to suffer appre- ciably from boiling, at least under conditions that do not lead to hy- drolysis. However, there seem to be few complete proteins that do not undergo coagulation on boiling in solutions at or near the neutral point. Those that have been investigated are found to retain their antigenic properties after brief boiling, these being: casein,* ® ovo-
ANTIGENS 29
mucoid,” *4 the so-called proteoses of plant seeds,?* mucins and sero- mucoids,?* and beta-nucleoproteins.® Of these, however, only the plant “proteoses” possess strong antigenic properties, although casein is far more active than some authors have maintained.**
Bacterial proteins seem to retain their antigenic capacity after being exposed to strong alcohol or to 1 per cent osmic acid.?®> On the other hand, red corpuscles, or suspensions of their stroma, are reduced in antigenic capacity after treatment with osmic acid.?¢
Ultraviolet light reduces the antigenic activity of proteins,?’ appar- ently affecting them in much the same way as heat, for coagulable pro- teins are converted into an irreversibly precipitated material when ex- .posed under suitable conditions to strong ultraviolet radiation,?*® or, if not coagulated, they form modifications that are readily precipitated by various salts ® and show other changes indicating decreased colloi- dal lability.*° Even sunlight from which infra-red and ultra-violet rays have been removed, causes changes similar to those of heat coagula- tion.** Ultraviolet light acting on meningococci alters somewhat their immunizing properties (Eberson).*?
B. Cleavage.—Free amino acids are not antigenic. In the hydro- lytic cleavage of proteins the power to invoke antibody formation is lost somewhere between the intact stage of the protein molecule and its complete separation into the constituent amino acids, but as yet we do not know just where. Although there have been many contradictory findings reported by various immunologists as to the antigenic capacity of proteoses, peptones and peptids, most of the statements in the litera- ture have evidently been made with little comprehension of the changes involved in protein hydrolysis or of the nature of the mixtures under investigation. Much of the recorded work has been carried out with commercial preparations of protein cleavage products, such as Witte’s “peptone” and the like. These are notoriously variable preparations, concerning the source, composition and purity of which the observer has no knowledge. To call work done with such materials scientific in- vestigation is, in my conviction, a prostitution of the word science, a parody of its standards and ideals. To make matters worse, much of the published evidence as to the antigenic value of protein frag- ments has been obtained with the anaphylaxis reaction. Here the criteria of reaction are furnished by the behavior of an injected animal, and as many of the symptoms of the anaphylaxis reaction (q.v.) may be es- sentially reproduced by intoxication with protein cleavage products
30 THE CHEMICAL ASPECTS OF IMMUNITY
independent of any immunological reaction, the value of the evidence becomes dubious indeed.
A critical review of the literature and a re-investigation of the subject in this laboratory by Fink ** emphasizes the worthlessness of much of the published evidence on this subject. He concluded that positive, corroborated evidence of the production of precipitins and complement fixing antibodies, the presence of which can be determined objectively, had not been presented. Fink himself studied the proteose fractions obtained by hydrolyzing coagulated egg white with steam under pres- sure and did find some evidence of slight antigenic activity by means of complement fixation, precipitin and anaphylaxis reactions in those fractions precipitated by three-fourths and full saturation with am- monium sulfate, but not with fractions precipitated at 144, 4%, and % saturation. Of course we do not know just what is in these several fractions that may be obtained by salting out the products of protein cleavage. It might be expected that the largest molecules would come out in the precipitates first obtained with the smallest con- centration of the salt, and that these should be more like the unhy- drolyzed protein molecules. Fink’s results, which seem to be definite, although slight, are therefore surprising in revealing that the first fractions contain no antigenic molecules, these appearing only in the last fractions. They emphasize our ignorance of the steps by which protein molecules undergo hydrolysis, and suggest another aspect of protein chemistry in which immunological methods may prove of value. There is no reason at present to assume that the so-called proteoses and peptones of various designations represent anything but the crudest of mixtures,** or to expect that constant results will be obtained in any investigation as to their properties, whether immunolo- gical or something else.
If no positive results can be obtained with most of the fractions of protein cleavage, including complete digestion mixtures of various pro- teins containing usually all the cleavage products, from proteoses through peptones and polypeptids to amino acids, it is not easy to ac- cept the statement that anaphylaxis can be produced by synthetic poly- peptids, even one containing 14 molecules of leucine and glycine (Abder- halden),*° still less the positive results of Zunz** with much simpler polypeptids (3-5 glycylglycine). ‘These reactions were obtained after intravenous injections of the peptids into sensitized rabbits, and con- sisted of fall of blood pressure, increased rate of respiration and ex- pulsion of feces and urine, but these reactions, slight as they are, were
ANTIGENS a0
by no means constant, and occurred under very particular conditions, e.g., positive results with animals sensitized with six intraperitoneal injections at weekly intervals, but not in animals that received six sub- cutaneous injections at similar intervals. Certainly, in view of the pos- sible sources of such reactions independent of true anaphylactic shock, and the large number of recorded negative results with peptones and digestion mixtures, the acceptance of the results yet obtained with synthetic polypeptids as proof of their antigenic activity will require much more convincing evidence than any that has yet been produced.
Plasteins——On the other hand, if digestion products are resyn- thesized into intact protein molecules they then may exhibit antigenic properties. At least this is the case if we accept the state- ment that the so-called “plasteins,’ formed supposedly from pro- teoses by synthesis through the reverse action of proteolytic enzymes, are true synthesized proteins. Gay and Robertson *” found that al- though the pepsin digest of casein is not antigenic, the plastein which they call paranuclein and believe to be formed by resynthesis of a pro- tein through the action of pepsin on the casein digest, was capable of sensitizing guinea pigs to itself and to casein; rabbits immunized with paranuclein produce a serum giving complement fixation reactions with paranuclein but not with casein, and precipitin reactions with neither. Landsteiner ** also prepared a “plastein” by the action of rennin on Witte peptone, which produced an immune serum giving precipitin reactions with the plastein as well as with the Witte peptone.
Hermann and Chain *® found that a plastein made from Witte pep- tone engendered a precipitin which reacted with the same plastein, also with plasteins from unrelated sources (proteoses from edestin, serum or egg albumin, almond globulin), but not with the original peptone or protein from which the plastein was derived. A similar lack of spec- ificity was observed with plasteins investigated by v. Knaffl-Lenz and Pick,*° who obtained precipitins but were unable to produce anaphy- laxis, either active or passive, with their plasteins.
COMPOUND PROTEINS AS ANTIGENS
Addition of elements or radicals to antigenic protein molecules seems not to impair their antigenic activity provided the resulting compound is still soluble. Casein, which is a phosphoric acid salt of a protein, is an effective antigen.?* Mucins, which seem to be compounds of proteins with chondroitin sulfuric acid,*! or mucotin sulfuric acid, are anti- genic (Elliott),?* as also are the so-called beta nucleoproteins ° in
22 eds CHEMICAL ASPECTS OF IMM ON TT
which the non-protein radical is guanylic acid. Ovo-vitellin, which, like casein, is a phospho-protein, and ovomucoid, a glycoprotein, are also antigenic.*4
Artificial Compound Proteins
Soluble artificial compounds, such as iodized and diazotized pro- teins * !* are also antigenic, although possibly the specificity (q.v.) may be altered by such chemical changes. There is indeed reason to think that sometimes the entrance of a new radical into the protein molecule may give it new properties that convert it virtually into a foreign pro- tein. At least this is a plausible hypothesis that has been advanced to explain certain cases in which an individual becomes hypersensitive to some simple chemical substance, not antigenic of itself, and reacts to this substance with the allergic phenomena characteristic of anaphy- laxis, because the chemical has combined with some of the body pro- teins and thus formed proteins that are sufficiently foreign in character to incite antibody formation. In support of this conception is the fact that many of the chemicals that produce such specific reactions are chemicals that might readily unite with protein molecules and produce foreign compound proteins, e.g., iodin, arsenic, mercury, formaldehyde, salicylic acid. Landsteiner has shown that rabbit serum treated with formaldehyde, when injected into rabbits, causes the production of an antiserum which will give precipitin reactions with this formaldehyde rabbit serum, and not with the formaldehyde treated sera of other species.*? Positive results have been obtained with the anaphylaxis reaction used as the means of demonstrating antigenic activity.** Simi- lar effects have been obtained with nitrated proteins (Obermayer and Pick). It is to be remembered that formaldehyde is a substance which notoriously leads to a specific hypersensitive state, many pathologists and others who work with formalin exhibiting a most marked sensitivity to formaldehyde. The properties of artificial compound proteins have been investigated by Obermayer and Pick, and especially by Land- steiner *® 44 who found that proteins united to various simple radicals, especially nitrogenous radicals, when injected into animals induce the production of antibodies which give precipitin reactions with any and all sorts of similarly treated proteins, no matter whether the proteins were related in origin to one another or not. For example, the serum of a rabbit immunized with horse serum, the proteins of which have com- bined with metanilic acid (m-aminobenzolsulfonic acid), gives strong precipitin reactions with other metanilic acid protein compounds, even
ANTIGENS 33
when the protein radicals are as unrelated to horse serum proteins and to each other as gliadin, globin, mucin, casein, legumin, silk, and rabbit serum,
Azotized albumoses give but slight precipitin reactions with an anti- serum for azotized proteins, and azotized peptones and amino acids none at all, but such azotized compounds do possess the capacity to unite with the antibodies present in the serum, even when they do not produce precipitin reactions with this serum. Apparently, then, these small, non-colloidal molecules share with the azo-proteins the capacity to react with the immune bodies of the serum, but fail to produce a precipitate. This assumption is supported by the fact that even simpler reactive substances possess the capacity, when in excess, to inhibit the precipitin reaction, just as an excess of antigen does (see p. 148), this inhibition being specific for the compound used in combining with the protein that served as antigen in producing the immune serum.
The explanation advanced is that the inhibiting substances contain a group which is identical with the specific group of the derived protein which reacts with the immune serum. This group can therefore unite with the antibodies of the immune serum quite as well as can the same group when attached to protein, and hence if there is an excess of the simple chemical compound it will bind most of the antibodies and pre- vent the precipitin reaction with the derived protein. Such simple radi- cals are not antigenic in the sense of being able to incite in animals the production of antibodies with which the chemicals react; only when bound in a protein molecule used in immunizing do they have the effect of determining the specificity of the resulting antibodies. They do exhibit one attribute of an antigen, namely, reacting specifically with antibodies, but this quality by itself does not constitute an antigen.*® It requires the large colloidal molecule of the proteins to stimulate the production of specific antibodies—which is the characteristic prop- erty of antigens. That any other large colloidal molecules than pro- tein can serve as antigens when combined with a specific reactive group has not yet been determined, but Landsteiner suggests that such a non- protein antigenic complex is a possibility.
Protein-Free “Antigens”
The foregoing facts may furnish the explanation of certain antigens obtained from bacteria which are said to be protein-free. Pick ** found in young typhoid cultures a material that did not give the ordinary protein reactions, which was resistant to heat and proteolytic enzymes
34 THE CHEMICAL ASPECTS OF IMMUNITY
and soluble in alcohol; it exhibited the power of giving specific precipi- tin reactions with immune serum, but had no capacity to produce anti- bodies when used for immunizing, Zinsser and Parker 4? found in pneumococci, influenza bacilli and staphylococci, a similar substance, free from gross amounts of proteins, precipitated by alcohol, and thermostable, which gave specific precipitin reactions with homologous antiserums but produced no antibodies when injected into animals. Heidelberger and Avery? have isolated from pneumococcus cultures a material which reacts specifically with antiserum for pneumococci, which appears to be a polysaccharide built up of glucose molecules, and is prob- ably related to the gums found in numerous other capsulated bacteria. This material seems to be responsible for the type specificity of the reac- tions between pneumococci and antipneumococcus serum, but it appears to be incapable of stimulating the production of antibodies when used for immunizing, and therefore it is not a true antigen. It seems to be similar to a complex carbohydrate isolated from yeast by Mueller and Tomesik,*”* which reacts with precipitins for yeast although not of itself antigenic. Perlzweig and Steffen ** prepared a pneumococcus antigen which immunizes mice to multiple lethal doses of pneumococci, and which contains only a trace of nitrogen, resists tryptic digestion, is soluble in 90 per cent alcohol but not in ether or absolute alcohol, and is of itself not toxic for mice. Too little is known about these antigens as yet to permit us to be sure that they are actually protein-free, and still less to make assumptions as to their exact nature.
Nucleoproteins
On the other hand, if the protein radical is non-antigenic the addition of other non-antigenic radicals does not produce an antigenic compound, at least as far as now known. This seems to be true for the important group of alpha nucleoproteins, which consist of compounds of nucleic acid with protamines or histones.’° Under the general term ‘“‘nucleopro- tein” has been included a variety of indefinite compounds. Inspection of the literature shows that the material usually has been obtained accord- ing to Woolridge’s method for isolating nucleoproteins, or some modifi- cation thereof. This method is delightfully simple, at least in principle. It consists essentially in an extraction of finely divided tissues with distilled water, filtration, and then acidifying slightly, usually with acetic acid ; a flocculent precipitate is now obtained which is labelled ‘“‘nu- cleoprotein,” used as such, and the results interpreted accordingly. Commonly, to facilitate extraction, the solvent is made slightly alkaline,
ANTIGENS 35
and for obtaining the “nucleoproteins” of bacteria, investigators have often used solutions made strongly enough alkaline to disintegrate the bacterial membranes, i.e., 1 per cent KOH, or even stronger solutions. If further refinement of the material is desired, which has not seemed necessary to many investigators, this is accomplished by redissolving the precipitate with weak alkali and reprecipitating with acetic acid, as often as thought necessary.
The assumption that this precipitate represents pure nucleoprotein may well be questioned. Cells extracted with an alkaline solution would certainly yield a multitude of substances, many of which would be precipitated with acids. Not only would nucleoproteins and nucleins be present in such a solution, but also mucin (which is said to be uni- versally present as an intercellular cement), “nucleoalbumins,” prob- ably various glycoproteins besides the mucins, simple globulins and al- bumins, and alkaline proteinate formed by the action of the alkali upon the native proteins. With slight acidification all these, with the ex- ception of simple albumin, might be precipitated more or less com- pletely, and such a mixture, together, probably, with many other un- determined cell constituents, would constitute the material which many investigators have called “nucleoproteins.”
As for extractions made with distilled water or physiological salt solution, these can contain only such nucleoproteins as are bound to bases, for the free nucleoproteins are insoluble in water or weak salt solutions; and, for the same reason, such an extract will consist of much the same material as the alkaline extracts but in lower concentra- tion. Even if by repeated solution and reprecipitation a purification of the material is sought, it is extremely doubtful if anything like a pure nucleoprotein will be obtained. In the first place, the various sub- stances enumerated above will behave much the same as the nucleopro- teins and accompany them in greater or less amount as long as the purification is continued. Secondly, the action of the acid and alkali will undoubtedly greatly alter the character of the original nucleo- protein, chiefly by denaturizing the proteins so that they become in- soluble, leaving the nucleins and nucleic acid in increasingly large pro- portions.
But most important of all is the question of the nature of the nu- cleoproteins themselves, The substances which have been isolated and designated under this title are, undoubtedly, salts of protein and nucleic acids. As pointed out by Osborne and Harris,*® the nucleic acids are multibasic, e.g., salmon and wheat nucleic acid are 6-basic, so that they
30 THE CHEMICAL ASPECTS. OF IMMUNITY
can unite with from one to six molecules of protein, which might be all either the same or different. Furthermore, as all protein molecules have the ability to unite with several acid radicals, the possible com- plexity is increased. It seems probable, therefore, that in the living cell the nucleic acids must exist bound to protein molecules, but it is highly doubtful that these compounds are the same as those which are precipitated from either neutral or alkaline extracts of the cells or tissues, a fact universally recognized by physiological chemists. With an abundance of proteins of many sorts and conditions present in such extracts and in view of the easy dissociation of the compounds of nu- cleic acid and proteins, it is to be supposed that the nature and pro- portion of the protein which is thrown down with the nucleic acid will depend entirely upon the conditions existing at the time. Variations in the concentration and character of the proteins, in the proportion of nucleic acid, of the concentration of salts and other solutes, of the de- gree of acidity or alkalinity of the solution, and perhaps even of the temperature, will all serve to cause variations in the composition of the precipitate which contains the nucleic acid. If the precipitation is repeated, more and more of the protein becomes insoluble in the form of albuminates, while the resulting material (as shown by Bang) °° becomes richer and richer in phosphorus, until it becomes of the same character as the “nucleins” which are formed by peptic digestion of nucleoproteins or of tissues containing them.
Neither is there any evidence that the nucleoproteins have for their protein elements any special and characteristic sort of proteins. Os- borne and Harris state that nucleic acids may combine with simple albumin and globulins, and that the resulting compounds behave as do the corresponding proteins when combined with any acids.
From these considerations of the nature of nucleoproteins it seems evident that we have to deal with three sorts of substances, as regards immunity reactions, One, the nucleic acid itself, which is non-protein, practically a glucoside in fact; the nucleins, which are compounds of doubtful character, but which seem to consist of nucleic acid bound firmly to proteins, especially to the most basic of the proteins, the histones, and sometimes, perhaps, to protamines; third, the nucleopro- teins, which would seem to be very indefinite and loose compounds of any or all the proteins of the cell with either nucleic acid itself, or with the nucleins.
Obviously, if the isolated nucleoproteins as such are considered, we are dealing with artificial substances which are of most uncertain and
ANTIGENS — 37
doubtful character, probably never alike in any two different prepara- tions, and owing their antigenic character chiefly if not entirely to the abundant and loosely bound proteins. To ascribe to these mixtures any particular cell—or organ—specificity would seem to be preposter- ous, for they must react as do the proteins they contain, in so far as these proteins have not been denaturized by manipulation. That any particular protein is specifically combined with nucleic acid to form nucleoproteins there is no evidence whatever, but, on the contrary, there is evidence that many sorts of proteins may be thus united; un- doubtedly this is the case in the living cell.
Perhaps the nearest thing we have to definite protein-nucleic acid compounds are the protamine and histone nucleinates. On account of the high content of diamino-acids in protamines and histones they are strongly basic, and therefore should be particularly firmly bound by such a polyatomic acid as nucleic acid, and some at least of the nucleins are of this nature. Therefore, the question arises—can a compound of nucleic acid and a histone or a protamine exist, which possesses specific antigenic properties, characteristic of the cell from which it is derived?
Nucleic acid, containing no protein, would not be expected to act as antigen. I have prepared sodium nucleinate from the sperm of the cod, and found that guinea pigs injected with this material were not sensitized to the same nucleinate nor yet to the albumin of the cod sperm.®? A. E. Taylor was unable to obtain any evidence of the forma- tion of a cytolytic immune body by injecting rabbits with nucleic acid prepared from the sperm of salmon.°? Abderhalden and Kashi- wado®? also failed to secure anaphylactic reactions with nucleic acid from thymus nucleoproteins.
As to the histones, I have found that the “Gadus histone” of Kos- sel and Kutscher, prepared from cod sperm, is of itself highly toxic, and that its toxicity is not decreased by heating at 56° for 30 minutes; however, guinea pigs previously injected with this histone showed no increased sensitivity to a second injection. Likewise, Taylor °* found no evidence of the formation of a cytolytic antibody in rabbits im- munized with the protamine of salmon sperm, although immunization with entire sperm produces an antibody.
Schittenhelm and Weichardt ** have called attention to the fact that a toxic histone becomes non-toxic when united with nucleic acid to form nucleohiston, as is also the case when toxic globin is combined in the form of hemoglobin. Gay and Robertson,°®® who found that the pro- tamine salmin is highly toxic and that this toxicity is removed by union
38 THE CHEMICAL ASPECTS OF IMMUNITY
of the salmin with casein, also found that salmin has no antigenic power as indicated by the complement fixation test, and that it did not alter the specificity or antigenic power of casein to which it is bound. Likewise they found a histone, globin, to be non-antigenic, both as re- gards anaphylaxis and complement fixation, and globin caseinate showed no properties other than those of casein, except that globin- caseinate antiserum gives a fixation reaction with globin.®
Although Browning and. 'Wilson *? maintain that globin is antigenic this opinion is not shared by other investigators °* and even Browning and Wilson often failed to secure antibodies, so that they “are forced to the conclusion that globin is not a potent antibody.” It seems prob- able that the antigenic inactivity of globin depends on its lack of solu- bility in the body fluids, by which it is immediately precipitated.
If, then, histones, protamines and nucleic acid are not antigenic, it is not surprising that nucleins composed of these radicals will not be anti- genic, and, according to the best evidence that I can obtain, this is the case. Preparations after Woolridge’s method, but reprecipitated several times, consist practically of nucleins, the albumin being denaturized and removed in insoluble form by the manipulation involved in puri- fication. Such preparations have been repeatedly tested in my labora- tory, and found to be incapable of causing anaphylactic reactions in guinea pigs, but if the original first precipitates, which are rich in pro- teins, are used, strong reactions will be obtained. Even the most care- fully purified preparations will, however, in large doses sensitize to the serum of the animal from which the nucleoprotein is derived, indicating that a certain amount of serum protein or equivalent tissue protein is still present in our preparations.
A series of experiments by Lake ®® has shown that even when nu- cleoproteins or nucleins have lost their power of causing anaphylactic reactions, they may still be able to cause the development of precipitins and complement-fixing antibodies. These antibodies are not, however, specific for the nucleins or nucleoproteins, but react with isolated globu- lin and albumin of the same organ, as well as with the serum of the same species, and without even any quantitative difference in favor of the homologous nucleoprotein.
From these observations it would seem that the antigenic properties of nucleoprotein preparations depend simply upon the proteins which may be present in these preparations, and which are not in any sense a characteristic integral part of a definite substance, nucleoprotein, but rather an adventitious impurity, the character and amount depending
ANTIGENS 39
entirely upon the method used in preparation. Such a conception serves best to harmonize the highly discordant results recorded in the literature.°°
Hemoglobin
Hemoglobin has been extensively investigated as to its antigenic prop- erties. According to our present knowledge of its chemistry, hemo- globin is a salt-like compound of a non-antigenic histone, globin, and a non-protein radical, hematin, which likewise is not antigenic. Schmidt and Bennett * reported that carefully purified hemoglobin is not anti- genic.*t Numerous previous observers, however, had reported that ‘hemoglobin is antigenic, and Hektoen and Schulhof ** © obtained active precipitins on immunizing rabbits with hemoglobin purified according to recognized methods, these precipitins being specific for hemoglobin of the same species ® °* and apparently distinct from the antibodies formed in response to immunization to corpuscle stroma. Conversion of the hemoglobin into carboxyhemoglobin, sulfhydrohemoglobin, or methemoglobin does not affect its specific antigenic properties. Re- peated crystallization or treatment with aluminum cream for purifica- tion does not diminish the proportion of active antigen in the hemo- globin. Hektoen and Schulhof also found that when the hemoglobin is split into hematin and globin by means of acetic acid the globin exhibits no antigenic property, the antigenic element present in hemoglobin as prepared in a supposedly pure condition remaining in solution after removal of the globin. “These experiments indicate that hemoglobin, isolated according to the methods now in use by chemists for this pur- pose, either contains some hitherto unrecognized component which is antigenic and species specific, or that hemoglobin thus prepared has ad- herent or adsorbed to it some antigenic substance which, according to immunological tests, is not present in serum free from red corpuscles.
Heidelberger and Landsteiner,®* however, believe that it is the hemo- globin itself which is the antigenic substance. It may be suggested that the reputed lack of antigenic effect of globin depends on its relative in- solubility in the body fluids, especially after it has been put through the necessary manipulations for isolation and purification. Since hemo- globin is soluble it may demonstrate antigenic effects due to the globin moiety held in solution as a compound with hematin, even if isolated globin is not capable of inciting antibody formation to a marked degree.
40 THE CHEMICAL ASPECTS OF IUMONTTY
RACEMIZATION OF PROTEINS
In experiments reported in 1909 it was found that treatment of crys- tallized egg albumin with hydrochloric acid in such a way as to form the so-called acid albumin, did not destroy its antigenic activity. On the other hand, the formation of alkali albuminate by the conventional methods completely destroyed the antigenic capacity (Wells).7* It was later suggested by Dakin,® that the effect of alkalies on proteins depends on a keto-enol tautomerism of the —CH—-CO— groups, result- ing in a loss of optical activity and a complete loss of the capacity to un- dergo hydrolysis by proteolytic enzymes, or even by putrefactive bac- teria. Ten Broeck® repeated my experiments, using Merck’s egg albumin rather than crystallized egg albumin, and found that the racem- ized mixed proteins of his preparation were also non-antigenic. Racem- ized zein, gliadin, casein and egg albumin, and proteoses from these racemized proteins, were likewise found to be non-antigenic by Land- steiner and Barron ®7 and by Kahn and ‘McNeil.®
The importance of this observation lies in its bearing on the ques- tion of the relation of antigenic activity to proteolysis. One theory of the fundamental principle of immunity reactions is that they all repre- sent par-enteral digestion of foreign proteins that have gained access to the tissues of the body. According to this conception, in an ideal normal condition foreign proteins never enter the body. Those taken with the food are all hydrolyzed in the alimentary canal until they have lost their original character, and are absorbed, not as foreign proteins but as simple amino acids and polypeptids which are the same as those already present in the tissues. Even the deadliest of snake venoms is harmless when taken into the stomach, for its poison is a protein which cannot be absorbed intact in harmful amounts. When bacteria and. other parasites enter the body, or when foreign serum or other proteins are artificially injected, the situation is met as it is when foreign proteins enter the alimentary canal—they undergo pro- teolytic hydrolysis until their character as foreign proteins is removed. The immunological reactions, under this hypothesis, consist in an aug- mentation of this capacity of the body to accomplish parenteral diges- tion and thus to destroy harmful foreign proteins. This view receives strong support by the demonstration of the fact that proteins lose si- multaneously their antigenic capacity and their digestibility when racem- ized by alkalies. Such racemized proteins are soluble, give the typical protein reactions, retain their coagulability by heat, and apparently
ANTIGENS 41
contain their complete quota of amino acid molecules. The only chemi- cal and physical change seems to be the enolization of such amino acids as have their amino radicals linked to a carboxyl radical within the protein molecule, while the terminal amino acid groups containing a free carboxyl group remain unchanged (Dakin).
This is readily seen by considering the change in an alanine radical forming part of a protein or peptid molecule, as shown below.
7, H © | | R—C—C—NH, en Oe nN | => | (Ele GEe Ketone form Enol form
The a carbon atom having by this enolization lost its asymmetry will form equal amounts of the two isomers and optical activity will be lost.
When such racemized proteins are hydrolyzed some of the amino acids recovered are in an optically inactive form while others retain their optical activity, suggesting that the latter have occupied a terminal position in the polypeptid linkage, and thus escaped racemization. With the racemization come two important alterations in the properties of the protein, loss of antigenic activity and of digestibility by proteolytic enzymes.®* Such racemized proteins injected subcutaneously are ex- creted unchanged in the urine (Dakin and Dudley) showing that, they cannot be attacked by the enzymes within the body. Therefore, it is altogether reasonable to make the assumption that their lack of anti- genic power depends on their lack of digestibility, the corollary being that antigenic activity is dependent on parenteral digestive proteolysis of foreign proteins.
The significance of digestibility as an essential factor in the antigenic action of proteins is brought into question by Landsteiner, who pro- duced acetylated protein preparations which are effective antigens de- spite complete resistance to trypsin and pepsin, at least in the reagent glass.°? However, he did not determine that the acetylated proteins also resist intestinal and parenteral digestion, as Dakin and Dudley did with their racemized proteins. Landsteiner proposes, as an alternative hypothesis, that the inert proteins merely lack the chemical structure necessary to stimulate antibody formation, independent of their digesti- bility, just as gelatin is digestible but non-antigenic.
42 THE CHEMICAL ASPECTS OF IMMUN
Landsteiner and Barron *? found that horse serum which had been treated with NaOH not only lost its antigenic function, but also would not bind antibodies for horse serum. If this racemized serum is treated with concentrated nitric acid and the alkali proteinates converted into xanthroproteins, this xanthoprotein is antigenic and behaves in all re- spects like a xanthoprotein prepared by nitration of natural horse serum which has not first been acted upon with alkali. The digestibility of this xanthoprotein was not tested. This indicates that racemization is reversible, at least as to its effect on antigenic action, but, as far as we can learn, the chemical reversibility of the racemization of proteins has not been tested. The fact that racemized proteins may be hydrolyzed by acids and yield mixtures of active and inactive amino acids indicates that the racemization is not reversible, or at least not completely reversi- ble, and throws confusion into the explanation of the observed facts. Furthermore, Kober * was unable to secure spectroscopic evidence of the keto-enol tautomerism suggested by Dakin as the explanation of racemization, and advances theoretical reasons indicating that the pro- teins cannot have as many free terminal amino and carboxyl radicals as Dakin’s hypothesis requires, for this makes the assumption that all the active amino acids obtained by hydrolysis of racemized proteins must have occupied terminal positions in peptid linkages. F. C. Koch has pointed out to me, however, that Kober’s criticism of ,Dakin’s hy- pothesis is not conclusive, because it is by no means certain, and even doubtful, that at every peptid linkage it has gone through the enol stage.
BACTERIAL TOXINS
In considering whether anything except intact or nearly intact pro- tein molecules can function as antigens, we meet at once an important problem in the bacterial toxins. These substances are characterized by their antigenic capacity, and yet it is not known whether they are pro- teins or not. Although they behave like electro-positive colloids,7 they diffuse more rapidly than proteins usually do, and very active toxins have been prepared that do not give the ordinary reactions for proteins. This last, of course, may merely mean that the toxins are so powerful that they are active in solutions so dilute that they do not respond to these tests. As an illustration of this possibility we have the case of ricin, the poisonous antigenic toxin of the castor oil bean.
This has been isolated ** in such purity that one-one thousandth of a milligram (0.000,001 gram) is a fatal dose per kilo of rabbit, and
ANTIGENS 43
smaller than lethal doses lead to antibody formation, i.e., are anti- genic. Nevertheless, ricin is inseparably associated with, and apparently identical with, the coagulable albumin of the castor bean. Because of the minuteness of the lethal dose some observers had thought that this vegetable toxin was not a protein, for they had obtained active solu- tions giving no protein reactions.
In favor of the view that toxins are proteins is the fact that they are colloidal molecules which are attacked by proteolytic enzymes. Kossel attaches some significance to the fact that, like proteins, they are precipitated by nucleic acid. They are adsorbed readily by animal charcoal and similar adsorbents,* pass slowly through even dense ‘dialyzing membranes but not at all through filters permeable only to ultramicroscopic particles of fine dimensions.
A study of the adsorption by charcoal of different toxins showed that a bacterial hematotoxin (which dissolves red corpuscles) when adsorbed to charcoal still retained the full power to neutralize anti- toxin, whereas diphtheria and tetanus toxins so adsorbed were par- tially inactivated (Eisler) ,*° indicating that probably there are consid- erable differences between different classes of toxins.
Heavy metals which precipitate proteins also precipitate toxins, which can be redissolved after removal of the metals; they come down in specific, we]l-defined concentrations of ammonium sulfate, and pro- tein precipitants inactivate toxin solutions. They are usually destroyed, or at least irreversibly inactivated, by heating; slowly at 45°, quickly at 80°, if in solution, but not at 100° when dried, although 150° is destructive; readily attacked by light and Roentgen rays, slowly altered by standing in solutions, not impaired by low temperatures, very susceptible to oxidizing agents and to alkalies, but much less so to acids. When the salts are removed by dialysis from a toxin solution, acidification precipitates the toxin (Glenny and Walpole). Many of the toxins when inhibited by treatment with acids of suitably low concentrations have their activity restored by neutralization (Doerr).7 But of course it is impossible to exclude the possibility that all these properties merely depend on the fact that the toxins are adsorbed to certain definite proteins.
There is reason to suspect that a toxin really is a protein to which is attached a toxic radical, which may or may not be an integral part of the protein molecule. This toxic radical of itself is not necessarily antigenic, and the production of antibodies is dependent entirely upon the colloidal protein radical which itself may not be toxic. Thus,
44 THE GHEMICAL ASPECLIS OF IMMUNITY
the venoms of different snakes of a single group (e.g., the vipers) may produce identical physiological and anatomical effects which sug- gests that the poisonous element is the same in each, but the immune serum against each shows specific differences which indicates that the protein radical of each venom is different.”
Particularly supporting this view of the complex structure of toxin are the observations of Landsteiner ** ** that proteins combined with various organic compounds act as antigens which produce antibodies reacting with any sort of protein to which the same or similar chemical groups are attached. The simple chemical radical when alone does not act as an antigen in the sense of stimulating the production of specific antibodies, although it will unite specifically with the antibodies engen- dered by the protein complex.
Resemblance of Toxins to Enzymes
The facts known concerning the properties of toxins serve to throw them into the same class as the enzymes, and Oppenheimer says of the toxins, ‘““we must be contented to assume that they are large molecular complexes, probably related to the proteins, corresponding to them in certain properties, but standing even nearer to the equally mysteri- ous enzymes with whose properties they show the most extended analo- gies both in their reactions and in their activities.” These similarities between toxins and enzymes are very striking. First of all we meet the same difficulty in isolating toxins that we do in isolating enzymes. “A pure toxin is as unknown as a pure enzyme” (Oppenheimer). At first both were believed to be proteins; now both are considered by many not to be proteins, but molecular complexes of nearly equally great dimensions. That toxins, like enzymes, are colloids, has been abundantly demonstrated.’* Both pass through porcelain filters, but both lose much of their strength in the process, and they are almost entirely held back by all but the most permeable dialyzing membranes. They behave similarly as regards adsorption by suspensions,’® and have similar effects on the physical properties of their solutions (Zunz).°° Neither will withstand boiling, and most forms are destroyed at 80° instantly or in a very short time; on the whole, however, toxins are more susceptible to heat, as well as to most other injurious agencies. Both stand dry heat over 100°, and extremely low temperature, with- out much injury. Left standing in solution for some time they gradu- ally lose their specific properties, and in each case this seems to be due to an alteration in the portion of the molecule that produces the de-
ANTIGENS 45
structive effects (toxophore or zymophore group in the Ehrlich nomen- clature), while the portion of the molecule that unites with the sub- stance that is to be attacked (haptophore group) remains uninjured, the toxin becoming a toxoid, the enzyme a fermentoid. On the other hand, enzymes and toxins seem to produce their effects according to different laws:—A small amount of enzyme can in course of time pro- duce an almost indefinite amount of effect, whereas toxins act more nearly quantitatively. It seems as if the enzyme were bound to the body upon which it acts, as is the toxin, but that after it has destroyed this body it is set free in a still active form, ready to accomplish fur- ther work, whereas the toxin is either not set free, or it becomes inac- ‘tive after it has once been combined. This similarity of toxins and enzymes brings up the question of
Tue ANTIGENIC CAPACITY OF ENZYMES
Although there is a considerable literature reporting positive results ** as well as some negative results,S? a critical review permits at the present time of no conclusion as to whether, in response to injection of fluids exhibiting proteolytic activity, there are formed specific anti- bodies capable of inhibiting the action of the proteolytic enzymes. The reported work is inconclusive because: (1) The existence of inhibiting substances in varying amounts in the normal serum has not always been properly considered; (2) inadequate quantitative meth- ods have been used for studying ferment action; (3) no proper con- trols have been made by immunizing with other tissue extracts not con- taining the enzyme under consideration; (4) the antisera usually give precipitin reactions with the proteins of the enzyme solution, which may readily adsorb the enzymes independent of any specific anti- enzyme action; (5) in some cases the antiserum inhibits the enzyme action by altering or buffering the reaction of the solution; (6) at best, the supposed positive results have generally concerned such slight quantitative effects on the enzymes that they may readily be due to unconsidered factors; (7) several experimenters have been unable to corroborate the reputed positive results.
Similar contradictory or defective evidence is presented in respect to the production of specific antienzymes ** for rennin ** and such non-proteolytic ferments as lipase,*’ emulsin,*® urease,*’ catalase,®” fibrin ferment, amylase, invertin,®** tyrosinase and laccase or phenolase.*® The latest investigation of the antigenic activity of these enzymes by Abderhalden and Wertheimer °° gave entirely negative results.
46 THE CHEMICAL ASPECTS OF IMMUNITY
It is evident, therefore, that despite the other points of resemblance, the enzymes are quite different from the toxins in respect to their anti- genic capacity, since this is certainly very slight if not altogether lacking in all the preparations of enzymes so far investigated.
‘TUBERCULIN
This presents much the same difficulties in respect to its nature and antigenic properties as the toxins and enzymes.® It has no appreciable toxicity of itself, and hence cannot be classed among the toxins. Only in the animal infected with tuberculosis does it exhibit any appreciable local or systemic toxicity, and hence it resembles the protein antigens in the anaphylaxis reactions. According to Zinsser,®* there are two fundamental types of sensitization to the products of the tubercle bacillus. One is the ordinary anaphylactic reaction to the proteins of the bacillus present in most preparations of tuberculin. The other, the typical skin reaction, is given even by preparations of tuberculin freed as far as possible from proteins removable by heat coagulation, the active solutions containing no proteins demonstrable by ordinary tests. That such preparations of tuberculin are entirely free from pro- tein has not been proved, and it is quite certain that often they are not, since heat coagulation is never complete. Neither has it been shown that such preparations are antigenic to the extent of inducing the formation of specific antibodies in non-tuberculous animals.°*# Hence our knowledge of tuberculin does not permit us to say that it does or does not contain a non-protein antigen.
MusHroom PoIsons
Another possible example of a non-protein antigen has been fur- nished by Ford,®* who found that rabbits can be immunized to extracts of Amanita phalloides, and that 1 cc. of the serum of such rabbits will neutralize five to eight times the lethal dose for guinea pigs and is anti- hemolytic for the hemolysin of Amanita when diluted to 1-1000. As he and Abel ** had found this hemolytic poison of Amanita to be a glucoside, this observation is to be interpreted as a successful produc- tion of an antibody for a non-protein poison, a glucoside. This work was further supported by successfully immunizing rabbits to extracts of Rhus toxicodendron, and finding that their serum in doses of 1 cc. will protect guinea pigs from 5-6 lethal doses of the poison, which was found by Acree and Syme * to be a glucoside.®°* Subsequent work by the same author confirms the main point, showing that an active
ANTIGENS 47
hemolysin can be obtained free front demonstrable protein, and that immunization with this protein-free hemolysin will result in strongly active (1-1000) antihemolytic serum.®* The antihemolysin unites with the hemolysin in simple multiple proportions.°* Another, non-hemo- lytic poison from Amanita, which Ford designates as Amanita toxin, was found to contain neither protein nor glucoside, and no antitoxic serum or definite artificial immunity can be obtained for it.
These observations of Ford are of so much importance in their rela- tion to the entire question of the nature of antigens that they should be repeated for verification. If accepted as they stand they constitute the strongest evidence yet presented as to the possibility of non-protein antigens. The newer developments in immunological research, more- over, make it seem entirely plausible that a complex glucoside, which can be hydrolyzed by enzymes, can act as an antigen. If we consider the evidence that immunity consists in the development of a special power to hydrolyze foreign colloidal substances, when these substances are of such a nature as to stimulate the cells to activity, and that Abderhalden and others have found evidence that specific enzymatic properties appear in the blood of animals injected with carbohydrates and fats, it seems entirely reasonable that a toxic glucoside can have antigenic properties.
Lipeorws ®® as ANTIGENS
There is a large literature on this topic *°° which reports most con- tradictory results, from complete denial of the possibility of antigenic action to the view that lipoids are of more importance as antigens than the proteins themselves. Before we can accept the idea that the lipoids are actually true antigens, capable on injection of inciting the production of antibodies which react specifically with them, we must overcome certain a@ priori objections that seem to render such a: con- clusion improbable if not impossible. As pointed out previously, an antigen must be a substance foreign to the body of the animal which is to furnish the antibody, and, as far as all the chemical evidence goes, with lipoids this is-never the case, at least as far as lipoids from animal sources are concerned, for when the lipoids of different animal species are investigated chemically it is found that they are usually the same in all species even when these are remotely separated in their zoological classification, Levene *°' says “it is significant that for the present, in our laboratory at least, we have failed to discover any distinction between lipoids derived from different tissues, or different
48 THE CHEMICAL ASPECIS. OF IMMUNITY
species.” Hence it is not to be expected that an injected lipoid can serve as an antigen since it is not foreign to the animal into which it has been injected, nor can it exhibit specificity, as Thiele and Em- bleton '°? pointed out long ago.
Despite these facts, the opposite view is strongly urged by not a few investigators. One source of confusion is the failure to distinguish between antigenic function and the capacity to react with antibodies, for, as pointed out elsewhere, these two properties are not always identical. Especially in the Wassermann reaction (q.v., Chapter VIII) we see lipoid mixtures successfully used as “antigen”? in the comple- ment fixation, despite the fact that such lipoidal “antigens” do not in- cite antibody formation when injected into animals (Fitzgerald and Leathes).1°? To be sure, numerous observers have found that tissue extracts made with fat solvents, do have more or less capacity to incite antibody formation. For example, Bang and Forssmann immunized with ethereal extracts of red corpuscles and obtained hemolysins, so they concluded that the antigenic constituent of the corpuscles is a lipoid, probably a phosphatid. This work has caused much contro- versy and many workers have failed to confirm their results.1°* It is a striking fact that when purified phosphatids, from sources favorable for obtaining pure materials, are used, the results are usually negative, while the positive results are generally reported with lipoids of more or less dubious purity.
Bacterial Lipoids
“Nastin,” the lipoid material from a streptothrix, has been used for immunizing by Much and others, who state that sera are obtained which give complement fixation reactions with nastin used as the antigen.1°° Similar results are described for the fatty materials from tubercle bacilli (“tuberculonastin”). Warden °° reports securing positive pre- cipitin and fixation reactions, not only with fatty complexes from bacteria and red cells, but also with artificial mixtures of soaps made up to resemble the cellular lipins. Furthermore, he claims to have produced specific antibodies by immunizing with these artificial fatty mixtures. Indeed, he states that the fat antigens are more specific than proteins, and infers that the specificity of antibodies is in part or wholly due to the fats of the cells. He maintains that the phos- phatids and cholesterol have no part in the process, but that only true fats and the salts of fatty acids are concerned, these acting as antigens when in a proper state of emulsification. Even diphtheria
ANTIGENS 49
toxin is looked upon by him as but a colloidal suspension of fats, and a simple suspension of 83.3 per cent oleic acid and 16.7 per cent palmitic acid with cholesterol is described as equivalent to the antigen of diph- theria bacilli. These statements, so heterodox as to seem almost fan- tastic, seem to have incited few published attempts to confirm them, but Dernby and Walbum? repeated the experiments with diphtheria toxin, and entirely failed to corroborate them in respect to the lipoid nature of diphtheria toxin or the toxin character of synthetic lipoid mixtures,
Despite the reported positive results, the antigenic power of bac- terial fats is by no means established. Borti¢é2°* has reviewed the literature and repeated some of the experiments of others, but failed to find that the lipoids of typhoid and diphtheria bacilli, cholera spirilla and staphylococci, freed from proteins by drastic measures, are capable of inciting agglutinin formation. Identical negative results were ob- tained by Beumer 1°° with the lipoids from tubercle bacilli and yeast, and with typhoid bacilli by Schmidt.11° These experiments give strong support to the opinion that the supposed positive results with bacterial fats depend upon the presence of antigenic proteins in the lipoid prepa- rations, and not on the lipoids themselves.
Lipoids as “Antigens” in Complement Fixation Reactions
Of course it is possible, and indeed probable, that bacteria contain lipoids not present in mammals, and these may possibly be sufficiently foreign to be capable of inciting antibody formation. The same rea- soning might be applied to the reported antigenic activity of lipoids from lower animals. Thus, Meyer? has reported the production of specific complement fixation antibodies by immunizing rabbits with acetone-insoluble lipoidal material obtained from tapeworms and echinococcus cysts. He has found the acetone-insoluble fraction of tubercle bacilli, presumably phosphatids, to serve as antigen in comple- ment fixation reactions with antibodies for tubercle bacilli,44? and much more effectively than the protein residue of the bacilli, wherefore he concludes that the reactions obtained with the lipoids certainly cannot be ascribed to adherent traces of protein. Here, however, the lipoid was not exhibiting an antigenic function by producing antibodies, but merely serving as an antigen in complement fixation reactions with antibodies engendered by immunizing with tubercle bacilli. In fact, in much of the reported work in which the antigenic function of lipoids is said to be established (e.g., Meyer'**), the observations are of this
50 THE CHEMICAL ASPECTS OF IMMUNITY
latter sort and not at all a demonstration of antigenic activity. How- ever, other observers have reported at least some antibody formation from immunizing with alcohol extracts of tubercle bacilli,* the purity of which is far from established, and Hoeden *® reports the production of complement-binding antisera by immunizing two guinea pigs with crude alcoholic extracts of echinococcus.
The number of reputed positive results with lipoids makes it im- possible at this time to state dogmatically that lipoids may not possess antigenic properties, but it must be taken into account that the success- ful use of lipoids as “antigens” in complement fixation reactions (q.v., Chapters VII and VIII) is not proof of their true antigenic nature. MacLean,'!® indeed, found evidence that even in the Wassermann reaction the active substance is not lecithin itself, but some other un- known substance which could be obtained practically lecithin-free. Ritchie and Miller ‘7 could find no antigenic activity in the lipoids of serum or corpuscles. Also Kleinschmidt,1’* who accepts the anti- genic nature of nastin, was unable to secure antibodies by immunizing rabbits with it. Neufeld found that rabbits immunized with lecithin developed no opsonins for lecithin emulsions.
Effect of Lipoids on Antigenic Activity of Proteins
A suggestive observation is that of Pick and Schwarz,'’® who found that the presence of lecithin increases the antigenic power of bacteria, which may help to explain the activity of possible traces of proteins in lipoid preparations used as antigens. Lipoids readily take up pro- teins, and it has been found that a solution of lecithin in chloroform will take up in fine suspension such colloids as cobra venom, trypsin, rennin and even oxide of iron (Dean).1*° It must also be taken into consideration that in some of the experiments in which lipoids have been thought to function as antigens, the reactions of the antiserum have not been specific.‘** Furthermore, it is necessary to take into account that the serum of normal rabbits and dogs often gives positive complement fixation with lipoidal antigens, despite the lack of any previous immunization, as Kolmer and Twist '*? observed.
Although it is entirely probable that compounds of lipoids and pro- teins are effective antigens, it is certainly incorrect to say, as Much 178 and others have done, that proteins owe their antigenic properties to admixed lipins, for the purest obtainable proteins, such as recrystallized egg albumin and the vegetable protein preparations of Osborne, are fully as active antigens as the unpurified proteins, and, in my experi-
ANTIGENS 51
ments, crystallized egg albumin was more active than corresponding amounts of unpurified egg albumin.*. Such observations entirely con- trovert the claims of Much and his school ??* that the chief function of proteins in immunization is to secure the necessary dispersion of the lipoid particles to which all the antigenic capacity is attributed.
Of interest in this connection are the heterogenetic antigens de- scribed by Forssman,'* which, present in the tissues of many animals of most varied species, nevertheless incite the formation of antibodies which cause hemolysis of sheep corpuscles.1° This antigenic agent has been found to be soluble in alcohol, and this feature has recently been re-investigated by Landsteiner and Simms (lit.).1°7 Landsteiner had suggested that these heterogenetic antigens consist of two parts: one, a protein, the real antigenic factor which is necessary for the production of antibodies; the other, alcohol-soluble and presumably lipoidal, has no antigenic capacity of itself but when united with the antigenic protein confers the peculiar heterogenetic specificity of this antigen. The experiments of Landsteiner and Simms supported this hypothesis, for they found that the isolated lipoidal element of the heterogenetic antigen, itself virtually non-antigenic, when merely mixed with normal serum produced an efficient heterogenetic antigen.1?* Such substances as these lipoids, which, devoid of antigenic activity, never- theless act specifically upon antibodies, Landsteiner has christened “haptenes,” and, as pointed out elsewhere (Chapter III), he and others have demonstrated the production of many such haptenes which are not lipoidal.
In view of all the foregoing contradictions and difficulties, it seems justifiable to say at this time that the capacity of fats and lipoids to serve as true antigens, capable of inciting the production of specific antibodies when injected into animals, has not yet been established. That such non-protein substances may, when united with proteins, modify the antigenic specificity of these proteins is, however, altogether probable. For example, the only recognizable chemical difference be- tween the immunologically distinct globulins of the blood, euglobulin and pseudoglobulin, seems to be the presence of a lipoid group in the former.
RECAPITULATION
In order that a substance may act as an antigen it must exist as a colloidal solution, it must be foreign to the animal producing the anti- bodies, and it must penetrate beyond the epithelial surfaces which protect the body effectively against foreign colloids, Apparently any
52 THE CHEMICAL ASPECTS OF IMMUNITY
complete foreign protein molecule soluble in the body fluids of an animal may serve as an antigen, except proteins that have undergone racemization by alkalies. Gelatin, and possibly globins, represent the largest soluble protein molecules known which are not antigenic. Such large complexes as the protamines and histones are not antigenic.
Coagulation of proteins removes their antigenic capacity only to the extent that it prevents their solution in the body fluids, for if the coagu- lation is reversible the redissolved protein possesses its original antigenic capacity.
Cleavage of the protein molecule practically destroys its antigenic capacity, even when all the fragments are used together for im- munizing. None of the isolated fragments of protein hydrolysis exhibits any considerable antigenic capacity, and a bare suggestion of antigenic capacity is exhibited by only a very small proportion of the larger fragments. It is not known just what step in the cleavage de- stroys the antigenic power of the protein molecule, but presumably the protein residue is antigenic as long as it is too large to diffuse readily into the cells and therefore requires the development of extra- cellular activities (1.e., antibodies) to accomplish its destruction. There- fore, when the fragments of protein cleavage are resynthesized to form colloidal molecules (plasteins) these are antigenic.
Compound proteins, whether natural or artificially prepared, are antigenic if soluble in the body fluids. The chief antigenic compound proteins occurring in nature are nucleoproteins and glycoproteins, and in these the non-protein radicals, being of very limited variety, seem to have little influence on the reactivity or specificity of the compound protein. The antigenic activity of hemoglobin is apparently not due to the globin moiety, but the active constituent has not been determined.
Addition of various non-protein radicals to protein antigens may alter their specificity, and simple chemicals uniting with the proteins of an animal may render them foreign to this animal so that it pro- duces antibodies to its own altered proteins. The antibodies to such artificial compound proteins may be capable of reactitig specifically with the non-protein radical, although the latter alone is not a true antigen for it cannot incite antibody formation when not combined with a protein. These facts probably explain the supposed antigenic activity of various non-protein substances, and it is still not proved that any non-protein substance can function as an antigen.
There is evidence that some toxic glucosides may serve as antigens, which is quite possible if they represent foreign colloids,
ANTIGENS 53
Although many reports exist which indicate that lipoidal suspen- sions may serve as antigens, this has not yet been satisfactorily estab- lished. The fact that, as far as now known, the lipoids are of limited number and not specific between different animals, makes it difficult to imagine that they can incite the tissues of an animal to antibody formation when they are not different from the lipoids already present in these tissues. Possibly lipoids from bacteria and lower animal forms, such as helminths, may be sufficiently foreign to the mammalian tissues to incite antibody formation, but it is not probable that the lipoids are so different in different species of bacteria or helminths as to account for the specificity of antisera obtained by immunizing with bacterial or helminth extracts. Purified lipoids are not anti- genic. Presumably the crude lipoids that are used successfully for immunizing also contain admixed protein antigens, and it is probable that lipoids may unite with proteins and modify the specificity of the resulting lipo-protein antigen.
There is some evidence that substances may be obtained from bac- terial cultures which are antigenic despite their failure to give any of the reactions of proteins in the solutions used for immunizing, but as yet these have been too little studied to permit the conclusion that they really do represent protein-free antigens. The typical soluble bacterial toxins, which are the antigens for antitoxin formation, come into this class of substances of unknown nature, actively antigenic even when purified so much that they do not give protein tests, but nevertheless these toxins are of colloidal dimensions. They bear numer- ous resemblances to the enzymes but the enzymes themselves apparently have little if any specific antigenic capacity.
REFERENCES
1The chief exception is the existence of antibodies bound in the non-striated muscles which may be demonstrated by the reactivity of such muscles to the specific antigen (anaphylaxis).
2 For example, pneumococci contain a complex carbohydrate which reacts specifically with antiserum for pneumococci, but which is not antigenic when injected into animals (Heidel- berger and Avery, Jour. Exp. Med., 1923 (38), 73).
3 Anat. Anzeiger, 1920 (53), 49.
4Wells, Jour. Infect. Dis., 1908 (5), 449.
5 Jour. Biol. Chem., 1916 (28), 11.
® Wells, Jour. Amer. Med. Assoc., 1908 (50), 527.
7 Starin, Jour. Infect. Dis., 1918 (23), 139; Landsteiner, Zeit. f. Immunitit., 1917 (26), 152.
8 Kahn and McNeil, Jour. Immunol., 1918 (3), 277.
®V. C. Vaughan, ‘Protein Split Products,” Philadelphia, 1913.
1 Wells, Zeit. f. Immunitat., 1913 (19), 599.
11 Schmidt and Bennett, Jour. Infect. Dis., 1919 (25), 207.
32 Wien. klin. Woch., 1906 (19), 327.
18 Infection and Resistance, 1923, p. I10.
54 THE CHEMICAL ASPECTS OF IMMOUNILY
14 Wells and Osborne, Jour. Infect. Dis., 1914 (14), 377.
15 Dale and Hartley, Biochem, Jour., 1916 (10), 408; Doerr and Berger, Zeit. f. Hyg., 1922 (96), 191; Ruppel, Deut. Med. Woch., 1923 (49), 40.
%Tsen, Jour. Med. Research, 1918 (37), 381.
17 Jour. Immunol., 1923 (8), 239.
18 Ergeb. Hyg., Bakt., Immunitaét. u. exp. Ther., Berlin, 1922 (5), 137.
1” Horse meat protein seems to become completely insoluble, at least it loses its antigenic power, if exposed long enough to strong alcohol—6o-120 days, according to Kodama (Zeit. Hyg., 1913 (74), 30).
20 Wells, H. G., Jour. Infect. Dis., ro09 (6), 506.
a bid. 190n 9), TAz.
22 Wells, H. G., and Osborne, T. B., Jour. Infect. Dis., 1915 (17), 259.
2 Piliott, @. Ei, Jour. Infect. Dis, soma (rs); 501.
24 Wells and Osborne, Jour. Infect. Dis., 1921 (29), 200.
2>Thorsch, Biochem. Zeit., 1914 (66), 486.
20 A, F. Coca, Biochem. Zeit., 1908 (14), 125; A. von Szily, Zeit. Immunitat., 1909 (3), 451; Kosakai, Jour. Pathol. and Bact., 1920 (23), 425.
27 Doerr and Moldovan, Wien. klin. Woch., 1911 (24), 555.
28 Schanz, Biochem. Zeit., 1915 (71), 406; Pfltigers Arch., 1918 (170), 646.
2? Burge, Amer. Jour. Physiol., 1916 (39), 335.
80 Mond, Arch. ges. Physiol., 1922 (196), 540.
31 Young, Proc. Royal Soc., London, 1922 (93B), 235.
82 Jour. Immunol., 1920 (5), 345.
88 Jour. Infect. Dis., 1919 C25) O77
84 See Haslam, Jour. Physiol., 1905 (32), 267; 1907 (36), 164.
85 Zeit. physiol. Chem., 1912 (81), 315.
86 Zunz, E., Biochem. Jour., 1916 (10), 160; Jour. physiol. et Path. gén., 1917 (17), 449; Arch Internat. de physiol., 1919 (15), 79, 92.
87 Jour. Biel, Chem:., ton2 (2), 233:
#8, Biochem. Zeit,, roro (93), ro6.
89 Zeit. physiol. Chem., 1912 (77), 289.
40 Arch. exp. Path. u. Pharm., 1913 (71), 298, 407.
41 Levene, Jour. Biol. Chem., 1918 (36), 105; Monograph No. 18, of the Rockefeller Inst. for Medical Research, July 7, 1922.
42 Landsteiner and Jablons, Zeit. Immunitat., 1914 (20), 618.
48 Landsteiner, Jour. Med. Res., 1924! (39), 631.
44 Biochem. Zeit., 1920 (104), 280.
45 Substances of this class are called haptenes by Landsteiner. Similar substances, non- antigenic but capable of reacting specifically with immune antibodies, have been found in bacteria by Zinsser (Infection and Resistance, 1923, p. 110) and others.
48 Beitr. Chem. Physiol. u. Pathol., 1902 (1), 397.
4“ Jour. Exp. Med., 1923) (37), 275.
#70 Jour. Exp. Med., 1924 (40), 343.
48 Jour. Exp. Med., 1923 (38), 163.
Zeit. ft. physiol, Chem., 1902 (36), 122.
50 Beitr. Chem, Physiol. und Path., 1904 (4), 115.
“Tour. Infect. Dis;, torr (o), 166.
52 Jour, Biol. Chem., ro08 (5), 311.
58 Zeit. £. physiol. Chem., 1912 (81), 285.
54 Zeit. f. exper. Pathol. u. Ther., 1912 (11), 69; Miinch. med. Woch., 1912 (59), 67.
SS Jour. Exper. Med., 1912 (16), 4709.
te Jour, Mxper. Med:, tor3. (07), 535.
7 Jour. Immunol., 1920 (5), 417.
68 Hektoen and Schulhof, Jour. Infect. Dis., 1922 (31), 32.
‘Jour, Unteet. Dis, tor (ia), 38s.
* Doerr and Pick (Biochem, Zeit., 1914 (60), 257) have ascribed antigenic activity to organ extracts, obtained by Pohl’s method, which they consider to depend on a nucleo- protein, but they submit no evidence to show that their antigen is the nucleoprotein rather than other tissue proteins associated with their preparation of ‘‘nucleoprotein.”’ Also it was not anaphylactogenic in guinea pigs.
ANTIGENS op
The functionally related hemocyanin is antigenic (C. L. A. Schmidt, Jour. Immunol., 1920 (5), 258). Hemocyanin is a copper protein compound of unknown composition found free in the circulating fluid of many invertebrates, where it functions as an oxygen carrier, but it is altogether different from hemoglobin in its chemical make up, the protein radical apparently being a globulin.
‘Jour: Infect. Dis; 1923% (33), 224.
6’ Confirmed by Higashi, Japanese Jour. Biochemistry, 1923 (2), 315.
®4 Heidelberger and Landsteiner, Jour. Exp. Med., 1923 (38), 561.
% Jour. Biol. Chem., 1912 (13), 357; 1913 (15), 263 and 271.
8 Jour. Biol. Chem., 1914 (17), 369.
@ Zeit. Immunitat., 1917 (26), 142.
®§ Racemized proteins may be hydrolyzed by the acids, however, and the proteoses thus obtained may have quite the same toxicity and physiological effects as proteoses from un- racemized proteins. (Underhill and Hendrix, Jour. Biol. Chem., 1915 (22), 453.)
® Landsteiner and Jablons, Zeit. f. Immunitat., 1914 (21), 193; Landsteiner and Praéek, Biochem. Zeit., 1916 (74), 388.
7 Jour. Biol. Chem., 1915 (22), 433.
71 Field and Teague, Jour. Exp. Med., 1907 (9), 86.
72 Osborne, Mendel and Harris, Amer. Jour. Physiol., 1905 (14), 259.
7 Kraus and Barbara, Wien. klin. Woch., 1915 (28), 524.
™ Kirschbaum, Wien. klin. Woch., 1914 (27), 289; Glenny and Walpole, Biochem. Jour., T915 (9), 2098.
™ Biochem. Zeit., 1923 (135), 416.
™ Wien. klin. Woch., 1907 (20), 5.
™ See Nicolle, Jour. State Med., 1920 (28), 293.
78 See Zangger, Cent. f. Bakt. (ref), 1905 (36), 239.
7 By flocculation of the colloids bearing adsorbed toxins it may be possible to secure them in comparatively pure condition (London, Compt. Rend. Soc. Biol., 1917 (80), 756).
Se Arch. di Histol, 1900) (7);. 137. '
81 Bertiau, Cent. f. Bakt., 1914 (74), 374. Stenitzer, Biochem. Zeit., 1908 (9), 382. Achalme, Ann. Inst. Pasteur., 1901 (15), 737. Bergmann and Guelke, Miinch. med. Woch., 1910 (57), 1673. Joseph and Pringsheim, Mitt. Grenz. Med. u. Chir., 1913 (26), 290. Von Eisler, Sitzungsber, Wiener. Akad. Wissensch., 1905 (114, Abt. 3), 119. Jochmann and Kantorowicz, Minch. med. Woch., 1908 (55), 728; Zeit. klin. Med., 1908 (66), 153. Levene and Stookey, Jour. Med. Res., 1903 (10), 217. Kirchheim and Reinicke, Arch. exp. Path. u. Pharm., 1914 (77), 412. Halpern, Zeit. Immunitat., 1911 (11), 609. Wago, Jour. Immunol., 1919 (4), 19.
82 Bergell and Schitz, Zeit. f. Hyg., 1905 (50), 305. Landsteiner, Zent. f. Bakt., 1900 (27), 357- Young, Biochem. Jour., 1918 (12), 499. Pozerski, Ann. Inst. Pasteur, 1909 (23), 205. Hamburger, Jour. Exp. Med., 1911 (14), 535. :
83 Farly literature given by Schtitze, Deut. Med. Woch., 1904 (30), 308.
84 Morgenroth, Zent. f. Bakt., 1899 (26), 349; 1900 (27), 721; Hedin, Zeit. physiol. Chem., 1912 (77), 229; Thaysen, Biochem. Jour., 1915 (9), 110.
85 Bertarelli, Cent. f. Bakt., 1905 (40), 231.
86 Bayliss, Jour. Physiol., 1912 (43), 455.
87 Jacoby, Biochem. Zeit., 1916 (74), 97.
87a Burnett and Schmidt, Jour. of Immunol. 1921 (6), 255.
88 Schiitze and Bergell, Zeit. klin. Med., 1907 (61), 366.
89 Bach and Engelhardt, Biochem. Zeit., 1923 (135), 39.
80 Fermentforschung, 1922 (6), 286.
1 Full review by Long in “Chemistry of Tuberculosis,’’ Wells, DeWitt and Long, Balti- more, Williams and Wilkins, 1923, p. 70,
62 Jour. Exp. Med., 1921 (34), 495-
92a See Adler, Wien. Arch. inn. Med., 1923 (7), 27.
%§ Jour. Infect. Dis., 1906 (3), 191; 1907 (4), 541.
% Jour. Biol. Chem., 1907 (2), 273.
®5 Jour. Biol. Chem., 1907 (2), 547.
6 Von Adelung (Arch. Int. Med., 1913 (11), 148) was unable to obtain antibodies for the poisons of Rhus diversiloba.
% Jour, Pharmacol., 1910 (2), 145.
56 THE CHEMICAL ASPECTS OF IMMUNITY
*8 Jour. Pharmacol., 1913 (4), 235.
%” The term lipoids is here used in the broad sense (better covered by the term lipins), including neutral fats, fatty acids, phosphatids, cholesterol, etc., and actually meaning in most immunological work the heterogeneous mixture obtained on extraction of tissues and blood with fat solvents. A valuable bibliography is given by Levene, Physiol. Reviews, 1921 (1), 327.
100 Early literature given by Landsteiner, Kolle and Wassermann’s Handbuch, 1913 (2), 124; Jobling, Jour. Immunol., 1916 (1), 491.
101 Jour. Amer. Chem. Soc., 1917 (39), 828.
102 Zeit. J. Immunitat., 1913 (16), 160.
108 Univ. of Calif. Publications, Pathol., 1912 (2), 39.
10 Review of literature by Landsteiner, Jahresb. Immunitatsforsch., 1910 (6), 209.
105 Literature in Beitr. Klinik d. Tuberk., 1911 (20), 343.
106 Jour. Infect. Dis., 1918 (22), 133; (23), 504; 1919 (24), 285; Jour. Bact., 1921 (6) 103.
107 Biochem. Zeit., 1923 (138), 505.
108 Biochem. Zeit., 1920 (106), 212.
109 Biochem. Zeit., 1921 (121), 127.
110 Zeit. f£. XImmunitat., 1924 (38), 511.
111 Zeit, Immunitat., 1910 (7), 732; 1911 (9), 530; 1912 (14), 355.
2 Zeit. Immunitat., 1912 (14), 359; 1912 (15), 245.
113 Biochem. Zeit., 1922 (129), 188.
114Dienes and Schoenheit, Amer. Rev. Tuberc., 1923 (8), 73.
115 Munch. med. Woch., 1924 (71), 77.
116 Lecithin and Allied Substances, Biochemical Monographs, 1918, p. 170.
447 Tour: Path. and Bact., 1o%3, (27), 420.
118 Berl. klin. Woch., 1910 (47), 57.
119 Biochem. Zeit., 1909 (15), 453.
nao Wancet, Jat. 135) 1OL7. Ds 456
121 Boquet and Negre, Ann. Inst. Pasteur, 1923 (37), 787; Schachenmeier, Biochem. Zeit., To2r (124), 165.
i22'Jour. Infect. Dis:, 1916 (18), 20.
128 Virchow’s Arch., 1923 (246), 292.
174 Full review given by Hans Schmidt, Zur Biologie der Lipoide, mit besonderer Beriich- sichtigung ihrer Antigenwirkung, Leipzig, Kabitzsch, 1922.
125 Biochem. Zeit., 1911 (37), 78.
126 See review in Jour. Amer. Med. Assoc., 1924 (82), 1465.
1% Jour. Exp. Med., 1923 (38), 127.
128 Other observers (Scimone and Torii, Zeit. Immunitit., 1923 (38), 264) have found that extraction of serum with ether before injecting it as an antigen leads to the production of more specific antisera than untreated antigenic sera.
Chapter III
Immunological Specificity
A property common to all biological processes, so striking in its exactness and its results in whatever way it manifests itself, specificity is especially clearly exhibited and advantageously studied in the im- ‘munological reactions. Investigations on the specificity of these reac- tions have added much knowledge to many and diverse problems of biology, and seem to a large measure to have explained the essential basis of biological specificity in all its forms.
We see the manifestations of specificity in so many and so varied processes that many of them are taken for granted, and others are not at first thought of as illustrations of the biological specificity which rules all the processes of life. It must obtain, from the fertilization of the ovum by the specific sort of spermatozoon which alone can enter the germ cell and stimulate division,’ on through the processes of growth which lead to the formation of specific structures charac- teristic of the species in form, composition and function, on to the specific reaction to injury which leads to repair by the proper tissue elements and the specific methods and means of defense against invad- ing enemies. The autumnal reddening of the maple leaf, the yellowing of the birch are as typical examples of chemical specificity as the secretion of a neurotoxic poison by the cobra or the production of antitoxin by a horse immunized with the poison of the diphtheria bacillus. The fact that a dog can follow the trail of his master and recognize him from all others by scent alone, shows that each of us is chemically individual.
RELATION OF IMMUNOLOGICAL TO BIOLOGICAL SPECIFICITY
When first the capacity of the animal body to produce antibodies in response to infection or artificial immunization was recognized, it at once became possible to settle an important biological problem in relation to bacteria. So simple were these forms, presenting so few possible points of structural differentiation, that there was doubt
37
58 THE CHEMICAL ASPECTS OF IMMUNITY
as to there being specific forms of bacteria with the capacity for in- definite reproduction of pure strains, which is a characteristic of more complex forms of life. But when in 1896 Gruber and Durham found that cholera spirilla and colon bacilli are agglutinated specifically by the proper immune serum, and when in 1897 Rudolf Kraus discovered that filtrates from cultures of plague bacilli and cholera spirilla gave a precipitate when added to the serum of a rabbit immunized with the corresponding filtrates, but that filtrates from other bacteria did not cause precipitation with these antisera, a final proof of the speci- ficity of these microscopic living forms had been furnished, which has since been reiterated by many other means.*
So, too, the specificity of the precipitin reaction threw light of im- portance on the evolutionary relationship of animals in a way that would have delighted Darwin and Huxley had they been alive to wit- ness it. In 1899 Tschistovitch found that the serum of rabbits im- munized with eel serum gave the precipitin reactions specifically with eel serum, and soon numerous studies by others had shown that such reactions could be obtained by immunizing with the serum of any sort of animal, and that the specificity exhibited was striking but of quantitative rather than of qualitative character. It was found to be a rule that with relatively slight immunization an antiserum would be obtained which would react almost solely with the same or homologous serum, but if the immunization was kept up to a maximum degree the range of reactivity would broaden until distinct reactions could be obtained with the serum of several, perhaps of many different species. It was also observed that these non-specific reactions were usually most marked with the sera of closely related species, and most certainly absent with the sera from widely remote species, and so it became apparent that immunological specificity is something closely related to biological specificity. This principle was brought out espe- cially by the extended studies of Nuttall * on precipitin tests measured quantitatively with blood from a vast number of animals. He found that, in general, immunological relationships correspond closely with the accepted zoological classifications, which of course rest on a mor- phological basis. Of particular interest was the evidence of relat.on- ship between man and the higher apes as brought out by the reaction between an antihuman serum and the blood of apes and monkeys. This antihuman serum gave slight or no reactions with the blood of mam- mals other than.the primates, but gave excellent reactions with the blood of primates, as indicated by the following table:
IMMUNOLOGICAL SPECIFICITY 59
ANTIHUMAN PRECIPITATING SERUM
Tested Against Precipitate SaeSpecimens= ume DlOOdincm nach stance as states e santas cies 100% * Soimiide.. sy species Of ANtTILONOIMS eelvaine o excises encis eis 100% 36 Cercopithecide (Common monkeys) ......c..ssseceees 92% 13 Cebide (Capuchins and Spider Monkeys).............. 7870 Am clenaticeom (Manin OCCUS emer rcepeate Mers deters alataie aus eraucyeraserers. = 50% PAMVeMirl Gee: MULSMIUES )) Sarees ew icine Gieis ¢ alae oe vie cine mere es (0)
* The percentages refer to the volume of precipitate formed on standing for a given time, the amount formed by the antiserum with its specific antigen being taken as 100 per cent. Antigen dilutions correspond throughout.
When the comparison is made by determining the dilution of either antigen or antibody which will give a reaction, instead of the less accurate estimation of bulk of precipitate used by Nuttall, it is found that the common monkeys are less closely related to man than appears in the above tables. For example, Hektoen* found that antihuman precipitating serum which reacted with human serum in a dilution of I-5,000, reacted with monkey (Macacus rhesus) serum only in a dilution of 1-100.
Another manifestation of specificity is seen in the differing capacity of a given antigen to incite the production of antibodies in different species of animals, for it is found that, at least in the case of im- munization with blood, the more closely related the species being immunized and the species furnishing the blood, the less formation of antibodies there will be. Thus, monkeys immunized with human blood produce little or no demonstrable antibodies in their serum,°® and the same difficulty is observed if we try to immunize dogs with the blood of wolves or foxes, or horses against ass blood. Immunization of an animal with blood of animals of the same species usually leads to no formation of antibodies, or, at the most, inconstant and slight anti- body formation is observed.
Although the first immunological studies were made either with bacteria or with blood, it was soon found that practically any soluble protein could be used as an antigen, and that immunological specificity was not always as definitely related to biological specificity, or better said, to zoological classification, as it had at first appeared. When egg white was used as antigen the resulting antiserum was found to react not only with the egg white of the same species, but with the eggs of many other species. Quantitative studies by Welsh and Chap- man °® led them to the conclusion that the white of bird’s eggs must contain an antigen common to all avian eggs, as well as antigens specific for each species of birds.* Likewise, with milk as antigen it
60 THE CHEMICAL ASPECTS OF IMMUNITY
has been found that there is a marked reaction between the antiserum for milk of any one species and the milk of all other species of mam- mals,’ although with quantitative differences in favor of the homo- logous milk. Further studies indicated that the casein is responsible for the common reactions, the other milk proteins for most of the specificity.
When the antigen is obtained from plants the same features of specificity are observed, and it is found that biologically related plants —e.g., wheat and rye, show marked immunological relationship, whereas entirely dissimilar plants usually show a corresponding immunologi- cal distinction.
On the other hand, after a time, when isolated proteins were utilized for immunization in place of such complex mixtures as blood, milk or egg white, the following facts were discovered :
(1). Even in one species of animals there may occur several different antigens readily distinguishable from one another.
(2). A common antigen may occur in many and quite unrelated species.
Distinct Antigens in a Single Species
To give an example of the first condition: In the hen’s egg there have been demonstrated by means of the anaphylaxis reaction, five distinct antigens and these correspond to five chemically distinct pro- teins that have been isolated from the egg. Again, although crystallized albumin from hen’s eggs shows practically no immunological distinc- tion from crystallized albumin from duck’s eggs, each of the three proteins separable from horse serum (euglobulin, pseudoglobulin and albumin) can be distinguished from each of the other two by their immunological behavior.‘° That is, two chemically similar proteins from different biological sources may be immunologically similar, whereas chemically distinct proteins from a single animal may be immunologically distinguishable. Another illustration is furnished by milk,® the casein of which is distinguishable chemically and immunologi- cally from the other proteins of the milk and also from the proteins of the serum of the same animal, whereas the globulin of the milk seems to be chemically and immunologically identical with that of the serum, On the other hand, casein from animals of different species is chemically similar and immunologically almost indistinguishable.
A certain but slight distinguishable specificity may be observed be- tween proteins from different organs of the same animal,!! which differentiation is still sharper between the tissue proteins and serum
IMMUNOLOGICAL SPECIFICITY 61
proteins of the animal? Sex cells especially seem to be distinct immunologically from the body cells.* Numerous instances of two separate proteins from the same plant seeds showing entirely distinct immunological specificity have been described (Wells and Osborne).'*
Common Antigens in Unrelated Species
As an illustration of this situation, in the crystalline lens of all animals there seem to be proteins which are the same for all species, since an antiserum for the lens of one animal will give good reactions with the lens protein of any species, including even that of the same species of animal furnishing the immune serum.’®
The lens contains two proteins, chemically distinguished by Morner,’® who identifies them as alpha and beta crystallin, there being more alpha crystallin in the outer and more beta in the inner parts of the lens. These two chemically distinct proteins were found by Hektoen and Schulhof *? to be immunologically distinct from each other even when obtained from the same lenses, but identical with the corre- sponding proteins of lenses from even remote species. On the other hand, precipitins for the albumin or globulin of beef serum, for example, do not give reactions with solutions of either alpha or beta crystallin from beef lens. The presence of these two proteins in the lens of various species, including certain fishes, explains the observed organ specificity of antibodies for lens proteins, since “the lens of different species always contains the same two chemically and antigenically distinct proteins.”
The. specific identity of lens proteins irrespective of species is ascribed by Krusius’?* to a process of denaturalization which takes place in the transformation of epithelium into the special form which constitutes the lens, analogous to the hornification of surface epithelium. Animals sensitized with whole mammalian lens will react somewhat with serum of the same species, but this sensitizing power of the lens for serum resides in the least differentiated external layers of the lens epithelium, whereas the sclerosed central part sensitizes only to lens and not to serum. According to this author, extracts of horse hoof, cow horn, and human hair show, by means of the anaphylaxis reac- tion, the same high degree of organ specificity and defective species specificity as is observed with lens extracts.’
Indeed it is altogether reasonable that lens proteins, keratin, mucin, and other proteins whose function is not metabolism, should be non- specific. Each of these proteins has quite the same function to per-
62 THE CHEMICAL ASPECTS OF IMMUNITY
form in every species, and is set off from the active tissues to per- form it. There is no more reason why they should be species specific than any other product of cell activity, e.g., trypsin, epinephrine, thyroxin, insulin. These all are products of cell activity with a definite function, and apparently alike in all species, just as lens proteins have been found to be. So, too, the characteristic protein of the thyroid, thyroglobulin, seems to be immunologically distinct from the blood proteins of the animal from which it is obtained, although not distinct from thyroglobulin of other animals.?? A similar distinction from tissue proteins is often seen in proteins set aside for nourishment of the offspring (e.g., casein, ovalbumin, seed proteins) and hence not participating in the metabolism of the organism from which they come.
Heterogenetic Antibodies. —A striking example of the existence of identical antigenic properties in materials of biologically unrelated origins, is furnished by the sheep corpuscle hemolysin discovered by Forssman 7 *? who found that many different tissue materials, when injected into rabbits, engender in the rabbits’ serum active hemolytic amboceptors for sheep corpuscles. This heterogeneous antigenic prop- erty has been demonstrated in the organs of the guinea pig, horse, cat, dog, mouse, chicken, turtle, and several species of fish,?* although not exhibited by organs of many closely related species (e.g., pig, Ox, rabbit, goose, frog, eel, man, pigeon, rat). It is not present in the red corpuscles of these animals, but is present in the corpuscles of the sheep, whose organs do not have this property. It has also been found in dysentery and other bacilli, mouse tumors and sheep sperma- tozoa. The Forssman antigen is soluble in alcohol and ether but not in water or acetone, so that it is of lipoid character and probably belongs in the phosphatid group. The lipins, being chemically simpler than the proteins, might differ little or none in various species, which would account for the presence of the same antigen in many of them. Further, this substance probably contains, besides the lipoid element, a non-lipoid fraction (protein?) with which the antigenic property is especially identified. Taniguchi** has shown that lipoid emulsions act as the specific antigen with Forssman antibodies and give in vitro precipitin and complement fixation tests, and even in vivo cause ana- phylactic shock, an antigen-antibody reaction ; but that pure lipoids never themselves generate the Forssman antibodies. Landsteiner and Simms 22 believe that the lipoid element of the Forssman antigen confers to the antigenic protein element the heterogenetic specificity.
IMMUNOLOGICAL SPECIFICITY 63
THE CHEMICAL BAsIs oF SPECIFICITY
It is to be presumed that immunological specificity must depend upon differences in the proteins, since, as indicated in the discussion of antigens, these are usually if not always, and chiefly if not exclu- sively, proteins. Indeed the very existence of specificity in such in- finite variety is of itself strong evidence in support of the view that only proteins are antigenic, for only the proteins show sufficient variety to account for the manifestations of specific immun'ty.
The blood and fixed tissues of all animals are made up of water, inorganic salts, carbohydrates, lipins and proteins. The water and inorganic salts are quite the same things throughout all the living tissues, and hence these cannot incite antibody formation since they are never foreign materials. Likewise there is no characteristic differ- ence in the carbohydrates found in different mammals, at least, and no one has been able to demonstrate antibody formation for even rela- tively complex carbohydrates found in the animal body. When we come to the lipins, using the term to include the fats, fatty acids and so-called lipoids, we find that here too the number is limited, and apparently decreasing rather than increasing with further investiga- tion of this class of substances (Levene).*° In any event, when the lipoids of different animal species are investigated it is found that they are usually the same in all species even when these are remotely separated in their zoological classification. Levene says “it is signifi- cant that for the present, in our laboratory at least, we have failed to discover any distinction between lipoids derived from different tissues, or different species.’”’ Hence it is not to be expected that any injected lipoid will serve as an antigen since it is not foreign to the animal into which it has been injected, nor can it exhibit specificity.
Although the chromatin elements of the cell seem to be the mor- phological structures which carry specificity in reproduction and cell multiplication, it is a striking fact that a characteristic constituent of chromatin, the nucleic acid, is quite the same substance in all animals so far investigated, although another type is found in plants. Hence nucleic acid is not an antigen and cannot be responsible for immuno- logical specificity.
As Levene also points out, the limited variety in the conjugated sulfuric acid radical of such gluco-proteins as mucin and chondrin, eliminates this as a factor in immunity. The radicals of this group that are now known differ only in the configuration of the nitrogenous
64 THE CHEMICAL ASPECTS OF IMMUNITY
sugar present in their molecules, and this difference seems to have no relation to differences in the character of the tissue from which the gluco-proteins were extracted.
The inadequacy of all the other known components of the body to account for the phenomena of specificity brings out all the more strikingly the adequacy of the proteins to account for specificity. In- finitely more complex in structure than any of these other tissue con- stituents, they occur in a virtually unlimited variety, and differences of chemical composition or physical properties are usually readily demonstrated between any two sorts of proteins that may be isolated from different sources or by different methods. Take, for example, so characteristic and actively functionating a protein as hemoglobin.
The Specificity of Hemoglobin
Having the same function in all species of carrying oxygen in loose combination for the benefit of other tissues, it might be expected that hemoglobin would be found to be a single uniform material. This, however, is not the case, for according to the crystallographic investiga- tions of Reichert and Brown * the hemoglobins of all species of ani- mals show readily recognizable differences. They say that the crystals obtained from different species of a genus are characteristic of that species, but differ from those of other species of the genus in angles or axial ratio, in optical characters, and especially in those characters comprised under the general term of crystal habit, so that one species can usually be distinguished from another by its hemoglobin crystals. But these differences are not such as to preclude the crystals from all species of a genus being placed in an isomorphous series. From their observations they reach the conclusion that the hemoglobins of any species are definite substances for that species. But upon comparing the corresponding hemoglobins in different species of a genus it is generally found that they differ the one from the other to a greater or less degree, the differences being such that when complete crystallo- graphic data are available the different species can be distinguished by these differences in their hemoglobins. As the hemoglobins crystallize in isomorphous series the differences between the angles of the crys- tals of the species of a genus are not, as a rule, great; butithey are as great as is the case with minerals or chemical salts that belong to an isoniorphous group. The immunological study of hemoglobins from animals of different species also indicates a distinct specificity.?” °8 Landsteiner and Heidelberger *® have also investigated the differentia-
IMMUNOLOGICAL SPECIFICITY 65
tion of hemoglobins by studying their effect on the solubility of one another, and found that the hemoglobins of horse, dog, rat and guinea pig show differences when tested by this method. But the hemoglobins of horse and donkey behaved as if isomorphous, which is interesting in view of the fact that it is difficult to differentiate these two hemo- globins by immunological tests.
The Complexity of the Proteins
Proteins being complexes of amino acids, most of the 20 or so which have been found in proteins being present in every protein, the possi- ble number of distinct and different proteins that might exist in na-
_ture is enormous, and entirely sufficient to explain what we know of
the range of immunological specificity. Abderhalden has calculated that 20 amino acids could form at least 2,432,902,008,176,640,000 different compounds, and this without including the enormously greater number that might be made by varying the proportion of the different amino acids in a single protein. Certainly this fact of itself is strong support for the hypothesis that immunological specificity, and most of the other features of biological specificity, must depend on the pro- teins. They alone could furnish the infinite variety of modifications found in the multitudes of different living creatures and in their life processes. And, moreover, we find that to guard the specificity of all living forms, elaborate provisions have been taken to see that foreign proteins do not enter the cells or the circulating fluids until they have been robbed of their foreign character, through hydrolysis into simple amino acid complexes that lack specificity. The perfec- tion with which the foregoing facts fit together to lead to the con- clusion that specificity depends on the proteins, adds convincing force to this conclusion.
Another contribution to the chemical basis of specificity has been made by Kossel,*° who finds certain relations in the proportions and groupings of the scanty number of amino acids that make up the prota- mines and histones of sperm, to-be characteristic of the sperm of cer- tain species and families. Unfortunately these simple proteins are not antigenic, so we are unable to compare their immunological specificity with their chemical composition.
The Evolution of the Proteins
The dependence of biological specificity on the proteins raises the question as to how the proteins came to be specific, for, on this basis,
66 THE CHEMICAL ASPECTS OF IMMUNITY
evolution resolves itself into the evolution of the protein molecule. Certainly during the vast period extending as far back as geological evidence goes, there cannot have occurred any great change in the character of the proteins. The earliest known forms of life are enough like existing forms to indicate that their chemical composition must have been essentially the same as that of their living descendants ; but they are not preserved in such a condition that their immunological relationships can be studied. The oldest proteins available for such purposes are those of the mummies and the mammoths, and the tests that have been made with mummy proteins show them to give reac- tions with antisera made by immunizing with human blood, and to sensitize animals to human proteins, demonstrating that in the rela- tively brief space of two to five thousand years no marked evolu- tionary alteration has taken place in human proteins, and also indi- cating that the specific character is fixed enough to remain unimpaired during all these years of preservation. Even more striking is the fact that the blood of a Siberian mammoth, possibly 25,000 years old, gave immunological reactions which indicated its relation to the mod- ern elephant (Friedenthal) .*4
Kennaway *? has considered this question of the evolution of the proteins, and asks:
“Why should the proteins that have been evolved possess so great a variety of amino acids, and why is it that none of the other related amino acids except these seventeen are found in proteins or elsewhere in nature? How are we to account for the fact that proteins usually contain amino acids that contain two, three, five or six carbon atoms, whereas four carbon amino acids are not found?” In the study of morphologic evolution it is customary to turn for light to simple forms in the scale of development, since the individual in its development recap.tulates the development of the species. Kennaway applies the same principle when he compares the chemical structure of proteins from mammals, protozoa, yeasts, moulds and bacteria; but it seems that the chemical structure of the unicellular organisms is not so much simpler than that of more complex forms as is their morphologic make-up. This, however, is to be expected, since the chemical structure of unicellular organisms is fully as complex as that of the individual cells of multicellular creatures, if not more so. In the yeast, the moulds and the five species of bacteria that have thus far been ana- lyzed, virtually all the known amino acids of animal and plant pro- teins have been found, with the exception of the sulfur-containing
IMMUNOLQGICAL SPECIFICITY 67
cystine. Apparently, then, “the simplest organisms now existing do not contain a series of amino acids any more primitive than that present in the higher organisms, except perhaps as regards the inclusion of cystine. One may suppose that the present apparently stereotyped series of utilizable amino acids represents the stable outcome of a struggle long ago among simple organisms in which those which made a less stiitable choice were beaten and have passed away, leaving no trace. We cannot know the biochemistry of the first organisms which appeared on the earth; the experiments and discarded compounds of that time are lost. The selection of amino acids must have taken place at an immensely remote period, for the earliest records which we have of the forms of life on the earth do not show us organisms which have any appearance of noteworthy difference in chemical com- position from those which exist at the present day. The doctrine of natural selection gives the impression that evolution proceeds through- out in a very gradual manner. But at the time when the amino acids were first being produced and tested, organic evolution must have proceeded very distinctly per saltum as each new compound was synthe- sized; natural selection would then act slowly and surely upon the organisms which made one or another choice, and thus the present series of amino acids was delimited.”
Nor can we learn the history of protein evolution by studying the steps by which the proteins are synthesized at the present day. All the amino acids of animal proteins are synthesized by plants and bac- teria; for although some experimental work suggests the possibility of a slight capacity to synthesize amino acids or proteins by animals, yet the results indicate even more strongly that this capacity is at the best very limited and of little or no significance as a usual source of proteins. Furthermore, no single amino acid has ever been found in animals that is not present in plants. Bacteria and plants can synthe- size proteins from extremely simple inorganic salts; but the inter- mediate steps cannot be followed, for the entire process is completed very rapidly and without demonstrable quantities of intermediates accumulating. Thus, a multiplying yeast cell growing on simple medi- ums produces an entire new cell from inorganic salts in a few hours, and bacteria accomplish this marvelous synthesis still more rapidly. In the time taken by a yeast cell to produce another by budding when growing on a medium such as Pasteur’s, containing no nitrogen but that of ammonium tartrate, it must synthesize each one of the score of amino acids and combine them as a series of polypeptids until
68 THE CHEMICAL ASPECTS OF IMMUNITY
yeast proteins are produced, and all the while carry on many other chemical operations within the compass of a yeast cell.
IMMUNOLOGICAL SPECIFICITY Is DEPENDENT ON CHEMICAL INDIVIDUALITY
So remarkable are the manifestations of specificity, so exquisitely delicate are many of the immunological reactions exhibiting specificity, and so wonderfully minute are the traces of proteins that may suffice to demonstrate it, that for some time it was assumed that specificity must depend on chemical or physical differences far too slight to be detected by other than immunological methods. More extensive ac- quaintance with the composition of the protein molecule and the appli- cation of more refined methods to its study, have suggested that m- munological differences between proteins are usually, and as far as now known always, associated with and presumably dependent upon chemical differences which can be detected by chemical or physical methods. A number of observations to justify this conclusion will be described in the following pages.
Evidence from Vegetable Proteins
A study of a great variety of proteins isolated from plant tissues in the purest condition possible, led Wells and Osborne ** to the conclu- sion that “since chemically similar proteins from seeds of different genera react anaphylactically in animals sensitized with one another, while chemically dissimilar proteins from the same seed in many cases fail to do so, we must conclude that the specificity of the ana- phylaxis reaction depends upon the chemical structure of the protein molecule.”
Among the observations on which this conclusion is based, we may cite the following examples: Seeds of peas, vetch, lentil and horse bean contain as their principal constituent a protein known as legumin, and no chemical difference can be found in the legumin obtained from these four seeds. Anaphylaxis experiments also failed to show any recognizable immunological difference between them. Here chemically similar proteins from different species of plants are found to be immunologically similar. A particularly striking case of this sort was reported by Jones ** and Wells in the fact that the globulin from. seeds of cantaloupe and of squash are chemically, crystallographically and immunologically identical. On the other hand, one often finds that two chemically different proteins from the same seed are readily
IMMUNOLOGICAL SPECIFICITY 69
distinguished by immunological reactions. Thus, the highly soluble seed proteins designated as “proteoses” are usually quite distinct from the other proteins found in the same seeds.*°
The proteins from seeds indeed offer a particularly favorable mate- rial for the study of specificity, because being merely storage proteins for nourishment of the embryo plant they are set aside in relatively pure form and of limited variety in the same seed. They also often offer unusual readiness of crystallization or peculiar solubilities which facilitate their separation in pure form. By investigating such purified materials rather than complex mixtures such as serum or tissue ex- tracts, much more exact information may be obtained. For example, the alcohol-soluble protein of wheat, gliadin, shows no recognizable chemical difference from the gliadin of rye, and these two proteins re- act immunologically as if identical, despite their derivation from plants of different species. Hordein of barley is chemically similar to gliadin but nevertheless definitely of different composition; immunologically hordein is found to be related to but distinguishable from gliadin. Sim- ilarly, another protein of wheat, glutenin, was found related to but dis- tinguishable from gliadin, despite the common source of the two pro- teins in the wheat grain. However, a relation between glutenin and hordein could not be shown immunologically.
Specificity not Dependent on Entire Protein Molecule
The above and similar results with other plant proteins led to the suggestion ** that the entire protein molecule is not involved in deter- mining the specific character of the immunological reaction, but this ts developed by certain groups or radicals of the protein molecule and a single protein molecule may contain two or more such groups. It was thought probable that the intact protein molecule is involved in the im- munological reaction, since apparently nothing less than an intact pro-
tein molecule can act as an antigen, but that certain groups determine
the specificity. The “group reactions” which are characteristic of biological reactions between proteins derived from closely related species, which usually have been interpreted as indicating the pres- ence in related organisms of identical as well as distinct proteins, can also be exhibited by single isolated proteins because of the presence of common and distinctive groupings or radicals in the protein mole- cules. These views, suggested by results obtained chiefly by the ana- phylaxis reaction, have received support from investigations carried
S.-
70 THE CHEMICAL ASPECTS OF IMMUNITY
out by others with different methods and materials, for example, the proteins of blood.
Specificity of Blood Proteins
The blood plasma contains four proteins that may be isolated by chemical methods, and which exhibit distinct chemical differences from one another. These proteins are fibrinogen, albumin, euglobulin and pseudoglobulin. When the fibrinogen is removed by coagulation the serum contains the other three proteins. Bauer and Engel *® found that fibrinogen from beef plasma is distinct from the proteins of the serum, since antiserum for the fibrinogen produces little or no com- plement fixation or precipitation with beef serum, and conversely.**
The three chief serum proteins, in turn, seem to be immunologically distinct from each other. As early as 1901, Leblanc ** found this to be true, and that these proteins are also distinguishable from hemo- globin by the precipitin reaction.
Dale and Hartley 1° found that a guinea pig which has received a sensitizing dose of any one of the purified horse serum proteins (albu- min, euglobulin, pseudoglobulin) is more sensitive to that one than to either of the other proteins from the same serum. In some cases, indeed, the reaction was rigidly specific.
Kato *® also separated carefully from horse serum the albumin, water-soluble globulin (pseudoglobulin) and water-insoluble globulin (euglobulin) fractions. He likewise found that while guinea pigs sensitized to one of these fractions did react to the other two, yet they reacted much more strongly to the homologous fraction. All these and other observers *° have noted that serum albumin is much less effective as an antigen than the other serum proteins. Doerr and Berger #1 not only found it possible to distinguish between euglobulin and serum albumin, but also between two fractions of the serum al- bumin obtained by precipitating with ammonium sulfate at 56-66 per cent and 66 to 99 per cent. According to them, five antigens may be distinguished in serum, namely, fibrinogen, euglobulin, pseudoglobulin, albumin C and D, to which we may add seromucoid. Each fraction is specific when tried against the corresponding protein of sera from other animals, unless possibly it be the seromucoid. To add to the number of antigens distinguishable in blood are those in the corpuscles, Hektoen and Schulhof *’ having found that hemoglobulin is distinct from both the corpuscular stroma, and the serum proteins, while the globulin of the red cells is immunologically distinct from the globulin
IMMUNOLOGICAL SPECIFICITY 71
of the serum (Bennett and Schmidt),‘? as well as from the hemo- globin.
As Doerr and Berger ** point out, each of the proteins of serum ex- hibits two kinds of specificity, one for the species and one within the species characterizing the particular protein itself. Hence there must be at least two different groups or factors determining specificity within a single protein molecule, as Wells and Osborne had previously sug- gested. Studies on bacterial antigens also have led Zinsser to a similar conclusion, namely, “it seems quite likely that the antigenic function of specific union with the antibody may be dependent upon a relatively small nucleus of the protein molecule.” *
Chemical Differences between Serum Proteins
It has been well established that these different serum proteins are chemically different. Long ago Abderhalden *° found that horse serum albumin contains no glycine, whereas serum globulin contains 3.5 per cent of glycine. Michaelis also found that the isoelectric point of serum albumin is much different from that of serum globulin, and of oxy- hemoglobin.*®
Hartley ** has shown that serum albumin differs from serum globu- lins in yielding, on hydrolysis, different proportions of cystine and diamino acids; however, he did not find corresponding differences be- tween the two globulins, and no difference can be found between them by racemization (Woodman).** Miss Chick and others have given reasons for considering euglobulin as being a compound of pseudo- globulin with a lipoid, which may account for the immunological differ- ences. In view of these facts the reputed transformation of serum albumin into globulin by various physical methods is obviously incor- rect, and Fanconi*® found that such “artificial globulins” show im- munologically that they are not globulins but merely albumins altered in solubility.
Of much significance in connection with the fact that different pro- teins from the same serum can be differentiated from one another, and also from corresponding protein fractions of sera of different species, is the study of Obermayer and Willheim °° on the formol titra- tion figures of serum proteins. The reaction in Sorensen’s serum titra- tion method seems to depend on the presence of terminal amino groups in the protein molecule. Taking as their “amino index” the quotient obtained by dividing the total nitrogen figure of the protein with the figure for the terminal amino groups as determined by formol titra-
F2 THE CHEMICAL ASPECTS OF IMMUNITY
tion (the result indicating actually how many nitrogen atoms the pro- tein molecule has for each terminal amino group), these authors in- vestigated the protein fractions of serum. In mammalian serum it was found that the amino index for the euglobulin fraction averages around 21.5, the pseudoglobulin gives a similar index, while the albu- min is constantly much lower, average about 12. In fowl serum the pseudoglobulin, with an average index of about 15, resembles more closely the albumin figure. It was found also that, according to the evidence of the amino index, the pseudoglobulin fraction and the albu- min fraction can each be separated by ammonium sulfate precipitation into two fractions differing distinctly in their composition, the latter corresponding presumably to the two albumin fractions which Doerr and Berger differentiated by immunological methods. The fact that the protein fraction of fowl and goose serum precipitated at 25-30 per cent saturation with ammonium sulfate has an amino index of from 28.5 to 32.5, whereas the corresponding fraction in horse and beef serum has an index of about 19, shows that these proteins are truly different chemically, just as they have been shown to be immunologically.
Bence-Jones and Noel Paton Proteins
As further evidence of the dependence of immunological specificity upon chemical composition of proteins, we may add the following examples : | Occasionally another protein, known as Bence-Jones protein, is
found in the urine and blood of persons suffering from extensive neo-
| plastic invasion of the bone-marrow. This protein is quite unique in
its chemical and physical properties. Hence, Massini* found that Bence-Jones proteins can be differentiated from human serum-proteins by complement fixation tests, and Hektoen * corroborated this with the precipitin reaction. Bayne-Jones and Wilson ** obtained similar results by these methods and also in anaphylaxis experiments. <A preparation of Bence-Jones protein which had been crystallized was found to be free from any of the normal serum proteins. It seems probable, to judge by their immunological behavior, that there may be more than one sort of Bence-Jones protein, and the crystalline Bence-Jones protein studied by Hektoen and Welker differed from one described by Krauss in containing less carbohydrate and loosely bound sulfur. Furthermore, another crystallizable protein, chemically distinct from Bence-Jones protein, found in urine by Noel Paton, has been found to be immunologically distinct from either Bence-Jones or
IMMUNOLOGICAL SPECIFICITY 73
serum proteins.°* Hektoen °4 has observed a case of excretion of still another crystallizable urinary protein, and demonstrated the immu- nological identity of this protein from both the Bence-Jones protein and the protein originally collected by Noel Paton.
Milk Proteins
A study of the chemistry and immunological behavior of milk pro- teins has furnished further evidence of the same purport (Wells and Osborne).§ Cow’s milk contains four chemically distinct proteins or protein fractions, namely, casein, lactalbumin, lactoglobulin and an alcohol-soluble protein. By means of the anaphylaxis test it can be shown that these four proteins are immunologically distinct. This fact furnishes another striking illustration of the dependence of immu- nologic specificity on chemical composition rather than biologic origin. Of these four proteins only one, the globulin, sensitizes to beef serum or causes reactions in animals sensitized to beef serum. This cor- responds to the observation of Crowther and Raistrick that milk globu- lin and serum globulin are chemically indistinguishable ; and of Wood- man ** that lactalbumin and serum albumin are two distinct proteins.
Biologically, casein is quite as distinct from the whey proteins and the serum proteins as it is chemically, whereas the biological reactions and chemical composition of caseins from different species of animals show close relationships. An immune serum for any one casein will give reactions with casein from any other species, even of remote relationship.°® The whey proteins as a whole are biologically similar to the soluble serum proteins of the animals from which they are derived. Casein from the milk of an animal of any given species shows a closer biologic relationship to the casein of another species than it does to either the whey proteins or to the serum proteins of its own species; the same is true of the chemical relations.
Although caseins from different species do not show demonstrable quantitative or qualitative chemical differences by ordinary methods, a study of the products of racemization by Dakin’s method yielded to Dudley and Woodman’*® evidence of some structural differences. Caseins from sheep’s and cow’s milk consist of identical amino acids, apparently in identical proportions, but when racemized the sheep’s casein showed all the tyrosine and most of the lysine unracemized, while both were completely racemized in cow’s casein, indicating that these amino acids occupy different positions in the caseins of the two species. Dale and Hartley °’ state that caseins from different species
74 THE CHEMICAL ASPECIS\OF IMMUNITY
show no clear disparity of antigenic properties, but it is possible that finer quantitative methods will disclose distinct discrepancies.
Egg Proteins
Ovomucoid, crystallized albumin and ovovitellin from hen’s eggs have been found to be readily distinguishable from one another by the anaphylaxis reaction.® °* By saturation tests it was shown that in crystalline egg albumin and in the globulin-fraction of hen’s egg white, there are specific and distinct antigens, as well as a common antigen which cannot be separated from them by fractional precipitation with ammonium sulfate. Therefore, there are distinguishable in the hen’s egg at least five distinct antigens which correspond to an equal number of proteins which chemists had previously distinguished. On the other hand, crystallized egg albumin of duck eggs corresponds perfectly within the limits of chemical methods with crystallized albumin of hen’s eggs, and each of these proteins sensitizes to the other appar- ently as well as to itself (Dale and Hartley).1° Egg yolk proteins from even such widely different species as fish and turtle may give precipitin reactions with antiserum for hen egg yolk proteins.°®
A particularly interesting observation bearing on the dependence of immunological differentiation on chemical variations has been furnished by the egg proteins. Dale and Hartley had found, in preliminary ex- periments, that crystallized albumins from hen eggs and from duck eggs sensitized animals mutually to one another, even with the guinea pig uterus method, but by using quantitative methods and by checking up with the desensitizing test, it was later possible for Dale to detect slight immunological differences between the two.°° The importance of this observation lies especially in the studies carried out by Dakin on the chemistry of these two albumins. When hydrolyzed, their amino acid content is found to be similar, but when the proteins are first racemized and then hydrolyzed it is found that there are distinct differences in the degree of racemization of the leucine, aspartic acid and_ histidine. There being reasons for believing that the amino acids which escape racemization occupy the terminal positions in the peptid chains of which the protein molecule is built, this result indicates that there are at least structural or spatial differences in these two proteins., In other words, a close relationship by immunological tests is here associated with chemical similarity, and a slight difference in chemical structure is found which presumably accounts for a slight immunological dif- ference that can be detected only by the most sensitive methods.
IMMUNOLOGICAL SPECIFICITY 75
IMMUNOLOGICAL BEHAVIOR OF ARTIFICIALLY Moptricp PROTEINS
The conclusion that immunological specificity depends on chemical differences in the proteins of a degree usually and perhaps always ca- pable of recognition by existing chemical methods, which, as outlined in the preceding paragraphs, has been obtained by the study of the immunological reactions of isolated proteins as they occur in nature, is corroborated by the results obtained by modifying the composition of the protein molecule by artificial means. The pioneers in this field were Obermayer and Pick,®' who found that specificity for species, as indicated by the precipitin react‘on, is altered when the aromatic radicals , are acted upon in various ways. This is most readily done by introduc- ing such radicals as iodin, NO, and N = N into the protein molecule, where they are known to enter specifically into the benzene ring of the aromatic amino acids. They report that such altered proteins produce precipitins that are specific for the form of protein used in immunizing, but not at all specific as regards the species from which the protein is derived; e.g., iodized bovine serum produces in immunized rabbits precipitins that will react only with iodized proteins, but they react with iodized proteins of any source, whether from egg white, horse serum, or even iodized rabbit proteins.** On the other hand, such modified pro- teins are said to be quite specific toward themselves, 1.e., precipitins for
iodized proteins will not react with diazotized proteins, and conversely. ;
Furthermore, if radicals were introduced into amino acids which do not possess the benzene ring, species specificity was not affected, al- though the resulting proteins exhibited specificity toward themselves.
These observations, as far as they agree, indicate that the aromatic radicals may be of importance in determining species specificity. This view received some support from the observation that gelatin, which of all proteins is poorest in aromatic amino acids, possesses no antigenic capacity (Wells, Starin).
Pick’s Conception of Specificity
The studies of Pick and his colleagues,°* amplified somewhat by other investigations, have led him to the following view of the chem- istry of specificity: There exist two sorts of specificity in each protein molecule; one of these is easily altered by simple physical measures, e.g., heat, cold, partial coagulation, etc., without essentially changing the chemical composition of the protein. When so altered the antigenic properties of the protein are likewise altered, in that ‘the antibody it
76 THE CHEMICAL ASPECTS OF IMMUNITY
engenders differs somewhat in the scope of its reactivity from the anti- body engendered by the original unaltered protein; but the alteration does not affect the species characteristics of the antigen. Thus, a heated antigen may engender precipitins that will react with this heated antigen, but not with similar heated proteins from other species of animals, while the antibodies engendered by the same but unheated antigen will not react with the heated protein. Bordet, who looks upon immuno- logical reactions as essentially adsorption phenomena, considers that this altered specificity of heated proteins supports the conception that ad- sorption affinity is the essential factor in the reactions, and that speci- ficity depends on physical rather than chemical relations of the ma- terials involved.**
The other sort of specificity is not so easily affected, according to Pick, only marked chemical alterations of the antigen modifying it, and this concerns the species characteristics of the antigen. This funda- mental species specificity seems to be closely related to the aromatic radicals of the protein antigen, for it is affected by introducing into the protein molecules substances which are known to combine with the benzene ring, e.g., iodin, diazo and nitro groups. Proteins thus chemi- cally altered will act as proteins foreign to animals of the species from which they are derived, and the antigens they develop are devoid of species specificity, although quite specific for proteins like themselves; e.g., a nitro-protein made by treating rabbit serum protein with nitric acid, will, if injected into even the same rabbit, cause the formation of antibodies which will react with this same nitro-protein, and also with nitro-proteins derived from entirely different species or even from plants,—but reacting only with nitro-proteins. It is also possible to cause chemical modifications analogous to the physical modifica- tions previously mentioned, which change only the scope of specificity of the antigen without altering its specificity for species. Appre- ciating that the number of different aromatic radicals in the protein molecule is not sufficient to account for the innumerable manifes- tations of specificity, Pick interprets the significance of these aromatic radicals as that of a central complex about which are the groupings which determine species specificity. It is not merely the number and proportion of amino-acid radicals in the protein molecule which de- termine its specificity, but, more important because presenting greater possibilities for variations, the arrangement of these radicals in the molecule. : |
IMMUNOLOGICAL SPECIFICITY 77
Landsteiner’s Observations
The extensive investigations of Landsteiner ®* and his colleagues, however, have led them to a different opinion. They found that species specificity may be altered in many other ways than by attacking the aromatic rings of the protein. This they accomplished by esterification with acid alcohol, acetylating with acid anhydride or acid chlorides, or
methylation by means of diazomethane. The common feature of these |
reactions is that the alterations occur in the salt-forming groups of the protein molecules, and they evidently do not depend solely on substi- tution of H-ions in the aromatic radicals. This artificial, structural
_ Specificity is marked, and traverses by far the biological lines; e.g., anti- |
serum for methylated protein of horse serum may react not only with methylated proteins from most diverse animal species, but also with methylated plant proteins.
It is also possible to produce artificial protein compounds which are antigenic when injected into the species from which the protein is de- rived, but which engender antibodies retaining species specificity; e.g., formaldehyde-treated rabbit serum protein when injected into rabbits produces antiserum reacting with itself, but not with formaldehyde- treated proteins from other species. In this case, then, a chemical change
a
which has made an animal’s own serum proteins antigenic for itself, has /
failed to alter the species specificity.°° It is believed that formaldehyde acts on the lysine radical of proteins, forming methylene compounds of
amino acid groups, R—N=CH;. Evidently, then, the occupation of an amido group of lysine by the methylene radical is almost devoid of any |,
effect on specificity, and a marked chemical change can take place in a protein without noticeable effect on the structural or species speci- ficity.
This lack of effect of formaldehyde may be contrasted with the marked effect on serological specificity which results from alkylization
or acetylation of the proteins, indicating that a relatively small radical, added to the large protein radical may exhibit its own characteristic | ; influence on the specificity of the entire complex. Thus it was found “‘~
that rabbits immunized with acetylated horse serum °’ produce com- plement fixation antibodies reacting with acetylated serum from other species than horse, e.g., fowl or rabbit serum, but not reacting with normal or diazotized horse serum. Here a simple added radical has de- prived the horse serum of its species specificity while giving it a new structural specificity depending on the presence of the acetyl group.
¥ 6
78 THE CHEMICAL ASPECTS OF IMMUNITY
An extensive series of protein compounds was made by combining horse serum with the diazonium derivatives obtained from the follow- ing amino-compounds: aniline, o-, m-, and p-aminobenzenesulfonic acids, a number of methyl, chlor, brom and nitro substitution products of the above compounds, p-aminophenylarsenic acid, o-, m-, and p-aminocinnamic acid, naphthionic acid and aminoazobenzenedisul fonic acid. The 23 different immune sera obtained by immunizing with these compound proteins were tested with 33 different azo proteins. Of these immune sera, only 6 (aniline, p-aminobenzenesulfonic acid, 0-, m-, and p-aminocinnamic acid and aminoazobenzenedisulfonic acid) were en- tirely specific, i.e., they reacted only with the homologous antigen and vice versa. Of the antigens, 15 reacted only with some one of the antisera and 9 of these 15 antigens were not homologous, antisera not having been prepared for them. The other immune sera and antigens showed a broader sphere of activity.
Cross reactions were obtained only with antisera that contained chem- ically closely related aromatic groups. The immune serum always reacted with its homologous antigen, but some of the sera also reacted with antigens that contained aromatic side chains that were either isomeric or very closely related chemically to the aromatic side chain of the homologous antigen. The important factor seemed to be the relative location, in the aromatic nucleus, of the acid and the diazo group. Cross reactions were obtained when these groups were identically located, irrespective of the other constituents of the nucleus, or when the groups were, say, meta instead of ortho to each other. For example, antigens prepared from diazotized o-aminobenzenesulfonic acid, p-toluidinesul- fonic acid and p-chlor, o-aminobenzenesulfonic acid and m-amino- benzenesulfonic acid, reacted with the diazotized p-toluidinesulfonic acid antiserum. However, the most pronounced reaction was obtained with the homologous antigen. In the first three antigens, the sulfonic acid group is located ortho to the diazo group. In the last antigen, the sulfonic acid group is in the meta position with respect to the diazo group.
The arsenic acid immune serum reacted only with the six arsenic acid antigens. The effect of the introduction of a substituting methyl group in general is not very considerable. The substitu- tion of Cl and Br have less influence on the serum reactions than carbonic, sulfonic and arsenic acids. The conclusion is reached that the specific behavior is determined by the chemical structure of rela- twely small portions of the large antigen molecule. The observation
IMICUNOLOGICAL SPECTFICILY 79
that the location in the molecule of a simple organic compound of definite groups can be indicated by immune serum when this compound is united to an antigenic protein, can best be explained as depending on certain spatial correspondence of antigen and antibody, just as Emil Fischer assumed for the specific action of ferments. Apparently other factors must also play a part in determining specificity, as there are some groups of reactions that do not correspond to this theory. Thus, the arsenic acid serum was strictly specific solely for arsenic acid anti- gens. Here the specificity must be determined by the arsenic acid group as such. Therefore, it is concluded that besides the configura- tion also the chemical characteristics of certain groups come to expres- sion in serological specificity. Landsteiner holds that it is probable that an antibody can react with various related but not identical antigens. Specificity is the expression of a quantitatively graded affinity which reaches a maximal value with a certain combination, designated as specific antigen and homologous antibody. Hence we have antibodies with varying range of reaction with related antigens.®*
Probably there is an enormous number of isomers of a protein and these cannot all be distinguished by the present serological methods. These isomeric differences can not well constitute the species’ differ- ences but correspond possibly to the numerous individual and racial differences that cannot be distinguished serologically. Landsteiner says that his own observations would place the species specificity on the spatial configuration of the protein and raise the question if the termi- nal acid groups are not of significance for specificity differentiation. He remarks that “it must be remembered that the question is often raised whether chemical properties are indicated by the serum reactions. After the investigations of Obermayer and Pick, Wells and Osborne and ourselves, it might seem superfluous to enter into this matter again, were it not that recently divergent opinions on this point have appeared. Thus the opinion of Traube, who assumes that the chemical nature is without significance for specificity, has found some acceptance, and in a review concerning specificity by Sleeswijk °° it is considered doubt- ful if specificity depends fundamentally on qualitative differences. It is clear that such opinions have become untenable. Much more should we take up for investigation the other question, if two substances of quite the same chemical properties can exhibit serological specificity
because of physical differences as Pick and Bordet have believed possible.”
80 THE CHEMICAL ASPECTS OF IMMUNITY
Reaction of Antibodies with Non-antigenic Compounds
Landsteiner also found that if he immunized with a compound of a protein with a non-protein radical, such as metanilic acid (m-amino- benzol sulfonic acid), para arsanilic acid (p-aminophenylarsenic acid) or p-amino benzoic acid, the resulting anti-serum would react with compounds of these radicals with many different proteins, no matter how unrelated the proteins are, e.g., rabbit serum, gliadin from wheat and casein from cow’s milk. These results also suggest that specificity depends on single groups in or attached to the protein molecule and not upon the protein molecule as a whole. With such artificial compound proteins, at least, the protein molecule functions only to incite antibody formation, not to determine the specificity, and it owes this antigenic capacity to its large molecular dimen- sions. The added groups used in these experiments are unable of themselves to incite antibody formation, although they can bind the Specific antibodies if they are added to the antisera engendered by proteins to which they had been united before serving as antigens.” For example, the specific precipitin reaction between a protein treated with metanilic acid and an immune serum made by injecting this com- pound protein, is inhibited by metanilic acid and closely related sulfonic acids but not by p-arsanilic acid, and p-amino-benzoic acid, or other unrelated aromatic and aliphatic organic compounds. The converse experiment with proteins treated with p-arsanilic acid or p-amino- benzoic acid shows the same specificity; e.g., the antiserum for protein treated with p-amino benzoic acid gives precipitin reactions for this protein which are inhibited only by compounds having a carboxyl group attached to the benzene ring. Apparently these inhibiting radicals can react with the antibodies of the wnmune serum just as well whether attached to a protein molecule or not, and hence by binding the antibody render it unable to react with the protein antigen containing the radical. That is, they saturate the specific antibody, although they are unable to produce a precipitation or other observable reaction when not attached to a protein radical. (See also pp. 32-34.)
Landsteiner makes the suggestion that possibly other, non-protein colloidal complexes might, when united with a suitable radical, serve in place of the protein molecule to incite antibody formation, but, as far as we know, such a phenomenon has not yet been demonstrated.
A natural case of specificity depending on the protein radical, while the toxicity depends on the non-protein radical, has been suggested by Nicolle. He calls attention to the fact that according to pharmaco-
IMMUNOLOGICAL SPECIFICITY 81
logical and pathological evidence the poisonous elements of several different sorts of snake venoms must be the same or similar, despite the fact that the protein-rich venom is immunologically specific, the anti- serum for one venom failing to protect against another. This may be explained by the assumption that the poisonous element in these venoms is the same, but attached to different and antigenically specific proteins.
Of particular significance is the fact that the experiments of Land- steiner indicate that non-protein substances may not only influence the specificity of immunological reactions but also that under certain con- ditions these reactions may give an indication of the nature of a non-protein radical. For example, it was found that if an antiserum was prepared by immunizing with a compound of para-arsanilic acid with protein, the precipitating power of the serum for this protein complex may be removed by saturating it not only with the homologous compound protein, but also with para-arsanilic acid: uncombined with protein, or with other arsenicals containing an aromatic group, but not with arsenicals which do not contain an aromatic group. Similarly, the antiserum for metanilic acid protein may be inactivated by numerous sulfonic acid compounds related to metanilic acid but not by unrelated compounds. Again, an immune serum for para-amino-benzoic acid protein is inhibited only by compounds that also have a carboxyl group in the benzene ring.
Influence of Physical Properties on Specificity
As Landsteiner has said, the evidence seems clear that chemical changes alone may be sufficient to account for specificity, but that the question is still open as to whether physical properties play any part in determining specificity. Apparently certain physical properties deter- mine whether a substance may serve as an antigen at all—that is, an antigen must be a large colloidal molecule. Can the physical configu- ration in this molecule determine any of the manifestations of speci- ficity? Among those who have taken the affirmative stand are Pick,® who has suggested that physico-chemical properties may come into play in determining the possibility of interaction of antigen and antibody. He believes that the electric charges on the amphoteric colloidal antigen and antibody, and perhaps also their surface configuration and their surface forces, all influence their reaction; these physico-chemical fac- tors greatly complicate the possibility of reaction between two colloids, and to these influences are added the influence of the chemical structure in determining subsequent chemical reactions. It would seem possible
82 THE CHEMICAL ASPECTS OF MMi y
that the existence of all these factors may account for specificity, it being necessary for each one of a long series of both physical and chemical adjustments to agree perfectly in order that reaction may take place—just as in a combination lock one lever after another is thrown by the proper manipulation of the dial, and only when all the long series of levers is in just the proper position does the bolt engage and the lock open.**
Nevertheless, we still lack a single certain example of specificity determined by purely physical properties (Doerr and Berger).’* Cer- tainly the evidence that the immunological reactions, at least those of agglutination and precipitation, resemble the union of two colloids of different electrical charges (see discussion of Agglutination and Precipi- tation, Chapter VI) fail to explain specificity. Michaelis and David- sohn *° have pointed out that such specific reactions will occur in solu- tions having a pH:of a wide range, indicating that the electrical charge of the components cannot be important in the production of specific immunological reactions.
NON-SPECIFIC REACTIONS
As stated before, there are numerous instances of apparently non- specific immunological reactions, that is, an antiserum for one antigen giving reactions with antigens from sources which seem to be unrelated, at least from a zoological standpoint. Undoubtedly the logical explana- tion is that there may exist proteins in different species which have chemical resemblances or identity, and this is scarcely to be doubted. We find identical lipoids, fats, nucleic acids, and carbohydrates in different species; many particular types of proteins show apparent chemical identity in different species (e.g., gelatin, keratin) ; some chemically similar, derived proteins also seem immunologically identical or closely related even when coming from unrelated sources (e.g., lens protein, casein). Therefore, it is highly probable that many tissue proteins may be identical in different species of animal cells, and even in animal and plant cells. Eberson 7° also obtained evidence that different species of bacteria may contain common antigenic proteins as well as specific antigens, accounting for group reactions and non-specific immunization,
Another sort of manifestation of apparently non-specific immunity reactions has been observed especially in therapeutic immunization,7"78 Beginning with the classical observation of Matthes that the tuberculin reaction could be produced with deutero-albumose, many similar non-
IMMUNOLOGICAL SPECIFICITY 83
specific reactions have been observed. Particularly the sharp reaction that follows intravenous injections of killed typhoid bacilli into typhoid patients has been found to result equally well if colon bacilli are used, or deutero-albumose. One possible explanation of this type of reaction is that the injected substance acts as a common antigen, which reacts with the common antibodies that were engendered by the antigens of the cause of the disease. Another possibility is that the foreign protein stimulates the tissues that form antibodies, presumably the red mar- row, so that they produce not only antibodies for this antigen, but also for the antigens of the specific etiologic factor of the disease that have been stimulating the bone marrow previously. Hektoen 7 has observed, for example, that if an animal that has previously produced precipitins for one foreign protein is reinjected with a different protein it will'then produce precipitins for both these proteins, and possibly for other proteins with which it has not been injected.®° It also has been found that animals previously immunized against one type of bacteria are capable of forming more antibodies for some entirely unrelated organism than are control animals.*‘ | Moreover, the febrile reaction, leucocytosis, increased lymph flow,*? and other phenomena, such as the altered ferment-anti-ferment balance of the serum (Jobling) ,*°* which follow injection of non-specific protein, may be responsible for favor- ably affecting the disease rather than actual antibody formation.
The opposite type of phenomenon, that is, non-specific interference with immunological reaction, is suggested by the observations of J. H. Lewis.** He found that small quantities of one protein injected into a guinea pig together with or shortly after large quantities of another protein (e.g., a weak solution of egg albumen in dog serum) would not sensitize the animal, although a similar amount injected alone would always sensitize. The suggested explanation is that the larger amount of foreign protein combines with so many of the available cell receptors that few of the small number of sensitizing protein molecules are able to be bound to the cells and to stimulate antibody formation; this ex- planation assumes a certain lack of specificity on the part of the cell receptors.
An interesting illustration of the fact that whatever stimulates the bone marrow may cause it to form, among other blood elements, specific antibodies, is furnished by the behavior of antitoxin-producing horses. If a horse that has been immunized to diphtheria toxin is bled as much as possible, it will be found to have regenerated the lost anti- toxin within 48 hours,*® although the last immunizing dose of toxin
84 THE CHEMICAL ASPECTS OF IMMUNITY
was received long before, presumably because bleeding powerfully stimu- lates the bone marrow to regenerate the lost blood elements. Also, it is stated that persons who have once had typhoid, but whose blood no longer contains much agglutinin, may show a high typhoid agglutinin content when infected by some other organism, or after any sharp febrile attack. It is highly possible that many therapeutic agents may similarly act by stimulating the tissues to increased formation of specific antibodies, e.g., arsenic, mercury and other metals, heliotherapy, hemorrhage *° or phlebotomy, hot baths.
RECAPITULATION
Immunological specificity is but one of the innumerable manifestations characteristic of all biological processes, from the fertilization of the ovum by a specific spermatozoon on through the development of specific structures and the specific cellular manifestations characteristic of each species. The study of specificity by immunological methods has thrown light on the entire subject of biological specificity, and demonstrates it to depend chiefly if not solely on the chemical structure of the proteins. None of the other constituents of the blood and tissues can account for the infinite variety of the manifestations of specificity, for they are limited in number and quite the same in many or all different species. The variety possible in the proteins is practically unlimited because of the number of different combinations that can be made with the score of amino acids found inthem. It has been found that the immunological differences that exist between different species of bacteria, between extracts of different plants, or between the blood and tissues of different species of animals, are dependent on differences in the composition of their proteins whenever isolated proteins are under observation.
As far as investigations have been made, any two proteins that are identical immunologically are indistinguishable chemically, those that are readily distinguished by one of these methods are also readily differen- tiated by the other. Proteins that are shown to be related but still dis- tinguishable by immunological tests are found to be similar to one another in chemical composition, but nevertheless show recognizable chemical differences when suitable methods are used.
Biological specificity depends on chemical individuality of proteins, and biological relationship is equally associated with the presence of chemically similar proteins. A single species of animals, or even a single animal, however, contains many different proteins which may be distinguished immunologically and chemically. A single protein, like-
IMMUNOLOGICAL SPECIFICITY 85
wise, may be found widely distributed through many different species, its identity being established readily by immunological methods, and equally certainly but less easily by chemical procedures. Although the antigenic capacity of a protein depends on the entire large colloidal molecular structure, its specificity seems to reside in certain of the radicals of the molecule. There is evidence that a single protein may exhibit more than one specific immunological reaction, by virtue of possessing more than one such active radical. Group reactions exhibited by complex antigens from biologically related species may therefore depend either on the presence in these antigens of both common and specific proteins, or by the presence in different proteins of common and specific reactive radicals.
The immunological specificity of proteins may be artificially modified by introducing into them various radicals. It is then found that a single group in the added radical may by itself determine the specific im- munological behavior of the entire molecule. Also, not only the char- acter of this specific group is of influence but also its place in the added radical (Landsteiner), indicating that spatial relations are of importance in determining specificity. Apparently an antibody can react with various related but not identical antigens (i.e., protein molecules), specificity being always a quantitative matter which reaches its maximum when the antigen is reacting with an antibody produced by immunizing with an identical antigen. It is not yet definitely determined whether purely physical alterations in antigenic proteins are associated with alterations in specificity, for the protein molecule is so labile that any physical alterations are likely to be associated with chemical changes. As yet, however, we have no proved case of immunological specificity determined by purely physical properties.
Non-specific immunological reactions may depend on the existence of identical proteins in different species, or of different proteins with iden- tical groupings determining a common specificity. Also, the stimulation of the mechanism of antibody formation by one antigen (A) may arouse the production of antibodies for other antigens (B and C), especially if the individual has previously developed a capacity to produce anti- bodies for these antigens (B and C).
REFERENCES
1For a discussion of specificity of Fertilization see J. Loeb, “The Organism as a Whole.” Putnam’s Sons, New York, 1916.
2 As Bordet (“Studies in Immunity,” Bordet-Gay, New York, Wiley and Sons, 1909, p. 497) points out, this specific antigenic property is apparently not always dependent on some- thing essential for life, since bacteria may lose their agglutindbility when cultivated on
86 THE CHEMICAL ASPECTS OF IMMUNITY
special media. Dawson (Jour. Bact., 1919 (4), 133) has also found changes in agglutinability in colon bacilli grown on different media.
3“Blood Immunity and Blood Relationship,” Cambridge Univ. Press, 1904.
4 Jour. Amer. Med. Assoc., 1918 (70), 1273.
5 Berkeley, Univ. of Calif. Publn. Pathol., 1913 (2), 105.
6 Jour. Hygiene, 1910 (10), 177.
™The reactions given with antiserum for hen egg yolk by yolk proteins from the most varied species (Seng. Zeit. Immunitat., 1913 (20), 355) suggest that yolks also have common and specific antigens.
8 Literature reviewed by Wells and Osborne, Jour. Infect. Dis., 1921 (29), 200.
9 Wells, Jour. Infect. Dis., 1911 (9), 147.
10 Dale and Hartley, Biochem. Jour., 1916 (10), 408.
11 Fleischer et al., Jour. Immunol., 1920 (5), 437; 1921 (6), 223.
12 See Salus, Biochem. Zeit., 1914 (60), 1.
18 Graetz, Zeit. Immunitat., 1914 (21), 150; Hektoen, Jour. Amer. Med. Assoc., 1922 (78), 704.
44 Jour. Infect. Dis., 1916 (19), 183.
1° Uhlenhuth, Zeit. Immunitat., 1910 (4), 761; Hektoen, Jour. Amer. Med. Assoc., 1921 (77), 323 Jour. Infect. Dis., 1922 (31), 72.
16 Zeit. physiol. Chem., 1894 (18), 61.
17 Jour. Infect. Dis., 1924 (34), 433.
18 Zeit. f. Immunitat., 1910 (5), 690.
22 An important utilization of this well established instance of specificity for a given protein is that by Guyer (Jour. Exp. Zool., 1920 (31), 171), who has shown that anti- rabbit lens serum injected into pregnant rabbits at a proper time may produce specific congenital eye defects which are transmissible through subsequent generations.
20 Hektoen and Schulhof, Jour. Amer. Med. Assoc., 1923 (80), 386.
21 Review by Doerr and Pick, Biochem. Zeit., 1914 (60), 257; also review in Jour. Amer. Med. Assoc., 1924 (82), 1465.
2 Landsteiner and Simms, Jour. Exp. Med., 1923 (38), 127.
28 Tsuneoka, Zeit. Immunitat., 1914 (22), 567.
24 Jour. Path. and Bact., 1921 (24), 122, 217, 241, 256.
25 Jour. Amer. Chem. Soc., 1917 (39), 828. ;
26“The Crystallography of Hemoglobins.’ Carnegie Institution of Washington, Publication No. 116, 1909.
*7 Hektoen and Schulhof, Jour. Infect. Dis., 1923 (33), 224.
*8 Heidelberger and Landsteiner, Jour. Exp. Med., 1923 (38), 561.
* Jour. Gen. Physiol., 1923 (6), 131.
%0 Zeit. physiol. Chem., 1913 (88), 163.
31 Deut. med. Woch., 1904 (30), 9o1.
52 Chemical News, 1920 (120), 13.
Jour, Infect, Dis. rong (rz), gars) toro (19), ze3.
*4 Jones and Gersdorff, Jour. Biol. Chem., 1923 (56), 79.
*° Wells and Osborne, Jour. Infect. Dis., 1915 (17), 259.
86 Biochem. Zeit., 1912 (42), 399.
* They did not compare fibrinogen with isolated serum proteins. Antiserum for beef fibrinogen did not react with fibrinogen from pig blood, or conversely.
88 La Cellule, 1901 (18), 335.
8 Mitteil. med. Fak., Univ. Tokyo, 1917 (18), 195.
40 Stern, Arch. f. Hyg., 1922 (91), 165; Ruppel, Ornstein and Lasch, Zeit. Hyg., 1922 (97), 188.
“| Zeit. f. Hyg., 1922 (96), 191.
“Jour. Immunol., 1919 (4), 20.
# Zeit. f. Hyg., 1922 (96), 258.
““Tnfection and Resistance,” 1923, p. 110.
° Zeit. physiol. Chem., 1903 (37), 495; 1905 (44), 17.
“°“Die Wasserstoffionenkonzentration,” Berlin, 1914, p. 56.
“Biochem. Jour., 1914 (8), 541.
4 Biochem. Jour., 1921 &15), 187.
IMMUNOLOGICAL SPECIFICITY 87
4° Biochem. Zeit., 1923 (139), 321.
50 Biochem. Zeit., 1912 (38), 3313; 1913 (50), 360.
5 Deut. Arch. klin. Med., 1911 (104), 209.
52 Jour. Amer. Med. Assoc., 1921 (76), 929; Hektcen and Welker, Journ. Infect. Dis., 1924 (34), 440.
5s Bull. Johns Hopkins Hosp., 1922 (33), 119.
54 Everett, H. S., Bayne-Jones, S., and Wilson, D. W., Bull. Johns Hopkins Hosp., 1923 (34), 385.
S4a Hektoen et al., Jour. Amer. Med. Assoc., 1924 (83), 1154.
*5 Fleischer, Roussky Wratsch, 1908 (7, pt. 2), 1638.
58 Biochem. Jour., 1915 (9), 97.
5 Biochem. Jour., 1916 (10), 431.
58 Wells, Jour. Infect. Dis., 1909 (6), 506.
6° Emmerich, Zeit. Immunitat., 1913 (17), 299.
® Biochem. Jour., 1919 (13), 248.
61 Wien. klin. Woch., 1906 (19), 327.
® Freund (Biochem. Zeit., 1909 (20), 503) obtained iodized serum and egg albumin containing 6.5-8 per cent of iodin, which produced precipitins that were not species specific but specific for iodized proteins. The proteins were, however, according to these figures, far from completely iodized, and with iodized serum, and with iodized crystallized egg albumin which was known to be completely saturated with iodin, no loss of species specificity was detected by the anaphylaxis reaction (Wells, Jour. Infect. Dis., 1908 (5), 449). To be sure Schittenhelm and Strobel (Zeit. exp. Path. Ther., 1912 (11), 102) state that iodized serum proteins sensitize to iodized egg white, and conversely, but their work is not reported in sufficient detail to establish that sensitization by traces of the intoxicating protein may not have occurred.
® FE. P. Pick, Kolle and Wassermann’s Handbuch d. path. Mikroorganismen, 1912 (1), 685.
® “Studies in Immunity,” Bordet-Gay, 1909, p. 525.
® Biochem. Zeit., 1918 (86), 343. Gives bibliography.
® Landsteiner and Lampl, Zeit. Immunitat., 1917 (26), 133.
® Landsteiner and Jablons, Zeit. Immunitat., 1914 (21), 193.
® See Landsteiner and van der Scheer, Jour. Exp. Med., 1924 (40), 91.
8 Ergebn. d. Immunitatsfr., 1914 (1), 395.
70 Biochem. Zeit., 1919 (93), 106.
71 Biochem. Zeit., 1920 (104), 280,
72 Jour. State Med., 1920 (28), 293.
™ The “resonance theory’ of Traube assumes that the surface forces of reacting substances must harmonize, just as the vibration of one tuning fork starts vibrations in another fork only when the two are in harmony, or as electromagnetic waves incite resonance phenomena (see Zeit. £. XImmunitat., 1911 (9), 246 and 779).
74 Klin. Woch., 1922 (1), 949.
75 Biochem. Zeit., 1912 (47), 59.
78 Jour. Immunol., 1920 (5), 345.
77 See reviews by Jobling, Jour. Amer. Med. Assoc., 1916 (66), 1753.
78 Petersen, Wm. F., “Protein Therapy and Non-specific Resistance.’”’ Macmillan, 1922.
79 Jour. Infect, Dis., 19127 (21), 279:
89 See also Herrmann, ibid., 1918 (23), 457.
81 Clark, Zellmer and Stone, Jour. Infect. Dis., 1922 (31), 215; Khanolkar, Jour. Path. aud Bacts, 1924. (27), 18%.
82 Clark, Brit. Med. Jour., Feb. 24, 1923, p. 315.
83 See review of this subject by Jobling, Harvey Lectures, 1917 (12), 181; also Cowie and Calhoun, Arch. Int. Med., 1919 (23), 69; also Petersen.”8
“Jour. Infect. Dis., 19oxs (17), 247.
85 O’Brien, Jour. Path. and Bact., 1913 (18), 89.
8° See Hahn and Langer, Zeit. Immunitat., 1917 (26), 199; Trommsdorf, ibid., 1921 (32), 379.
Chapter IV The Nature of the Antibodies
After immunization with a given antigen, whether artificially or through the natural processes of infection, the blood of the animal, and possibly all the fixed tissues as well, usually come to exhibit the capacity to react in some way or other with the antigen, in a manner qualita- tively different or in a degree quantitatively greater than previously. We attribute this altered reactivity to the presence of “antibodies,” despite the fact that we have absolutely no knowledge what these anti- bodies may be, or even that they exist as material objects. Like the enzymes, we recognize them by what they do without knowing just what they are. We do not know whether they are specific molecular aggregates or merely physical forces dependent on altered surface energy of the same substances which were already present in the blood before the process of immunization was ever begun.
The methods for their recognition are several, and according to the procedure employed we designate the antibodies as precipitins, agglu- tinins, antitoxins, complement fixation antibodies, opsonins, cytolysins, anaphylactins, and so on. These names assume the existence of several different and distinct reactive substances or antibodies, and this assump- tion has been currently accepted as if it were an established fact, al- though this is far from the case. Of late there has been a growing tendency to doubt that there is any such considerable variety of anti- bodies, for we usually find after immunizing even with a single purified protein as antigen that we can demonstrate several if not all of these properties in the serum of the immunized animal.
ARE THERE DIFFERENT Types oF ANTIBODIES?
An attractive hypothesis is that there are two fundamental types of immunity reactions." One, having to do with substances which are essentially active poisons, neutralizes or inhibits their toxicity by direct chemical action. In this group come the antitoxins for bacterial toxins
(diphtheria, tetanus, etc.), and the antibodies for venoms, vegetable 88
LHE NATURE-OF THE ANTIBODIES 89
toxins (ricin, abrin, etc.) and various bacterial hematotoxins. It is to be noted that these toxic substances, the true toxins, are all similar to one another in being classed as large colloidal aggregates, resembling proteins, but not yet identified as proteins.
The other group of immunity reactions is concerned with defense against foreign proteins, whether toxic or non-toxic and whether in solution or aggregated into cells (bacteria, corpuscles, tissue cells). In all the reactions of this group we deal with processes that tend to alter the colloidal state of the foreign proteins, by making them larger aggre- | gates (precipitation, agglutination), or smaller aggregates (proteolysis, hemolysis, bacteriolysis, cytolysis), and in each case the reaction con- sists of two separable steps, sensitization and reaction. It is tempting to accept the view, championed especially by Friedberger, Zinsser, Dean, and Nicolle, that in this second group of reactions but one and the same antibody is concerned, and that all the reactions are essentially the same, differing merely in the method by which the reaction is demon- strated. There are, indeed, many facts capable of interpretation as supporting this hypothesis, but there are still other observations that do not harmonize with it, notably the lack of constant quantitative relations between the different reactions produced by the same immune serum, and as yet it is neither completely established nor disproved.
Evidence Favoring Unity of Antibodies
Zinsser has christened this the “unitarian” hypothesis, and presented much of the evidence in its favor. It is indeed a fact that the identity of precipitin and agglutinin has been commonly accepted, it being recog- nized that the precipitin reaction is obtained when the colloidal particles of a dissolved antigen are brought together into aggregates too large to remain suspended, and that the agglutinin reaction occurs when the colloidal antigen is already in large undissolved masses which also are brought together and precipitated. As Zinsser? says:
“Tf the antibody comes in contact with a very finely divided antigen, as in a bacterial extract or in, let us say, horse serum, if electrolytes are present and perhaps other necessary physical factors furnished by the presence of serum, etc., precipitation occurs,
“When we are dealing with whole bacteria of relatively large mass and correspondingly small surface exposure, agglutination is the result, and quantitative parallelism with the precipitin reaction is not to be expected because of the much greater dispersion of the antigen in the latter test.
90 THE CHEMICAL ASPECTS OF IMMUN TT Y
“When alexin is present complement fixation or. hemolysis or bacteri- cidal effects result, since the changes produced by the sensitization have permitted union with the complement.
“When there are leucocytes present the union makes possible the phagocytosis of the antigen, and when the antibody is absorbed by the
_cells of an animal, anaphylactic ‘sensitization’ occurs.”
The chief fact which led to the view that there are as many different antibodies as there are ways of demonstrating them, was the common observation, especially in the early days before the best technical methods were developed, that an antiserum might exhibit marked quantitative differences in its, activity in respect to the different sorts of reactions when tested with the same antigen. Sometimes, indeed, one reaction could be demonstrated when another could not be obtained at all. In immunizing with soluble protein antigens it is usually observed that the complement fixation reaction may first be demonstrated, but that later, when good precipitin reactions are obtained, the serum also exhibits the capacity to sensitize animals to the antigen, or, in the nomenclature of immunology, the precipitins and the anaphylactins appear together. This and other observations have given much support to the doctrine that the precipitin and the anaphylactic sensitizing antibody are one and the same. ‘The relative proportion of these two antibodies is usually found to run parallel in immune sera, and Richard Weil ° added
-much in support of their identity by finding that the precipitate obtained
when immune serum reacted with the specific protein, if injected into guinea pigs, conferred passive sensitization to the specific antigen. Since the precipitate formed in the precipitin reaction is known to consist chiefly if not entirely of precipitin, this experiment indicates that prob- ably the precipitin is also the immune body responsible for the anaphy- laxis reaction. | That there should not be a similar quantitative parallelism between the complement fixation antibodies and the precipitins is no argument against their identity, and there is much evidence in support of the assumption that both these reactions depend on the same antibody. As Dean has pointed out, the reason why the two reactions do not run a parallel course is not that they are caused by two different sets of antibodies, i.e., precipitins and amboceptors, but because they represent two phases or two stages of the same reaction, and it may not be pos- sible to demonstrate both stages under the same conditions. Went ° found also a parallelism between agglutinin and opsonic ac-'
THE NATURE OF THE ANTIBODIES gt
tivity of immune sera, which he interprets as indicating that both depend on the same antibody.
There is also a growing tendency to suspect that even antitoxins and precipitins are identical, since Ramon®* developed a method for assaying the strength of diphtheria antitoxic sera by determining its capacity to produce precipitin reactions with diphtheria toxin. Kraus ® has ob- served that although when fresh untreated toxin and antitoxic serum are used the precipitin reaction does give quantitative results approximately parallel to the antitoxic strength, yet if either the toxin or the antitoxin has been altered by preservation, aging, etc., the precipitin tests do not parallel either the toxic or antitoxic strengths of the solutions. This indicates the possibility that the antitoxin is not the same as the pre- cipitin, or the diphtheria toxin the same as the precipitinogen.
Significance of Quantitative Discrepancies
The failure to secure parallelism in titration of the several antibodies in a given serum lacks force as negative evidence in the face of the known great quantitative inaccuracy of most immunological measure- ments, and the ready inhibition of the reactions by factors which are often unknown, for the variables in all these tests are far too many and too uncontrollable to permit of exactitude. Thus, as Zinsser points out, we must not forget that agglutination and precipitation are actually only secondary phenomena, after the union of antigen and antibody has taken place, and are dependent upon a great many environmental factors which may not, to the same degree, influence phenomena in which alexin (complement), the leucocytes or the body cells of animals are involved. The flocculation reactions depend upon the presence and the concentration of electrolytes, upon the pH, upon mutual relations of concentration, and perhaps upon viscosity. The suspension equilibrium of the sensitized antigen must to some extent depend upon the varying factor of the inactive serum constituents carried into the union with the antibody, for we know that in precipitation reactions the bulk of the precipitate comes from the immune serum, and yet antibodies relatively free from protein can be split off from such a complex; this proves that, in the union, much inactive protein substance is carried along, which inevitably must influence reactions of flocculation.
It is not to be wondered at, Zinsser says, that in view of the factors mentioned above, agglutination and precipitation curves should not run parallel with the curves of other antibody functions. In regard to such lack of parallelism, while it has frequently been seen that agglutinating
92 THE CHEMICAL ASPECTS OF IMMUNITY
and precipitating functions are often weaker than other antibody effects, or even absent entirely in such sera, it has rarely been observed that they are powerfully and specifically present when other effects are lacking.
He concludes, “We do not wish by any means to convey the im- pression that we consider the ‘unitarian’ view as absolutely and rigidly proved. We do believe, however, that the denial of such a view neces- sitates the assumption that the injection of a pure antigen calls forth five or six fundamentally different reactions on the part of the tissue cells, a theory which would be justified only on the basis of incontro- vertible proof.”
Unity of Antibodies Agrees with Bordet’s Theories
The ideas of the French school of immunology, sponsored by Bordet, naturally fit into the unitarian hypothesis of the antibodies, for instead of recognizing the existence of definite antibodies as postulated by Ehrlich, they have always sought to explain the phenomena of im- munity as the result of simple physico-chemical reactions between the colloids of the serum and the antigen. These views are reviewed by Nicolle,*? who holds that there is but one sort of antibody which has the property, as first shown by Bordet, of rendering the antigen sus- ceptible to coagulation by the electrolytes present in the mixture, con- stituting the agglutinin and precipitin reactions, as well as one phase of the opsonic and anaphylactic sensitizations. If complement is also present in the mixture we get the lytic phenomena of immunity, i.e., hemolysis, bacteriolysis, cytolysis, and possibly the formation of the anaphylactic poison. When the coagulation reaction is especially vig- orous the aggregation of the antigen into dense masses seriously inter- feres with the lytic phenomena, through reducing the surface for attack by the enzymes; this explains why a strong precipitin serum may seem to be weak in lytic activity, without postulating the existence of separate antibodies.
Nicolle would even include the toxin-antitoxin neutralization in the same class. He maintains that true toxins and their specific antitoxins precipitate each other, and has based on this principle methods of titra- tion of antigens and antibodies. No evidence seems to have been advanced, however, that the precipitate obtained when antitoxic serum is added to toxin depends on the reaction with the toxin itself rather than with the proteins of the heterogeneous solution which contains the
toxin.
THE NATURE OF THE ANTIBODIES 93
Objections to the “Unitarian” Hypothesis
Despite all the arguments that may be advanced in favor of the “uni- tarian”’ hypothesis, at least as covering the antibodies for protein anti- gens other than true toxins, it is by no means universally accepted, for
arguments of more or less weight have been advanced against it. For
example, Longcope® found that white rats produce precipitins for foreign proteins but do not themselves become anaphylactically sensi- tive to the antigenic protein, nor does the rat serum which contains precipitins confer passive sensitization to guinea pigs into which it is injected, as does a rabbit serum which contains precipitin. This work,
however, failed of confirmation by the Parkers,’° who produced ana- . phylactic shock in white rats after both active and passive sensitization. ©
Dean *™ has called attention to the fact that in active tuberculosis the serum shows a decrease in opsonins which sensitize tubercle bacilli to phagocytosis, while at the same time there is an increase in the comple- ment fixation power of the serum. This would suggest that these anti- bodies, opsonins and complement-binding amboceptors, are not the same, but it is possible that there is some other explanation for the discrep- ancy; for example, in active tuberculosis substances may appear which interfere with phagocytosis or which prevent the effect of opsonins on tubercle bacilli without interfering with complement fixation.
Hektoen ?? has observed opsonins in the spinal fluid of dogs when agglutinins could not be demonstrated, although the blood of the animal exhibited both opsonic and agglutinative activity. This finding might suggest that the opsonin had been selectively secreted into the spinal fluid, while the agglutinin had been held back, supporting the view that these are distinct antibodies.
Landsteiner 1° has also pointed out that there are peculiarities in the specificity manifested by agglutinogens and precipitinogens which sug- gest that there is an essential difference in the chemical structures which determine the specificity of the two kinds of antigens, and hence presumably the antibodies must also differ. There are also such obser- vations as that of Mackie '* who found that a highly potent agglutinat- ing serum obtained by immunizing with a certain strain of colon bacilli did not agglutinate other strains of colon bacilli which appeared iden- tical with the antigenic strain by every other test; neither did these heterologous strains of colon bacilli absorb the agglutinins. But this same agglutinating antiserum gave complement fixation reactions with many strains of colon bacilli, even when these were in other respects
94 THE CHEMICAL ASPECTS OF IMMUNITY
very different from the original antigenic strain. These results suggest the existence of some distinct difference between agglutinins and com- plement fixation antibodies. Furthermore, Singer‘ reports the separa- tion of hemolytic amboceptors from the hemagglutinins in an immune serum.
One fact that may be advanced as failing to harmonize with the unitarian hypothesis, is the observation, repeatedly made, that the dif- ferent antibodies are contained in different fractions of the serum pro- teins. It has been found by some observers that antitoxin is in the water-soluble globulin fraction or pseudoglobulin, along with opsonins, hemolysins and antibacterial immune bodies, while in the euglobulin (water-insoluble) fraction have been found the precipitins, agglutinins, anaphylactins, and complement fixation antibodies. However, these find- ings are far from convincing for two reasons:
(1) The results of different investigators do not agree as to which fraction contains which antibody. For example, it has been found that the agent which carries the capacity for passive sensitization is in the albumin fraction,’® although it is generally accepted that the precipitin is the sensitizing antibody and that this is found in the globulin frac- tion (Funck).1° Again, Ruppel+* found the hemolytic amboceptor in the pseudoglobulin and the complement fixation antibodies in the euglobulin, although these are ordinarily considered as identical. He considers all the antibacterial protective elements of serum to be in the pseudoglobulin fraction, but Homer *® found the antidysenteric and antimeningococcic agents of the specific immune serums to reside chiefly in the euglobulins.
(2) The chemical nature of pseudoglobulin and euglobulin is not well defined. Chemical and physical studies have shown no appreciable dif- ference between the two unless it be the presence of lipoids in the euglobulin fraction, although immunological tests show them to be dif- ferent from one another. Ruppel and others claim that pseudoglobulin tends to pass over to the euglobulin fraction, or at least to become less and less water-soluble, so the purity and constancy of these fractions is open to question.’® But in any case the antibodies seem definitely associated with the globulins rather than the other proteins of the serum, and Dean maintains that in all the serum reactions there is an aggregation of globulin particles about the antigen, the main phenomena of which process are most readily explained by Bordet’s adsorption theory.
THE NATURE OF THE ANTIBODIES 95
THE NATURE AND PROPERTIES OF ANTITOXINS
Accepting the prevailing view that the antitoxins are different from all the other antibodies, in that they manifest the property of specifically neutralizing certain types of poison in a quantitative manner, we may consider what is known about them. Because of their great practical importance they have received much more study than the other anti- bodies, and enough has been learned of their properties to permit of considerable concentration of the antitoxic activity of sera. It is gen- erally agreed that the antitoxins are associated chiefly if not entirely with the water-soluble pseudoglobulin fraction of the serum. Adolf ?® has found that if antitoxic serum is subjected to electrodialysis until the conductivity is reduced to that of distilled water, the antitoxin dis- appears entirely, only albumin being left in solution and this is free from antitoxin. This experiment raises doubt as to the existence of a true water-soluble “pseudoglobulin” carrying the antitoxin, but it was found that during such dialysis the antitoxin comes out with the lat- ter part of the precipitated globulin, which undoubtedly corresponds to the globulin fraction commonly considered as pseudoglobulin whether it is truly water-soluble or not. But we do not know whether the anti- toxin is a definite substance adherent to the pseudoglobulins, or a form of pseudoglobulin itself which differs from normal pseudoglobulin in some change in physical properties or addition of active radicals, whereby it has obtained the capacity of neutralizing toxin.
In any event the antitoxins behave as colloids, moving toward the cathode in an electrical field,” diffusing little or not at all, their reac- tion curve resembling more an adsorption curve than the reaction curves of crystalloids, and being influenced by all conditions that influ- ence colloids.24. Whether the active groups (receptors) are secreted in a free condition in antitoxin formation, or combined with a large molecule, is unknown.
It is an interesting fact that the antitoxins formed by different ani- mals which have been immunized with a given toxin seem to be the same—horse serum, or sheep serum, or goat serum will neutralize diphtheria toxin if the animals have been made immune to this toxin; *? and, furthermore, their serum when introduced into the body of an entirely different animal, e.g., a guinea pig or a child, will neutralize diphtheria toxin within its body. Fqually important is the fact that the antitoxin for one toxin will not neutralize any other toxin; e.g., diphtheria antitoxin will not neutralize tetanus toxin, or conversely,
96 THE CHEMICAL ASPECTS OF IMMUNITY
but we have not the slightest idea just what chemical or physical differ- ence determines this specificity of the antitoxins.
According to Eisler,?* toxin injected in a single moderate sized dose produces about the same amount of antitoxin as the same amount divided into several injections, although the less soluble antigens noto- riously give much more antibody formation if they are injected in several small doses (Maisin) .*4
Are Antitoxins Globulins?
The fact that during immunization of an animal the proportion of pseudoglobulin in the blood is increased does not establish that anti- toxin actually is this new-formed globulin, for Meyer ** and his col- leagues found in their study of the blood proteins during immuniza- tion, that the proportion of globulins increases according to the severity of the intoxication, and not in any definite relation to the degree of immunity or antitoxin production. Not only bacterial poisons but also any protein antigen causes an increase in the globulin content of the immune serum (Doerr and Berger, lit.),?° but if the immunization is carried out carefully with small quantities of antigen there may be no increase of the globulin (Glaesner 7%"), indicating that antibody in- crease is not essentially associated with globulin increase, nor is the relation of globulin and antitoxin increase a constant one.?” The aver- age antitoxic horse serum contains 12 per cent albumin, 78 per cent of soluble globulin containing the antitoxin, Io per cent euglobulin ; whereas in normal non-immunized horses the proportion is 40 per cent albumin, 42 per cent pseudoglobulin and 18 per cent euglobulin. Homer 1° found that there seemed to be a slight difference between diph- theria and tetanus antitoxin, more of the latter being in the portion of the protein in the pseudoglobulin-euglobulin zone. As far as we know the only attempt made to determine whether the antitoxin-containing pseudoglobulin is chemically different from normal pseudoglobulin showed no very distinct differences, although the histidine nitrogen
figure seemed to be lower.?®
Although the immunization of an animal is usually accompanied by a marked rise in the proportion of globulin in the serum, studies of the physical properties of the serum have not shown much change that could be correlated with the presence of specific antibodies. Du Nouy *° found no change in the refractive index of the serum in immunized animals, but did observe a difference in the effect of time on the sur- face tension. When serum is diluted the drop in surface tension which
THE NATURE OF THE ANTIBODIES 97
normally takes place on standing becomes more marked, and this “time drop” shows a maximum in certain dilutions, usually about 1 to 10,000. Immune serum shows 50 to 100 per cent greater time drop than before immunization. The meaning of this change is not known, beyond indi- cating that there is “in certain immunity states a decrease in the surface energy of the substances normally absorbed in function time in the surface layer.”
A difference in the absorption band between the wave lengths 2950 and 2400 has been observed in the serum of different species of ani- mals when immunized,*° but the significance of this change is also unknown.
An interesting hypothesis advanced by Ostromuislenskii *' is that antitoxin is merely normal serum globulin physically altered by the toxin which it has adsorbed. He claims that normal serum globulins combine with toxins, and if after they have been thus united for some time they are dissociated by the action of acids, free specific antitoxin may be obtained. If this is true it should be possible to prepare anti-
toxins im vitro without animal immunization, and the practical impor- |.
tance of this suggestion entitles it to further consideration.
The Resemblance of Antitoxins to Proteins
The relation of antitoxins to proteins has also been investigated by permitting digestive enzymes to act on antitoxic serum. Pick digested the antitoxin-containing globulin of horse serum for several days with trypsin; after five days, when part of the protein was still not digested, the antitoxin was but little impaired in strength; after nine days, when most of the protein was digested, the antitoxin had lost two-thirds of its strength. This indicates a considerable resistance of antitoxin to trypsin, but also shows that it is affected in much the same way as the globulin (which is itself very resistant to trypsin) and therefore is presumably of similar nature. Antitoxin seemed to be much more , rapidly destroyed by pepsin-HCl digestion than by trypsin, in which | respect it again resembles the serum globulin. Berg and Kelser °?” found that trypsin and pepsin destroy the antitoxin and serum pro- teins at about the same rate, but their failure to observe “significant chemical changes” in the proteins of serum acted upon by weak acid or alkali that slowly inactivate antitoxin, does not seem to warrant their deduction that antitoxin is non-protein.
In favor of the view that antitoxin is a definite protein body or attached to one, is the fact that it is not carried down in indifferent
f
j
98 THE CHEMICAL ASPECTSVOP IMMUNITY
precipitates, as are the enzymes, but comes down always in a certain fraction of the protein precipitates, e.g., we can precipitate all the serum albumin from an antitoxic serum, and it does not carry down with it any of the antitoxin. Another important point has been brought out by Arrhenius and Madsen,** who determined approximately the molecular weight of toxin and antitoxin by means of their rate of diffusion, and found that the toxins (diphtheria toxin and tetanolysin) diffused ten or more times as rapidly as the corresponding antitoxins. Gelatin filters also hold back antitoxin and let toxin pass through, and toxins diffuse into cells which seem to be impermeable for the antitoxin. This indicates that the antitoxin molecules are much larger than the toxin molecules, agreeing with the idea that antitoxin is of protein nature and that toxin either is not protein or is smaller than most protein molecules. It is difficult, however, to accept the figures of Arrhenius,** which give the size of the antitoxin molecule as 100 times as great as that of the toxin molecule, which in turn may have a molecular weight of 15,000.
Another point in favor of the protein nature of the antitoxins is that precipitin reactions with antitoxic serum throw down the anti- toxin in the precipitate, and may even make the antibody inactive when there is no visible precipitate (Eisler).*° The same is true of