Advances in Immunology Volume 26

ADVANCES IN Immunology V O L U M E 26 CONTRIBUTORS TO THIS VOLUME J . DONALDCAPRA K. EICHMANN TONY E . HUGLI JAN ...

Author: Dixon | Kunkel

11 downloads 952 Views 18MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

ADVANCES

IN

Immunology

V O L U M E 26

CONTRIBUTORS TO THIS VOLUME

J . DONALDCAPRA K. EICHMANN TONY E . HUGLI JAN KLEIN NORMANR. KLINMAN HANS J . MULLER-EBERHARD NOLAN H. SIGAL ELLEN S. VITETTA

ADVANCES IN

Immunology EDITED BY

FRANK J. DIXON

H E N R Y G. K U N K E L

Scrippr Clinic and Research Foundation La Jolla, California

The Rockefeller University New Yo&, New Yo&

V O L U M E 26

1978

ACADEMIC PRESS

New York

Sun Francisco

A S u b s i d i a r y of Harcourt Broce Jownovich, Publishers

London

COPYRIGHT @ 1978, BY ACADEMIC PRESS, INC. ALL RIGHTS RESERVED. NO PART OF THIS PUBLICATION MAY BE REPRODUCED OR TRANSMITTED IN ANY FORM OR BY ANY MEANS, ELECTRONIC OR MECHANICAL, INCLUDING PHOTOCOPY, RECORDING, OR ANY INFORMATION STORAGE AND RETRIEVAL SYSTEM, WITHOUT PERMISSION IN WRITING FROM THE PUBLISHER.

ACADEMIC PRESS, INC.

111 Fifth Avenue, New York, New York 10003

United Kingdom Edition priblished by ACADEMIC PRESS, INC. (LONDON) LTD. 24/28 Oval Road, London NWI 7DX

LIBRARY OF CONGRESS CATALOG CARD NUMBER:61-17057 ISBN 0-12-022426-7 PRINTED IN THE UNITED STATES OF AMERICA

CONTENTS

.

LIST OF CONTFUBUTORS . . . PREFACE

.

. .

.

. .

.

. .

. .

. .

. .

. .

. .

vii ix

C3a a n d C5a TONYE. HUGLI AND HANS J . MULLER-EBERHARD

Anaphylatoxins:

I. Introduction

.

.

.

.

.

11. Historical . . . . . . 111. Fonnation and Control of Anaphylatoxins

.

.

.

.

.

. .

.

. .

. .

.

.

.

.

,

. . .

. .

. . .

,

.

48

. .

56

,

63 64

IV. Clinracterizatioii of the Anaphylatoxins . V. Function of Anaphylatoxins . . . \'I, Roles of C3a and C5a i n Inflanlnlatioli and Acute V I I . Corrcluding Remarks . . . . . References , , , . , , ,

.

.

Shock

. .

. ,

.

1

-1 6 14 30

44 47

H-2 Mutations: Their Genetics a n d Effect on Immune Functions JAN

KLEIN

I. Introduction. . . . . . . . . 11. T h e H - 2 Syllabus . . . . . . . 111. Basic Terms in Mutation Genetics . . . . IV, Methods of Histocompatibility Mutation Study . . V. Mutation Rates , . . . . . . . VI. Genetics of Available H-2 Mutations . . . VII. Biochemistry ofH-2 Mutations . . . . . VI11. Origin of H - 2 Mutations . . IX. Mutations arid Polymorphism . . . . X. Effect off1-2 Miitations on Immune Functions . XI. What Have H - 2 Mutations Contributed to Immunology? XII. Perspectives , . . . . . References . . . . . . . . .

. . . . .

. . . .

.

. .

.

.

.

,

.

,

. . . .

. . ,

. .

fjo

73 91 96 99

.

106 108

. . .

137 139

. . . . . . . .

148 148 1M 157 167

.

141

The Protein Products of the Murine 17th Chromosome: Genetics a n d Structure

ELLEN s. VITETTA

I . Introduction 11. &MicrogIol)ulin 111. ThrT/t Region .

AND

.

.

,

.

.

J. DONALDCAPRA

. .

. .

IV. Thcb K and D Regions . \I. T h e I Region . . VI. T h e S Region . . VII. Th c~Region 1)etween H-2D and TLa . V I I I . mj . .

. .

. .

.

. .

. .

.

,

,

.

177 181 182

vi

CONTENTS

IX. Perspectives References .

.

.

.

.

. .

.

.

. .

. .

.

.

.

.

.

.

.

.

.

.

. . . .

.

.

.

.

. .

. .

. .

.

.

.

.

.

. .

.

.

186 188

Expression and Function of ldiotypes on lymphocytes

K. EICHMANN I. Introduction . . . . . . . 11. Anti-idiotypic Reagents . . . . . 111. Analysis of B- and T-cell Receptor Idiotypes . IV. Fnnctional Role of Lymphocyte Receptor Idiotypes V. Summary . . . . . . . . References . . . . . . . .

. . .

. .

195 199 211 234 247 248

The 6-Cell Clonotype Repertoire

NOLAN H. SIGALAND NORMAN R. KLINMAN I. 11. 111. IV. V. VI.

Introduction . . . . . The Clonal Selection Hypothesis . Methods of Clonotype Identification . Defining the Adult B-Cell Repertoire . Defining the Neonatal B-Cell Repertoire Conclusion . . . . , References . . . , . .

SUBJECTINDEX . . . . CONTENTS OF PREVIOUS VOLUMES

. .

. .

. .

.

.

. .

.

,

.

.

.

. .

.

.

.

.

.

.

.

. .

.

.

.

.

.

.

. . . .

.

.

.

.

. .

.

.

.

. .

,

255 256 276 294 316 327 328 339 34 1

LIST OF CONTRIBUTORS Numbers in parentheses indicate the pages on which the authors’ contributions begin.

J. DONALDCAPRA,Depurtment of Microbiology, Unioersity of Texas Southtuestern Medical School, Dallm, Texus 75235 (147) K. EICHMANN, Institiit f u r lniinuriologie und Genetik, Deutsches Kreb.~forschuiigsZe12tri~?n, Heidelberg, Federul Republic of Gerniuiq (195)

TONYE. HUGLI,Departmelit of Moleculnr lninwrzology, Research Institute of Scripps Clinic, La Jolla, California 92037 (1) JAN KLEIN, Department of Microbiology, Universit!/ of Texas Health Science Center, Dallas, Texas, and Dey?artnic>ritof lmmunogenetics, The Mcix Planck Institute for Biology, Tiihitigen, Federal Republic of Gerinun!j (55) NORMANR. KLINMAN, Department of PathohgtJ, University of Penn.s~\lvciriici School of Medicine, Philadelphia, Perins~ylvunia(255)

HANS J. MULLER-EBERHARD, Departnaent of Molecular lmmunology, Research Institute of Scripps Clinic, La Jollm, California 92037 ( 1 ) NOLANH. SIGAL,Depnrtment of Patholog!\, Uniz;er,sity of Penns[ylouniu School of Medicine, Philutlel~rhiuPeniisyl.uaizia (255)

ELLENS . VITETTA,Depcirtinent c!f Microbiology, Unimrsity of Texas Southtuestern Medicul School, Dnllns, Texas 75235 (147)

vii

This Page Intentionally Left Blank

PREFACE

The articles in the present volume illustrate the pervasiveness of current immunology. This discipline extends from the specialities of genetics, both fundamental and applied, to stnictural chemistry, particularly of antigen receptors and cellular antigens-the molecules responsible for individuality and for the integration of the immunologic network-and finally to those biological sciences that deal with the production and action of effector molecules which carry out immunologic design. Just as immunologists have contributed much to the understanding of gene structure and function, cell-cell interactions, and the process of inflammation, so have the talented investigators in these separate fields contributed much to immunology. The present volume is the product of such a diverse group of researchers, whose quite separate interests come together to advance our knowledge of immunology. The first article, by Drs. Tony Hugli and Hans Muller-Eberhard, deals with the structural and functional characterization of anaphylatoxins, spasmogenic substances released during complement activation. These low molecular weight peptide fragments of C3 and C5 elicit a variety of cellular responses, which implies that they play a significant role in inflammation and acute allergic reactions. Their stimulation of multiple cell types at pico- and femtomolar concentrations suggests a hormone-like action via specific cell surface receptors. By using the information now available defining the amino acid sequences of these molecules, much of which has come from the authors’ laboratory, it is now possible to synthesize oligopeptides with the properties of anaphylatoxins and these compounds should serve as effective tools for analyzing further the mechanism(s) by which these potent substances function. The use of H-2 mutations to study functions of the murine major histocompatibility complex began only a few years ago, but already has yielded much information. For example, this line of endeavor has shown that (1)a single H - 2 locus can control a variety of immunologic phenomena, (2) strong MLC reactions can occur to H - 2 K and D antigens as well as to Z region antigens, ( 3 )histogenetically and serologically detectable antigens are of comparable complexity, and (4)small changes in the H-2 molecule, perhaps a single amino acid substitution, can have drastic effects on phenotype. Moreover, this work has provided numerous leads for further research on the molecular determinants of immune function. In the second contribution, Dr. Jan Klein, an outstanding authority in the field of immunogenetics, sumix

X

PREFACE

marizes this progress and puts this experimental approach in appropriate perspective. Since this field will undoubtedly increase in activity and importance, his review serves well to set the stage and orient the reader for what is to come. Another approach to the study of the MHC, viz., via the analysis of the protein products of the murine 17th chromosome, is treated in the third article by Drs. Ellen Vitetta and Donald Capra, both of whom are important contributors to this field. Many of the gene complexes that map between the centromere and TLa appear to play an essential role in cell-cell interactions and the control of immune responsiveness, embryogenesis, and differentiation. All the products of these genes are glycoproteins, most of which are expressed on cell surfaces presumably as receptors and often in association with p-microglobulin. This review analyzes the current state of biochemical knowledge about these glycoproteins (including their primary structure), emphasizes their common features and interrelationships, and explores the implications of these qualities for the evolutionary origins of the associated genes. It is suggested that this segment of chromosome 17 acts " in some ways as a super gene" in which the genes and gene complexes coding for a number of functionally interrelated molecules are closely linked, presumably as a result of selective pressures. One of the most significant of recent developments in immunology has been the recognition of idiotypes of antibodies and lymphocyte receptor molecules and of the formation of anti-idiotypic antibodies, which may serve as regulators of immune function. Since the individual antigenic specificity of an antibody, i.e., its idiotype, must depend upon its unique variable region sequences, the idiotype is an expression of antibody V region genes and as such has been successfully exploited as a probe in a variety of fundamental immunologic and genetic studies. The article by Dr. K. Eichmann provides an excellent background for this subject and emphasizes two new areas of research to which the author is a major contributor. The first concerns the molecular nature and identity of cell surface receptors and the structure of their idiotypic determinants. The second is the possible function of these determinants as key elements of regulation within a network of immunocytes that are interlocked to one another through idiotype-anti-idiotype interactions. Implicit in the clonal selection hypothesis is the existence of a repertoire of genetically determined B-cell clonotypes. T h e genetic determination, the development, and the regulation of this library of immunologic responses are the subjects of the last article, by Drs. Nolan Sigal and Norman Klinman. These authors draw on their exten-

PREFACE

xi

sive experience in this field to put into perspective the clonal selection scheme with its more than lo7clonotypes and multiple forms and levels of control. In keeping with the genetic basis of the scheme, all members of an inbred strain express the same repertoire of major clonotypes, but it is not yet possible to prove that every clonotype is shared, an observation that would rule out a role for random somatic events in immunologic development. The authors also trace the maturation ofthe repertoire from its restricted state in the neonate to its full development in the adult and note the many levels for control of this expression. We wish to thank the authors for their long and painstaking efforts in the preparation of these excellent reviews, and the publishers, who present this product with great skill.

FRANKDIXON HENRYKUNKEL

This Page Intentionally Left Blank

ADVANCES IN IMMUNOLOGY, VOL. 26

Anaphylatoxins: C3a a n d C5al TONY E. HUGLI AND HANS J. MULLER-EBERHARD~ Department of Molecular Immunology, Research Institute of Scripps Clinic, l a l o l l a , California

111. Formation IV. Characteri C. Chemical Properties ...................................................................................... V. Function of Anap ....................................................................... A. Ceilular Effect ....................................................................... B. Chemotaxis ..... ........................................................................... C. Effects on Smooth Muscle and Other Tissues ..... .... D. Systemic Effe ...................................................................................... VI. Roles of C3a and d Acute Shock ................................ VII. Concluding Remar ..... ...... References ............................................................................................................

1 2 6 14 14 17 21 30 30 34 38 42 44 47 48

I. Introduction

Activation of the coagulation, fibrinolytic, and complement systems leads to the formation of primary and secondary reaction products, both of which play important biological roles. While the primary products enter the main pathway of their respective systems, the secondary products constitute activation peptides such as the kinins, fibrinopeptides and anaphylatoxins. Anaphylatoxins are low molecular weight, biologically active polypeptides that are released during complement activation from C33 and C 5 and are commonly denoted as C3a and C5a. For the purpose of this review, anaphylatoxins will be defined as spasmogenic factors derived from serum complement components. They are functionally defined by their actions on the vascuThis is publication number 1433 from Research Institute of Scripps. This work was supported by U.S. Public Health Service Program Project Grants A1 07007 and HL 16411 and by Grant HL 20220 from the National Heart, Lung and Blood Institute. Dr. Muller-Eberhard is the Cecil H. and Ida M. Green Investigator in Medical Research, the Research Institute of Scripps Clinic. The terminology for the complement components conforms to the recommendations of the World Health Organization Committee on Complement Nomenclature (1968). 1 Copyright" 1976: hy Academic Press, Inc. All rights of reproduction in any form reserved. ISBN 012-02242fi-7

2

TONY E. HUGLI AND HANS J. MBLLER-EBERHARD

lature, smooth muscle, mast cells, and certain types of peripheral blood cells. C3 and C 5 are each composed of two polypeptide chains, the alpha chain (a)having a molecular weight of 100,000-120,000 and the beta a molecular weight of 70,000-80,000. Anaphylatoxin formachain (/I) tion results from a highly selective, proteolytic scission of the a-chain of the parent molecules. Current evidence indicates that C3a and C5a originate from the NH2-terminus of the a-chains of C3 and C5, respectively. Human blood and the blood of other mammals have two distinct mechanisms of complement activation. These mechanisms can be set into motion by immune complexes, by microbial or fungal polysaccharides, and, presumably, by certain virus particles (Cooper et al., 1976).The two mechanisms are called the classical and the properdin, or alternative, pathways of complement. The classical pathway involves complement components C1, C2, and C4, which form specific enzymes capable of cleaving C 3 and C5 and liberating anaphylatoxins (Muller-Eberhard, 1975).The corresponding alternative pathway involves enzymes that are constituted from C3, C 3 proactivator (Factor B), C3PA convertase (Factor D), and properdin (Medicus et aZ., 1976) and produce an analogous selective cleavage of components C 3 and C5. Current evidence suggests that, whether activation of complement proceeds via the classical or the alternative mechanism, the resulting C 3 or C 5 fragments are identical both biologically and chemically. The C3a and C5a anaphylatoxins have been isolated and characterized in recent years, permitting their detailed molecular definition. Their multiple activities imply that anaphylatoxins play a significant role in inflammation and possibly in some acute allergic reactions. A recent comprehensive review written by Ryan and Majno (1977)examines the action of numerous plasma mediators involved in the acute inflammatory response. The remarkable potency of the anaphylatoxins makes them likely candidates for figuring prominently in the various processes of inflammation. I I . Historical

Sixty-seven years ago at the University of Berlin, Friedberger (1910) observed that a potent toxin was formed when fresh serum was treated with aggregated immunoglobulin. H e demonstrated that serum taken from a guinea pig and incubated with an immune precipitate was toxic, often fatal, when re-injected into the donor animal. Armed with the knowledge that the principal in serum producing this toxin was

ANAPHYLATOXINS

3

thermolabile and could be generated in serum without adding foreign protein (Novy and deKruif, 1917) it was concluded that the toxic factor might be derived from the complement system, at that time poorly understood. Since the physiological effects of the toxic substance in guinea pigs mimicked the symptoms of anaphylaxis or anaphylactic shock as defined by Portier and Richet (1902), Friedberger termed the factor an “anaphylatoxin.” Subsequent studies showed that anaphylatoxin formation could be initiated by substances other than immune precipitates. Among these additional activators were complex polysaccharides, such as agar (Bordet, 1913a), inulin (Bordet and Zunz, 1915),starch (Nathan, 1913a,b), and dextran (Hahn et al., 1954). All the activating agents mentioned are capable of generating anaphylatoxin activity in sera from rats, rabbits, guinea pigs, or pigs. Conceptualizing his observations, Friedberger proposed the “humoral anaphylatoxin theory,” which claimed that humoral factors were directly responsible for the shock-inducing behavior of activated serum. Later, when histamine was found to be released from circulating cells b y the action of anaphylatoxins, a “cellular histamine theory” became accepted. Proposal of a cellular release mechanism proved to b e particularly attractive for explaining the shock phenomenon, since the anaphylatoxin mediates cellular vasoamine release and thereby may act indirectly to promote its spasmogenic effect. However, rats did not express the typical shock syndrome when activated rat serum was injected, even though rat serum proved to be considerably more toxic in guinea pigs than did activated guinea pig serum. This phenomenon was attributed to the fact that anaphylatoxin causes very little histamine liberation in rat serum (Hahn et d.,1954), but excites liberal histamine release in both guinea pig serum and lung (Hahn and Oberdorf, 1950; Hahn, 1954). Although it is now well established that the anaphylatoxins can release histamine from mast cells and basophils, the issue of whether the systemic and tissue responses to these factors are entirely direct or indirect effects still remains unresolved. A review by Hahn, which appeared in 1960, accurately summarized the accumulated knowledge concerning the anaphylatoxins at that time. It contained fewer than 45 references that covered more than 50 years of investigating the biological actions of anaphylatoxin. Near the end of his treatise, Hahn addressed himself to a discussion of the differential role that anaphylatoxins might play in the phenomena of anaphylactic and Forssman shock. It had been established that the ability to form anaphylatoxin did not diminish in the serum of an animal experiencing anaphylactic shock, whereas it diminished b y about 50%

4

TONY E. HUGLI AND HANS J. MULLER-EBERHARD

in the serum of animals with Forssman shock (Giertz et d.,1958). Hahn concluded from the difference between the two types of systemic shock that Forssman shock is mediated by humoral mechanisms, probably involving anaphylatoxin, and that anaphylactic shock is probably mediated by cellular events. Hahn’s conclusion has been borne out by investigations of the past 10 years. The acute allergic reaction is now attributed to antibody of the IgE class, which in conjunction with allergens effects mediator release from mast cells or basophils via a complex cellular process that is independent of complement and hence of anaphylatoxin (Ishizaka et al., 1970; Kaliner and Austen, 1973). However, there are acute allergic reactions caused by antibody classes other than IgE that are capable of activating complement. Thus, the extent of participation of the anaphylatoxins in allergic reactions remains to be defined. As mentioned, in 1917 Novy and d e h i f had already speculated that anaphylatoxin activity was a function of a protein arising from serum, not a direct interaction of immune complexes with the tissues. Osler et al. in 1959 furnished the first experimental evidence indicating that anaphylatoxin is indeed derived from complement, specifically from its “third component.” The classical “third component” has since been shown to consist of six proteins, namely C3, C5, C6, C7, C8, and C9. Vogt and Schmidt (1966) provided further evidence demonstrating that the classical rat anaphylatoxin was generated by a serum protease activated by contact with Sephadex and that protease seemed similar to the anaphylatoxin-forming activity of cobra venom. Toward the end of the 1960s Lepow and associates (1969), working with human complement proteins, showed anaphylatoxin to be a low molecular weight product of C3, and Jensen (1966), working with guinea pig complement, showed it to be a derivative of C5. The discrepancy was resohed by the demonstration that a low molecular weight fragment with anaphylatoxin activity can be enzymtically cleaved both from human C3 and from human C5 (Cochrane and Muller-Eberhard, 1968). These two anaphylatoxins were unique and were shown to have distinct biological specificities. At about the same time, similar fragments were cleaved from analogous complement proteins of guinea pig serum (Jensen, 1967),and of pig serum (Vogt et al., 1971b), and the corresponding C3 and C5 fragments were subsequently isolated directly from activated pig and human serum (Vallota and Muller-Eberhard, 1973; Hugli et al., 1975a; Corbin and Hugli,

1976). As realistic amounts of highly purified C3a and C5a were made available, it became possible to delineate similarities and differences

ANAPHYLATOXINS

5

of their effects on cells, including mast cells, basophils, neutrophils, monocytes, macrophages, and smooth muscle cells. It also became possible to tackle the comparative structural analysis of the two peptides, a feat that led to the complete elucidation of their respective covalent structures (Hugli, 1975a; Fernandez and Hugli, 197713) and to the realization that they share a measure of sequence homology (Fernandez and Hugli, 1977a,b). Thus, the similarity in biologic function was reduced to a similarity in chemical structure. It is anticipated that their functional differences will eventually be attributed to discrete structural differences. The use of synthetic oligopeptides with anaphylatoxin-like activities (Hugli and Erickson, 1977) will aid in identifying the chemical basis of anaphylatoxin specificity and activity. The remarkable biological potency of the anaphylatoxins emphasized the probable existence of very strict control mechanisms. A major regulatory principle was found in serum; it appeared to be an enzyme with carboxypeptidase B specificity (Bokisch et al., 1969).Removal of the COOH-terminal arginine residue from either C3a or C5a totally abrogates anaphylatoxic activity (Bokisch and Miiller-Eberhard, 1970; Fernandez and Hugli, 1976). Discovery of the serum enzyme and, subsequently, its effective inhibition allowed both anaphylatoxins (e.g., C3a and C5a) to b e produced in whole serum without being inactivated (Vallota and Miiller-Eberhard, 1973). With a growing understanding of the classical complement pathway in the 1960s and the alternative pathway in the 1970s, the enzymes directly responsible for anaphylatoxin liberation became definable. Although the convertases of the classical pathway differ from those of the alternative pathway in subunit composition, their overall structures are similar and their substrate specificities are identical. T h e essential zymogen of the alternative pathway [i.e., C3PA (Factor B)] was detected because it was required for cleavage of C 3 or C 5 (Jensen, 1967)b y cobra venom factor, a process later described almost without reference to anaphylatoxin formation (Gotze and Miiller-Eberhard, 1971). On the other hand, Stegemann et al. (1964, 1965) and Vogt and associates (1966, 1969a) employed cobra venom factor extensively as an activator of anaphylatoxin in serum without further reference to complement activation. Today it is clear that the anaphylatoxinforming enzymic site of the alternative pathway resides in Factor B, as that of the classical pathway resides in component C2. The present view deals primarily with the studies of the past 15 years, which permits us the advantage of hindsight in explaining many of the questions posed during the first 50 years of anaphylatoxin research. That is not to say that the early contributions go unnoticed or

6

TONY E. HUGLI AND HANS J. MfjLLER-EBERHARD

unappreciated, for they have indeed provided valuable insights and have stimulated many of the studies in progress today. Ill. Formation and Control of Anaphylatoxins

Investigations following the discovery of anaphylatoxin activity in serum led to the description of two well characterized polypeptides, the C3a and C5a anaphylatoxins. However, our present knowledge of their chemical nature and their mechanisms of formation in serum did not come easily. For example, we now know that serum contains an indigenous, enzymic control mechanism capable of rapidly inactivating both anaphylatoxins. The resulting inactive products are little changed chemically and are immunologically indistinguishable from their active forms. Until recently this anaphylatoxin inactivator remained unrecognized. Ignorance of the existence of an inactivator caused uncertainties that lasted for more than 50 years. During that time, investigators puzzled over the presumed number of different anaphylatoxins and the circumstances prompting the differential expression of anaphylatoxin activity in sera taken from various animals. The anaphylatoxin inactivator is now known to be an exopeptidase that functions in the serum of all higher mammals studied to date. Since the apparent efficiency of the inactivator varies with the source of the serum, it must be concluded that the inactivator constitutes a major factor in the expression of anaphylatoxin activity in vitro and in vivo. Earlier implication of complement in the generation of anaphylatoxin activity had been based on circumstantial evidence. Then Osler and his colleagues (1959) performed experiments designed to answer this very question. By demonstrating that anaphylatoxin formation occurred concurrently with complement fixation and C3 consumption but failed to appear when complement activation was prevented by heating, by metal chelation, or b y phlorizin, an inhibitor of C3 activation, they concluded that anaphylatoxin forms as a direct consequence of complement activation. Later, observations by Ratnoff and Lepow (1963) and others (Smink et al., 1964; Dias da Silva and Lepow, 1965) showed that isolated human C1 esterase can induce a response characteristic of anaphylatoxin when injected into guinea pig skin, proving that an established complement enzyme was indeed capable of generating anaphylatoxin activity. In addition, these studies provided some of the earliest evidence that anaphylatoxin formation might be an enzymic process, a concept questioned at the time. In experiments that followed, the active factor induced by C1 esterase was compared with that generated in serum by the introduction of enzymically inert

ANAPHYLATOXINS

7

substances such as complex polysaccharides. Both activities were biologically identical (Dias da Silva and Lepow, 1965).Such comparison is possible because smooth muscle is specifically desensitized by repeated applications of the anaphylatoxin, a phenomenon known as tachyphylaxis. Stimulation of smooth muscle by C3a is followed by a state of unresponsiveness to a subsequent stimulation b y C3a regardless of the animal origin of the C3a. The same phenomenon applies to C5a. However, C3a will stimulate smooth muscle rendered unresponsive to C5a, and vice versa. These investigators also showed that formation of anaphylatoxin by C 1 esterase was inhibited either by EDTA or by salicylaldoxime and phlorizin, all of which prevent C3 activation, a finding that implicated additional complement proteins in the process of anaphylatoxin formation. In Germany, Vogt and Schmidt (1966) independently concluded that anaphylatoxin formation was an enzymic process. Their conclusion was based on their ability to separate the contact-activation event from the step that actually produced a functional anaphylatoxin. They succeeded in isolating a plasma fraction, reportedly not C1 esterase, that was activated either b y Sephadex or Bio-Gel. The activated plasma fraction could then be mixed with a fraction containing the anaphylatoxin precursor “anaphylatoxinogen,” in the complete absence of contact activators, with the resultant formation of anaphylatoxin. Although Vogt and Schmidt correctly concluded that the activation was enzymic, they erroneously surmised that the enzymes involved were unrelated to complement. Unknowingly, they had demonstrated that anaphylatoxin had formed via the alternative (properdin) pathway of complement activation. This realization would come later, when the C3 activator system was described and the process was shown to activate C 3 without involving components C1, C2, and C4 (Brade and Vogt, 1971a; Gotze and Muller-Eberhard, 1971). Regardless, the impact of the observations of the 1960s was far-reaching, since thinking at that time was divided between ascribing anaphylatoxin formation to an enzymic process or simply to physical aggregation of selected serum components. Giertz and Hahn (1961)addressed this very problem. In their discussion they considered the formation of toxic aggregates, action of denatured toxic proteins and assorted physical interactions which may lead to enzyme activation as possible modes of anaphylatoxin formation. However, by late 1966 complement was convincingly implicated in the formation of anaphylatoxin and the process was understood to be enzymic. Definitive studies identified C 3 and C 5 as the anaphylatoxinogens after these proteins were recognized as complement components and

8

TONY E. HUGLI AND HANS J. mLLER-EBERHARD

became available in purified form. Dias da Silva et al. (1967)detected anaphylatoxin activity in a mixture containing purified C E , C2, C3, and C4. Since they found that the activity resided in a small fragment derived from the C3 molecule, they defined C3 as the precursor of human anaphylatoxin. Jensen ( 1967) reached a different conclusion working with functionally purified components from guinea pig serum. H e found the anaphylatoxinogen to be associated with a serum fraction containing C5. The activity generated from this fraction corresponded in specificity to classical anaphylatoxin activity, as evidenced by desensitization of the guinea pig ileum. [Jensen’s serum fraction contained C3, and presumably C3a was formed; however, this activity might well have been overshawdowed by the more potent C5a.l In contrast, the C3-derived fragment did not cause cross-desensitization to classical anaphylatoxin, and therefore Dias da Silva and co-workers (1967) interpreted this behavior as evidence for a “new” anaphylatoxin. Their results with isolated human C5 were inconclusive since no classical anaphylatoxin activity could be demonstrated when they added C5 to the reaction mixture containing C E , C2, C3, and C4. In retrospect, the disparate results obtained by Jensen and Lepow and his group are entirely explicable. While C3 convertase functions efficiently in both a particle-bound and soluble form, C5 convertase is an efficient enzyme only in its bound form. Jensen used the cell-bound enzyme, while Lepow’s group relied on the soluble system. It was Jensen (1967) who first demonstrated that anaphylatoxin activity could be generated by treating complement components with trypsin. Cochrane and Muller-Eberhard (1968) finally identified both human C3a and C5a after isolation directly from C3 and C5 that had been treated previously with either the appropriate complement enzyme or with trypsin. They showed that C5-derived anaphylatoxin (C5a) was a low molecular weight polypeptide (approximately M , 10,000) that did not desensitize ileal strips to C3a. Thus, for the first time it was demonstrated that there are two biologically and chemically distinct anaphylatoxins. The current view on the properties of their precursors and the mechanisms of their formation may be summarized as follows. C3 is an M , 180,000 &-globulin that is present in serum with the greatest abundance of all complement proteins, approximately 1500 pglml. In man, the plasma disappearance half-time of the protein, t1,2,is 50 hours, its fractional catabolic rate is 2.12% per hour, and its calculated rate of synthesis is 1.16 mgkg. The protein contains 2.5% carbohydrate, including neutral hexose, hexosamine, and neuraminic acid. It consists of an M , 110,000 a-chain and a 75,000 P-chain, both linked by disulfide bonds

ANAPHY LATOXINS

9

and noncovalent forces. The protein is exceedingly susceptible to proteolytic attack, which causes fisson of the molecule into several immunochemically defined fragments having a variety of biological activities. Fragmentation proceeds according to the following expression: c3 convertases

c3' c3u + C3b C311

C3h INA

+ /3lH

>

C311,(disulfide-linked split product)

c3b,tryphclike serum enzymes

>

c3c

+ C3d

The C3a anaphylatoxin is composed of the NH,-terminal 77 amino acid residues of the a-chain of C3. C3bi represents a C3b fragment containing a single scission in the a-chain portion without resultant dissociation. Both C3b and C3d are capable of reacting with specific cellular complement receptors. C3b also functions as a subunit of three different complement enzymes. C3d contains the site through which nascent C3b binds to nonspecific acceptors present on biological membranes, immune complexes, and a great variety of natural particles. C3c is the immediate precursor of a low molecular weight fragment with leukocytosis evoking activity (Ghebrehiwet and Muller-Eberhard, 1977). There are two C 3 convertases. The classical pathway C 3 convertase is a bimolecular complex composed of the major proteolytic fragments of C2 and C4. C4b functions as the acceptor of nascent C2a, the ftision of which yields C4b,2a. The fragment C4b is made up of three polypeptide chains, a' ( M , 88,000), p ( M , 78,000) and y ( M , 33,000). The fragment C2a consists of a single polypeptide chain of M , 85,000. Accordingly, the quaternary structure of C 3 convertase comprises four chains, and the cumulative molecular weight of the enzyme is 284,000. The active site resides in the C2a subunit, as evidenced by the fact that C2a, when dissociated from the complex, hydrolyzes N acetylglycyllysine methyl ester (AGLME) and that AGLME inhibits C3 cleavage by C4b,2a. The enzyme is also inhibited by diisofluorophosphate (DFP).The complex is relatively labile, having a half-life of 10 minutes at 37". The alternative C 3 convertase is a bimolecular fusion product of C 3 and C 3 proactivator (Factor B). More precisely, C3b functions as acceptor and modulator of Bb, the major physiological fragment of Factor B. The complex C3b,Bb is composed of three polypeptide chains and has a molecular weight of 234,000. The active site resides in the Bb subunit. The enzyme is inhibited b y DFP and the Bb fragment

10

TONY E. HUGLI AND HANS J. MOLLER-EBERHARD

binds approximately 1 mol of 14C-labeledDFP. Both native Factor B and the Bb fragment hydrolyze AGLME. The enzyme decays with a half-life of 1-2 minutes at 37°C. Association of the subunits is enhanced and the half-life of the enzyme is increased severalfold by the attachment of properdin. C5 is an M , 180,000 pl-globulin composed of two polypeptide chains linked by disulfide bonds and noncovalent forces. The site of the a- and p-chain is similar to that of the C3 chains. C5a is derived from the a-chain through the action of either the classical or alternative C5 convertase. The b-fragment in its nascent state has the ability to combine with C6 and to form a stable bimolecular complex that constitutes the first intermediate in the formation of the multimolecular membrane attack complex. The C5 convertase of the classical and the alternative pathway arises from addition of C3b to the respective C3 convertase. C3b appears to be essential for the handling of C5 as substrate. The regulator for the anaphylatoxins was not cited in the literature until 1969, and, because anaphylatoxin activity could be shown in the sera of rats, guinea pigs, pigs, and dogs, but not of humans, the anaphylatoxins were purported to play no major role in man. Lepow et al. (1969) and Bokisch et al. (1969) helped to dispel this belief by designing simple but enlightening experiments that demonstrated that an anaphylatoxin inactivator exists in human serum. When active C3a obtained from purified C3 was mixed with serum the activity was rapidly destroyed. These results stimulated a search for the serum factor responsible. The inactivator was isolated from serum and identified as an a-globulin having a molecular weight of approximately 280,000 and a serum concentration of approximately 50 pg/ml. Its ability to function is abolished by heating at 56” for 30 minutes or by treatment with metal chelators such as EDTA and phenanthroline (Bokisch and Miiller-Eberhard, 1970). The inactivator was shown to have carboxypeptidase B activity and to release arginine from C3a concomitant with destruction of anaphylatoxin activity. Since arginine had been assigned to the carboxy-terminal position in human C3a, these results showed that the natural mechanism for C3a inactivator involved enzymic removal of the arginyl residue from the carboxyl end of the polypeptide. Human C5a was also inactivated by the isolated serum carboxypeptidase, and it was concluded that this fragment too was controlled by a similar mechanism. The susceptibility of human C5a to the action of the serum exopeptidase provides an explanation for the scarcity of “classical” anaphylatoxin in complement-activated human serum (Vogt et al., 1969a). Since anaphylatoxin activity is readily demonstrable in various animal sera that also contain the inac-

ANAPHYLATOXINS

11

tivator, it is likely that a distinct structural feature renders these anaphylatoxins resistant to the enzymic control mechanism. Isolated anaphylatoxin inactivator, like pancreatic carboxypeptidase B, specifically cleaves COOH-terminal basic amino acid residues. A serum carboxypeptidase previously identified as a kininase (carboxypeptidase N) by Erdos and co-workers (Erdos and Sloane, 1962; Erdos et al., 1964) is probably identical to the anaphylatoxin inactivator. The importance of this serum enzyme is illustrated by the fact that the potent biological activities of kinins, anaphylatoxins, and fibrinolytic peptides are all partially or totally controlled by this single exopeptidase. Therefore, while investigators continued to confirm that the “classical” anaphylatoxin generated in sera from different animals by a variety of enzymic and nonenzymic activators appeared to be closely related both structurally and functionally (Kleine et al., 1970; Vogt et al., 1971a; Brade and Vogt, 1971a,b), there was a growing suspicion that only a fraction of the potential activities of anaphylatoxin had been uncovered. In light of this possibility an attempt was made to find inhibitors that would permit activation to proceed in serum but would block the serum carboxypeptidase. Although anaphylatoxin can be generated from purified anaphylatoxinogen, the process is inefficient and cumbersome. Studies by ValIota and Muller-Eberhard (1973) showed that most conventional inhibitors of pancreatic carboxypeptidase B also significantly inhibited the serum carboxypeptidase. One of these inhibitors, eaminocarproic acid (EACA), proved to be particularly compatible with complement activation in serum. Addition of relatively high concentrations of EACA to human serum allowed subsequent generation of both active C3a and C5a. These studies provided the first observations of stable C3a and C5a activity in activated human serum. Later, a more highly active form of C5a was also produced in animal sera treated with EACA (Vallota et al., 1978). Such observations confirmed that enzymic removal of the carboxy-terminal residue represents the primary control mechanism for these factors. Introduction of EACA as a suitable carboxypeptidase B inhibitor afforded isolation of sizable quantities of human C3a (Hugli et al., 1975a),human C5a (Fernandez and Hugli, 1976), porcine C3a (Corbin and Hugli, 1976), and porcine C5a (Vallota et al., 1973) directly from complement-activated serum. Proof that the carboxy-terminal position is indeed occupied by arginine in both C3a and C5a was obtained by carboxypeptidase B digestion of the peptides isolated from human and porcine serum. We formulated a general scheme for the formation and control of the

12

TONY E. HUGLI AND HANS J. MULLER-EBERHARD

anaphylatoxins based on the foregoing information as illustrated in Figs. 1 and 2. Both anaphylatoxins (e.g., C3a or C5a) are single products produced by a highly selective cleavage of the parent molecule. Scission of a single, specific bond in C3 or C5 is required for liberation of the respective anaphylatoxin. Human C3a generated either by the classical or the alternative pathways or by trypsin appear to be virtually identical in size, charge, and biological activity (Hugli et al., 1975a). The previous report that C3a anaphylatoxin activity is generated by treating human C3 with 0.5 M hydroxylamine (NH,OH) represents the only evidence of a chemical cleavage leading to anaphylatoxin formation (Budzko and Muller-Eberhard, 1969). Considering that a COOH-terminal arginine seems to be essential for anaphylatoxin activity, either NH20Hproduces an uncharacteristic scission between an Arg-X bond or an entirely novel interpretation is required to explain the generation of anaphylatoxin by NH,OH. A fragment released from human C3 by human leukocyte elastase (HLE) was characterized recently and found to be very similar to the C3 activation:

Human or Porcini C3

C3a

C3b

1 lhr

I1

or Plasmin

14.16

X-amma-Arg .X +TI

t

C i m

[s.jln Porcine C5

,

14-11 X + z m m - A r g + X i v i a chain

Isy, I

Porcine C5a

rgthain

C5b

FIG.1. The third (C3) and fifth (C5)components of complement are cleaved at highly selective sites by enzymes (convertases) of the classical and alternative pathway. Trypsin and certain of the serum enzymes that exhibit trypsinlike specificity also release anaphylatoxins from C3 and C5. Cleavage of the C3 a-chain occurred at position 77 and the C5 a-chain was cleaved between residues 74 and 77 in both human and porcine components. These results suggest that indigenous serum convertases cleave C3 and C5 in all higher mammals selectively, and at nearly an identical location in the NH,terminal region of the respective a-chain. Note that only human C5a contains significant quantities of carbohydrate (CHO).

13

ANAPHYLATOXINS

1 I1

(humon)

leukocyte slatme

2) t h m b i n

C3a4ike fragments (,nodr.)

humoral ond

CCllYlOr

c5 ( hc3 umml

.nzyms,

I 1 Ibukocyle

21

pro,.ale,

C5a.like fragments

thrombin

(inoctiv.')

FIG.2. The primary mechanism for controlling C3a and C5a activity involves enzymic removal of an essential arginine from the COOH-terminus of the anaphylatoxins. A serum carboxypeptidase (SCPB) exhibiting a specificity for basic amino acids provides efficient enzymic control. A secondary control mechanism is provided by certain cellular and humoral enzymes which cleave the parent C3 and C 5 molecules at locations other than the selective convertase sites. For instance, thrombin and lysosomal enzymes from polymorphonuclear leukocytes are known to cleave C3 and C5 forming fragments that are structurally similar to the anaphylatoxins but without biological acand other C5a-like fragments produced tivity. The asterisk (*) indicates that C5adesArs, by nonconiplement enzymes that are inactive as anaphylatoxins but remain active as chemotactic factors.

C3a molecule (Tayloret al., 1977).However, HLE cleaves the a-chain of C3 amino terminal to the essential arginine at position 77 and thus releases a biologically inactive fragment that nonetheless mimics C3a both immunologically and chemically. This generation of an inactive, anaphylatoxin-like fragment by HLE suggests a second possible control mechanism, one mediated by noncomplement protease. For instance, leukocyte-derived enzymes attack both C 3 and CS but produce inactive C3a-like and C5a-like fragments. T h e primary and secondary mechanisms of control of the anaphylatoxins are outlined in Fig. 2. In human serum C3a and CSa are rapidly inactivated by SCPB. In porcine, rat, and guinea pig serum C5a partially escapes inactivation, and so residual activity remains. The residual anaphylatoxin activity found in complement-activated sera from certain animal species can be explained by the following observations. Porcine CSa, "classical" anaphylatoxin, contains a COOH-terminal arginyl residue that resists attack by SCPB or pancreatic CPB, and only after extensive ther-

14

TONY E. HUGLI AND HANS J. IdJLLER-EBERHARD

ma1 denaturation does the molecule become susceptible to enzymic removal of the terminal arginyl residue (Vallota et al., 1978). It is therefore postulated that in serum the porcine C5a molecule rapidly rearranges when released from the parent molecule and that rearranged C5a molecules, being no longer susceptible to cleavage by the carboxypeptidase, thereby represent a serum-stable product. Since this rearrangement proceeds in competition with an efficient enzymic control mechanism, only a fraction of the C5a molecules generated in serum avoid inactivation by SCPB. The nature of this rearrangement has yet to be determined, but this scheme appears to accurately fit the empirical results.

IV. Characterization of the Anaphylatoxins

A. ISOLATION Before the mid-1960s, the anaphylatoxins were known only as biological activities. Activity was detected in activated serum or plasma primarily by assaying for contraction of the guinea pig ileum or by observing a lethal shock reaction in the animal. Initial attempts to isolate anaphylatoxin were performed by using complement activated rat or porcine serum (Stegemann et al., 1964; Vogt, 1968),because they contained stable anaphylatoxin activity. Since only the C5a anaphylatoxin is stable (see Section 111), early investigators dealt solely with the isolation of this “classical” anaphylatoxin. Stegemann et al. (1964) prepared small quantities of rat and porcine C5a from plasma that had been activated either by dextran or by cobra venom factor. Although these investigators attained a 3000-fold purification, their product is only 5-20% as active as that recently reported for electrophoretically homogeneous porcine C5a (Vallota and Miiller-Eberhard, 1973). Nonetheless, these initial studies are important contributions, because the investigators recognized the cationic nature of the anaphylatoxin and demonstrated its remarkable stability to heat (1hour at SOOC) and resistance to acid denaturation. Later, Vogt (1968) devised a purification scheme combining gel filtration with a variety of chromatographic steps employing various cation exchangers. Vogt et al. (1969a) were also successful in demonstrating low levels of anaphylatoxin activity (C5a) in human serum, made possible by a rapid acidification of the serum during the course of yeast cell activation. Before 1970, most attempts at preparing isolated anaphylatoxiiis involved the direct use of serum. Considering the specific activity re-

ANAPHYLATOXINS

15

cently reported for purified C5a anaphylatoxin (Vallota and MullerEberhard, 1973), many of the earlier preparations must have contained inactive (C5adesAIg) material. For example, Vogt (1968)reported a dose requirement of 0.2-0.5 Fg of purified porcine C5a per milliliter of bath fluid for eliciting contraction of guinea pig ileum. Vogt et al. (1969a) later reported that 1.5 pg/ml of purified human C5a was required to induce smooth muscle contraction. Human C5a prepared by the procedure of Vallota and Muller-Eberhard (1973), who used EACA, required doses ofjust 5-10 ng/ml of bath fluid to elicit an ileal contraction. The activity of anaphylatoxin isolated from serum containing EACA was clearly greater, b y approximately 50- to 100-fold, than that obtained from serum without EACA. This new approach replaces a more arduous one that required purification of the precursor anaphylatoxinogen followed by its conversion and finally, purification of the active factor (Cochrane and Muller-Eberhard, 1968). In addition, the need to supply converting enzymes such as the C3 or C 5 convertase can be avoided by the direct serum activation approach. Despite the obvious disadvantages in isolating the anaphylatoxinogen for producing anaphylatoxin, this approach played a major role in the early identification and characterization of anaphylatoxins, particularly that of C3a. Bokisch et aZ. (1969)demonstrated that a basic polypeptide, similar to the C3a fragment generated b y C 3 convertase, was released from C3 upon exposure to trypsin, plasmin, thrombin, or the cobra factor-Factor B complex. The trypsin-generated C3a-like fragment was easily separated from C3b by Sephadex gel filtration or by preparative gel electrophoresis. This trypsin-produced C 3 fragment compares favorably both in biological and physicochemical properties with C3a released from isolated C 3 by C 3 convertase (Hugli et al., 1975a). The relatively large quantities of peptide required for an exhaustive chemical characterization demanded that a highly efficient preparation scheme be developed. Therefore, isolation of anaphylatoxin directly from activated serum has again become a major means of obtaining this material. For instance, human C3a has been purified to homogeneity, in a fully active form, directly from senim that contained the carboxypeptidase inhibitor EACA during complement activation (Hugli et d.,1975a). This four-step isolation procedure employs gel filtration on Sephadex G-100, CM-Sephadex adsorption followed by elution with 0.1 M HCl, lyophilization, a second gel filtration on G-100 and finally passage over DEAE-Sephadex. This comparatively simple method, reminiscent of Vogt’s earlier procedure ( 1968), affords a 30% final recovery of homogeneous C3a. It can be easily ex-

16

TONY E. HUGLI AND HANS J. MULLER-EBERHARD

panded for processing liters of activated serum to obtain milligram quantities of the active material from either human or animal sources (Corbin and Hugli, 1976). The major purification steps involved in the isolation of both C3a and C5a have presently been combined into a single preparative scheme (Fernandez and Hugli, 1976),as outlined in Table I. Close attention to chromatographic conditions and to the order in which CMcellulose and CM-Sephadex are employed will yield consistently homogeneous preparations of C5a. Since C5a occurs in activated serum in minute concentrations 3 to 5 x lo-' M , very precise techniques must be used throughout the course of the purification if satisfactory yields are to be realized. The outline in Table I describes an isolation scheme that has been applied only to the human anaphylatoxins; however, with minor modification, a similar approach should prove successful for isolating anaphylatoxins from a number of animal species. TABLE I ANAPHYLATOXINPUFUFICATIONSCHEME" Procedure

I. Serum activation

11. Acidification and dialysis

111. CM-cellulose chromatography

IV. Gel filtration

V. CM-cellulose chromatography

VI . CM-Sephadex chromatography

Comments e-Aminocaproic acid is added to 1 M for SCPB inhibition. Boiled yeast cells (20 g/liter) are introduced initiating complement cascade and activation proceeds for 45 minutes at 37°C. Both C3 and C5 are efficiently and selectively converted. Activated serum is titrated to pH 3.8 with cold 1M HCI and EDTA (10 mM) is added prior to dialysis against ammonium formate buffer at 4°C. C3a and C5a are coeluted by 0.5 M ammonium formate at pH 5.0. Anaphylatoxins are recovered with other low molecular weight substances from a column of Sephadex G-100. C5a is eluted from CM-cellulose by 0.15 M buffer. The C3a is eluted by a gradient of 0.15-0.5 M ammonium formate, pH 5.0. The C3a obtained by this procedure is in a highly purified form. Electrophoreticall y homogeneous C5a is eluted from the CM-Sephadex (2-25 column with ammonium formate (0.1 M) at pH 7.0.

' Outline of a purification procedure described by Fernandez and Hugli (1976).

ANAPHYLATOXINS

17

B. PHYSICALPROPERTIES C3a and C5a can be readily distinguished on the basis of their physical properties; this distinction applies to all animal species studied to date. Consequently, C3a and C5a will first be considered independently and then compared, since their structural and functional similarities strongly suggest a close mutual relationship. Active C3a was generated by treatment of C 3 with CoF-Factor B complex, C4b,2a (C3 convertase) or trypsin and was estimated to have molecular weights from 6000 to 17,000 by gel filtration on Sephadex G-100 (Cochrane and Miiller-Eberhard, 1968).Another estimate using analytical ultracentrifugation indicated that C3a had a molecular weight range of 7000 to 8700 (Bokisch et aZ., 1969). Recent gel filtration and chemical structure studies (Hugli et al., 1975a) have placed the molecular weight of both human and porcine C3a at approximately 9000. One physical property of C3a that is perhaps more distinctive than size is its unusual cationic nature. The relative electrophoretic mobility of human C3a on cellulose acetate at pH 8.5 is +2.1 x 1 0 - ~omz V-' sec', indicating that the pZ for this protein is above 8.5 (Vallota and Muller-Eberhard, 1973).Human C3a electrofocused in acrylamide gels assumes an equilibrium position at approximately pH 9.7 (T. E. Hugli, unpublished observation). The conformational arrangement of the C3a polypeptide has been examined by circular dichroism (CD) studies (Hugli et d., 1975b). Spectra of native human and porcine C3a are nearly identical and indicate a 41-45% a-helical content. The extent of helical structure implies that native C3a may assume a very compact and stable conformational arrangement. Total disruption of the secondary structure of native C3a, as judged by C D measurements, required both 6 M guanidinium chloride and 0.02 M mercaptoethanol. However, removal of the reducing and denaturing agents permitted the molecule to refold and its biological activity to be restored. Furthermore, adding mercaptoethanol alone induced only partial unfolding, at a relatively slow rate (t1,*= 15 minutes). When the disulfide bonds of C3a were reduced and then alkylated, only about half of the CD absorbance at 222 nm was eliminated, but more than 90% of the biological activity was destroyed (Hugli et nl., 1975b). These results clearly established that the secondary conformation of the C3a molecule contributes significantly to expression of the anaphylatoxin activity. C3a anaphylatoxin is remarkably stable to acid since no activity is lost after long-term exposure to 0.1 M HCI at room temperature (Hugli

18

TONY E. HUGLI AND HANS J. MfjLLER-EBERHARD

et al., 1975b; Corbin and Hugh, 1976). The shape of the CD curve of human C3a in 0.1 M HC1 is similar to that of C3a at neutral pH; however, the magnitude of the absorbance at 208 and 222 nm is diminished by about 25%, which suggests a small but reversible change in secondary structure. Thermostability is yet another remarkable property exhibited by C3a. It retained nearly full activity after 5 minutes at 100"and it became fully inactivated only after exposure to 100"for 2-3 hours. The C D spectrum of heat-inactivated C3a was equivalent to that obtained for a random coil polypeptide, suggesting that the thermally induced unfolding had progressed to completion. Because classical anaphylatoxin (C5a) was recognized earlier than C3a, its biological properties have been studied more intensely. However, the C5a molecule still remains less well defined physically, primarily because of the scarcity of C5 protein in serum. Relative serum concentrations of C5 are approximately 5% those of C3 (Miiller-Eberhard, 1975), which means that the maximum quantity of C5a available in activated serum does not exceed 4-6 mg per liter. Just 10 years ago, such limited quantities of a protein would have made it virtually impossible to characterize physical or chemical parameters. Today, milligram quantities of a protein are adequate for obtaining an extensive characterization of the molecule. Early estimates of the size of the C5a anaphylatoxin varied widely, primarily because investigators had not yet purified C5a and could determine such parameters only by monitoring activity. In one of the earliest attempts to determine the size of classical anaphylatoxin, before it was known to be the C5a molecule, Stegemann et al. (1965) estimated the molecular weight to be 15,00030,000. Cochrane and Miiller-Eberhard (1968) later reported that the anaphylatoxin derived from isolated human C5 was a polypeptide ranging from M , 9000-11,000, as estimated by gel filtration. Shin et al. (1968) characterized the corresponding fragment of guinea pig C5 as an M , 15,000 molecule based on ultracentrifugation in a sucrose density gradient. More recently Wissler (1972),who used gel filtration, estimated the molecular weight of highly purified rat C5a to be 9500. Porcine C5a was similarly estimated to be a 9000 fragment, based primarily on the compositional analysis of the purified material (Lieflander et al., 1972).Vogt (1968) had earlier estimated porcine anaphylatoxin to have a molecular weight of 7000 to 8000. Vallota et al. (1978) found that quantitative end group analysis of porcine C5a supported the 9000 molecular weight value previously reported by Lieflander et al. (1972). In addition, Vallota et al. (1978) were the first to directly compare human and porcine C5a on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gels, and the

ANAPHY LATOXINS

19

relative migration clearly demonstrated that the porcine C5a was a slightly smaller molecule than human C5a. Since porcine C5a did not stain for carbohydrate, and the human C5a was known to contain carbohydrate (Fernandez and Hugli, 1976), the apparent dif'ference in size between human and porcine C5a quite obviously could involve only a carbohydrate moiety. The carbohydrate moiety of human C5a readily distinguishes it from rat and porcine C5a, both of which are essentially devoid of carbohydrate (Vogt, 1968; Wissler, 1972). The molecular weight of human C5a has been carefully reinvestigated as a result of detecting carbohydrate on the molecule, and the relative contributions of carbohydrate and protein were assessed chemically. Analysis of human C5a established that the polypeptide portion is of approximately M, 8200 and the carbohydrate moiety 2500 -3000, totaling approximately 11,000 for the intact molecule (Fernandez and Hugli, 1976). Clearly the difference in size between human C5a and C5a obtained from rats and pigs appears to be associated with variations in carbohydrate content. When porcine C5a was examined by CD, a substantial helical content was also found (Morgan et al., 1974). In fact, the C D spectra of porcine C5a between 200 and 250 nm is quite similar to that of human and porcine C3a. Calculations of the helical content in C5a indicate that approximately 40% of the amino acid residues participate in an a-helical conformation. As with C3a anaphylatoxin, the regular secondary structure of porcine C5a in only partially disrupted by 2-mercaptoethanol, and the time-dependent unfolding takes place with a half-life of about 20 minutes in 0.07 M reducing agent. More than 90% of the biological activity is eliminated in 0.25 M 2-mercaptoethanol after 2 hours, but full activity is restored if the reducing agent is removed. Rat C5a is also reportedly inactivated in 0.1 M 2-mercaptoethanol after 24 hours (Wissler, 1972). Human, rat, and porcine C5a are functionally stable over a wide range of p H conditions. Human C5a has been examined for acid and base lability, and it retains activity over a p H range of 1 to 13. Similarly, it withstands temperatures up to loo", and boiling for 15 minutes produces only partiaI inactivation (H. N. Fernandez, unpublished observation. The physical properties of the C3a and C5a are summarized for comparison in Table 11. The polypeptide portions of C3a and C5a are similar in size. The carbohydrate content of human C5a suggests that this anaphylatoxin should b e slightly larger than C5a from other animals and also larger than the C3a molecule, which is entirely devoid of carbohydrate. When estimated by gel filtration, ultracentrifugation,

TABLE I1 COMPARISON OF THE PHYSICAL PROPERTIES OF C3a AND C5a ANAPHYLATOXINS C3a Property Molecular weight Chemical estimate" Physical estimateb Electrophoretic mobility" (x cmz V-' sec-') Estimated pl pH stabilityd Thermal lability Mean residue ellipticitye Percent a-helix :c", Absorbance'

4 0

C5a

Human

Porcine

9083 9OOo +2.1

9165 9000

+ 1.9

!3000-17,000 - 1.7

9.7 1-10 loo", > 15min - 16,000 41-43 6.30

> 8.6

= 8.6

1-10 loo", > 15 min - 16,000 42-45 6.45

1-13 loo", > 15 min ND ND 10.1

Human

11,Ooo

Porcine

Rat

9000 9000-16.000 - 1.2

NM 9500 ND

18.6 4-9 ND - 15,000 40 ND

9.0 4 -9 ND ND ND 4.10

Values taken from Hugli et nl. (1975a), Corbin and Hugh (1976), Fernandez and Hugli (1976), and Lieflander et al. (1972). Values taken from Hugli et al. (197%) Cochrane and Miiller-Eberhard (1968),Vallota and Muller-Eberhard (1973), Vogt (1968),ValIota et al. (1973), and Wissler (1972). ' Values taken from Vallota and Muller-Eberhard (1973). Mobility determined at pH 8.6 on cellulose acetate strips. Minimum range of pH conditions over which the anaphylatoxin activity was unaffected. Ellipiticity [el at 222 nm was estimated from the circular dichroism spectra. Absorbance values given for protein at 276 nm. Protein concentrations estimated for human C3a, porcine C3a and human C5a by amino acid analysis. ND, not determined. a

'

2

m

ANAPHYLATOXINS

21

or SDS -PAGE, both porcine and human C5a appear somewhat larger than indicated by chemical determinations (Fernandez and Hugli, 1976; Vallota et al., 1973). Although the anomalous hehavior of human CSa can b e explained b y its carbohydrate content, the physical behavior of porcine CSa has yet to be fully understood. Little is known about the three-dimensional structures of the anaphylatoxins, except that all C3a and CSa molecules studied to date contain three disulfide bridges and have approximately 40% of the residues in an a-helical conformation. This would seem to imply that both anaphylatoxins exist as compact and rigid molecules. C3a and C5a readily undergo reversible conformational rearrangements, including an apparent reformation of the three disulfide lionds after reductive scission. Although the capacity of proteins to unfold and then to spontaneously resume native structure is not generally considered unusual, one is reminded the polypeptides in question here are but fragments of the larger parent protein. C. CHEMICALPROPERTIES Not until 1971 was the C3a anaphylatoxin obtained in quantities adequate for chemical characterization. At that time Budzko et al. (1971) managed to isolate milligrams of the active anaphylatoxin by subjecting purified human C 3 to limited digestion with either C4b,2a (C3 convertase) or trypsin. C3a, which is produced by either enzyme, was shown to contain an NH,-terminal seryl and a COOH-terminal arginyl residue. Recent studies by Tack et al. (1976)have confirmed that the C3a fragment is derived from the NH, terminus of the a-chain of C3. That the COOH-terminal arginine is essential for C3a spasmogenic activity was confirmed when isolated human C3a was inactivated by digestion with pancreatic carhoxypeptidease B or with serum carboxypeptidase B (SCPB), also known as anaphylatoxin inactivator (Bokisch and Miiller-Eberhard, 1970). Budzko et al. (1971) established that the human C3a molecule contained no free sulfhydryl groups; they also reported the first complete amino acid composition of C3a. Human C3a was later isolated directly from inulin-activated serum that contained the SCPB inhibitor EACA (Hugli et nl., 1975a). These studies provided further -chemical evidence that the C3a molecules generated by the C4b,2a enzyme of the classical pathway, by trypsin or by the C 3 convertase of the alternative pathway were structurally identical. Porcine C3a was also isolated from inulin-activated serum containing EACA, making possible a direct comparison between the chemical structures of these two anaphylatoxins (Corbin and Hugli,

22

TONY E. HUGLI AND HANS J. MULLER-EBERHARD

1976). Both human and porcine C3a were devoid of carbohydrate, and both were of approximately the same size (about M , 9000), the latter indicated by gel filtration on Sephadex G-50. However, amino acid compositions of human and porcine C3a (see Table 111) demonstrate a nominal, but definite, difference between these two molecules. For instance, there are fewer arginyl and threonyl residues in porcine C3a and more histidyl, glutamyl, and aspartyl residues in human C3a. As a result, there is a net charge difference at pH 8.6, porcine C3a being the more anionic of the two polypeptides (Corbin and Hugli, 1976). The absence of threonine in porcine C3a distinguishes it from human C3a; however, all the residues that have hydrophobic side chains, except methionine, appear in identical proportions in both molecules. Identical half-cystine contents, which accounts for 8%of the total residues in each molecule, suggest that certain structural aspects of C3a have been conserved during evolution. However, common antigenic TABLE I11 AMINO ACID COMPOSITIONS OF HUMANAND PORCINE C3a AND C5a ANAPHYLATOXINS~ C3a Amino acid Lysine Histidine Arginine Aspartic acid Threonine Serine Glutamic acid Proline GIycine Alanine Half-cystine Valine Methionine Isoleucine Leucine Tyrosine Phenylalanine Total residues: Molecular weight:

C5a

Human

Porcine

Human

Porcine

7 2 11 5 3 4 9 2 4 4 6 3 3 2 7 2 3 77 9083

7 3 9 6 0 4 11 2 4 5 6 2 4 2 7 2 3 77 908 1

8 2 5 6 3 4 9 1 3 8 7 5 1 5 4 2 1 74 8285

11.3 (11) 1.0 ( 1 ) 4.2 (4) 7.5 (7-8) 1.5 (2) 1.1 ( 1 ) 11.5 (11-12) 1.1 (1) 3.1 (3) 9.1 (9) 6.0 (6-7) 2.2 (2) 1.7 (2) 5.2 (5) 3.3 (3) 5.1 (5) 1.1 (1) 74-76 9004-9248

a Data were compiled from the following sources: human C3a (Hugli, 1975a),porcine C3a (Corbin and Hugli, 1976), human C5a (Fernandez and Hugli, 1976), and porcine C5a (Lieflander et ol., 1972; Vallota et al., 1978).

23

ANAPHYLATOXINS

determinants between the human and porcine C3a molecules appear minimal; only very weak cross-reactivity is observed when either human or porcine C3a is chaIlenged with the heterologous antibody. The linear amino acid sequence of both human and porcine C3a has now been established (Fig. 3 ) .Comparison of the two sequences indicate 23 amino acid replacements, homology was maintained for approximately 70%of the residues in human and porcine C3a. While the hydrophobic residues were quantitatively conserved, placement within the two anaphylatoxin molecules was not vigorously conserved. Replacements involving hydrophobic amino acids occur at positions 12 (Val/Leu),45 (Leu/Gln), 50 (LysNal),and 52 (Val/Ala). Two structural features seem to be conserved in the C3a molecule. The se-

Pacine

C3a

Humon C3a Humon C5a

10

20

30

40

60

50

Leu

- Gly - Leu - Alo - Prg - COOH

Leu

- Gly

70

- Leu - Alo - Prg

- COOH

74

FIG.3. A comparison of the amino acid sequences of human C3a, porcine C3a, and human C5a. The primary structures of these three polypeptides were arranged arbitrarily to maximize sequence homology. Human C5a was indented by two residues and several gaps were required in both the C3a and C5a sequences to obtain optimal identity. Human and porcine C3a contain approximately 70% homology, and 40% of the residues in human C5a were identical with the primary structure of human C3a. Note that the six half-cystine positions in human C3a have been conserved in both porcine C3a and human C5a.

24

TONY E. HUGLI AND HANS J. mLLER-EBERHARD

quence of residues adjacent to the essential arginine at position 77 (Leu-Gly-Leu-Ala-Arg) has been preserved, and each of the six halfcystines are located at identical positions in both molecules. The two repeating cysteinyl sequences located at positions 22,23 and at 56,57 represent distinctive structural features of the C3a molecule. Chemical evidence has shown that in both human and porcine C3a the three disulfide bridges linking the six half-cystines are interconnected in such a manner that a virtual knot of the polypeptide chain is formed (Hugli et al., 1975a; Corbin and Hugli, 1976). Although the disulfide bonds have not been chemically assigned, it is apparent that a single disulfide bond links the two Cys-Cys sequences and that a second bond connects each of the Cys-Cys sequences with either the Cys residue at position 36 or 49, as illustrated in Fig. 4. These additional covalent linkages presumably contribute greatly to the unusual stability exhibited by C3a. The disulfide bonds are known to influence greatly the biological expression of C3a, since its functional activity decreases by more than 90% after the S-S linkages have been ruptured with mercaptoethanol. It has been suggested that repeating cysteinyl sequences in C3a play an important structural role as has been found for cysteinyl residues in the pituitary hormones, including the subunits of chorionic gonadotropin (human a),ICSH (ovine a and ovine p), lu-

@Ale

@Phe

@ Lyr

@Pro

OThr

@Cyr

@Gly

@Leu

@)In

@Val

@ASP

@His

@ M ~ I

@ A ~ Q

@Tip

@ Glu @IIa

@ Arn @ Sar

@Tyr

FIG.4. A representation of the secondary structural features of the human C3a molecule. Disulfide bonds in C3a have not yet been chemically assigned; however, it is known that the six half-cystines are interconnected by three S-S linkages. The amino and carboxyl regions of the C3a polypeptide chain are shown folded to represent an a-helical arrangement. More than 40% of the residues in C3a were estimated to assume a regular helical conformation according to the circular dichroism spectra. The C5a molecule can be depicted in a similar manner, since it also contains three disulfide bonds and approximately a 40% a-helical content. Drawing by Dr. Bruce Erickson.

ANAPHYLATOXINS

25

teinizing hormone (porcine a and ovine a ) , and the neurophysins of bovine, porcine, and ovine origins (Hugli, 1975a). Comparison of anaphylatoxins with these polypeptide hormones is reasonable, considering that both factors interact specifically with cell surface receptors. Additional modes of generating functionally active C3a and inactive C3a-like fragments have recently been reported. They include the treatment of C 3 with streptokinase-activated plasmin, thrombin (Bokisch et d . , 1969), and human leukocyte elastase (HLE) (Taylor et d., 1977),as well as the yet unexplained release of a C3a-like fragment b y hydroxylamine (Budzko and Miiller-Eberhard, 1969). In these studies the plasmin-derived C3a was active, but the fragments generated by thrombin and b y HLE were not. The C 3 products released by these various enzymes have now undergone preliminary chemical characterization. While proteases, like thrombin and HLE, generate C3a fragments physically similar to C3a produced b y highly specific convertases of the complement system, biological and chemical evidence has shown that definite structural differences between these fragments do exist. For example, the C3a-like fragment produced by HLE was further examined and found to be a mixture of two inactive fragments resulting from cleavage of the C 3 a-chain at positions 68 and 71 (Taylor et ul., 1977). These results may explain the observation by Taubman et al. (1970) that lysosomal enzymes derived from human leukocytes fragmented human C3 without producing biologically active anaphylatoxin. Preliminary evidence indicates that thrombin preferentially cleaves the a-chain of C 3 at arginine 69, likewise producing an inactive C3a-like fragment lacking arginine 77 (Hugli, 1977). Because the COOH-terminal arginine at position 77 is required for biological activity, polypeptides were synthesized with structures based on the COOH-terminal sequence of human C3a. Our understanding of anaphylatoxin function has been greatly increased because of the recent synthesis of an octapeptide, C3a (70-77), with a sequence identical to residues 70-77 in human C3a (Hugli and Erickson, 1977). Peptide C3a (70-77) exhibits all the biological activities of intact C3a, including contraction of guinea pig ileum, production of skin erythema and edema, enhancement of vascular permeability, and release of histamine from rat mast cells. Synthetic peptide C3a (7077) also has a high degree of C3a specificity, producing total desensitization of the guinea pig ileum toward C3a, but not toward C5a or bradykinin. Conversely, intact C3a desensitizes the ileum to C3a (70-77). C3a and C3a (70-77) can be distinguished biologically only by the relative potency each expresses in various bioassays. The octa-

26

TONY E. HUGLI AND HANS J. M~~LLER-EBERHARD

peptide is approximately 1-2% as active as natural C3a in all biological assays that have been performed to date. A glycyl derivative of the octapeptide was synthesized [C3a (70-77)-glycine], and this modification resulted in a marked depression in the peptide’s biological activity; C3a (70-77)-glycine was approximately 50- to 100-fold less active than C3a (70-77) in contracting isolated smooth muscle. These findings show that the carboxyl group of the terminal arginyl residue is not functionally essential, but greatly contributes to full expression of biological activity by C3a. Furthermore, partial masking of C3a (70-77) activity by chemical modification with glycine begins to explain how the covalent linkage between a bulky C3b moiety and the carboxyl terminus of C3a completely abrogates anaphylatoxin activity. Extending the length of the COOH-terminal peptide analog of C3a to a tridecapeptide [C3a (65-77)] did not enhance the biological activity over that observed with C3a (70-77). However, peptides shorter than the octapeptide were progressively less active than C3a (70-77). The relative activities for an entire series of synthetic C3a peptides are compiled in Table IV. Tridecapeptide [C3a (65-77)l is approximately as active as the octapeptide, and compared to the octapeptide, activities of the pentapeptide [C3a (73-77)] and the tetrapeptide [C3a (74-77)l are reduced about 5- and leO-fold, respectively. Except for the tripeptide, all these peptides were capable of specifically desensitizing the guinea pig ileum toward natural C3a or toward other synthetic C3a peptides. Therefore, the minimal size required for a synthetic peptide to express C3a activity is between a tripeptide [C3a (75-77)] and a tetrapeptide [C3 (74-77)]. This important structurefunction relationship can be related directly to the known linear structures of C3a anaphylatoxins. Amino acid replacements occur near the COOH-termini of human and porcine C3a at positions 70,71, and 72 (see Fig. 3). This fact suggests that these residues contribute minimally to C3a function. Otherwise, human and porcine C3a might not be expected to exhibit identical biological activities, which they do. Furthermore, enhancement in activity between the pentapeptide [C3a (73-77)l and the octapeptide [C3a (70-77)] was only 5-fold. Conclusions were also derived from the fact that synthetic C3a peptides that express C3a specificity do not recognize the C5a receptor. Several of the residues in the COOH-terminal region of C5a (see Fig. 3) are identical to those in C3a and, therefore, a peptide from the COOH-terminus of either molecule containing fewer than 3 or 4 residues might not permit differential receptor recognition. Obviously, the essential structural information required for the induction of a variety of cellular responses is contained in a very short linear portion of

HUMAN C3a Peptide Native C3a (1-77) C3a-(65-77) C3a-(65-77)-Gly N-Ac-C3a-(70-77) C321-(70-77) C3a-(71-77) C3a-(72-77) C3a-(73-77) C3a-(73-76)-Lys C3a-(74-77) C3a-(75-77)

AND

TABLE IV SYNTHETIC C3a PEPTIDES: MINIMAL CONCENTRATIONS EFFECTINGSMOOTHMUSCLE CONTRACTION Activity rangea (mM)

Structure

Arg-Cln-His-Ala-Arg-Ala-Ser-His-Leu-Gly-Leu-Ala-Arg Arg-Gln-His-Ala-Arg-Ala-Ser-His-Leu-Gly-Leu-Ala-Arg-Gly Ac-Ala-Ser-His-Leu-GIy-Leu-Ah-Arg Ah-Ser-His-Leu-Gly-Leu-Ah-Arg Ser-His-Leu-Gly-Leu-Ala-Arg His-Leu-Gly-Leu-Ah-Arg Leu-GIy-Leu-Ala-Arg Leu-Gly-Leu-Ala-Lys GI y-Leu-Ala-Arg Leu-Ala-Arg ~~

0.011-0.012 0.4-0.7 28-60 0.8-1.2 0.5-0.8 0.6-1.0 1.0-2.0 2.0-4.0 Inactive* 120-240 Inactiveb

~~

Values represent concentrations of peptide that stimulate a contractile response in guinea pig ileal strips. Data taken from Hugli and Erickson (1977) and Caporale et u1. (1977). '' C3a-(73-76)-Ly\ was inactive at the level of 15 i d , and C3a-(75-77) expressed no activity up to a concentration of 480 mM.

*

F

x

3 3 0

E

2

28

TONY E. HUGLI AND HANS J. mLLER-EBERHARD

the C3a molecule, Whether this is also true for C5a has yet to be ascertained. The bulk of the C3a molecule is arranged in a highly regular secondary conformation and apparently acts to amplify, by a factor of about 50-100, the stimulatory response induced by the COOH-terminal portion of the molecule. Chemical characterization of C5a anaphylatoxin has been impeded until recently because of a scarcity of material. Four to five micrograms of human C5a can be generated per milliliter of activated human serum; however, only about 20-30% of that amount can presently be recovered by classical isolation procedures (Vallota and Muller-Eberhard, 1973; Fernandez and Hugli, 1976). Wissler (1972) described a ten-step, 5000-fold purification procedure for isolating rat C5a, and the isolated protein has reportedly been crystallized. The purified rat C5a was judged to be homogeneous, to have a molecular weight of 9000-10,000 and to lack carbohydrate. Unfortunately, no other chemical characterizations such as end-group or compositional analyses of the polypeptide were reported. Preliminary studies of purified porcine C5a isolated from activated serum (Lieflander et al., 1972) revealed a leucyl residue at the COOH-terminal position- however, when the porcine C5a was isolated from serum containing EACA, the COOH-terminal residue was found to be arginine with a leucine in the penultimate position (Vallota et al., 1978). These results can be easily interpreted if one assumes that the majority of the C5a molecules, prepared without the aid of EACA, undergo limited degradation by the indigenous SCPB. Porcine C5a anaphylatoxin prepared without EACA is definitely less active than C5a recovered from EACA-containing serum. It was therefore concluded that, as in human C3a and C5a and porcine C3a, the COOH-terminal arginine is a functionally essential residue of porcine C5a. Human C5a is totally inactivated in either human or animal serum, removal of the COOH-terminal arginine results in inactivation, and the arginine can be released either by SCPB or by pancreatic carboxypeptidase B (Fernandez and Hugli, 1976). These observations support the generalization that all C3 and C5 convertases from human or animal sources appear to exhibit trypsinlike specificity and selectively cleave the respective C3 and C5 molecule’s carboxyl to a specific arginyl residue. Presumably, the structural property that permits a fraction of the C5a molecules in porcine, rat, and guinea pig serum to escape total inactivation by SCPB is similar in each of these species. The chemical data currently available for human and animal C5a may well hold the key to explaining the differences in susceptibility between human C5a and C5a from other mammalian species toward SCPB (see Fig. 1).

29

ANAPHY LATOXINS

Human C5a has recently been shown to contain a rather sizable carbohydrate moiety (Fernandez and Hugli, 1976). This structural feature seems to be unique to human C5a anaphylatoxin. Earlier studies provided evidence that porcine C5a contains at most a single galactosamine residue per molecule of' polypeptide, which was assumed to b e 75-77 residues in length (Lieflander et al., 1972). Wissler (1972) reported that carbohydrate is absent in porcine, rat, and guinea pig C5a. Conversely, carbohydrate in human C5a accounts for as much as 25% of the molecule's total weight. Four different carbohydrate moieties were identified and quantitated. An estimated four residues of mannose, two of galactose, and four of glucosamine were detected per molecule of C5a. Sialic acid was present at 3 to 4 mol per mole of C5a. Therefore, the carbohydrate content in human C5a is quantitatively significant, but its functional role remains unknown. Regardless, the complex carbohydrate associated with human C5a represents one feaTABLE V EFFECTS OF ANAPHYLATOXIN O N CONTRACTILE TISSUE Human C3a Guinea pig Ileum Uterus Colon Intestine Trachea Papillary muscle Arterial strips Atrium Rat Ileum Uterus Intestine Arterial strips Atrium Rabbit Duodenum Hamster Colon

Porcine C5a

5 x 10-9" 5 x 10-8 2 x 10-7

4 x

1x

8 x lo-"

+

10-10

1 x 10-8

+ +

2 x 10-6

+ +

1 x 10-7 1 x 10-8

> 3 x 10-8 2 x 10-8

C3a

7x

10-9

Rat

Guinea pig

C5a

C5a

C5a

3 x io-'"

+ +

+ +

+ + + + + + +

+ + + +

+ +

+ +

" Values represent molar concentrations of purified anaphylatoxin capable of inducing a contractile response. Tissues stimulated by crude anaphylatoxin are indicated by a plus sign (+). Data were accumulated from Kleine et al. (1970), Greeff et al. (1959), Friedberg et a!. (1964), Vallota and Miiller-Eberhard (1973), Hugli et al. (1975a), Strandberg et nl. (1977), and Bernauer et al. (1972).

30

TONY E. HUGLI AND HANS J. WLLER-EBERHARD

ture not shared by the anaphylatoxins from any other mammalian species studied to date and may be related to the important difference observed in their resistance to serum carboxypeptidase. Immunological comparisons support the chemical data given for C3a and C5a in Table 111, which implies that human and porcine C3a differ more than do the respective C5a molecules. Porcine C3a and human C3a produce barely detectable lines of cross-reactivity by immunodiffusion analyses when aIlowed to react with a heterologous antibody raised in rabbits. However, rabbit antiporcine C5a reportedly produces a marked reaction of immunological identity when allowed to react with human, mouse, rat, or porcine C5a (Vogt et al., 1971a). Compositional analyses of human and porcine C5a (see Table 111) indicate a gross similarity, particularly in the content of half-cystine and hydrophobic residues that are nearly identical. The most characteristic difference between C3a and C5a is that C5a contains more isoleucine than leucine and more tyrosine than phenylalanine; this relationship is exactly reversed for C3a. The rationale for directly comparing C3a and C5a is derived from the partial identity observed between their primary structures (see Fig. 3). When the linear sequence of C5a is aligned with that of C3a in a manner that maximizes the degree of identity (i.e., threonine at position 1in C5a aligned with glutamine at position 3 in the C5a), the relationship is marked (Fernandez and Hugli, 197713). One must impose gaps and insertions in order to align the two Cys-Cys sequences occurring in both anaphylatoxins. Given the arbitrary manipulations necessary to attain maximal sequence alignment, 29 of 74 (38%)residues in human C5a are identical with the human C3a sequence. This degree of identity is most compelling and supports the conclusion that C3a and C5a share a common genetic ancestry. These provocative results provide the first chemical evidence that certain complement components are indeed genetically related. Demonstrating such a genetic relationship between C3 and C5 molecules carries particular significance, since extensive chromosomal linkage studies have yet to detect association between C3 and C5 or C3 and any of the other complement components (Alper and Rosen, 1976). V. Function of Anaphylatoxins

A. CELLULAR EFFECTS Perhaps the best characterized cellular effect attributable to anaphylatoxins is the release of vasoactive amines from mast cells. Early studies demonstrated that the classical anaphylatoxin stimulated his-

ANAPHY LATOXINS

31

tamine release in isolated lung tissue from guinea pigs (Rocha e Silva et al., 1951; Rocha e Silva, 1952; Rocha e Silva and Aronson, 1952). Mota (1959a) examined the mast cells in guinea pig mesentery after the cells had been exposed to anaphylatoxin (e.g., rat C5a) and found that the metrachromatic granules had disappeared, indicating release of the vasoactive amines. Mota observed that degranulation of mast cells by anaphylatoxin resembled the effect observed previously when antigen was presented to sensitized (e.g., IgE-coated) mast cells (Mota, 1959b), but that it is quite different from the degranulation produced by chemical histamine liberators, such as compound 48/80 or octylamine. In recent experiments the purified human factors C3a and C5a were shown to effect degranulation of mast cells in animal tissue (Dias da Silva and Lepow, 1967; Dias da Silva et al., 1967; Cochrane and Muller-Eberhard, 1968). Biopsies of human skin taken from the site of anaphylatoxin injection clearly indicated mast cell degranulation (Lepow et al., 1970; Wuepper et al., 1972; Vallota and MullerEberhard, 1973). Histamine release by human C3a and C5a from isolated rat mast cells was measured over a wide range of anaphylatoxin concentrations (Johnson et al., 1975).C5a released histamine over a range of4 x lo-' to 8 x lop6 M , approaching a maximal level at 2 to 4 x lop6 M , whereas C3a induced maximal release at M . Maximal histamine release from mast cells never exceeded 30-40% of the total histamine in the cell, and release could be blocked either by EGTA or by prior incubation of the cells at 45°C. Effects of purified human C3a and C5a on isolated mast cells were independent of one another as evidenced by an additive release when the factors were applied consecutively. Coincidentally, the des Arg form of C3a failed to release histamine from mast cells as this form of the C3a molecule was also unable to stimulate smooth muscle contraction. After measuring the uptake of human [lz5I1C3aand [lz5I1C5aby rat mast cells over a range of 0.2-2.0 nmol of peptide per lo6cells, Johnson et al. (1975) reported an apparent saturation of binding at the higher levels. Their calculations indicate that 0.5 to 1.0 x lo7 molecules of human C3a or porcine C5a were bound per rat mast cell, assuming that all cells bind the anaphylatoxin. This assumption, however, is probably not valid, since ter Laan et al. (1974) found human C3a bound to only about 20%of the rat mast cells treated with purified C3a. In their studies using immunofluorescence techniques, ter Laan et al. (1974) showed that, while degranulated mast cells were relatively devoid of bound C3a, the surfaces of cells in which degranulation had been inhibited by EDTA or disodium cromoglycate were heavily coated with anaphylatoxin molecules. In neither the uptake

32

TONY E. HUCLI AND HANS J. mLLER-EBERHARD

nor immunofluorescence experiments has anaphylatoxin binding been found unequivocally associated with a specific mast cell receptor. However, the observed binding is believed to be directly associated with the phenomenon of histamine release. Basophilic leukocytes from human plasma released histamine in the presence of human C5a (Dupree et al., 1974; Grant et al., 1975).Petersson et al. (1975)obtained a similar result by using purified porcine C5a and human leukocytes (basophils), and the latter investigators concluded that the histamine release was noncytotoxic, since cells pretreated with C5a in the absence of calcium remained fully responsive to challenge with anti-IgE. Repetitive addition of C5a resulted in a desensitization of the basophilic leukocytes, a response reminiscent of tachyphylaxis in smooth muscle tissue. Grant et al. (1975)claimed that C3a was not capable of releasing histamine from the human basophils. However, purified human C3a did induce histamine release when added to isolated human basophilic leukocytes in the complete absence of the serum enzyme SCPB (Glovsky et al., 1977). If anaphylatoxins are instrumental in mediating the local inflammatory response, their effect on leukocytes, particularly neutrophils, is of paramount importance. The studies summarized below concern the effects of anaphylatoxin on a variety of white cell types. When Goldstein et al. (1973a) treated isolated human leukocytes with cytochalasin B, the cell’s phagocytic ability was frustrated, although the cells remained viable and could be induced to release granular enzymes. These cytochalasin-B treated cells responded to complement-activated human serum by releasing a variety of lysosoma1 hydrolases, The granular hydrolases commonly monitored during leukocyte activation are P-glucuronidase and myeloperoxidase, both of which are released in zymosan-activated human serum by a low molecular weight factor identified as C5a (Goldstein et al., 1973b). In other studies, lysosomal enzymes were reported to be released from rabbit peritoneal neutrophils by human C3a, C5a, and C567 (Becker et al., 1974) and cytochalasin-B was shown to enhance enzyme release (Becker and Showell, 1974), as did the microtubular dissociating agents, vinblastine and colchicine. In the latter studies, the metal ion Ca2+,but not Mg2+,was shown to be required for enzyme release and release was inhibited by DFP. It was therefore suggested that a serine esterase may be involved in the release mechanism. Sulfhydryl modifying reagents such as iodoacetic acid and organic mercurials are strong inhibitors of enzyme release, indicating that the process may require free sulfhydryl groups. Both untreated and cytochalasin B-treated human leukocytes responded to human

ANAPHYLATOXINS

33

C5a by generating almost equal quantities of superoxide [02-] (Goldstein et al., 1975). Therefore, generation of superoxide appears to be associated with surface stimulation and does not depend on phagocytosis or lysosomal degranulation. Schorlemmer et al. (1976) recently observed that guinea pig C3a may induce macrophages to release acid hydrolases, but this response was associated with a loss of cell viability. Little information is presently available concerning the involvement of either C3a or C5a in macrophage activation or release reactions. However, the evidence accumulated thus far clearly demonstrates that anaphylatoxins at least in uitro, act as potent mediators or effectors of various release reactions on cell types intimately associated with acute and chronic inflammation. Cell types other than those immediately associated with the inflammatory process are now known to be influenced by anaphylatoxins. Preliminary evidence that guinea pig C3a exerts a cytolytic effect on tumor cells, lymphocytes, and macrophages has been reported (Ferluga et al., 1976). The levels of C3a required to lyse lymphocytes and macrophages is relatively high, certainly greater than expected in circulation even during extensive complement activation. However, lysis of cultured mastocytoma cells and transformed fibroblasts occurred with lower levels of C3a, indicating that preferential cytolysis was indeed possible. Participation of the anaphylatoxins in a biologically significant cytolytic event, like an attack mechanism for limiting proliferation of transformed cells, is consistent with other functions of C3a and C5a, such as accelerating the processing of cells damaged during tissue injury or inflammation. These provocative observations sggest that anaphylatoxins may play a role in the body’s natural defenses against tumor growth and certainly deserve further investigation. Platelets were reportedly altered in the presence of porcine C3a or C5a, resulting in aggregation (Grossklaus et ul., 1976).Effects of these anaphylatoxins were species dependent, but in an unexpected manner. Porcine C3a induced aggregation of platelets from guinea pigs only, and C5a induced aggregation of platelets from cats but did not aggregate the platelets from humans, rabbits, or even pigs. Rat C5a also caused aggregation of cat platelets, but was more effective in aggregating guinea pig platelets (Schumacher et al., 1975).The platelet aggregation was Ca2+-dependentand reversible, but renewed addition of C5a indicated that the platelets were desensitized. Preliminary studies concerning the mechanism of platelet aggregation by rat C5a suggested that the anaphylatoxin might be involved in an energy-dependent reaction (Benner et al., 1975). Both the sulfhydryl blocking

34

TONY E. HUGLI AND HANS J.

MULLER-EBERHARD

agent p-chloromercurihenzoate and the enzyme inhibitor tosyl arginine methyl ester (TAME) diminished aggregation induced by the anaphylatoxin. Although the effect of anaphylatoxins on platelets is of questionable physiological significance, it does conform to the general scheme in which anaphylatoxin binding to the cellular surface results in a functional alteration of cellular behavior. B. CHEMOTAXIS Chemotaxis is most simply characterized as a unidirectional locomotion of living cells. A comprehensive and learned treatment of this subject is contained in a monogram by Wilkinson (1974). The prevailing concept of chemotaxis proposes that cells respond to a concentration gradient of the stimulatory factor. Boyden (1962) introduced a new technique for measuring the chemotaxis of leukocytes b y employing chambers containing two compartments separated by a micropore filter. Cells are placed in the upper chamber; the migration of cells either into the filter or through the filter and into the lower chamber containing the test factor provides an index of activity. Boyden used his technique to show that there are heat-labile factors in serum that are responsible for the generation of chemotactic activity. This early observation provided the initial evidence that complement may participate in the generation of chemotactic activity in serum. Ward et ul. (1965)were the first to suggest that it was the complement components themselves that served as a source of the humoral chemotactic activity. Initially it was proposed that the chemotactic activity resided in a complex of C 5 and C6 (Ward et al., 1966);they later demonstrated that activity was associated with a complex of C5,6,7. In light of the present knowledge that the C5a portion of C5 is a potent factor in chemotaxis, the activity earlier described for C5-7 requires a careful reinvestigation. It is conceivable that C5a was associated with this complex and that this factor was responsible for the observed biological activity. The fact that the C5-7 complex does not represent a major chemotactic factor in man was proved by the finding that normal chemotactic activity could be generated in serum obtained from a patient with a homozygous C6 deficiency (Leddy et al., 1974). Proteolysis in vitro of selected complement components can produce chemotactically active factors. Plasmin was reportedly capable of generating a low molecular weight (= 6000) chemotactic factor from C 3 but not from C5 (Ward, 1967a,b). This result was the first real evidence that direct proteolysis of isolated complement components could generate chemotactic activity. The report that plasmin treatment of human C 5 did not generate chemotactic activity is surprising

ANAF’HY LATOXINS

35

in light of the evidence reported by Shin et al. (1968), who found that chemotactic activity was generated when guinea pig C 5 was treated with E A C m , another enzyme with trypsinlike specificity. An explanation for Ward’s inability to demonstrate chemotactic activity in a plasmin digest of C5 might relate to the quantity of digest tested, for it is now known that chemotactic activity is expressed only over a narrow concentration range (Fernandez et d.,1978). In fact, Ward and Becker (1968) later showed that prior addition of the C.W complex desensitized neutrophils when tested for chemotactic competence. Thus, Ward’s earlier results (1976b)could be explained if the quantity of plasmin-generated C 5 factor tested was simply too large or too small. In 1969 Ward and Newinan reported that treatment of isolated human C 5 with either trypsin or E A C W 3 produced a chemotactic factor that has a molecular weight of approximately 10,000 and is therefore presumably related to C5a. Jensen et al. (1969) reported a comparison between the anaphylatoxin and chemotactic factor derived from guinea pig C 5 and concluded, as did Ward and Newman (1969),that the two materials, if not identical, were at least closely related. Hurley (1964) and later Hill and Ward (1969) demonstrated that tissue proteases from a wide variety of tissues could generate chemotactic activity in fresh rat serum. The latter authors concluded from their studies that the protease from heart tissue was a serine esterase with trypsinlike specificity, since trypsin inhibitors blocked release of the chemotactic factor by the tissue protease. Their work implicated C3 as the primary source of the chemotactic activity, and the size of the active factor was estimated at M , 14,000. At approximately the same time Bokisch et al. (1969) presented evidence that C3a and C5a generated from purified human components exhibited chemotactic activity. Shortly thereafter, Taubman et al. (1970), reported that lysosomal enzymes from human leukocytes generated a chemotactic factor from purified human C5, and Ward and Hill (1970)found an enzyme in lysosomal granules of rabbit leukocytes that also formed a chemotactically active fragment from C5. Ward and Hill suggested that such cellular enzymes may play a role in augmenting the acute inflammatory response. Snyderman et al. (1972) demonstrated that bovine macrophage proteases produced a chemotactic factor when incuhated with guinea pig C5. Chemotactic activity also appeared in guinea pig serum when purified lipopolysaccharide (LPS), a substance that activates the complement cascade, was added (Snyderman et nl., 1969). The active factor generated by LPS had an estimated molecular weight of 15,000. The

36

TONY E. HUGLI AND HANS J. MOLLER-EBERHARD

authors compared the size of the active factor generated by LPS with that associated with the C567 complex and estimated that the relative activity of the M 15,000 factor was far greater than that associated with any other fraction in activated serum. Activity of the factor was inhibited by anti-guinea pig C5 but not by anti-guinea pig C3. Since, like LPS, immune precipitates (Jensen et al., 1969), CoF (Shin et al., 1968),and (21423 (Ward and Newman, 1969) each generate a chemotactic factor from C5, one might conclude that the resulting low molecular weight fragments are chemically similar or identical to the C5a anaphylatoxin molecule that is generated in a similar fashion. Although generation of the C5 chemotactic factor is presumably limited to specific enzymic events, the origin of these enzymes or enzyme activators varies widely. They are found, for instance, in tick salivary gland extracts (Berenberg et al., 1972), in secretions of other biting insects, and in snake venoms (Shin et al., 1968). All these substances presumably contribute to the inflammatory response observed at the site of the inflicted tissue injury by inducing a local chemotactic response. Since local viral infections commonly result in an accumulation of inflammatory cells, it is not surprising to find that the supernatant from virus-infected rabbit kidney cells induces chemotactic activity in rabbit serum and that the active factor is immunologically identifiable as C5a (Brier et al., 1970).Some viruses confer, simply by inducing cell lysis, the potential for generating chemotactic factors on the supernatants of cells they infect. Other viruses, such as Newcastle disease virus or mumps virus, evoke both a cellular leukotactic factor and a leukotactic factor-generating enzyme from infected cells, but only the latter is known to depend directly on complement (Ward et al., 1972). Phlogistic mediators have been associated with a variety of diseases involving tissue injury and inflammation (Ward, 1971).For instance, extracts of infarcted myocardium contained chemotactic activity immediately after surgical ligation of the coronary artery (Hill and Ward, 1971). The chemotactic activity seemed to be associated with a fragment of rat C3, since anti-rat C3 supressed most of the activity. The inflammatory exudate recovered from guinea pigs previously injected with endotoxin also contained a chemotactic factor, that was identified as C5a (Snydennan et al., 1971a). Further evidence that the C 5 fragment was partly responsible for accumulating an exudate after endotoxin injection came from studies with normal and C5-deficient sera from mice. The C5-deficient serum gave no chemotactic response when treated with either immune complexes or endotoxin, but the activity was restored by adding purified C 5 (Snyderman et al., 1969; Ro-

ANAPHYLATOXINS

37

senfeld and Leddy, 1974). Danierau and Vogt (1976)observed marked leukocyte emigration into the pleural cavity of a guinea pig after injection of 10-20 p g of purified porcine C5a. Synovial fluids from patients with rheumatoid arthritis proved to b e chemotactically active, and this activity was characterized as derived from C3 and C5, based on inhibition experiments with the respective antibody (Ward and Zvaifler, 1971).In addition, the fluid was capable of generating chemotactic factors either in whole serum or from the purified components C3 or C5. Evidence is rapidly mounting that C5a stimulates migration of inflammatory cell types other than just neutrophils. Snyderman et al. (1971b) reported that C5a chemotactically attracted mononuclear leukocytes (macrophages) obtained from guinea pigs. Kay and Austen (1972) showed that normal human basophilic leukocytes responded chemotactically to human C5a, which might explain why these leukocytes are also present at the site of delayed cutaneous hypersensitivity (DCH) reactions (Boetcher and Leonard, 1973). The latter investigators also reported a lymphocyte-derived chemotactic factor (LDCF) that specifically enhances the capacity of isolated basophils to respond chemotactically to C5a. The eosinophils are responsive to at least two chemotactic factors: (a) ECF-C, a complement-dependent factor that has been characterized as C5a (Kay, 1970; Kay et al., 1971) and (b) ECF-A, presumably a cellular factor released from mast cells (Kay et ul., 1973). When introduced together, C5a and ECF-A show greater activity than the sum of each applied separately, indicating that these two factors are synergistic in attracting eosinophilic leukocytes. Wissler et ol. (1972) described a serum factor (cocytotoxin) that apparently is essential for C5a from hog, rat, and guinea pig to function as a chemoattractant of leukocytes. This certainly was not the case for purified human C5a, which was active for attracting human neutrophils without the aid of any foreign protein (Fernandez et al., 1978). Highly purified human C3a, as well as trypsin or HLE digests of human C3, did not stimulate chemotaxis of human neutrophils either in vitro or in vivo when examined over a wide range of concentrations (Taylor et al., 1977; Fernandez et al., 1978).Human C5a, under conditions similar to those used for examining C3 products, stimulated neutrophil chemotaxis over a rather limited range of concentrations. Thus, neither human C3a nor other C 3 digestion products studied were capable of stimulating neutrophil chemotaxis at physiologically meaningful concentrations. Earlier observations that proteolytic digests of C3 gave rise to chemotactic activity (Bokisch et ul., 1969)might be explained if the earlier C 3 preparations were contaminated with C 5 as

38

TONY E. HUGLI AND HANS J. mLLER-EBERHARD

Tack and Prahl(l976) have suggested. In fact, variations among preparations might explain earlier problems in demonstrating chemotactic activity in trypsizined C3 (Ward, 1969). When human C3 is prepared by the procedure described by Tack and Prahl (1976), and functional C5 activity is eliminated, trypsin digestion of the C3 generates no chemotactic activity with human leukocytes (Fernandez et al., 1978). Fernandez et a2. (1976) demonstrated that although the C5adesArg is totally inactive as a spasmogen, it is still a chemoattractant for human neutrophils. These results were obtained with isolated C5adesAm, but are consistent with earlier reports that chemotactic activity is expressed in complement-activated human serum that should contain little C5a (i,e., anaphylatoxin activity). This observation indicates that active chemotactic factor remains in serum long after the anaphylatoxin activity has been abolished by serum carboxypeptidase B. It remains to be demonstrated whether CSadesArg is also a chemoattractant for basophils, eosinophils, and macrophages. The chemotactic response of human neutrophils, as estimated in Boyden chambers, reaches an optimum at levels between lo+ and lo-’ M C5a and between 5 X and lo-’ M C5adesAlgt but the response falls rapidly when higher levels of either factor are used. This apparent feedback mechanism might be explained by a desensitization of the neutrophils when presented with an excess of the stimulant. In fact, such a desensitization response has been postulated by Ward and Becker (1968) for rabbit neutrophils. From this hypothesis, one could also explain why neutrophils do not progressively accumulate at an inflammation site where cellular proteases and other complement activators may maintain persistently high levels of the chemotactic factor (Taubman et al., 1970). Realization that C5adesArg, the major form of the C5a molecule in activated human serum, functions as a potent chemotactic factor suggests that it participates actively in the inflammatory process. This chemotactic factor may also function in other pathobiological reactions in which sequestration of circulating macrophages or leukocytes is intimately associated with the disease process. The chemotactic activities of purified C5a and C5adesArg, as compared with the total activity measured in complement-activated serum, suggest that these molecules represent the predominant humoral chemotactic factors.

c. EFFECTSON SMOOTH MUSCLEAND OTHER TISSUES Pharmacological actions of the anaphylatoxins have been studied with a variety of isolated tissue preparations. Initially, the biological activities of anaphylatoxin (e.g., C5a) present in complement-acti-

ANAPHY LATOXINS

39

vated plasma or serum were examined, and the most comprehensive shidies o f this type were performed by Kleine et al. (1970). They examined contraction of isolated tissues from various animal species upon introduction of contact (Sephadex) or cobra venom activated plasma derived from rats, pigs, and guinea pigs. More recently, the biological activity associated with anaphylatoxin isolated from activated serum was examined; however, unless an inhibitor to serum carboxypeptides is introduced during activation, no C3a and only small quantities of functionally active C5a can be recovered. Procedures were developed for purifying porcine C5a (Stegemann et nl., 1964; Lieflander et al., 1972) and human C5a (Vogt, 1974), and, although each exhibited smooth muscle contracting activity, it was later concluded that these preparations consisted of a mixture of both active C5a and inactive C5adesArg. Hence, no valid quantitative assessment of C5a activity can be obtained using material purified from serum activated without benefit of an inhibitor to SCPB. Furthermore, C3a anaphylatoxin activity remains undetected using these activation conditions. Quantitation of the activity of C3a and C5a on isolated tissues became possible only after techniques for isolating fully active anaphylatoxins were developed. The procedures currently used in obtaining fully active anaphylatoxin are enzymic conversion of the isolated parent components C 3 and C 5 to produce the respective anaphylatoxins (Cochrane and Miiller-Eberhard, 1968)and direct isolation of C3a and C5a from serum activated in the presence of EACA (Vallota and Miiller-Eberhard, 1973). With these methods it has been possible to obtain enough C3a and C5a to quantitatively compare their potency on various tissue prepar at’i o n s Tissues responsive to anaphylatoxins include ileum, intestine, colon, uterus, trachea, arterial strips, and atrium (see Table V). Although anaphylatoxins can induce contraction of numerous muscle tissues taken from a variety of animal species, quantitative data still remain sparse. From these limited results, it appears the human and porcine C3a express equivalent activities on guinea pig ileum. Although they are more active than C3a, the corresponding C5a molecules from human and porcine sources also appear to function with equal potency. These results indicated that despite their chemical differences, C3a or C5a anaphylatoxins act similarly and are independent of species variations. Another apparently constant property among anaphylatoxins from any source is the relative specific activity of C3a compared to C5a. On a molar basis, C5a was generally 10-20 times more active than C3a for inducing contraction of guinea pig ileal and

40

TONY E. HUGLI AND HANS J. MfjLLER-EBERHARD

tracheal tissue. On the other hand, the potential concentration of C3a in serum is 10-20 times greater than that of C5a, which makes the potential activities of C3a and C5a nearly equivalent, Both C5a generated in activated serum and partially purified C5a (Kleine et aZ., 1970) produce tachyphylaxis (desensitization). Human, porcine, guinea pig, rat, and mouse C5a all induce tachyphylaxis in ileum, and each exhibits cross-tachyphylaxis when challenged by C5a from any other species. Human C3a also produces tachyphylaxis in ileum and is cross-tachyphylactic to porcine C3a (Hugli et al., 1975a), but not toward C5a (Cochrane and Muller-Eberhard, 1968). Despite Jensen's (1972) earlier objections that C3a is not a true anaphylatoxin, two notable properties qualify C3a as a spasmogen, its ability to induce tissue (mast cell) histamine release and its ability to produce tachyphylaxis of smooth muscle. Furthermore, tachyphylaxis induced by the anaphylatoxins is highly specific, as evidenced by the lack of C3a to produce cross-tachyphylaxis toward C5a, histamine, or kinins, although a synthetic octapeptide based on the COOH-terminal sequence of C3a (C3a 70-77) (Hugli and Erickson, 1977) desensitized the ileum selectively to natural C3a. Arguments formerly presented for the existence of C5a tachyphylaxis seemingly apply as well to C3a. For example, Randall et aZ. (1961) showed that guinea pig ileum rendered unresponsive to C5a regains responsiveness after 15-30 minutes. Friedberg et al. (1964) and Vogt and Zeman (1964) likewise concluded that tachyphylaxis induced by C5a was reversible. They also concluded that the refractory period was too short to permit the histamine stores to become replenished, as Rothschild and Rocha e Silva (1954) had also proposed. Bodammer and Vogt (1970)measured ['4C]histamine release from guinea pig ileum as a function of smooth muscle contraction induced by porcine C5a; they found that available substrate. [14C]histidine,was depleted only b y about 20% during tachyphylaxis. They further observed that, after responsiveness to C5a returned, little additional histamine was released by successive contractions induced by C5a. Tissues other than ileum, particularly guinea pig and rat uterus, respond normally to anaphylatoxins (C3a and C5a) when desensitized to histamine or treated with chemical antihistamines (Vogt and Zeman, 1964; Conrad and Mutsaars, 1949; Strandberg et al., 1977). Also, isolated rat intestine, which is distinctly unresponsive to histamine, can be contracted by rat C5a (Vogt et al., 196913).This evidence, along with the demonstration that C3a receptors on ileum undergo differential proteolysis without affecting the tissue response to histamine or bradykinin (Hugli, 1975b), suggests that the contractile response and

ANAPHYLATOXINS

41

tachyphylaxis are, at least in part, effects of the anaphylatoxins and that these processes are only partially mediated by, or dependent on, histamine release. At one time it was thought that disulfides might play a role in tachyphylaxis, as suggested by Vogt (1967), that (a) oxidation or alkylation of the anaphylatoxins caused more than a 90% loss of activity, (b) thioglycolate enhanced the effect of C5a on guinea pig ileum and reduced the period of recovery from tachyphylaxis. Other reducing agents, such as 2-mercaptoethanol, also inhibited the action of anaphylatoxin on ileum (Morgan et al., 1974). Nevertheless, since sulfhydryl-free, synthetic peptides can mimic the effects of C3a on smooth muscle, including tachyphylaxis, it is most unlikely that the disulfidelinked sulfhydryl groups in anaphylatoxins play any significant role in producing tachyphylaxis. The pharmacological action of the anaphylatoxins has not been examined systematically, but several drugs do affect the action of anaphylatoxin on selected tissue preparations. As mentioned, antihistamines are effective blocking agents in guinea pig ileum but not in uterine tissue. Greeff et al. (1959)reported that epinephrine and serotonin inhibitors (bromlysergic acid diethylamide and dihydroergotamine), ganglionic blocking agents (hexamethonium and pendiomide), and reserpine, which depletes catecholamines stores, were ineffective for blocking the tachycardia or enhanced strength of contractions produced by the action of C5a on a guinea pig auricle preparation. Whether these drugs will prove equally ineffective for reducing anaphylatoxin stimulation of smooth muscle contractions is still unknown. However, it is known that HI-type histamine antagonists are effective for blocking smooth muscle contraction induced by anaphylatoxins but not H,-type agents (e.g., rnetiarnide) (Strandberg et al., 1977). Furthermore, phentolamine, which reportedly blocks the constrictive response of C3a on rabbit omentuni (Mahler et al., 1975), does not inhibit the action of human C3a on guinea pig ileum, whereas phenoxybenzamine, another a-adrenergic blocking agent, does. In lung tissue anaphylatoxin (rat C5a), exerts two predominant effects: histamine release (Sackeyfio, 1971) and a prostaglandin-like perivascular and peribronchial edema (Sackeyfio, 1975). Sackeyfio found that histamine release was only partially abolished by the antihistamine rnepyrarnine, but the prostaglandin-like effect, which mimics PGFza, was abolished by the anti-inflammatory agents indomethacin and phenylbutazone. Hence, anaphylatoxins may induce respiratory distress by releasing mediators such as histamine and PGF2,. Prostaglandin release is apparently not involved in the con-

42

TONY E. HUGLI AND HANS J. MULLER-EBERHARD

traction of other smooth muscle tissues induced by anaphylatoxin, since indomethacin reportedly had no effect on the ileum (Strandberg et al., 1977). However, these pharmacologically different responses may only mean that the action of anaphylatoxins is not accomplished by a single or even a common mechanism in these various tissues.

D. SYSTEMIC EFFECTS Of the wide variety of biological effects attributed to the anaphylatoxins, perhaps the most vague and certainly the most difficult to demonstrate are the systemic effects. Few studies pointedly address themselves to the general effects of anaphylatoxin’s action in live animals, primarily because of two major limitations: (a)anaphylatoxins are readily inactivated by the carboxypeptidase in plasma and, hence, soon after injection become ineffective for producing gross systemic changes; (b) the quantity of anaphylatoxin, particularly C5a, available from serum prevents the accumulation of sufficient material for extensive studies with experimental animals. Now that various synthetic polypeptides that express anaphylatoxin activity are available (Hugli and Erickson, 1977), the existing limitation imposed by the use of natural anaphylatoxins may be eliminated. While the potency of purified anaphylatoxins in vivo remains poorly characterized, the few publications at hand indicate tht systemic functional behavior coincides with the bioactivities observed in uitro. For instance, intravenous injection of purified porcine C5a in large doses was lethal in guinea pigs, but smaller amounts caused distinct physiologic changes (Bodammer and Vogt, 1967). Microgram quantities of C5a induced bronchospasm in guinea pigs, resulting in a selective impairment of expiratory air flow. Bodammer and Vogt (1967) observed that if the response is severe, the animal’s respiration ceases, leading to the “shock” phenomenon originally described after injection of complement-activated serum. As these authors pointed out, the term “shock” probably has been misused in reference to effects of the anaphylatoxins, since the apparent lethal effects are not caused by acute circulatory failure but instead are associated with respiratory failure. A biphasic circulatory effect is seen immediately after C5a injection, in which the initial hypotensive response is followed closely by a hypertensive action. The bronchoconstriction and hypertensive effects are tachyphylactic, and the hypotensive response can be induced repeatedly by administering C5a. The systemic effect of circulating C5a may relate to the response obtained with C3a on rabbit omentum, where the vasoconstrictive effect on arterioles does not appear to b e tachyphylactic (Mahler et at., 1975). Other more general studies with animals involved the use of

ANAPHYLATOXINS

43

complement-activated serum and not purified anaphylatoxin as the source of the active factor. These studies were originally born of attempts to test whether cellular or humoral factors were principally responsible for the phenomenon of anaphylaxis. Hence, the early workers examined lethal systemic effects without much attention to sublethal responses. Friedberger (1909, 191l ) , Bordet (1913b), and Nathan (1913b) each demonstrated a toxic factor in animal blood treated with a variety of substances capable of activating complement. The histamine theory of anaphylaxis became the predominantly accepted explanation of shock in the 1930s and was further supported by evidence that antihistamines prevented anaphylatoxin-induced shock in the guinea pig model (Hahn and Oberdorf, 1950). Rocha e Silva and his collaborators (1951; Roche e Silva and Aronson, 1952) proposed a general theory of anaphylaxis that incorporated aspects of both the cellular and the humoral theories. They suggested that humoral principles (anaphylatoxins) serve as intermediate releasing agents, which in turn release cellular and tissue mediators (e.g., histamine, serotonin). Later, however, certain shock phenomena were found not to involve the anaphylatoxins as mediators. In particular, active anaphylaxis proceeded without significant anaphylatoxin formation even though in Forssman shock anaphylatoxins were formed (Giertz et al., 1958). Passive anaphylaxis seems to involve complement activation, after which presumably the anaphylatoxins participate in mediating the shock phenomenon (Frick et d . , 1962). Frenger et ul. (1958)showed that guinea pigs administered prednisone were insensitive to anaphylatoxin shock, and Frick et al. (1962) demonstrated that repeated treatment with sublethal doses of anaphylatoxin protected guinea pigs from passive anaphylactic shock for several weeks. Both methods of protection were interpreted as relying on partial depletion of tissue histamine stores. Hahn et al. (1969)studied the protective effect of antihistamine treatment in animals and found that, although protection was not absolute, its efficacy was enhanced when antihistaminic was introduced before the anaphylatoxin was injected. More recently, Garbe and Friedberg (1972) refuted much of the earlier work concerning prior treatment with anaphylatoxin as a means of depleting tissue histamine stores and protecting animals from lethal anaphylactic shock. They found that prior treatment of sensitized animals with anaphylatoxin provided little protection against shock provoked by a secondary antigen challenge; however, repeated anaphylatoxin treatment presumably protects an animal via the tachyphylactic route from anaphylatoxin-induced shock. In addition to the implications of histamine release by anaphyla-

44

TONY E. HUGLI AND HANS J. mLLER-EBERHARD

toxins, epinephrine and norepinephrine levels are elevated in guinea pig blood after injection of activated rat serum (Bundschu et aZ., 1973). This result led to the speculation that sympathetic neuronal activity is induced during anaphylatoxin shock. Studies by Hicks and Sackeyfio (1972) pointedly examined the possible adrenergic mechanisms involved in anaphylatoxin (rat C5a) activity. These authors concluded that adrenergic mechanisms operative during anaphylatoxin-induced bronchoconstriction and cardiovascular pressor responses were primarily associated with histamine and, to a lesser extent, with catecholamines released b y the agent. Numerous inflammatory responses have been suggested to require complement as the basis for the reaction, implying that the factors C3a and C5a may be actively involved. Among these disease states are acute gouty arthritis (Byers et al., 1973), acute immunologic arthritis (DeShazo et al., 1972), and crystal-induced inflammation of the joint (Phelps and McCarty, 1966). Other experimental conditions, such as the Arthus (Jensen et al., 1971; Ward and Hill, 1972) and Swartzman reactions have long been suspected to involve anaphylatoxin, but no firm evidence has been acquired to incriminate these serum factors directly in the pathogenesis of such tissue injury. Considering the growing list of potent mediators described in biomedical literature, one might anticipate that the anaphylatoxins play a significant role in shock, tissue injury, and inflammation, but that this role is integrated with the functions of numerous other mediators. Taken together, these mediators comprise a complex, interdependent effector system in which no single factor now known exerts predominance over all others. Our present-day understanding of the systemic effects of C3a and C5a would lead one to conclude, as did Vogt (1974), that lability of anaphylatoxins in plasma probably prevents major expression of many of their potential systemic responses. Anaphylatoxins are undeniably capable of expressing major spasmogenic activity in any animal; however, under most conditions the strict control mechanism probably limits them to transient or local effects except for chemotaxis (Jensen et aZ., 1969). VI. Roles of C3a a n d C 5 a in Inflammation a n d Acute Shock

Major involvement of anaphylatoxins in pathological conditions most probably lies in their participation in the acute inflammatory response. T h e principal reason for assessment derives from the known biological activities of C3a and C5a; in uitro they may affect contrac-

ANAPHYLATOXINS

45

tion of various isolated smooth muscle and vascular preparations, induce vasoamine release from mast cells, and stimulate the directed migration of PMN. Correspondingly, these same potent effects of C3a and CFja are readily illustrated in zjivo. For instance, both of these human anaphylatoxins produce immediate vasoconstriction and enhancement of vascular permeability within the microcirculatory system, cause localized degranulation of mast cells, and, in the case of C5a, excite chemotaxis of leukocytes through a barrier such as the vascular endothelium. The nature of these activities both suggest and infer that either C3a, C5a, or both play significant roles as mediators in acute inflammation, yet firm evidence that this is indeed the case remains elusive. Although the human anaphylatoxins are known to b e under strict enzymic control in the circulation, it has recently been shown that radiolabeled C3a that is injected intravenously escapes into the extravascular space at a very rapid rate ( t l l P of approximately 1 minute) (T. E. Hugli, unpublished observations). This behavior further supports an hypothesis that anaphylatoxins may induce a potent but transient tissue response, adjacent to their site of generation, by translocating over the vascular barrier thus avoiding inactivation. Anaphylatoxins may also be generated in the intravascular space. The effect would then be analogous to the responses of enhanced vascular permeability and/or leukocyte localization caused experimentally by sulxutaneous injection of the anaphylatoxins (Wuepper et al., 1972; Damerau and Vogt, 1976). Ratnoff and Lepow (1963)strongly implicated the anaphylatoxins as potential mediators of the inflammatory process when they designed experiments that mimicked this response by intradermally injecting purified C 1 esterase into guinea pigs. A primary reaction believed to be caused b y local activation of complement ensued. This reaction led to an immediate secondary response attributed to newly formed anaphylatoxin that was characterized by a visible enhancement in vascular permeability. A similar effect has been observed in acute graft rejection reactions (Jensen et d.,1972). For example, in the extreme case of xenogenic rejections, noninimunologic complement activation is induced followed by extensive liberation of vasoactive substances (Linn et d.,1971; Pratschke et d.,197s). The acute inflammatory response in gouty arthritis correlates with the appearance of sodium urate crystals and is preceded by a localization of PMN (Phelps and McCarty, 1966).Since the migration of PMN is stimulated by a mediator signal, complement is considered to be a candidate for inducing the response. Moreover, sodium urate crystals cause pronounced complement conversion in serum, resulting in a

46

TONY E. HUGLI AND HANS J. mLLER-EBERHARD

marked reduction in hemolytic activity and in the functional levels of both C 3 and C5, which diminish to less than 30% of normal after 90 minutes of exposure (Naff and Byers, 1973). Similarly, synovial fluids from patients with rheumatoid arthritis tend to show depressed C3 activity as compared with normals (Ruddy and Austen, 1970), possibly as a result of complement activation by immune complexes (Winchester et al., 1970).Therefore, complement conversion is readily demonstrated in association with various forms of inflammation, and as a consequence, the anaphylatoxins are implicated although their direct involvement remains unproved. The chemotactic activity associated with the C5a molecule is not abolished b y carboxypeptidase digestion and therefore the chemotactic activity (i.e., C5adeSAm) survives in vivo for a considerably longer time than does the spasmogenic activity. This fact suggests how a complement-derived fragment may amplify the sequestration of leukocytes at or near the site of anaphylatoxin generation, indicating a major role for C5a in inflammation, namely in the process of directed cellular mobilization. Resistance of C5a chemotactic activity to serum inactivators suggests that chemotaxis may be one of the most significant in vivo functions associated with the anaphylatoxin. Acute hypotensive shock can be induced in animal models given microbial endotoxin or extracted LPS, and, although complement has been implicated (Lachmann and Nicol, 1974; Mergenhagen et al., 1969; Morrison and Kline, 1977), the overall part played by complement in systemic shock phenomena remains unclear. Major discrepancies exist in the literature concerning the possible involvement of complement or complement-derived fragments in the acute shock of various animal species. For instance, the early hypotensive changes induced by injecting LPS into dogs (From et al., 1970; Garner et al., 1974) were abrogated by pretreatment with CoF, yet pretreatment with CoF in cats had little effect on LPS-induced hypotensive changes (Kitzmiller et al., 1972). Ulevitch et a2. (1977) reported that CoF depletion of C3 and terminal complement components in rabbits and monkeys failed to inhibit the prolonged, and often fatal, hypotension induced by LPS. However, since the initial hypotensive changes observed in dogs were difficult to demonstrate in rabbits and monkeys, the effects of complement depletion on the early phase of hypotension cannot be compared in all animals. Differences demonstrated among complement-related systemic shock responses of certain animals may be associated with the variable resistance of anaphylatoxins to individual control mechanisms in their sera. A case in point is the residual C5a activity found in complement-activated rat, porcine, and

ANAPHY LATOXINS

47

guinea pig sera. General agreement then seems to exist on the point that elimination or depletion of the anaphylatoxinogens C 3 and CS does not abolish pathologic changes such a s disseminated intravascular coagulation, prolonged disappearance of circulating platelets or an immediate fall in the number of mononuclear leukocytes. However, questions still remain concerning the involvement of complement in the various physiologic stages of hypotension, because some, if not all, species of animals develop an immediate hypotensive response after CoF treatment (Ulevitch and Cochrane, 1977). In human beings, several states of shock have been associated with a fall in complement levels. Serum levels ofhuman complement components C3, CS and certain factors of the properdin pathway are significantly depressed in patients experiencing shock from gram-negative bacteremia as compared with bacteremic patients without complications (McCabe, 1973; Fearon et ul., 1975). One must conclude that active complement conversion is underway in patients experiencing shock, but there is little evidence that anaphylatoxins contribute directly to the phenomenon of bacteremic shock. Rapid depression in serum levels of numerous components of the complement system has also been correlated with the progressive severity of the viral infection dengue hemorrhagic fever (Bokisch et al., 1973). Dengue fever is caused by a Group B arbovirus, and the course of this disease often advances to a characteristic hemorrhagic shock syndrome. The C3 turnover rate in a patient with dengue fever was estimated to be nearly twice that in a normal control. Perhaps even more significant is the observation that in sera from patients with dengue shock, the controlling enzyme of the anaphylatoxins (i.e., serum carboxypeptidase B) registers 50% less activity than that present in nomial serum (Corbin et al., 1976). While it is then clear that activation and consumption of complement is associated with a number of inflammatory conditions, and is undoubtedly accompanied b y generation of the anaphylatoxin, it is less obvious that these products play an active role i n the pathogenesis of the disease. VII. Concluding Remarks

Anaphylatoxin research spans a time period of about 70 years. Originally heginning as serological phenomenology, it has advanced to the stage where, on the basis of the known primary structure of both anaphylatoxins, oligopeptides may be synthesized that exhibit anaphylatoxin activity. The major contributions were made during the past 10

48

TONY E. HUGLI AND HANS J. WLLER-EBERHARD

years, when the anaphylatoxins were shown to be low molecular weight activation peptides of complement proteins, capable of eliciting responses from a variety of cells in vitro and in vivo and when their complete structure became known. Their ability to express activity in pico- and femtomolar concentrations pointed out their strong interaction with presumably specific cell surface receptors. These findings have paved the way for an inquiry into their mechanism of action on the cellular level, including the nature of their respective cell surface receptors. Perhaps the most important contribution of anaphylatoxin research to biology is the realization that complement reaction products act as hormonelike messengers that profoundly affect the function of certain cells. It is now known that a variety of cell types are activated by the complement reaction products C3a and C5a, whether they be acting as spasmogens, chemotaxins or cause release reactions. And it is likely that the number of cell types so affected will grow as a result of future research. The future task is to determine the principal function of these complement products in the intact animal organism. An availability of chemically defined natural and synthetic anaphylatoxins will greatly facilitate this task.

ACKNOWLEDGMENTS The authors thank Dr. Horacio N. Fernandez for allowing them to quote from his unpublished results. The expert editorial assistance of Mrs. Phyllis Minick and Mrs. Maureen Shumate is gratefully acknowledged. The authors would also like to express their thanks to Mrs. Carol Davis for her excellent assistance in preparing the manuscript. Dr. Joerg A. Jensen and Dr. Walter Vogt critically reviewed this manuscript and made valuable suggestions that are incorporated in the final text.

REFERENCES Alper, C. A., and Rosen, F. S. (1976).Ado. Hum. Genet. 7, 141. Becker, E. L., and Showell, H. J. (1974)J.Zmmunol. 112,2055. Becker, E. L., Showell, H. J., Henson, P. M., and Hsu, L. S. (1974).J.Zmmunol. 112, 2047. Benner, K. U., Schumacher, K.-A, and Classen, H.-G. (1975).Arzneim-Forsch. 25,1635. Berenberg, J. L., Ward, P. A., and Sonenshine, D. E. (1972).J.Zmmunol. 109,451. Bernauer, W., Hahn, F., Nimptsch, P., and Wissler, J . (1972). Znt. Arch. Allergy Appl. 42, 136. Bodammer, G., and Vogt, W. (1967).Int. Arch. Allergy Appl. Immunol. 32,417. Bodammer, G., and Vogt, W. (1970).Int. Arch. Allergy Appl. Zmmunol. 39,648. Boetcher, D. A., and Leonard, E. J. (1973).Zmmunol. Commun. 2,421. Bokisch, V. A., and Miiller-Eberhard. H. J. (1970).J.Clin. Znoest. 49,2427. Bokisch, V. A., Miiller-Eherhard, H. J., and Cochrane, C. G. (1969).J . E x p . Med. 129, 1109. Bokisch, V. A., Top, F. H., Jr., Russell, P. K., Dixon, F. J., and Muller-Eberhard, H. J. (1973).N . Engl. J . Med. 289,996. Bordet, J. (1913a).C . R . Seances Soc. Biol. Ses Fil. 74,225.

ANAPHYLATOXINS

49

Bordet, J. (1913b).C. R . Seances Soc. Biol. Ses Fil. 74,877. Bordet, J., and Zunz, E. (1915).Z. Zrnrnunitaetsforsch. E x p . Ther., I 23,42. Boyden, S. (1962).J . E x p . Med. 115,453. Brade, V., and Vogt, W. (1971a).Eur. J . Zmrnunol. 1, 290. Brade, V., and Vogt, W. (1971b).Eur. J . Zmmunol. 1, 295. Brier, A. M., Snyderman, R., Mergenhagen, S. E., and Notkins, A. L. (1970).Science 170, 1104. Budzko, D. B., and Miiller-Eberhard, H . J. (1969). Science 165, 506. Budzko, D. B., Bokisch, V. A., and Miiller-Eherhard, H. J. (1971). Biochemistry 10, 1166. Bundschu, D., Bernauer, W., and Filipowski, P. (1973). Znt. Arch. Allergy Appl. Zmmunol. 45, 385. Byers, P. H., Ward, P. A., Kellermeyer, R. W., and Naff, G. B. (1973).J.Lab. Clin. Med. 81, 761. Caporale, L. H., Erickson, B. W., and Hugli, T. E. (1977). In “Proceedings of the Fifth American Peptide Symposium” (M. Goodman and J. Meienhofer, eds.), p. 225. Halstead Press, New York. Cochrane, C. G., and Miiller-Eberhard, H. J. (1968).J.E r p . Med. 127,371. Conard, V., and Mutsaars, W. (1949). C. R. Seances Soc. Biol. Ses F i l . 143, 129. Cooper, N. R., Jensen, F. C., Welsh, R. M., Jr., and Oldstone, M. B. A. (1976).J.Exp. Med. 144,970. Corbin, N. C., and Hugli, T. E. (1976).J.Zmmunol. 117, 990. Corbin, N. C., Hugli, T. E., and Miiller-Eberhard, H. J. (1976). Anal. Biochem. 73,41. Damerau, B., and Vogt, W. (1976).Naunyn-Schniiedeberg’s Arch. Pharmacol. 295,237. DeShazo, C. V., Henson, P. M., and Cochrane, C. G. (1972).J . Clin. Inoest. 51,50. Dias da Silva, W., and Lepow, I. H. (1965).J. Zmmunol. 95, 1080. Dias da Silva, W., and Lepow, I. H. (1966). Zrnmuriochemistry 3,497. Dias da Silva, W., and Lepow, I. H. (1967).J . E x p . Med. 125, 921. Dias da Silva, W., Eisele, J. W., and Lepow, I. H. (1967).J. E x p . Med. 126, 1027. Dupree, E., Goldman, A. S., and Grant, J . A. (1974).J . Allergy Clin. Immunol. 53, 75. Erdos, E. G., and Sloane, E. M. (1962).Biochem. Pharmacol. 11,585. Erdos, E. G., Sloane, E. M., and Wohler, I. M. (1964).Biochetn. Pharmucol. 13,893. Fearon, D. T., Ruddy, S., Schur, P. H., and McCabe, W. R. (1975).N . Engl. J. Med. 292, 937. Ferluga, J., Schorlemmer, H. U., Baptista, L. C., and Allison, A. C. (1976). B. J . Cancer 34,626. Fernandez, H. N., and Hugli, T. E. (1976).J.Zmmunol. 117, 1688. Fernandez, H. N., and Hugli, T. E. (1977a).J. Biol. Chem. 252, 1826. Fernandez, H. N., and Hugli, T. E. (1977b).In “Proceedings of the Fifth American Peptide Symposium” (M. Goodman and J. Meienhofer, eds.), p. 228. Halstead Press, New York. Fernandez, H. N., Henson, P., and Hugli, T. E. (1976).J.Irnmunol. 116, 1732. Fernandez, H. N., Henson, P., Otani, A,, and Hugli, T. E. (1978).J.Zmmunol. 120, 109. Frenger, W., Scheiffarth, F., and Gemahlich, M. (1958). Z. Zrnrnunitaetsforsch. E x p . Ther. 115, 367. Frick, 0. L., Halpern, B. N., and Liacopoulos, P. (1962).J. Physiol. (London) 163, 191. Friedberg, K. D., Engelhardt, G., and Meineka, F. (1964).Znt. Arch. Allerg!/ Appl. Zmrnunol. 25, 154. Fnedberger, E. (1909).Z. Znimunitaetsforsch. E x p . Ther., 1 2, 208. Fnedberger, E. (1910).Z. Imniunitaets forsch. E x p . Ther., 1 4, 636. Friedberger, E. (1911). Dtsch. Med. Wochenschr. 37,481.

50

TONY E. HUGLI AND HANS J. MULLER-EBERHARD

From, A. H., Gewurz, H., Gruninger, R. P., Pickering, R. J., and Spink, W. W. (1970). Infect. Zmmun. 2,38. Garbe, G., and Friedberg, K. D. (1972).Naunyn-Schmiedeberg’s Arch. Pharmacol. 273, 401. Gamer, R.,Chater, B. V., and Brown, D. L. (1974).B . 1.Haematol. 28,393. Ghebrehiwet, B., and Miiller-Eberhard, H. J. (1978).J.Zmmunol. 120, (in press). Giertz, H., and Hahn, F. (1961).Int. Arch. Allergy Appl. Zmmunol. 19,94. Giertz, H., Hahn, F., Jurna, I., and Lange, A. (1958).Int. Arch. Allergy Appl. Immunol. 13,201. Glovsky, M. M., Hugli, T. E., Ishizaka, T., and Lichtenstein, L. M. (1977).Fed. Proc., Fed. Am. Soc. E x p . Biol. 36, 1264. Coldstein, I. M., Brai, M., Osler, A. G., and Weissmann, G. (1973a).J.Zmmunol. 111,33. Goldstein, I. M., Hoffstein, S., Gallin, J,, and Weissmann, G. (1973b).Proc. Natl. Acad. Sci. U . S . A. 70,2916. Goldstein, I. M., Roos, D., Kaplan, H. B., and Weissmann, G. (1975)J Clin. Inoest. 56, 1155. Gotze, O., and Miiller-Eherhard, H. J. (1971).J.E x p . Med. 134,90s. Grant, J. A., Dupree, E., Goldman, A. S., Schultz, D. R., and Jackson, A. L. (1975).J . Immunol. 114, 1101. Greeff, K., Benfey, B. G., and Bokelmann, A. (1959).Naunyn-Schmiedebergs Arch. E x p . Pathol. Pharmakol. 236,421. Grossklaus, C., Damerau, B., and Vogt, W. (1976). Naunyn-Schmiedeberg’s Arch. Pharmacol. 295, 71. Hahn, F. (1954). Naturwissenschaften 41,465. Hahn, F. (1960). Polypeptides Which Affect Smooth Muscles Blood Vessels, Proc. Symp., 1959 p. 275. Hahn, F., and Oberdorf, A. (1950).2. Zmmunitaets forsch. 107,528. Hahn, F., Lange, A., and Giertz, H. (1954).Naunyn-Schmiedebergs Arch. E x p . Pathol. Pharmakol. 227, 12. Hahn, F., Ebner, C., and Giertz, H. (1969).Znt. Arch. Allergy Appl. Immunol. 35,434. Hicks, R., and Sacheyfio, A. C. (1972). B. J . Pharmacol. 46,260. Hill, J. H., and Ward, P. A. (1969).J.Exp. Med. 130,505. Hill, J. H., and Ward, P. A. (1971).J . Exp. Med. 133,885. Hugli, T. E. (1975a).J.Biol. Chem. 250,8293. Hugli, T. E. (1975b).In “Proteases and Biological Control” (E. Reich, D. B. Rifkin, and E . Shaw, eds.), p. 273. Cold Spring Harbor Lab., Cold Spring Harbor, New York. Hugli, T. E. (1977). In “Chemistry and Biology of Thrombin” (R. L. Lundblad, J. W. Fenton 11, and K. G. Mann, eds.), p. 345. Ann Arbor Science, Ann Arbor, Mich. Hugli, T. E., and Erickson, B. W. (1977).Proc. Natl. Acad. Sci. U . S . A . 74, 1826. Hugli, T. E., Vallota, E. H., and Miiller-Eberhard, H. J. (1975a). J . Biol. Chem. 250, 1472. Hugli, T. E., Morgan, W. T., and Miiller-Eberhard, H. J. (1975b).J . B i d . Chem. 250, 1479. Hurley, J. V. (1964).Ann. N . Y. Acad. Sci. 116,918. Ishizaka, T., Ishizaka, K., Orange, R. P., and Austen, K. F. ( 1 9 7 0 ) Zmmunol. ~ 104,335. Jensen, J. A. (1966).Zmmunochemistry 3,498. Jensen, J. A. (1967).Science 155, 1122. Jensen, J. A. (1972). In “Biological Activities of Complement” (D. G. Ingram, ed.), p. 136. Karger, Basel. Jensen, J. A., Snyderman, R. and Mergenhagen, S. E. (1969).In “Cellular and Humoral Mechanisms in Anaphylaxis and Allergy” (H. Z. Movat, ed.), p. 265. Karger, Basel.

ANAPHYLATOXINS

51

Jensen, J. A., Garces, M. C., and Iglesias, E. (1971).J.Infect. Imniun. 4, 12. Jensen, J . A,, Davies, D., Linn, B. S., Snyderrnan, R., and Franklin, L. (1972).Circ. Res. 30, 332. Johnson, A. R., Hugli, T. E., and Muller-Eberhard, H. J. (1975).Immunolog!/ 28, 1067. Kaliner, M., and Austen, K. F. (1973).J.E x p . Med. 138, 1077. Kay, A. B. (1970).Clin. Exp. Immunol. 7, 723. Kay, A. B., and Austen, K. F. (1972).Clin. E x p . Zmmunol. 11, 557. Kay, A. B., Stechschulte, D. J., and Austen, K. F. (1971).J.E x p . Med. 133, 602. Kay, A. B., Shin, H . S., and Austen, K. F. (1973).Immunology 24, 969. Kitzmiller, J. L., Lucas, W. E., and Yelenosky, P. F. (1972).Am.J. Obstet. Gynecol. 112, 414. Kleine, I., Poppe, B., and Vogt, W. (1970).Eur. J. Pharmacol. 10, 398. Lachmann, P. J., and Nicol, A. E. (1974).Ado. Biosci. 12, 262. Leddy, J. P., Frank, M. M., Gaither, T., Baurn, J., and Klernperer, M. R. (1974).J.Clin. Invest. 53, 544. Lepow, I. H., Dias da Silva, W., and Patrick, R. A. (1969). In “Cellular and Humoral Mechanisms in Anaphylaxis and Allergy” (H. Z. Movat, ed.), p. 237. Karger, Basel. Lepow, I. H., Willms-Kretschrner, K., Patrick, R. A., and Rosen, F. S. (1970).Am. J. Puthol. 61, 13. Lietlinder, M., Dielenberg, D., Schmidt, G., arid Vogt, W. (1972). Hoppe-Seyler’s Z. Ph!piol. Chem. 353,385. Linn, B. S., Jensen, J. A., Pardo, V., Davies, D., and Franklin, L. (1971). Transplant. Today Proc. Int. Congr. Transplant, Soc., 3rd, 1970 p. 527. McCabe, W. R . (1973).N . Engl. J. Med. 288,21. Mahler, F., Intaglietta, M., Hugli, T. E., and Johnson, A. R. (1975). Microuasc. Res. 9, 345. Medicus, R. G . , Schreiber, R. D., Gotze, O., and Muller-Eberhard, H. J. (1976).Proc. Nutl. Acad. Sci. U . S . A. 73, 612. Mergenhagen, S. E., Snyderman, R., Gewurz, H., and Shin, H. S. (1969). Curr. Top. Microbiol. lmniirnol. 50,37. Morgan, W. T., Vallota, E. H., and Miiller-Eberhard, H . J. (1974). Biocheni. Biophys. Res. Commuri. 57, 572. Morrison, D. C., and Kline, L. F. (1977).J. Zmmunol. 118, 362. Mota, I. (1959,). Immunology 2,403. Mota, I. (1959b).J. Physiol. (London) 147,425. Muller-Eberhard, H . J. (1975).Annu. Reu. Biochem. 44,697. Naff, G . B., and Byers, P. H. (1973).J . Lab. Clin. Med. 81, 747. Nathan, E. (1913a).Z. 1mniuliitaetsfor.sch. E x p . Ther., 1 17,478. Nathan, E. (1913h).Z. Immunitaetsforsch. Exp. Ther., Z 18,636. Novy, F. G., and deKniif, P. H . (1917).J.Infect. Dis. 20,776. Osler, A. G., Randall, G. H., Hill, B. M., and Ovary, Z. (1959).J.E x p . Med. 110, 311. Petersson, B.A., Nilsson, A., and Sdlenheirn, G. (1975).J . Immunol. 114, 1581. Phelps, P., and McCarty, D. J., Jr. (1966).J.E x p . Med. 124, 115. Portier, D., and Richet, C. (1902).C . R. Seances Soc. B i d . Ses Fil. 54, 170. Pratschke, E., Land, W., Schilling, A., Pielsticker, K., and Brendel, W. (1975). Langenbecks Arch. Chir., Suppl. p. 129. Randall, H. G., Talbot, S. L., Neu, H. C., and Osler, A. G. (1961). Immunology 4, 388. Ratnoff, 0. D., and Lepow, I. H. (1963).J.E x p . Med. 118,681. Rocha e Silva, M. (1952). Br. Med. J. 1, 779. Rocha e Silva, M., and Aronson, M. (1952). Br. J. E x p . Pathol. 33,577. Rocha e Silva, M., Aronson, M., and Bier, 0. G. (1951).Nature (London) 168,465.

52

TONY E. HUGLI AND HANS J. MULLER-EBERHARD

Rosenfeld, S. I., and Leddy, J. P. (1974).J.Clin. Znoest. 53,67a. Rothschild, A. M., and Rocha e Silva, M. (1954). Br. J. E x p . Pathol. 35,507. Ruddy, S., and Austen, K. F. (1970).Arthritis Rheum. 13, 713. Ryan, G. B., and Majno, G , (1977).Am. J . Pathol. 86, 185. Sackeyfio, A. C. (1971).Br. J. Pharmacol. 43,424P. Sackeyfio, A. C. (1975). Br. J. Phurmacol. 55,240P. Schorlemmer, H. U., Davies, P., and Allison, A. C. (1976).Nature (London)261,48. Schumacher, K. A., Benner, K. U., Classen, H. G., Hagedorn, M., and Frede, K. E., (1975). Bibl. Anat. 13,267. Shin, H. S., Snyderman, R., Friedman, E., Mellors, A., and Mayer, M. M. (1968).Science 162, 361. Smink, R. D., Jr., Ahernathy, R. W., Ratnoff, 0. D., and Lepow, I. H. (1964).Proc. SOC. E x p . Biol. Med. 116,280. Snyderman, R., Shin, H. S., Phillips, J. K., Gewurz, H., and Mergenhagen, S. E. (1969). J. Zmmunol. 103,413. Snyderman, R., Phillips, J. K., and Mergenhagen, S. E. (1971a).J.E x p . Med. 134, 1131. Snyderman, R., Shin, H. S., and Hausman, M. H. (1971b). Proc. Soc. E x p . B i d . Med. 138,387. Snyderman, R., Shin, H. S., and Dannenberg, A. M., Jr., (1972).J.Zmmunol. 109, 896. Stegemann, H., Vogt, W., and Friedberg, K. D. (1964).Hoppe-Seyler’sZ. Physiol. Chem. 337,269. Stegemann, H., Hillebrecht, R., and Rien, W. (1965).Hoppe-Seyler’sZ . Physiol. Chem. 340, 11. Standberg, K., Miiller-Eherhard, H. J., and Hugli, T. E. (1977).Fed. Proc., Fed Am. Soc. E x p . B i d . 36, 1282. Tack, B. F., and Prahl, J. W. (1976). Biochemistry 15,4513. Tack, B. F., Morris, B. F., and Prahl, J. W. (1976). Fed. Proc., Fed. Am. Soc. E x p . B i d . 35,494 (Abstr. 1594). Tauhman, S. B., Goldschmidt, P. R., and Lepow, I. H. (1970).Fed. Proc., Fed. Am. Soc. E x p . Biol. 29,434. TayIor, J. C., Crawford, I., and Hugli, T. E. (1977).Biochemistry 16,3390. ter Laan, B., Molenaar, J. L., Feltkamp-Vroom, T. M., and Pondman, K. W. (1974).Eur. J . Zmmunol. 4,393. Ulevitch, R. J., and Cochrane, C. G . (1977).InJammation 2, 199. Ulevitch, R. J., Cochrane, C. G., Morrison, D. C., and Henson, P. M. (1977). In “Proceedings of the 7th International Symposium on Immunopathology” (P. A. Miescher, ed.), pp. 262-281. Schwabe and Co., Basel. Vallota, E. H., and Miiller-Eherhard, H. J. (1973).J.Exp. Med. 137, 1109. Vallota, E. H., Hugli, T. E., and Miiller-Eberhard, H. J. (1973).J . Zmmunol. 111, 294. Vallota, E. H., Fernandez, H. N., Hugli, T. E., and Miiller-Eberhard, H. J. (1978).in preparation. Vogt, W. (1967). Ergeb. Physiol. Biol. Chem. Exp. Phnrmakol. 59, 160. Vogt, W. (1968). Biochem. Pharmacol. 17,727. Vogt, W. (1974).Pharmacol. Reo. 26, 125. Vogt, W., and Schmidt, G. (1966). Biochsm. Pharmacol. 15,905. Vogt, W., and Zeman, N. (1964).Naunyn-Schmiedeberg’s Arch. E x p . Pathol. Phannakol. 247,328. Vogt, W., Bodammer, G., Lufft, E., and Schmidt, G. (1969a). Experientia 25,744. Vogt, W., Zeman, N., and Garbe, G. (1969b).Naunyn-Schmiedebergs Arch. Pharmokol. Exp. Pathol. 262,399.

ANAPHYLATOXINS

53

Vogt, W., Lieflander, M., Stalder, K. H., Lnfft, E., and Schmidt, G . (1971a). Eur. J . Immunol. 1, 139. Vogt, W., Lufft, E., and Schmidt, G . (1971b). Eur. J. Zrnmuriol. 1, 141. Ward, P. A. (1967a).Protides Biol. Fluids, 15,487. Ward, P. A. (1967b)./. E x p . Med. 126, 189. Ward, P. A. (1969).In “Cellular and Huinoral Mechanisms in Anaphylaxis and Allergy” (H. Z. Movat, ed.), p. 279. Karger, Basel. Ward, P. A. (197 1)./. E x p . Med. 134, 109s. Ward, P . A,, and Becker, E. L. (1968).J.E x p . Med. 127,693. Ward, P. A., and Hill, J. H. (1970)./.Immunol. 104,535. Ward, P. A,, and Hill, J. H. (1972).1.Imrnutiol. 108, 1137. Ward, P. A., and Newman, L. J. (1969)./. Immunol. 102, 93. Ward, P. A., and ZvaiHer, N. J. (1971).J.Clin. Invest. 50, 606. Ward, P. A,, Cochrane, C. C., and Miiller-Eherhard, H. J. (1965)./. E x p . Mecl. 122,327. Ward, P. A,, Cochrane, C. G., and Muller-Eherhard, H. J. (1966).Immunology 11, 141. Ward, P. A., Cohen, S., and Flanagan, T. D. (1972)./. E x p . Med. 135, 1095. Wilkinson, P. C. (1974).In “Cheniotaxis and Inflammation.” Churchill, London. Winchester, R. J., Agnello, V., and Kunkel, H. G . (1970).Ann. N . Y. Acud. Sci. 168, 195. Wissler, J. H . (1972).Eur. J. Zmmunol. 2, 73. Wissler, J. H., Stecher, U. J., and Sorkin, E. (1972).Eur. J . Immunol. 2, 90. World Health Organization Committee on Complement Nomendature. (1968).Bull. W. H . 0. 39,939. Wuepper, K. D., Bokisch, V. A., Miiller-Eherhard, H. J., and Stoughton, R. B. (1972). Clin. E x p . Zmmunol. 11, 13.

This Page Intentionally Left Blank

.

AWANCCS IN IMMUNOLOGY. VOL 26

H-2 Mutations: Their Genetics and Effect on Immune Functions JAN KLElN Doportment of Microbiology. University of Texor Heolth Science Center. Dolfos. roxor. ond Deportment of lmmunogenoticr. The Mon Plonck Institute for Biology. Tubingen. Fedorol Republic of Gennony

I . Introduction ........................................................................................................

56

11. The H-2 Syllabus ............................................................................................... A . Chromosome 17............................................................................................

B. The H-2 Map ................................................................................................ C. The H-2L Locus ........................................................................................... 111. Basic Terms in Mutation Genetics ................................................................... IV. Methods of Histocompatilility Mutation Study .............................................. A . Detection ...................................................................................................... B . Progeny Test ................................................................................................. C. Isolation ........................................................................................................ D . Linkage Test ................................................................................................. E. Naming the Mutation ................................................................................... F. Determination of the Mutation Type ......................................................... G . Complementation Test ................................................................................ V . Mutation Rates ................................................................................................... A . General Considerations ............................................................................... B . Rates of Histocompatibility Mutations ..................................................... C. H-2 Mutation Rates ...................................................................................... D Interpretation of H-2 Mutation Rates ......................................................... VI . Genetics of Available H-2 Mutations ............................................................... A . Distribution of H-2 Mutations among Regions ........................................ B. Genetic Localization of H-2 Mutations ...................................................... C. Description of Available H-2 Mutations .................................................... VII . Biochemistry ofH-2 Mutations ......................................................................... VIII . Origin of H-2 Mutations .................................................................................... A . Intragenic Crossing-over (Gene Conversion)............................................ B . Regulator-Gene Mutation ............................................................................ C. Point Mutations ............................................................................................ IX. Mutations and Polymorphism ........................................................................... X . Effect of H-2 Mutations on Immune Functions .............................................. A . Allograft Reaction and Immunological Tolerance .................................... B. CML to Alloantigens ................................................................................... C. CML to Associative Antigens .................................................................... D. MLR and GVHR ..........................................................................................

60 60 60 62 63 64 64 66 66 66 67 68 69 73 73 79

84

.

55

87 91 91 91

95 96 99 99 101 105 106 108 108 112 118 121

.

COpynghtO 1978 by Academic Press Inc. All rights of reproduction in any form reserved. ISBN &1!&022426-7

56

JAN KLEIN

E. Production of an Allogeneic Supernatant .................................................. F. Craft-versus-Host Disease .. G. Serology of H - 2 Mutants ....... ....................... H. Genetic Control of Immune XI. What Have H - 2 Mutations Contributed to Immunology? ................. .................................... XII. Perspectives ............................. References ........................................ .........................................................

126 127

141

I dedicate this review to the Herculi of H - 2 mutation genetics: Donald W. Bailey, Igor K. Egorov, and Henry I. Kohn. To graft some 50,000 mice is, I think, no smaller task than to bring three golden apples from the Hesperides. I . lnrroduction

One of the most fascinating features of the H-2 system, the major histocompatibility complex (MHC) of the mouse, is its pleiotropism, or the fact that it controls a large number of traits. At latest count, some 60 traits have been claimed to be controlled by this system (Table I). The pleiotropism raises the question of whether there is one locus for each of the 60 traits, or whether single loci control groups of traits. Until recently the approach in attempting to answer this question has been based on the use of intra H-2 recombinants (Fig. 1). In this approach, a loss of anH-2-associated trait accompanying the loss, through crossing-over, of a part of the H-2 complex is interpreted as evidence

H-?,. .tkFLo-

H-2 REGIONS

RESPONSE

.I

K

A

B

C

S

D

000000 LOW

a

h4

ooonnn

111111 001111

HIGH

LOW HIGH

FIG. 1. Principle of H - 2 mapping. In the upper part of the panel, strains carrying H-2" haplotypes are high responders and H - 2 q strains are low responders to a synthetic polypeptide (H,C)-A-L. A recombinant (H-2"') carrying the K region of H - 2 a and the remainder of the H - 2 complex from H-2" is a high responder, indicating that the K region is not involved in the genetic control of the responsiveness. In the lower part of the panel, H-2" is a high, and H-2b a low responder to the same antigen. A recombinant (H-Zk4)derived from the two by crossing over between A and B regions is a high responder, indicating that the response to (H,G)-A--Lis controlled by theA region. (Based on data of McDevitt et al., 1972.)

H-2

MUTATIONS

57

that the missing part controls the particular trait. With the use of this approach, genes controlling several traits have been mapped into the same regions of the H-2 complex (Fig. 2). For convenience, the genes controlling the various traits were assigned different names, but the question of whether they were genes at different loci within one region or alleles at the same locus remained unresolved. The resolution of this question through the recombinant approach would require a complete saturation of the H - 2 map, that is definition of all loci composing the map-a goal that at the present time is impossible to accomplish. Recently, however, an alternative approach has become available to resolve the problem of the H-2 pleiotropism. The approach is based on the study of H-2 mutations, and it assumes that if a point mutation alters several traits, then the traits must be controlled by the same locus. This review evaluates what mutations have contributed to the study of the H - 2 pleiotropism and what conclusions can be drawn from such a study about the function of the H-2 complex. The review is divided into twelve sections. Following this introduction and a brief section summarizing the current status of H-2 genetics, Section I11 explains basic terms used in mutant studies so as to provide the necessary genetic background for nongenetically trained readers. Section IV then describes methods of detection and analysis of histocompatibility mutations in general, and H-2 mutations in particular. The fifth section is devoted to mutation rates, and here againbecause of the relative unfimiliarity of many immunologists with the subject-fairly extensive background information precedes the dis-

FIG. 2. Pleiotropism of the H-2 complex. On the extreme left, some of the traits known to be associated with H-2 are listed. The remainder of the figure shows genes coding for these traits and the position of these genes on the H-2 map, as determined by recornbinational analysis. Genes in the same region could be at different loci or alleles at the same locus.

58

JAN KLEIN TABLE I Th6irtre de l’absurde: LISTOF H-2 ASSOCIATEDTRAITS Trait

Reference

Serologically detectable class I antigens‘ Antigens eliciting acute allograft reaction Hybrid resistance to parental tumor grafts Hybrid resistance to peritoneal exudate cells 5. Antigens eliciting graft-vs-host reaction 6. Serum-protein (Ss) level 7. Hybrid resistance to bone marrow transplants 8. Resistance to oncogenic viruses 9. Antigens eliciting cell-mediated lympholysis 10. Allogeneic inhibition-syngeneic preference 11. Antigens eliciting mixed-lymphocyte reaction 12. Immune response to T cell-dependent antigensb 13. Allogeneic effect 14. Serum-protein allotypes (Slp) 15. Androgen levels

Corer, 1936 Gorer, 1937 Snell, 1958 Boyse, 1959

16. Reactivity with sheep erythrocytes 17. T-B cell collaboration

Sabolovic et a/., 1971 Kindred and Shreffler, 1972 Ivinyi et al., 1972

1. 2. 3. 4.

Simonsen and Jensen, 1959 Shreffler and Owen, 1963 Cudkowicz and Stimpfling, 1964 Lilly et a/., 1964 Brondz, 1964 Hellstrom et al., 1964; Moller and Moller, 1965 Dutton, 1965 McDevitt and Chinitz, 1969 Hirst and Dutton, 1970 Passmore and Shreffler, 1970 Hampl et d . , 1971, Ivanyi et d . ,

1972

18. Weight of testes, vesicular gland, thymus, and lymph node 19. Level of hormone-binding proteins

Ivinyi et ul., 1972; Goldman et

a/., 1977 20. Autoimmunization in parent + F, combina-

Gleichmann et al., 1972

tions

21. Complement activity’ 22. Serologically detectable class I1 (Ia) antigensd

23. Shape of the mandible 24. Expression of Thy-1 on lymph node cells 25. Phytohemagglutinin responsiveness by

Demant et ul., 1973 David et ul., 1973; Hauptfeld et

al., 1973 Festing, 1973 Mickova and Ivanyi, 1973 Donner et a/., 1973

spleen cells

26. Cell-mediated immunity to viral antigens 27. Cell-mediated immunity to bacterial anti-

Zinkernagel and Doherty, 1974 Zinkernagel, 1974

gens

28. Complement-receptor expression 29. Production of B-cell activating (helper) factorse

30. Production of T-cell suppressing factorse

Gelfand et a/., 1974 Taussig and Munro, 1974; and others Tada, 1974

H-2

59

MUTATIONS

TABLE I (Continued) Trait 31. 32. 33. 34. 35. 36. 37.

Immune suppression Enhancement of allografts Expressivity of Brachyury (T) gene Duration of Brucella-induced splenomegaly Rate of DNA synthesis Effector cell induction Cell-mediated immunity to hapten-modified cells 38. Cell-mediated immunity to minor H antigens 39. Cyclic AMP levels 40. T-cell-macrophage collaboration 41. Cortisone-induced cleft palate 42. Sensitivity to testosterone in orchidectomized mice 43. Spermatogenic failure of translocation heterozygotes 44. Ontogenesis of H-2 antigens on erythrocytes 45. Developnient of delayed-type hypersensitivity 46. Production of natural killer cells 47. Mating behavior 48. Antigen-induced T-cell proliferation 49. Spleen cellularity 50. Corticosteroid effects on thymus cells 51. Expressivity ofH-Y on thymus and skin cells 52. Proportion of T and B lymphocytes 53. Castration effect on thymus weight 54. Suppression of mixed lymphocyte reaction 55. Cell-mediated immunity to tumor-associated antigens 56. Cytostatic effect of immune lymphocytes 57. Production of macrophage-migration-inhibition factor 58. Helper factor for generation of effector T lymphocytes 59. Body weight

Reference Kapp et al., 1974 Shines et al., 1974 Mickova and Ivanyi, 1974 0 t h et al., 1974 Rychlikova and Ivanyi, 1974 Festenstein et al., 1974 Shearer et al., 1975 Bevan, 1975; Cordon et al., 1975 Meruelo and Edidin, 1975 Erb and Feldmann, 1975 Bonner and Slavkin, 1975 Mickovi and Ivlinyi, 1975 Forejt, 1975 Boubelik et d.,1975 Miller et al., 1976 Petranyi et al., 1976 Yamazaki et al., 1976 Schwartz and Paul, 1976 Ferreira and Nussenzweig, 1976 Pla et d . , 1976 Kralovi and Demant, 1976 Donner and Wioland, 1976 Viklicky et al., 1976 Rich and Rich, 1976 Gomard et al., 1976 DeCiorgi et nl., 1976 Suslov et al., 1976 Plate, 1976 Gregorova et al., 1977

Some 80 H-2 antigens have been identified. Immune response to some 40 different antigens has been shown to he H - 2 controlled. ' At least four complement-associated traits seem to be H - 2 linked. Some 30 Ia antigens are controlled by H - 2 . ' At least 10 helper and suppressor factors have been associated with H - 2 .

60

JAN KLEIN

cussion of the H - 2 mutation rates. The sixth section describes the available H - 2 mutants and summarizes the present knowledge concerning their genetics. In the seventh section, the limited biochemical data on H - 2 mutations are described. Section VIII discusses in some detail the possible origin of H - 2 mutations and attempts to dispel certain misconceptions regarding this topic. In the ninth section, the knowledge of H - 2 mutations is related to the perennial problem of polymorphism. In the tenth and largest section, the individual immune functions are discussed one by one, and data are presented on how mutations change these functions. And in the final two sections, the possible contributions-past and future-of H - 2 mutations to immunology are considered. II. The H-2 Syllabus

There is no need for a detailed description of H - 2 genetics here, since the topic has been dealt with in several recent reviews (Klein, 1975; Snell et al., 1976; Gotze, 1977). The purpose of this section is merely to bring up to date these summaries and to provide a reference point for further discussion. For the updating purpose, two figures are presented, one showing chromosome 17 (Fig. 3 ) , in which the H - 2 complex is located, and the other showing the genetic map of the H - 2 complex and its vicinity (Fig. 4).A few explanatory comments on these graphic summaries follow.

A. CHROMOSOME 17 The map of chromosome 17 (Fig. 3 ) has recently been enriched by several markers that should prove useful for H - 2 genetic studies. Among these, markers coding for isozymes are particularly promising. They are Ce-2, or kidney catalase-2 (Hoffman and Grieshaber, 1977); Pgk-2, or phosphoglycerate kinase-2 (Cherry and Eicher, 1976; VandeBerg and Klein, 1978);ApZ, or acid phosphatase of the liver (Womack and Eicher, 1976); and M a p - 2 , or mannosidase processing-2 (Dizik and Elliott, 1977).The four loci have been placed at the D end of H - 2 , but their exact locations in relation to each other and to other loci in that region remain to be established. On the centromeric side of H - 2 , one more gene-Kb, or knobbly (Lyon, 1977)-has been added to the growing list of tail-affecting loci. B. THE H - 2 MAP The H - 2 map (Fig. 4 ) remains as unstable as always: regions are added to it, and others are removed. The figure shows a relatively con-

'qk M-1' La

T

0

2

1

3

'

1

3 "

w~0p-2 mf

n-2 00-

Kb

11 2 I

5

I

rE0-2 I

Ill

I

Ill

7

II I

'

6

I,-& 1

2 '

FIG.3. Genetic map of chromosome 17. Loci from the left to the right are: T, brachyury or short tail; 9k, quaking; Hst-1, hybrid sterility-1; Low, low transmission ratio; Fu, fused; tf, tufted; Kb, knobbly; H-2, histocompatibility-2; Qa, Q-region antigens; Tla, thymus-leukemia antigens; Ce, liver catalase; Pgk-2, phosphoglycerate kinase-2; Apl, acid phosphatase-liver; Map-2, mannosidase processing-2; thf, thin fur; Eo-2 erythrocyte antigen-2; 17-3, immune response-5. References for loci not mentioned in the text can b e found in Klein (1975).Brackets indicate that the order of loci is not known. The centromere is depicted by a full circle.

Regions 8 Subregions: LOCI

H-39

I K A H-33 H-2K 10-1

J E C S G D L Ia-2 10-4 10-5 10-3 %Sip H-2G H-2D H-2L Pa-1

T

B

H-31

00-2 Qo-3

H-32

TI0

FIG.4. Genetic map of the H-2 complex and its vicinity. Only one locus for each region or subregion is shown. The loci are as follows: H , histocompatibility; Ia, I-region associated antigen; Ss, serum serological; S l p , sex-limited protein; Qa, Q-region antigen; Tla, thymus-leukemia antigen.

62

JAN KLEIN

servative view of the map, particularly with respect to the Z region. In addition to H-2 loci, Fig. 4 also shows loci immediately adjacent to the H-2 complex. These are of two types: loci detectable by histogenetic (H-31,H-32, H-33, andH-36, cf. Flaherty and Wachtel, 1975; Flaherty, 1975; Artzt et al., 1977) and by serological (Qa-1,Qa-2, Qa-3, and Tla, cf. Flaherty, 1976; Stanton and Boyse, 1976; Old et al., 1963) methods. Of particular interest here is the H-2L locus, mapping in close proximity to H-2D in the D region. The locus is described in more detail below. At this point it is important to define the terms “region” and “ locus.” Region is a segment of chromosome of unspecified length, into which at least one locus has been placed b y recombinational analysis. A region could contain just this one locus, or several additional loci about which we may have no information. Locus, on the other hand, is a unit of genetic information coding for a single polypeptide chain. Thus, the K region, for example, contains the H-2K locus, and we have no way of knowing as yet whether this is the only locus in that region or whether additional loci might later be discovered in the region. Alternative forms of both loci and regions are referred to as alleles, and a particular combination of alleles at H-2 loci is known as the H-2 haplotype. C . THEH-2L Locus The H-2L locus is a new addition to the H-2 map, and so it might be of some interest to describe how it was discovered. In 1961 Stimpfling and Pizarro described an AKR (H-2k) anti-AKR.M ( H - 2 9 antiserum that contained at least five H-2 antibodies: anti-H-2.13,27,28,29, and 30. Of these, H-2.13 and H-2.30 were relatively easy to identify, but the identification of the remaining three antigens posed some problems for later investigators. Sometimes certain strains were typed variously as positive or negative for one or all three of these antigens, depending on what antiserum was used for the typing. Snell et al. (1974)tried to resolve the confusion by producing several anti-H-2.27, 28,29 sera and analyzing them in a systematic manner. They came to the conclusion that the three antigens constituted a family in which the individual members had a similar, though not identical, haplotype distribution. The authors postulated two duplicated sites coding for the three antigens, one in the K region and one in the D region. Among the antisera that they used was one that later became crucially important for the interpretation of this complex antigenic situation: (B1O.BR X A.CA)F1anti-A.SW (later known as D-28; haplotype combination k lf anti-s). Another important antiserum, designated D-28b, was later produced at The Jackson Laboratory in strain combination

H-2

MUTATIONS

63

(B1O.BR x LP.RII1) F1 anti-BlO.A(2R) or H-2k/H-2' anti-H-2h2(Snell, 1974). The first indication that the H-2.28 family could be even more complex than had been thought came from a cocapping study of Lemonnier et al. (1975). These authors reported that capping of H-2.28 was aIways accompanied by concomitant capping of H-2.4, a typical D-region antigen, but redistribution of H-2.4 always left a significant portion of H-2.28 in a diffused state. They concluded, therefore, that there were two types of molecule controlled by D-end genes: one type-presumably controlled by the H-2D locus-carrying both H-2.4 and H-2.28 antigens, and another type carrying only H-2.28 antigens. The latter type was thought to be controlled by another locus, which Neauport-Sautes and Demant (1977) designated H-2L. The hypothesis of two loci was supported by biochemical studies of Demant et al. (1975),which showed that the H-2.4 and H-2.28 molecules had similar molecular weights but different peptide composition. However, in 1975, Hauptfeld and Klein reported that when the D-28 antiserum was used on HTG ( H - 2 9 cells, antigen H-2.28 cocapped with a K-region antigen, H-2.31. The finding clearly established that in H-2O, antigen H-2.28 was controlled by the H-2K locus. Thus, there appear to be at least three loci coding for H-2.28: H-2K, H-2D, and H-2L. The H-2K and H-2D loci encode, in addition to H-2.28, also private and other public antigens, whereas no antigen other than H-2.28 is known to be encoded by the H - 2 L locus. Ill. Basic Terms in Mutation Genetics

Mutation is a hereditable change of genetic material not caused by recombination. Depending on the extent of the change, mutations can be placed in two categories, chromosomal and gene mutations. Chromosomal mututions involve entire chromosomes or chromosomal segments. They may lead to a deletion or duplication of individual genes or strings of genes, an insertion of a chromosomal segment into a new position, or an exchange of genetic material between nonhomologous chromosomes (translocation). Gene or point mutations, on the other hand, are alterations of one or more nucleotide pairs within a single gene. The alteration can result in a replacement (substitution), addition (duplication), or omission (deletion) of one or more nucleotide pairs. In substitutions, the new nucleotide pair can change the sense of a codon (nucleotide triplet) so that the new codon then determines an amino acid different from the one encoded by the original triplet; or the replacement can result in a loss of sense, and the nonsense codon thus generated can then provide a signal to terminate the message prematurely. In the latter case, amino acids controlled by the

64

JAN KLEIN

DNA segment past the nonsense codon will be absent in the resulting polypeptide chain, Addition or deletion of a single nucleotide pair results in the shift of the reading frame (frameshift mutation), so that the nucleotide sequence following the mutant site is then read with an altered meaning, and the amino acid sequence is changed. On the other hand, addition or deletion of an entire codon may result in addition or deletion of a single amino acid, without affecting amino acids in the rest of the polypeptide chain. A gene (allele) altered by a mutation is referred to as a mutant gene and is contrasted with a wild-type gene, from which it arose. A second mutation restoring the wild-type phenotype is known as reverse mutation. The restoration can, in principle, occur in two ways: the second mutation can restore the original nucleotide sequence of the gene (= genuine reverse mutation), or the second mutation can cancel out the effect of the first mutation (= suppressor mutation). Genuine reverse mutation must occur in the same nucleotide (or nucleotides) as the forward mutation, whereas a suppressor mutation can occur at a different position in the same gene or a different one. A wild population usually contains one or two alleles at most loci; these alleles are then rightly considered to be the wild type. With H-2 the situation is different. Here, a wild population may contain a large number of alleles, presumably including those having arisen by mutation in the laboratory (cf. Section IX), Therefore, it makes no sense to refer to those alleles from which mutant alleles have been derived as “wild-type.” Instead, I shall refer to such alleles (haplotypes) as standard, since they are the yardstick against which mutant alleles are compared. Mutations obtained after treatment of an organism with an agent (mutagen) known to increase the mutation frequency are referred to as induced, and are contrasted to spontaneous mutations, occurring in the absence of such a treatment. For further information regarding the general genetics of mutations, the reader is referred to the monographs by Auerbach (1976). IV. Methods of Histocompatibility Mutation Study

A. DETECTION

1 . Skin Grafting To detect mutations at H - 2 loci, groups of 10-20 mice are selected, and the individual animals are skin-grafted in a reciprocal-circle fashion (see upper part of Fig. 5 ) . In the circle, each mouse donates two

H-2 BALB/c

MUTATIONS

65

C57BL/6(:B6)

LF,/

I

BCn x BCn

I

86 - H - 2 mu'on'

FIG.5. Method of mutation-detection by skin grafting in a reciprocal-circle fashion. Explanation in text. [Reproduced from Klein (197614 with permission of Plenum Press, New York.]

grafts, one to each of its two neighbors, and receives two grafts, one from each of the two neighbors. The grafting is most rapidly done using the tail-to-tail technique of Bailey and Usama (1960),but other techniques can also be used. Mice in the circle can be either F, hybrids derived by crossing of two inbred strains ( H - 2 congenic lines), or mice of any standard inbred strain ( H - 2 congenic line). The use of F, hybrids has the advantage of allowing the detection of certain types of mutation that would otherwise b e undetectable (see Section IV,F). An alternative way to detect H - 2 mutations is to graft skin from two inbred strains or congenic lines onto F, hybrids derived by crossing the two strains. The hybrids should accept the grafts unless one of their H loci has mutated.

2 . Other Methods Since many other traits have been associated with the H - 2 complex, it should be possible, at least theoretically, to detect H - 2 mutations b y methods other than skin grafting. Among these, typing for serologically detectable antigens might seem most suitable. However, when this method was used, no mutants were found among several thousand screened mice ( J . Klein, unpublished data). In view of what is now known about the H - 2 mutant serology (see Section X,G), the failure is not surprising, and the method cannot be recommended for mutation detection. Other methods for mutant screening, such as typing for the mixed lymphocyte reaction (MLR) response or typing for the H - 2 controlled immune response to antigens are far more cumbersome and time-consuming than skin grafting or serological typing.

66

JAN KLEIN

Such methods, however, could turn up mutations that might be undetectable by skin grafting, and their use for H - 2 mutant detection would therefore be highly desirable.

B. PROGENY TEST When a mutation appears to be the cause of an unexpected graft rejection, the genetic nature of the suspected mutation must first be established by a progeny test. The anomalous animal must be mated to another mouse and the progeny examined. The examination consists of exchanging skin grafts between the progeny and mice of the original inbred strain or of the F, hybrid, depending on the origin of the mutation. If the ability to reject grafts proves to be transmitted to the progeny, two mice carrying the mutation are mated inter se and a line homozygous for the mutation is established from the progeny of this mating.

C. ISOLATION If the mutation occurred in an inbred strain, the establishment of a line homozygous for the mutation is all that is needed to make the mutation available for further testing. The new line is then coisogenic with the original (standard) inbred strain, since it differs from this strain (at least until it accumulates additional mutations) at a single locus. If, on the other hand, the mutation occurred in an F, hybrid, the mutant line, although homozygous for the mutation, is heterozygous at other loci. The quickest way to make the new line homozygous at all loci is to backcross it repeatedly to one of the parental strains from which the original F, hybrid was derived (see lower part of Fig. 5). In each backcross generation, mutation-bearing animals are detected by transplanting their skin to the F, hybrids. (An F, hybrid is a universal recipient: it accepts grafts from all backcross animals, except those carrying the mutation.) After 10- 12 backcross generations, the mutant heterozygotes are intercrossed and a congenic line is established. (A congenic line, in contrast to a coisogenic line, may differ from its inbred partner in more than the differential locus.) The backcrossing thus “isolates” the mutation from the rest of the genome and results in a congenic line, which can then be used to study the effects of the mutation without the interfering influences of other H genes.

D. LINKAGETEST Since a mutation at any of t h e 40 or so H loci may lead to graft rejection, the mutation detection methods will pick H mutations indis-

H-2

MUTATIONS

67

criminately. Although one can get some inkling as to whether the mutation occurred in H - 2 or in one of the minor H loci from the speed of the graft rejection, one finally has to determine genetically where the mutation occurred. This determination is usually accomplished by a linkage test, in which a segregating population is tested simultaneously for the presence of the mutation (by skin grafting) and for the serologically detectable H-2 antigens. If the mutation originated in an F, hybrid, the linkage test could be done during the backcrossing designed for the production of the congenic line; otherwise, a special backcross mating must be set up. To avoid the time-consuming linkage test, a complementation study using H - 2 recombinants may b e done. If H - 2 genes can be shown to complement the mutation, obviously the mutation must have occurred in H - 2 . This method is described in detail below. E. NAMINGTHE MUTATION Once it is established that a mutation has occurred in one of the H - 2 genes, the mutation is given a genetic name. T h e haplotype carrying the mutation is assigned a double-letter superscript symbol, in which the first letter indicates the haplotype from which the mutation is derived, and the second letter represents a serial designation of the mutation. For the serial designations, letters in the first half of the alphabet are used. (These letters are reserved for minor variants of major haplotypes, and mutations are considered to be such variants.) For example, the first mutation in the H-2* haplotype was given the symbol H-2”l (Bailey and Kohn, 1965), the second, H-2bb (Bailey and Kohn, 1965), the fourth, H-2*d (Egorov and Blandova, 1968), etc. [The H-2bCsymbol was given to a naturally occurring minor variant of H - 2 * ; cf. Snell et ul. (1971b).] If two mutations which by most criteria appear to be identical occur, they are given identical double-letter symbols and are distinguished by arabic numerals following the letters. For example, the two H-2*g mutations, which are indistinguishable by immunological tests but may differ biochemically, have been assigned symbols H-2bg1 and H-2*g2 (Melief et nl., 1975; Melvold and Kohn, 1976). Congenic or coisogenic lines carrying H - 2 mutations can be designated in two ways: by the mutant haplotype (preferred way) or b y some other symbol. For example, the C57BL/6 line carrying the H-2hd mutant haplotype can be referred to either as B6-H-26d or B6.M505 (abbreviated M505), where M505 is the serial number of the mutant animal (Egorov and Blandova, 1968).Congenic lines derived from F, hybrids are designated by symbols that also reflect the origin from the

68

JAN KLEIN

two parental strains. For example, the H-26Q-bearingline, which is derived from (C57BL/6 x BALB/c)F, hybrid, is designated either B6.C-H-26Q(where C stands for BALB/c) or BG.C(Hz1) [where H z l was a tentative designation of the locus in which the mutation occurred, before the locus was identified as H - 2 K ; cf. Bailey and Kohn (19631.

F. DETERMINATION OF THE MUTATIONTYPE In the allograft reaction, the reaction on which the detection of H-2 mutations is based, antigens carried by the transplant are recognized

by specific receptors on T lymphocytes. As will be discussed later in this chapter, in vitro studies indicate that the H-2 antigens recognized by T cells are complex in the sense that different T cells presumably recognize different portions of the H-2 molecule. Following the practice used by H-2 serologists, the different portions of the molecule can be considered as distinct antigenic determinants and can be assigned numerical symbols. A mutation changing only a portion of the H-2 molecule can then, in formal terms, have one of the following effects. First, the mutation can result in a loss of one or more determinants while leaving other determinants on the same molecule unaffected: standard-type determinants: mutant-type determinants:

1,2,I,-,-

This situation will result in the rejection of the standard-type graft by the mutant, whereas the mutant-type graft will be accepted by the standard-type recipient. (The basic genetic rule governing graft rejection is that transplants carrying extra antigens, that is, antigens absent in the recipient, are rejected, whereas transplants lacking antigens present in the recipient may be accepted.) Second, the mutation can result in a gain of one or more antigenic determinants: standard-type determinants: mutant-type determinants :

1,2,1,2,3

In this case, the standard-type host will reject the mutant transplant, while the mutant will not reject the standard-type graft. Third, the mutation can result in a loss of one or more determinants and a gain of other determinants: standard-type determinants: mutant-type determinants:

1,2,1y-J

H-2

69

MUTATIONS

In this situation, grafts exchanged reciprocally between standard and mutant strains will be rejected.

G. COMPLEMENTATION TEST

1. Principle

The final step in the genetic analysis of an H-2 mutation is to determine in which of the H-2 loci the mutation occurred. This step is accomplished by utilizing the genetic complementation test. T h e principle of the test is this (Fig. 6). Consider two different mutations, m , and m2,one carried by one homologous chromosome, and the other by the second chromosome. Such mutations are said to b e in a transconfiguration: m1+

+m

2

affecting two different loci (cistrons, Fig. 6a) or two different sites (mutons) at the same locus (Fig. fib). In the former instance, a cell in a trans configuration will produce four types of gene product-one mutant and one wild type at each of the two loci. The wild-type products will lie able to carry out normal cell functions, and the overall pheno-

a > w

&&

&..f-t ]-I.= +

I-

s Ba

n w

W z

CI

m2

36 0% +

PHENO TYPE

WILD

m2

3 m2 MUTANT

FIG.6. Principle of genetic complementation. Two situations are considered: In the left-hand panel, mutations m , and m2 occurred i n two different genes so that one wildtype copy of each of the two genes is present in each cell; in the right-hand panel, the two mutations occurred in the same gene, so that only mutant copies of this gene are present in each cell.

70

JAN KLEIN

type of the cell (or individual) will be of the wild type. If, however, the two mutations were to affect the same locus, the locus would produce only mutant gene products and the cell would display a mutant phenotype. Thus, b y placing two mutations into a single cell in a transconfiguration, a geneticist can tell whether the mutations OCcurred at two different loci or at a single locus. If the phenotype of such a cell were of the wild type, he would conclude that the mutations affected different loci; if it were of the mutant type, he would assume that the two mutations affected the same locus. It should be emphasized, however, that the complementation test is far from being foolproof in assigning mutations to loci. The problem is that two mutations at the same locus sometimes produce wild type, instead of a mutant phenotype. In some organisms and with some genes this intragenic complementation is quite frequent. Intragenic complementation often occurs between genes coding for multimeric proteins. It is explained by the assumption that when two monomers, carrying defects in different portions of their polypeptide chains, assemble into a dimer (polymer), the two defects cancel each other out. As a result, the dimer functions normally or almost normally.

2 . H - 2 Mapping

To determine in what region (and, if possible, in what locus) an H - 2 mutation has occurred, one produces an F, hybrid between the mutant and an H - 2 recombinant strain, and transplants skin from the standard strain to the heterozygous animals. The recombinant strain is chosen in such a way as to share with the standard strain part of the H - 2 complex, yet differ in another part. If the mutation has occurred in the unshared segment, the graft will be rejected, because the recombinant strain will not complement the defective region. In all other circumstances, the graft should be accepted. An example of such a complementation test is shown in Fig. 7. Here, a mutation has occurred in the H-26 haplotype and resulted in a new haplotype, H-26a. When the H-26a and H - 2 i haplotypes were combined in an F1 hybrid, the F, recipient failed to reject H-26 grafts; however, when the same mutant haplotype was combined with H-2h, the F1hybrid rejected the standard-type grafts. Since H-2' carries H - 2 6 genes in the K end of the H - 2 complex, the failure of the H-26a/H-2' heterozygotes to reject H-261H-26 grafts indicates that the mutation occurred somewhere in that end. The failure of the H-2' haplotype to complement the mutation indicates that the D end was unaffected by the mutation, since the H-2' is H-26 at this end. By using other H - 2 recombinants in combina-

H-2

MUTATIONS

71

FIG.7 . Mapping of an H - 2 mutation into a region. Arrows indicate the direction of skin grafting; broken arrow signifies rejection of a graft in a given direction. Solid rectangle represents the mutant site; shaded and open bars represent chromosomes of different genetic origin. See the text for explanation. [Reproduced from Klein (1976b) with pennission of the Plenum Press.]

tion with the mutant haplotype, the mutation can be mapped more precisely to one of the H - 2 regions. It might seem, from this description of complementation mapping, that the resolution of the mutational analysis depends on the resolution of the recombinational analysis, and so could not contribute much to the study of the H - 2 pleiotropism. One can argue that even after mapping a mutation into a particular locus, one can never be sure that a second mutation has not occurred in an adjacent locus in the same region. Theoretically, such an argument is, of course, valid. But in practice the probability that two mutations would occur simultaneously in adjacent genes is very low, and the probability of picking up-without selection-a second mutation later is negligible. In addition, there are ways-as will be discussed later-of identifying the mutant genes without reliance on H - 2 recombinants. Thus, despite the theoretical handicap, the resolution power of the mutational analysis is incomparably higher than that of the recombinational analysis.

3. M a p p i n g of M u l t i p l e Mutations When more than one mutation maps into the same region, the question arises whether these mutations affect the same locus or different loci within this region. T o answer this question, the genetic complementation test is performed between the different mutants. I n this test two types of skin grafting are performed: in one type, grafts from the standard strain are placed on an F, hybrid between two mutant strains; in the second type, grafts are exchanged between the two mutant strains. Theoretically, the following three situations could occur (Fig. 8). First, the two mutations affect the same site (muton) at a single locus

JAN KLEIN

SAME CISTRONSAME MUTON

SAME CISTRONDIFFERENT MUTONS

T W O DIFFERENT CISTRONS

I 0 0

0 0

0 0

0 0

0 0

0

i

11

++

1

0 0

1

+i

0

0

0

0'0 00

00 OQ

FIG. 8. Principle of establishing the genetic relationship between two mutations. Open rectangles represent genes (cistrons);solid rectangles, mutant sites (mutons); circles (stars), gene products. Direction of skin grafting is indicated by arrows, rejection by doubly crossed arrows. See explanation in text. [Reproduced from Klein (1976b) with permission of Plenum Press.]

(cistron; see left panel in Fig. 8). In this situation, the F1 hybrid will reject grafts from the standard strain, but the two mutant strains will accept each other's grafts. Such is the case, for example, with the H-2b81 and H-2b02 mutants (Melief et d.,1975; Melvold and Kohn,

1976). Second, the two mutations affect different mutons in the same cistron (see middle panel in Fig. 8).This situation is exemplified by the H-2ba and H-2bd mutants, which reject each other's grafts; their F, hybrid rejects grafts from the standard strain (Egorov and Blandova,

1968). Third, the two mutations affect two different loci (see right panel in Fig. 8). One mutation in the H - 2 K b locus (e.g., H-2Kba) and another (still hypothetical) in H - 2 D b would be an example of this situation. T h e F, hybrid should then accept H-2O grafts but the two mutants should reject each other's grafts. The above considerations are based on the assumption that intragenic complementation does not occur in the H - 2 genes. However, this assumption might be incorrect, since one can imagine a situation that, in terms of antigenic determinants, might look like this: standard strain: mutant strain No. 1: mutant strain No. 2:

-9293,1,-,3,-,2,-,4

H-2

MUTATIONS

73

In this situation, the F, hybrid would accept standard-strain grafts and the mutant strains would reject each other’s grafts, despite the fact that both mutations occurred at the same locus. Whether such a situation does occur remains to be determined. V. Mutation Rates

A. GENERALCONSIDERATIONS

The number of mutations occurring per unit of genetic information and per unit of time is referred to as the mutution rate. The unit of genetic information usually used is the gene, but mutation rates can also he expressed per group of genes or-at the opposite extremeper nucleotide (muton). The type of time unit used depends on the type of mutation studied. Mutations occurring independently of DNA or RNA replication (and such mutations do, indeed, occur, though usually with frequencies lower than those of replication-dependent mutations) are measured in terms of “clock time,” that is, in hours, days, or months. Replication-dependent mutations, on the other hand, are measured in units of “biological time,” usually in “per generation” terms. In an ideal situation, one generation would be equivalent to one replication cycle, as is the case in unicellular organisms. In bacteria, for example, mutation rates can be expressed on a “per cell” basis, because here one cell represents one generation or one replication cycle. In multicellular organisms, counting mutations on a per-replication-cycle basis is not possible, and so the generation time must b e measured in some other way. In the mouse, mutation rates are usually expressed in “per gamete” terms, although the mutations are not detected directly in the gametes, but in the progeny derived from them. The fact that each individual mouse is derived from a zygote that arose by fusion of two gametes must, therefore, be taken into account. Whether a mutation in both gametes forming the zygote or in only one of them can be detected depends on the mating system used. I n the former case, the mutation rate ( p ) is calculated from the formula: p = m/2n, where m is the number of mutation-carrying mice, and n is the number of mice screened. If mutation can be detected in only one of the two contributing gametes, the formula for the calculation of mutation rates is p = mln. The choice of a mating system used for mutation detection is dictated by the type of mutation sought. The types of nonhistocompatibility mutation that have been studied in the mouse are recessive visible, dominant visible, dominant lethal, and recessive lethal. For

74

JAN KLEIN

comparison, the methods used for the detection of these mutation types and the methods of mutation-rate calculations are described briefly below.

I. Recessive Visible Mutations To detect recessive visible mutations, one instructs the animal caretaker to watch for any variation in the phenotype of mice maintained in the colony. This approach was used by Schlager and Dickie (1971), who, over a 6-year period, screened over 7 million mice at The Jackson Laboratory, Bar Harbor, Maine. The screening was done on standard inbred strains homozygous for one or more recessive coat color genes, or on F, hybrids derived by mating recessive homozygotes at a particular coat color locus with homozygotes for the wild-type alleles at this locus. In the inbred strains, only reverse mutations can be detected. For example, strain C57BL/6 is an ala homozygote (where a is a recessive mutant allele at the agouti locus), and therefore of nonagouti (black) color. If, however, a reverse mutation a + were to occur, and the +- bearing gamete would participate in fertilization, the resulting +/a heterozygote would be easily recognized because of its agouti color. Since a mutation in any of the two contributing gametes would be detected in such animals, the mutation rate would be estimated from the formula p = m/2n. The inbred-strain matings do not provide any information regarding forward recessive visible mutation. However, such information can be obtained from F, hybrids produced by a mating of two different inbred strains. To illustrate the latter situation, let us consider an F, hybrid between strains A/ J (genotypically ala, blb, and clc, where a, b, and c are different coat color loci) and C57BL/6 (genotypicaly a/a, B/B, and C I C ) . Normally, the F, hybrid would have an ala, Blb, and Clc genotype, and would be black. If, however, a forward mutation in the C57BL/6 strain were to change B to b or C to c and the mutationbearing gamete were to participate in fertilization, a brown (ala, blb, Clc) or an albino (ala, B/b, clc) mouse would appear among the F, hybrids. Since only mutations contributed by one (C57BL/6),but not the J), parent are detected in this mating (of the two zygote-formother (A/ ing gametes, only one has the potential of contributing a detectable mutation), the mutation rate is estimated from the formula p = mln. (Since the F, hybrid is homozygous at the a locus, reverse mutations at this locus can be detected in the same way as in standard inbred strains .) The frequencies of mutations at individual loci estimated in inbred strains and their F1 hybrids are given in Table 11. The overall muta-

-

TABLE I1 FREQUENCYOF SPONTANEOUS MUTATIONS AT VISIBLE RECESSWE LOCI OF THE MOUSE Frequency found by investigator Locus and allele

Lyon et al., 1972

Russell, 1951

Schlager and Dickie, 1971

Batchelor et al., 1969

Lyon and Morris, 1966

a+

0/157,42 1

0137,868

019328

41157.42 1

0137.868

C+

0/157,42 1

0137,868

3167,395 3418,167,854 31919,699 013,092,806 51150,391 013,423,724

0137,813

a b+ b C

0137,813

CCh

d+

21157,421

1137,868

d se+ d+se+ P+ S+

In+ In

bP+

fi

+

Pa pe+ Total +

0137,813

Ol157,42 1 21157,421 11157,421 21157,421

101839,447 912,286,472

0137,868

0137,8 13 0/37,8 13

0137,868 1137.868

0/37,8 13 0137,8 13 41243,444 01266,122

019328 019328 019328 019328 Off614

Total

Mutation rate ( x per locus per gamete

31309,825 3418,167,854 7/1,152,80 1 0/3,092,806 51345,680 013,423,724 0137,813 1311,034,736 912,324,285 01233,102 21157,421 11233,102 31233,102 41252,772 01266,122 019,328 019,328 019,328 O f f ,614 8 1/21,300,743

9.6 4.1 6.0 < 0.97 14.4 < 0.88 < 79.3 12.5 3.8 < 12.8 12.7 4.2 12.8 15.8 < 11.3 C321.0 <321.0 <321.0 3.8 3.8

? tQ

5

2

t! 0

5

76

JAN KLEIN

tion rates for the five coat color loci studied by Schlager and Dickie (1971) are 11 x 10+ per locus per gamete for mutations from wild type and 2.5 x for mutations from recessive alleles. Another way to detect recessive visible mutations is the specijk locus method (Russell, 1951), which is similar to the F, hybrid method. In the specific locus method, wild-type males or females are mated to partners of the opposite sex, homozygous for several (up to seven have been used) recessive tester genes, purposely placed together into one (tester) stock. The progeny of the mating are then scored for individuals with mutant phenotypes controlled by the tester loci. Most offspring are of the wild type, since they are heterozygous at the tester loci. However, whenever a mutant gamete is contributed to the zygote by the wild-type parent, the resulting individual becomes homozygous for the recessive mutation and expresses the mutant phenotype. For example, the appearance of a black mouse among the progeny of a mating between a stock homozygous at a, b, crh,p , d, se, and s loci [the seven loci tested in the original Russell (1951)experiment] and a strain homozygous for the wild-type alleles at these loci, indicates that the wild-type mouse contributed a mutation that changed a+ + a , so that the hybrid became a h , +lb, +lcrh, / p , /d, + h e , /s. As in the F, hybrid method, in the specific locus method (the difference between the two is only that in the latter a special tester stock is procured for the testing), the mutation rate is calculated using the formula p = mln. Large specific-locus experiments, designed to test the effects of X-irradiation on mutation induction, were carried out at the Oak Ridge National Laboratory in the United States (Russell, 1951) and at Harwell in the United Kingdom (Lyon and Morris, 1966; Lyon et al., 1972). The results from these experiments with respect to the control group (unirradiated mice) are shown in Table 11. The mutation rates obtained were 7.5 x lop6(Russell, 1951) and 10.0 x lop6(Lyon et al., 1972) per locus per gamete. In a third experiment (Lyon and Morris, 1966),in which a different set of recessive loci (In, bp, fi, pa, p e ) was screened for visible mutations, no mutation was found among 9328 mice tested. The overall spontaneous rate for the recessive visible mutations is 3.8 x lop6per locus per gamete, or 5.9 x lop6for forward mutations and 9.8 x for reverse mutations.

+

+

+

2 . Dominant Visible Mutations Dominant visible mutations can b e detected by external or internal examination of the mouse body. The externally visible mutations are those affecting tail length, coat color, hair and whisker growth, toe

H-2

77

MUTATIONS

growth, and the like. The study of internally observable mutations has been limited primarily to skeletal defects and variations. Because of the dominant nature of these mutations, they can be detected in practically any mating type and no special method is needed for their discovery. The number of loci at which dominant visible mutations can occur is not known, and it is therefore impossible to estimate an overall mutation rate for these loci. One can, however, calculate the average mutation rate for those loci at which mutations have been detected. Since mutation in either parent can result in a mutant offspring (i.e., mutations in both contributing gametes are detected), the mutation rate is calculated using the formula p = m/2n. Extensive data on externally detectable mutations have been compiled by Schlager and Dickie (1971), who detected mutations at 12 loci (Table 111) and estimated the average mutation rate for these loci to be 4.4 x lop7per locus per gamete. The most mutable locus was W (white spotted coat color), at which 23 mutations were detected among some 5.2 million per locus per gamete, mice examined (mutation rate of 2.2 x comparable to the rate of the recessive visible genes). The least mutable loci had a mutation rate of 7 x lo-* per locus per gamete, which is two levels of magnitude lower than the rate of the recessive visible mutations. Judging from the number of independent occurrences of short-tail ( T / + )animals in different laboratories, the brachyury of T would seem TABLE 111 FREQUENCY OF SPONTANEOUS MUTATIONSAT VISIBLE DOMINANT LOCI"

Allele W SP TO"

s1

TU" Xt MoSr Re Bn Hx Ru Lm

Frequency

23/5,226,531 12/5,226,531 6/7,010,732 515,226,531 117,010,732 1/7,010,732 1R,010,732 1/7,010,732 1/7,010,732 117,010,732 1/7,010,732 1/7,010,732

Average

" After Schlager and Dickie (1971). Sex-linked; calculation assumed 50 :50 sex ratio.

Mutation rate per locus per gamete ( x

2.20 1.15 0.77 0.48 0.13 0.07 0.13 0.07 0.07 0.07 0.07 0.07 0.44

78

JAN KLEIN

to be another highly mutable locus (Bennett, 1975). However, Schlager and Dickie (1971) did not detect any mutation at this locus. In laboratories in which T mutations had occurred, no records were kept of the number of animals that were examined, and so no estimate of the spontaneous mutation rate at T locus is available. However, in the progeny of X-irradiated mice (100 R 500 R), Selby and Selby (1977, and personal communication) detected five brachyury mutations among 133,351 mice examined (induced mutation rate of 1.8 x lop5per locus per gamete). In the same experiment, no T mutations were observed among 38,448 control, unirradiated mice. An extensive search for skeletal mutations was carried out b y Ehling (1966), who detected in his control group one skeletal variant among 739 mice examined. Assuming that this variant was a true mutation (of which one could not be sure because the animals were killed for examination and no progeny test was done), the spontaneous mutation rate for the group of loci affecting skeletal variation is 2.9 x lop4 per gamete.

+

3 . Dominant Lethal Mutations In this category of mutations, a gene alteration in one of the parents causes the death of the offspring inheriting the altered gene. Since the mutations usually affect development in utero, it is in the embryos that they are found. The method for the detection of dominant lethal mutations consists of dissecting pregnant females at some advanced stage of gestation and counting live and dead embryos (Rohrborn, 1970). Since deciduomata resulting from early embryonic death remain until late pregnancy, counting is possible in advanced stages of gestation, allowing detection of lethality during the entire period of embryonic development. The death of an embryo is presumed to be the consequence of a dominant lethal gene action. Of course, no proof exists in this type of study that the defect is genetic in nature, and the genes involved cannot be identified. Many of the dominant lethal effects are caused by chromosomal aberrations, such as translocations or inversions. The average frequency for the group of loci causing dominant lethality is 6-15% (as measured by counting dead implants; cf. Ehling et al., 1968).

4 . Recessive Lethal Mutations Of all mutations, the recessive lethals are the most difficult to detect. Because their effect is discernible only in the homozygous state, a special mating system must be designed for their detection. In this system, a newly arisen mutation must first be multiplied to produce at

H-2

MUTATIONS

79

least two heterozygotes, which can then he mated together and their progeny tested. A further complication arises from the fact that each individual carries a certain number of preexisting recessive lethals, and it is therefore necessary to differentiate between the preexisting and the newly arisen lethals. The detection is somewhat easier for lethals present in sex chromosomes, since in the male these chromosomes carry only one dose of each gene, and so there is no need to produce homozygotes for the detection of the lethal effect. Several systems for the detection of autosomal, recessive lethal mutations have been designed. The most common method is to mate the animal suspected of carrying a mutation to another individual, and to backcross the resulting F, hybrids to this carrier. If the original animal has indeed carried a mutation, it would have transmitted it to about onehalf of the F, hybrids; and if several of the F, hybrids were backcrossed, some of the backcrosses should have produced lethal progeny. Lethal effect can then be inferred from the presence of deciduomata in pregnant females, from reduced litter size, or from distorted segregation ratios of a marker gene, if the mutation occurred in a chromosome carrying this marker. Appropriate control mating must be carried out to show that the parents of the original animal did not carry the mutation and that it was in fact a newly arisen mutation. For details on the recessive-lethal methodology, see Luning (1971). Because of the complicated mating systems used for the detection of recessive lethal mutations, estimating the mutation rate is not simple. The number of loci at which recessive lethals can occur is not known, so the rate can b e expressed only per group of loci. Luning and Searle (1971) estimated the overall mutation rate of the recessive lethals to be 2.9 x per gamete.

MUTATIONS B. RATES OF HISTOCOMPATIBILITY An active search for H mutations has been carried on by three principal investigators: Donald W. Bailey, T h e Jackson Laboratory, Bar Harbor, Maine; Igor K. Egorov, Institute of General Genetics, Academy of Sciences, Moscow, U. S. S. R.; and Henry I. Kohn, Harvard Medical School, Boston, Massachusetts. In the laboratories of these investigators, 50,394 mice were skin-grafted, and 230 mutant mice were detected (Table IVA). In other words, about 0.5% of the tested mice carried an H mutation. In several instances a single litter had two or more individuals carrying the same mutation. It is extremely improbable that these mutation clusters represent independent occurrences of the same mutation among closely related individuals. It is more likely that they signify the fact that one parent of the mutant

TABLE IVA FREQUENCY OF HISTOCOMPATIBILITY MUTATIONS IN INBRED AND F, HYBRIDMICE

Strain (C57BU6 X BALB/c)F,

(A.CA X A)F, (B10.D2

X

C57BL/10)F1

(C57BU10 x B1O.D2)F1 C57BU6 BALBIc

(C57BU6 x BALB/c)F, (C57BU6 x BALB/c)F, Totals

-

Treatment of the parent

00

0

No. of mutants found

Total No. tested

Percent of mice with mutation

None

68

2,609

2.6

X-rays None Diethyl sulfate None

57 0 1 1

3,156 204 514 519

Diethyl sulfate Diethyl sulfate None Triethylenemelamine None

14 1 7 0

864 154 4,003 670

1.6 0.6

2

4,889

5

1,211

"'r

Melvold and Kohn, 1975

35 3

17,424 563

0.5 o.2

Kohn, 1973

8 28 121 85 24 230

2,335 11,279 31,983 14,435 3,976 50,394

0.2 0.3 0.4 0.6 0.6 0.4

Triethylenemelamine None Triethylenemelamine None X-rays None X-rays Chemical mutagens Treated + untreated

Reference Bailey and Kohn, 1965; Bailey, 1966

0.2

0.21 1

Egorov and Blandova, 1972 Egorov and Blandova, 1972 u

Egorov, 1967 Kohn, 1973

o.2

E

0.4

1 I

i2 P

Kohn et al., 1976

H=2

MUTATIONS

81

progeny was either a mutation carrier or a gonadal mosaic. The former would arise if the mutation occurred in one of the grandparents or great-grandparents so that the parent would be heterozygous for the mutation. The latter would occur if the mutation were to arise in the parent, but in an early stage of germ-cell differentiation. For instance, a mutation in one of the germ cells colonizing the testes during embryonic development would produce patches of seminiferous tissues carrying the mutation. Such a genetically mosaic male would then constantly release a certain number of mutant speim from the patch, and the probability of mutation detection in the progeny would b e correspondingly increased. Since one is usually interested only in independently occurring mutations, clusters are either excluded from calculations of mutation frequency, or each cluster is counted as a single mutation. The frequency of H mutations after the adjustment for mutation clusters (and exclusion of data in which such an adjustment was not possible) is 0.2% (105independent mutations in 45,287 tested mice, cf. Table IVB). It will be the corrected data in Table IVB with which we will concern ourselves in the remainder of this section. Two attempts have been made to increase the frequency of H mutations by X-irradiation (Bailey, 1966; Kohn et al., 1976),but both failed. X-rays were found to have no effect on H mutation induction-in fact, in some experiments they even seemed to decrease mutation frequency. This finding, which applies to a wide dose range of X-rays (Kohn et al., 1976), is in sharp contrast to experiments with non-H loci. For example, in the specific-locus test, mutations at loci affecting pigmentation and cartilage growth increase significantly after the exposure of spermatogonia to X-rays, and the increase is dose-dependent (for a review, see Searle, 1974).Why X-rays fail to increase H-mutation frequencies is not known, but at least two possibilities can be considered. Some geneticists believe that X-rays induce primarily chromosomal abnormalities, such as deletions, duplications, or inversions, and it is possible that such changes in the H genes either result in inviable gametes or escape detection because of the nature of the screening test. The second possibility is that X-ray-induced lesions in H loci, for some reason, are very likely to be repaired (Kohn et al., 1976). Limited attempts have also been made to increase H mutation frequency by the treatment of males with chemical mutagens (alkylating agents, cf. Egorov and Blandova, 1972; Egorov, 1967; Kohn, 1973). A modest, although apparently significant, increase in the mutation rate was observed (Table IVB). The majority (about 90%) ofH mutations (both spontaneous and in-

TABLE IVB FREQUENCY OF MUTATIONS AT HISTOCOMPATIBILITYLOCI IN INBRED AND F, HYBRIDMICEAFTER EXCLUSION OF MUTATION CARRIERS AND/OR MUTATIONCLUSTERS

Strain combination

Treatment of parent

No. mutants foundltotal No. mice tested

Percent of mice with mutation

(C57BU6 X BALBlc)F, (C57BL/6 x BALBlc)F, (C57BU10 X BlO.D2)Fl

None X-rays None

1711,567 41656 1/519

0.7 0.6 0.2

(BlO.D2 X C57BWl0)F1

Diethyl sulfate

101864

1.1

(A.CA x A)F,

Diethyl sulfate

1/514"

0.2

C57BU6

None Triethylenemelamine None

514,003 01670

0.1

214,889

0.04

511,211

0.4

28117,454 31563

0.2 0.5

711,128 22111,279 60129,530 26111,935 1913,822 105145,287

0.6 0.2 0.2 0.2 0.5 0.2

BALBlc

(C57BU6 x BALBlc)F, (C57BU6 X BALBlc)F,

Totals

Triethylenemelamine None Triethylenemelamine None X-rays None X-rays Chemical mutagens Treated + untreated

The one mutation was of spontaneous origin.

Reference Bailey, 1966 Bailey, 1966 Egorov and Blandova, 1972 Egorov and Blandova, 1972 Egorov and Blandova, 1972

-

+ %

Z

P Melvold and Kohn, 1975; Kohn, 1973

Kohn et al., 1976

fl

2:

H-2

MUTATIONS

a3

duced) is of the gain or the gain-and-loss type, and only a small fraction (10%)is of the loss type. Three explanations can be offered for the preponderance of gain and gain-and-loss mutations. First, the preponderance could be an artifact of the screening method (Bailey and Kohn, 1965). In the reciprocal circle, loss mutations can be detected at heterozygous, but not at homozygous, H loci. The reason for the failure to detect loss mutation in homozygotes is this. Imagine an H locus coding for two antigenic determinants, 1 and 2. Assume further that one of the two determinants has been lost as a result of a mutation. The segment of the reciprocal circle involving the mutant animal will then look as follows:

- 1,2 - 1,2 - 1,2 -

.-@-&~-.&l&-

In other words, grafts from and to the mutant animal will not be rejected and the mutation will pass undetected. In a situation where mice in the reciprocal circle are of an inbred strain, and thus presumably homozygous at all H loci, no loss mutation will be detected. When F, hybrids are used to form the reciprocal circle, loss mutations will be detected only at loci at which the two parents of the hybrid differ. Since two inbred strains differ at some 30 or 40 H loci and share alleles at perhaps several hundreds of H loci (Bailey, 1968), it follows that the majority of loss mutations will escape detection even in an F, hybrid. Second, Bailey (1966) speculated that H mutations may result from the integration of proviruses into chromosomes, and that H antigens are, in fact, viral antigens capable of inducing allograft reaction. The observation cited in support of this speculation was that in “dirtier” environments (mouse colonies with high possibility of exposure to viral infections), H mutations seemed to be more frequent than in cleaner, more isolated environments (Bailey, 1966; Kohn and Melvold, 1974). Recently, the speculation has received a new impetus from the findings of Zinkernagel and Doherty (1975), Bevan (1975), and Gordon et al. (1975).The first two authors demonstrated that cellular immunity to cells infected with lymphocytic choriomeningitis (LCM) virus is H-2-restricted; that is to say, effector T cells generated against a particular combination of viral and H-2 antigens are able to kill cells carrying the same combination of antigens, but not cells in which the viral antigens occur with different H-2 antigens. Bevan (1975) and Gordon et al. (1975) demonstrated that the same H-2 restriction applies to non-H-2 histocompatibility antigens. The latter obser-

84

JAN KLEIN

vation would make sense if non-H-2 histocompatibility antigens were, in fact, viral antigens. Third, by far the simplest explanation of the observed disproportionality between gain and loss H mutations could lie in the nature of the H mutations themselves. It is likely that no matter how one alters the H molecule, the T cell will almost always recognize the alteration as foreign and the mutation will, therefore, classify as the gain type. A true loss mutation would-according to this explanation-result mostly from the physical deletion of the entire H gene, an event that could be expected to be rather infrequent. A word should also be said about the expression of H mutation frequencies. Since the number of H loci is not known, the frequencies cannot be expressed on the “per locus per gamete” basis. Instead, one can express the frequency as percent of mutations, as is done in Tables IVA and IVB, or as “mutants per 1000” (Kohn et ul., 1976). Once the mutated loci are identified, it should be possible to calculate mutation rates for the individual H genes.

C. H - 2

MUTATIONRATES

The data that one can use for calculation of H - 2 mutation rates are given in Table VA. The rates can be calculated in several ways, depending on how one interprets the data, and the resulting figures vary accordingly. The calculations in Table VB are based on the following assumptions. First, the reciprocal-circle skin-grafting technique detects mutations at only two H-2 regions, K and D. As mentioned earlier, this assumption might be incorrect since, at least theoretically, one should also be able to detect mutations at theH locus or Ioci in the I region. However, no I-region mutation has been detected yet, and so either the technique fails for some reason in this particular instance, or the Z-region loci mutate at rates far lower than those of the K - or D-region loci. Second, the K and D regions are treated as though each contained only a single locus analyzable by the skin-grafting technique. This assumption, again, is probably incorrect, at least so far as the D region is concerned. The latter is known to contain at least two loci, and the number may increase with further studies. Third, H-2 mutations that do not map in the K region’ are presumed to be D-region mutations. This assumption is made only for simplicity’s sake, At the time when the first H-2 mutations were discovered, they could not be assigned to the known H-2 loci. For this reason the mutated genes were given tentative designations H z l , Hz2, etc. (Bailey and Kohn, 1965). Since it is now clear that the H z l locus is identical to H - 2 K , the former designation should be dropped.

TABLE VA MUTATION RATES OF H - 2

LOCI

No. of mutations detected Strain of origin (C57BL/lO x BIO.DZ)F, (C57BLi6 X BALB/c)F, (C57BW6 X BALB/c)F, (A.C.4 x A)F, L4.C.4 X A)F,

Treatment of parent Diethyl sulfate None X-rays

No. of mice tested

Including clusters

Without clusters

1

1

154

2

2

H-2 mutations recovered

154'

d

H-2D"

da

2,572

2,572*

b

H-2Kb

ba, bb

-

0 1

3,156 204 514

0

0

519

519*

-

3

4,003

8,m*

b

H-2Kb

b

H-2Kb

b

non-H-ZK

None

5

(C57BL/6 x BALB/c)F,

None

8 17,454

f

H-2K'

-

17,454+

1

411

0 2

0 2

4,889 10,010

9,778* 10,010*

0 1 4

0

2,335 11,279

2,335* 11,279

i]

-

-

-

d

64.9

-

-

0

C57BL/6

a

H-2 gene affected

0 0 1

(C57BW10 x BlO.DZ)F,

None None None X-rays

H-2 haplotype affected

3,156; 204* 514*

None Diethyl sulfate None

BALB/c (C57BU6 X BALB/c)F, (C57BU6 x BALB/c)F, (C57BW6 x BALB/c)F,

No. of H-2 haplotypes testedu

Mutation rate per locus per gamete ( x 10')

fu

7'8

19'4

-

b g l, bg2, hh hi, bg3, bj, bk bin

db, dc

-

Egorov, 1967

]

1

Ba;z5and Kohn.

Egorov and Blandovn, 1972

2

3.7 ' 2.3 0.57

Melvold and Kohn, 1975, 1976

-

2.0 ,

-

-

?

?

-

-

b

?

?

3.5

d

Reference

o,g]

K T g r d Melvold,

Entries used for calculations in Table VB are marked by asterisks. a3 CJI

86

JAN KLEIN

TABLE VB SUMMARY OF H-2 MUTATIONRATES (CALCULATED FROM *ENTRIESIN TABLE VA) Allele

Mutation type"

Mutation freqnency"

Mutation rate ( X lo-").

9140,896 0/3,310 1140,896 013,310 0142,668 013,310 2142,862 113,824 01204 11514 01204 01514 01204 01514

2.20

~

K b

Db

K 'I D" K'

D' K k

S I S I S I S I S I S I S I

0.24

0.47 2.61 19.45

S = spontaneous; I = induced. Number of mutations detectedlnumber of genes tested. ' Per gene per gamete.

but very few non-K mutation have actually been mapped into the D region. Fourth, in an F1hybrid the four alleles (KxKvDxDy)have an equal chance to mutate. Hence, when a mutation is discovered at, say, K", the positive as well as the negative results (i.e., no mutation at K' D" and D") are used for mutation rate calculations. Unmapped H - 2 mutations were excluded from the calculations. Finally, we assume that data from different laboratories are comparable and so can be pooled for mutation rate calculations. Once again, this assumption may not be correct, but one justification for making it is the fact that the pooled data do not differ significantly from the most extensive data obtained in a single laboratory (that of Kohn and Melvold). Because of the problems inherent in H - 2 mutation rate calculations, any conclusions based on such calculations must be considered tentative. Keeping this qualification in mind, one can say this about the data in Tables VA and VB. The spontaneous mutation of an average H - 2 gene is 4 x per locus per gamete, or lower. One striking exception is the H - 2 K b gene, which mutates at a rate of about 2.2 x per locus per gamete, and thus, as pointed out by Melvold and

H-2

MUTATIONS

87

Kohn (1975), is probably the most mutable mammalian gene known. Whether the mutation rates can significantly be increased b y mutagen treatment is not clear at this time. Although several H - 2 mutations have been obtained from experiments wherein mutagens were used, only one mutation ( H - 2 9 could be traced back to a mutagenized parent. X-rays seem to have little effect on N - 2 mutation induction, and this finding is in keeping with the ineffectiveness of this agent in H mutation induction in general. Chemical mutagens, alkylating agents in particular, may be more promising, but more data are needed to evaluate their effectiveness.

D. INTERPRETATIONOF H-2 MUTATION RATES

The overall data on mutation rates as discussed in the preceding sections may be summarized as in the following tabulation:

Type of mutation

Rate or frequency

Recessive visible Dominant visible Dominant lethal Recessive lethal Histocompatibility

3.8 x per locus per gamete 4.4 x lo-’ per locus per gamete 6 to 16%per group of loci per gamete 2.9 x per locus per gamete 0.2%per group of loci per gamete 4 x lo+ per locus per gamete

H-2

The mutation rate of the H - 2 loci appears to be two or three levels of magnitude higher than the rate of recessive or dominant mutations (the only two types of mutation for which specific locus information is available). The “appears to be” in the previous sentence is not just a verbal padding, but is, rather, an expression of genuine doubt. At least three considerations justify caution in drawing this important conclusion about the high mutation rate of H - 2 loci. First is the fact that mutation rates for any type of mutation vary from locus to locus and among alleles at the same locus. The high mutation rate of the W locus (a rate approaching that of some H-2 loci) was already mentionedand undoubtedly there will be other loci with similarly high rates. Thus, the high mutation rate is probably not an exclusive characteristic of the H-2 loci. The second consideration concerns the magnitude of the screening procedures, Although screening of some 50,000 mice

88

JAN KLEIN

is an impressive achievement, the number is not near the sample size that would be needed to draw firm conclusions about the rates. To obtain reliable data, one would have to test several million mice-a task that goes beyond the capabilities of all H-2 laboratories combined. The present rate estimates must therefore be viewed with reservation, because they may be altered significantly by the next set of data. The third consideration concerns the possibility that the difference in the mutation rates of H - 2 versus non-H genes might be only at the phenotype level, while at the gene level both types of mutation are equally frequent. Consider, for example, a mutation at a coat-color locus. The pigment responsible for the coat color is produced by complex enzymic reactions, different enzymes being controlled by different genes. The change in the coat color is the result of an inactivation of one of these enzymes, which, in turn, is the result of a mutation in the portion of the gene coding for the enzyme’s active site. Conceivably, many mutations might occur that do not change the active site and so escape detection. In the H-2 molecule, on the other hand, there may be no active site and mutations at any portion of the H - 2 gene may conceivably be detected. According to this interpretation, the coatcolor gene mutation rates would be grossly underestimated, whereas the H - 2 gene mutation rates would reflect more closely the actual number of changes occurring in genes. Two other factors might also be contributing to the seemingly lower mutation rates of the coat-color loci. One such factor could be the observation that many of the coatcolor mutations are homozygous lethals and so many escape detection. In the H - 2 complex, on the other hand, no lethal mutation has been detected so far. The second factor is the finding that in most coatcolor genes studied, the rate of forward mutation (i.e., -+m ) is significantly higher than the reverse rate (i.e., m + +). In fact, among the genes causing visible mutations, true reverse mutations have been detected only at the d locus, where the reverse rate is 3.8 x in comparison to the forward rate of 1.2 x (Schlager and Dickie, 1971).The preponderance of forward mutations can be interpreted as suggesting that the forward change is of a complex nature, requiring several mutational steps before it becomes phenotypically detectable. The probability that the same steps would occur in the same order in the reverse direction is therefore extremely low, and the chance of reconstituting the original phenotype is negligible. In the H - 2 genes, accumulation of so many mutations might not be necessary for the detection of the mutation’s phenotypic effect. In fact, the available data are compatible with the notion that a single point mutation of an H - 2 locus

+

H-2

MUTATIONS

89

can be detected by the presently used screening methods (cf. Section IX,A). This last consideration leads one to conclude that the mutation rates of nonhistocompatibility genes might in fact be higher than the available estimates would seem to indicate. Arguing uguinst this view is the observation that more sensitive methods of mutation detection at enzyme-controlling loci have not resulted in any dramatic increase in mutation rate estimates. For example, electrophoretic methods of protein analysis are generally believed to detect two-thirds of all molecules altered by mutations (the remaining one-third escape detection because the particular amino acid substitutions do not alter the molecule’s net charge), yet the mutation rates estimated b y these methods are still at the level of at least in DrosophiZu (Mukai and Cockerham, 1977). However, let us disregard for a moment the three considerations discussed above and assume that the mutation rate of H - 2 loci i s higher than the rate of most other loci. Perhaps the difference is not so large as two or three orders of magnitude, but nevertheless it does exist. What could be the reason for this difference? One obvious answer to this question is that H-2 mutations may not be true mutations, but some other type of genetic change that normally occurs with much higher frequency than genuine mutations. I shall discuss the nature of H-2 mutations in Section VIII; here it suffices to state that the conclusion of this discussion is that H-2 mutations are probabIy point mutations of H-2 structural genes. Therefore, the interpretation of the mutation rate caused by some unorthodox genetic mechanism does not apply. What, then, could be the explanation of the high H-2 mutation rate? The most attractive is the possibility that the high rate is somehow connected with the extensive polymorphism of the H - 2 loci. Mutation rate, like any other trait, is subject to natural selection, which can favor either a high or a low rate, depending on the need for genetic variants in a particular environment. Since the demand for H-2 variants in natural populations seems to be extraordinary (see Section IX), the H-2 genes have evolved in such a way as to permit high numbers of genetic variants to be constantly generated. T h e high mutation rate could be accomplished in two principal ways. One would b e that the H-2 genes themselves would become hypermutable; the second way would be that a mutator gene outside the H-2 complex would increase the mutation rates of the H - 2 structural genes. Examples of both these situations are known in other organisms (see, for instance, Green, 1976). How a gene becomes hyper-

90

JAN KLEIN

mutable is not known. However, Benzer’s classical studies on bacteriophage T4 established that, within a single gene (cistron), certain sites mutate with much higher frequency than others (Benzer, 1961), with some of these “hot spots” being 500 times more mutable than most other sites. One could imagine, therefore, that a gene consisting of a series of hot spots would itself become a “hot gene,” particularly if the “hotness” were favored by selection. H-2Kb might be an example of such a hot gene. What makes a particular site hot has not been determined, but, very likely, the high mutability is somehow determined by the structure of the site. The nature of the mutator-gene effect is somewhat better understood. Studies in prokaryotes have established that the effect can be generated by a variety of mechanisms (for discussion and references, see Auerbach, 1976). For instance, mutation can alter replicating enzymes so that they make more mistakes than is normal when copying certain stretches of DNA. In another instance, enzymes that normally correct replication errors become-because of a mutation-less efficient in their “editing” function and leave many mistakes in certain genes uncorrected. In both these instances, a mutation in one gene (presumably the gene coding for the replication-related enzyme) causes a change in the mutation rate of another gene. Another way of generating gene instability is by moving a foreign genetic element (episome, exosome, or controlling element) to this gene’s vicinity. For a long time, genes were believed to be more or less permanently fixed to their positions in the chromosome, and Barbara McClintock’s (1950) quiet insistence that the true situation might be more complex was looked upon with the type of curiosity with which one views a person claiming to have seen a pink elephant. Recently, however, more and more geneticists have begun to realize that genes in the genome move around much more than one would have been willing even to consider a few years ago. Could one of these exotic mechanisms play a role in the generation of H-2 mutations? The answer to this question must be an emphatic no, since the orthodox mechanisms explain the mutations quite adequately (see Section VIII). In summary, the high mutation rate of H - 2 genes can best be explained by combining two assumptions: first, that the techniques for the selection of H - 2 mutations detect more mutations than techniques available for the detection of other mutants; and second, that the high mutation rate of H - 2 is favored by natural selection, and that the primary structure of the H-2 DNA molecule has evolved in such a way as to make the high rate possible.

H-2

MUTATIONS

91

VI. Genetics of Available H-2 Mutations

A. DISTRIBUTION OF H-2 MUTATIONS AMONG REGIONS Of the 22 H - 2 mutations described in the literature (Table VI), 12 occurred in the K region and 3 in the D (or L ) region; the mapping of the remaining 7 mutations is still incomplete. No mutation affecting the Z region has been detected so far. Of the K-region mutations, all but two occurred in the H - 2 K b gene. The reasons for this disproportional distribution of H - 2 mutations among regions and alleles are not clear. It could be that the disproportionality, particularly the apparent lack of Z-region mutants, is the result of chance variations in the small sample scored. If, on the other hand, the disproportionality were to hold even on a larger sample, one could argue that it reflects some bias in the selection procedure. At the present time, however, it is difficult to see where such bias may lie, since all three regions ( K , D , and Z ) are equally strong in terms of graft rejection, the method by which H - 2 mutations are detected. ,4nother possibility is that for some loci, particularly the H - 2 A locus in the ZA subregion, a more complex genetic change is required to generate antigens that can be detected by skin grafting. However, the most likely explanation of the disproportionality is the assumption that there are differences in the mutation rates not only among different H - 2 genes, but also among different alleles at the same locus, and that the different mutability is the consequence of differences in the primary structure of the corresponding cistrons. Alleles at H - 2 loci differ considerably more than alleles at any other known locus, in some instances as much as genes at two different loci (for a review, see Hood and Silver, 1977).It should not be surprising, therefore, to find some degree of not only intergenic, but also interallelic, variability in mutation rates. According to the hypothesis put forward in the preceding section, one could simply assume that certain alleles at certain H - 2 loci accumulated more hot spots than other alleles at the same or at different loci. The reason that no Z-region mutants have been found thus far could therefore be that, in the haplotypes scored, the Z alleles happened to have a relatively low mutation rate. Screening of other haplotypes with different Z-region alleles might produce some surprising results.

B. GENETICLOCALIZATION OF H - 2 MUTATIONS Genetic mapping based on the use of H - 2 recombinants can place mutations into regions, but not into loci. Fortunately, there are several other ways of placing a mutation into a gene. The conclusion from the

ED

t G

TABLE V1 LISTOF KNOWNH-2 MUTANTS Origin Strain designation Full

Synonym

Abbreviation

H-2 haplotype

H-2 region affected

Spontaneous induced (I)

Gain (G) or loss (L)

6)or

Reference

Derived from H - P

B6.C-H-2'"

B6.C (Hzl)

Hzl

bu

K

S

G+L

B6.C-H-2bb

B6.C (Hz49)

Hz49

bb

K?

S

G+L

B6-H-2'"

B6.M505

M505

bd

K

S

G+L

S

G+L

B6-H-2bg'

S

G+L

B6-H-2bu2

S

G+L

B6-H-2bg3

S

G+L

B6-H-2bh

S

G+L

B6-H-2"'

S

G+L

B6-H-2*'

S

G+L

B6.C-H-2"

B6.C(Hz170)

Bailey and Kohn, 1965 Bailey and Kohn, 1965 Egorov and Blandova, 1968 Bailey and Cherry, 1975 Kohn and Melvold, 1974 Kohn and Melvold, 1974 Melvold and Kohn, 1976 Kohn and Melvold, 1974 Melvold and Kohn, 1976 Melvold and Kohn, 1976

9

1:

r

E z

BlO.M513

M513

hli

K

S

G+L

b 171

Not K

S

G+L

19 11

Not K

S

G+L

b0

?

S

G

I> ?

?

S

L

Melvold and Kohn, 1976 Melvold and Kohn, 1976 Melvold and Kohn, 1976 D. W. Bailey, personal communication Egorov and Blandova, 1972

Derived from H-Zd BlO.D2-H-2"' BALBlc-H-2"'

B lO.D2(504)

M504

rln db

D D(L)

I S

G+L L

BALBIc-H-2 "

dC

Not-D

S

L

BA LBIC-H-2 ' ' I

dd

NotD

S

G+L

Egorov, 1967 Melvold and Kohn, 1976 Melvold and Kohn, 1976 Melvold and Kohn, 1976

Derived from H-2' A.CA-H-2 "'

A.CA (M506)

M506

B 10.M-H-2fi

.f

K

S

G+L

fb

D

S

G+L

S

G+L

Egorov and Blandova, 1972 Mobraaten and Bailey, 1973

Derived from H - P CBA-H-2"'

CBA (M523)

M523

ku

K

Blandova et a l . , 1973, 1975

94

JAN KLEIN

variety of these tests is that all known K-region mutations have occurred in the H - 2 K locus, as defined by serological methods and by tissue grafting. Of the three known D-region mutations, one (H-2da) definitely affected the H - 2 D locus, one (H-2db)probably occurred outside of the H - 2 D locus (probably in the H - 2 L locus), and the position of the third mutation (H-2'9 remains to be determined. The evidence for these assignments is summarized briefly below; for a detailed discussion of each individual point, see the corresponding section in this review. First, molecules isolated from mutant strains with anti-H-2K or antiH-2D sera differ from the corresponding standard molecules in their peptide compositions. Biochemical data of this sort are available for H - 2 mutations ba, bd, bg, bh, da, fa, and ka (Brown and Nathenson, 1977a,b; S. Nathenson, personal communication, S. Ewald, L. Hood, and J. Klein, unpublished data). Since bb, bf, bi, bj, and bk of the remaining H - 2 b mutations are known ( from genetic complementation tests) to have occurred at the same locus as H-2ba or H-2bd, one can conclude by extrapolation that these five mutations also affected the H - 2 K locus. Second, in mutant strains in which the effect of the mutation can be detected serologically, antigens controlled by the H - 2 K or H - 2 D loci are altered (Dishkant et al., 1973; Klein et al., 1974a, 1975, 1976~). Whenever new antigens appear in a mutant (or a standard) strain, they cocap with the H-2K or H-2D antigens (Klein et al., 1976~). Third, T cells stimulated to viral or minor H antigens in the context of standard-type H-2K or H-2D molecules fail to react with some of the corresponding mutant strains, even though the strains carry the stimulating antigens (Blanden et al., 1976; Zinkernagel and Klein, 1977; J. Klein, unpublished data). Thus, since the stimulating antigens are the same in the standard and mutant strains, these strains must differ in their H-2K or H-2D molecules. Fourth, in the cell-mediated lymphocytotoxicity (CML) assay, the reactivity between the standard and mutant strains can be blocked by the addition of anti-H-2K sera to the reaction mixture (Nabholz et al., 1975). Fifth, in the mixed lymphocyte culture, the reaction between the mutant and the standard strain is accompanied by the adsorption of H-2K-stimulating antigens onto the responding blast cells (Geib and Klein, 1978). In the face of these data, a great deal of intellectual juggling would be required to deny that the mutations occurred in the H - 2 K and H - 2 D loci. Furthermore, since the placement into the H-2K or H-2D loci has

H - 2 MUTATIONS

95

been accomplished by three different immunological methods (serological, CML, MLR), the logical conclusion is that these loci are responsible for the effects measured by the three methods. There is, therefore, no need to postulate additional mutations at loci adjacent to H - 2 K or H - 2 D . In summary, the well-characterized H - 2 mutations occurred, with one exception, in either the H - 2 K locus in the K region or the H - 2 D locus in the D region. There is no evidence that any other loci were affected by these mutations. C. DESCRIPTION OF AVAILABLEH - 2 MUTATIONS

A detailed description of individual H - 2 mutant strains can be found in Klein (1975), or in McKenzie et al. (1977b).The available mutants are listed in Table VI. The mutant strains were derived in one of two ways. One consisted of screening an inbred strain (e.g., C57BL/6, or B6) for histoincompatibilities by skin grafting and then establishing homozygotes for a mutation by interbreeding the progeny of the deviant individual. Coisogenic strains carrying H - 2 mutations bd, bgl, bg2, bg3, bh, bi, bj, bk, bm, bn, bo, da, db, dc, dd,f a , f i , and ka were derived in this way. The second method of mutant strain derivation was the production of F, hybrids between two inbred strains, the selection of mutants in these hybrids, and the production of congenic lines by backcrossing the H - 2 mutation onto the background of one or the other parental strain. These congenic lines, therefore, probably differ by more than one locus. This second procedure was used to derive congenic lines carrying H - 2 mutations ba, bb, and bf. In addition to H - 2 variants recovered by mutation, “natural” H - 2 variants have also been described (Table VII). The latter were detected by serological or histogenetic typing of standard inbred strains. The H-2 haplotypes of the natural variants resemble, but are not identical to, haplotypes carried by other strains. For instance, the C57BL/LiA and the C57BL/10 strains reject each other’s skin grafts rapidly, and the rejection can be shown to be H-2-controlled (Dux et al., 1971). Yet, no serological difference has been found between the two strains (Dux et al., 1971), and the strains do not stimulate each other in the mixed lymphocyte culture (Melief et al., 1976).The haplotype of the C57BL/LiA strain, therefore, appears to be a minor variant of H-2*, and as such it is designated H-2be. In general, natural H - 2 variants are poorly characterized with respect to their genetic, immunological, and biochemical properties. The little that is known about them suggests that they might be tran-

96

JAN KLEIN TABLE VII LISTOF SPONTANEOUSH-2 VAFUANTS H-2 haplotype

Strain

B 10.129(6M) C57BLLiA BN

bc be bP

WBiRe

Variant H-2 allele ? ? Kb?

Db

?

B1O.Y SMIJ B 10.SNA57

Pa w21

DO D?

BlO.KPA42

wl

D''?

0

Reference Snell et al., 1971a Dux et al., 1971 Czarnomska and Demant, 1975 Snell et al., 1971b; Demant et al., 1971a,b Snell et al., 1971b Snell et al., 1971b Klein and Zaleska-Rutczynska, 1977 Klein and Zaleska-Rutczynska, 1977

sient forms between genuine H - 2 mutations and unrelated H - 2 haplotypes, and as such, they could become a valuable material for the study of H - 2 gene evolution. Preliminary data suggest that a particularly rich source of natural H - 2 variants may be found in wild mouse populations (Klein and Zaleska-Rutczynska, 1977, and unpublished data). VII. Biochemistry of H-2 Mutations

Peptide maps are available for four of the 22 known H - 2 mutations (Table VIII): H-2ba and H-2bd (Brown and Nathenson, 1977), H-2da Brown et al., 1978), and H-2ka (S. Ewalds, L. Hood, and J. Klein, unpublished data). Preliminary biochemical data are also available for H-2b0, H-2bh, and H-2fa (S. Nathenson, personal communication; S. Ewald, L. Hood, and J. Klein, unpublished results). An example of a comparison between peptide maps of mutant and standard strains is shown in Fig. 9. Several important conclusions can be drawn from the available biochemical data. First, the fact that mutant molecules can be precipitated with antiH-2K or anti-H-2D sera establishes unequivocally that the mutations occurred in the H - 2 K or H - 2 D loci. Second, since in all four mutations the altered products have approximately the same molecular weight (M,) as the products of the corresponding standard strains (M,45,000), it is unlikely that large deletions or duplications occurred in the mutant genes. And third, since less than 10%of the peptides have been

H-2

97

MUTATIONS

TABLE VIII NUMBEROF PEPTIDE DIFFERENCES BETWEEN MUTANT AND STANDARDH-2 MOLECULES No. of peptide differences

H-2 mutation

"

Lysine peptide

Arginine peptide

Reference

1

2 -3

hd

1

1-2

da ka

6

4 NT"

Brown and Nathenson, 1977 Brown and Nathenson, 1977 Brown et al., 1978 S. Ewalds, L. Hood, and J. Klein, unpublished data

1

Not tested.

altered in three of the four thoroughly studied mutations, and since any two products of unrelated H-2K or H - 2 D alleles differ in some 40% of their peptides (Brown et al., 1974), the peptide mapping establishes that the H-2 mutations are far more closely related to the standard allele than two alleles of independent origin are related to each other. This conclusion is important for considerations of the origin of the H-2 mutations, since it virtually rules out the possibility that the mutations are derepression of previously silent H-2 genes (see Section VII1,B). What kind of alteration in the primary structure of the H-2 molecule

I

I ----

ARCININE PA505 " - 2 3 3

I

FRACTION NUMBER

FIG.9. Tryptic peptide map of the H-2K glycoprotein of 3H-arginine (Hzl) and I4Carginine (C57BU6). (From Brown and Nathenson, 1977; reproduced with permission.)

98

JAN KLEIN

is needed to produce the 10% difference in the peptide map is not clear. However, considering that the procedures of peptide mapping tend to exaggerate biochemical differences by uncontrolled cleavages, and, furthermore, that it is possible to obtain two-peptide map differences even with a single amino acid substitution, the data are compatible with the possibility that H-2ba, H-2", and H-2ka are each the result of a single point mutation. Whether they are point mutations will be known only after the amino acid sequence of the mutant molecules has been established. If more than one amino acid substitution were to be found, such data could then be explained by assuming the occurrence of two or more point mutations in sequence. Where in the H - 2 molecule the amino acid substitutions took place is not known. However, S. Nathenson (personal communication) favors the possibility that they occurred in the N-terminal half of the polypeptide chain. The peptide mapping data also tentatively suggest that the substitutions in different mutations occurred at different positions along the polypeptide chain. In the H-2da mutant, three types of biochemical alteration are found (Brown et al., 1978): reactivity of anti-H-2Dd serum with H-2Dda molecules is only 25% of that with H-2Dd molecules; 6 0 4 0 % of the H-2Dda molecules either lack carbohydrate chains or have chains of drastically different composition; and 30-40% of the peptides have been altered by the mutation. The first finding is consistent with the fact that H-2da is serologically quite different from H - 2 d . Therefore, the most likely explanation of the reduced reactivity is that the serological site in the H-2Ddamolecule has been so drastically altered that the anti-Dd antibodies react with it only with low affinity. However, an alternative explanationthat the reduced reactivity reflects reduced content of H-2Ddamolecules-cannot be ruled out. But it should be possible to distinguish between the two alternatives by comparing the reactivity of anti-Dd and anti-Ddaantisera with Ddaantigens. The second observation (the lack of carbohydrate chains) can be explained by postulating that the H-2da mutation altered the amino acid to which these chains normally attach. To explain the third observation (30-40% difference in the peptide map) one has to postulate at least five to six amino acid changes in the primary structure of the H-2Ddamolecule. Accumulation of so many point mutations within a short time-period is unlikely, so one must conclude that the mutational event was, perhaps, of a more complex nature-for example, a frame-shift mutation. The latter explanation is supported by the fact that the H-2da mutation arose in a mouse whose parent was treated with diethyl sulfate-an alkylating

H-2

MUTATIONS

99

agent that increases the frequency of frame-shift mutations (Egorov, 1967). The fact that the molecular weights of the H-2Dda and H-2Dd products do not differ drastically again excludes the possibility of a large deletion, or a deletion accompanied by a fusion with the adjacent (H-2L?)gene. VIII. Origin of H-2 Mutations

How do H-2 mutations arise? Are they true mutations, or the result of some other event mimicking mutations? T o answer these questions, let us consider various mechanisms that might explain the origin of H - 2 mutations and weigh them against the available data. A. INTRAGENIC CROSSING-OVER (GENECONVERSION) When two homologous chromosomes (for example, the two chromosomes 17), each consisting of two chromatids, pair at meiosis and the two DNA duplexes (one per each chromatid) align in a parallel fashion, an exchange of genetic material (crossing-over) may occur between them. The precise mechanism of the exchange is not known, but one possibility is that the two adjacent, identically oriented DNA strands in the aligned duplexes break at exactly opposite points and then rejoin in a crisscross fashion so that the left-hand portion of one strand connects to the right-hand portion of the other strand, and vice versa (Fig. 10). A cross-bridge (cytologically visible as a chiasma) is thus formed (Fig. 10c). The bridge can move along the chromosome b y a process in which hydrogen bonds between nucleotides in the original duplexes break and new bonds are established between nucleotides of nonsister chromatids (Fig. 10d). Since, for the most part, the two DNA duplexes have identical nucleotide sequence, the movement of the exchange point does not alter the coding ability of the genetic material; it merely results in equivalent nucleotides exchanging places on the adjacent strands. However, there are sites in the DNA molecule where the base-pair sequence is different in the two homologous chromosomes-sites altered by mutations in genes for which the animal is heterozygous. When the exchange point reaches such a site, no bond can be formed between the juxtaposed nucleotides because of lack of complementarity between them. Leaving nucleotides at the mutant site unpaired, the exchange point then moves on, pairing additional parts of the molecule, until its action is stopped by another break and rejoining event. The unpaired portion of the DNA duplex, known as the heteroduplex, can b e retained until the next round of DNA replication when comple-

100

JAN KLEIN

Chromatid ! >romosoma Chromatid 2

(a 1 Chromatid 3\hromosorna

\ /~Homo'oq'

Chromatid 4'

FIG.10. Principle of gene conversion (intragenic crossing-over). The figure depicts a short segment of DNA in paired homologous chromosomes, each chromosome consisting of two chromatids joined at the centromere (extreme left). Letters stand for the four nucleotides of the DNA chain. For explantation see the text.

mentary nucleotides are inserted at all positions of the DNA duplex. Or the small unpaired loop can be recognized by repair enzymes, one of the two unpaired nucleotides excised, and a nucleotide complementary to the second nucleotide inserted (Fig. IOe).The repair event thus converts the site of genetic difference on the DNA molecule to either the wild or the mutant type, depending on which of the two mispaired nucleotides has been excised from the duplex.

H-2

MUTATIONS

101

This gene conversion may, therefore, alter the chromatid segregation so that instead of the two wild and two mutant chromatids expected to result from a reciprocal recombination event, three wild and one mutant, or three mutant and one wild chromatids (and later chromosomes and gametes) may appear. With most genes, gene conversion is a rare event, since either the animal is homozygous for them, or-if it is heterozygous-the site of genetic heterogeneity is usually restricted to a single nucleotide pair. In the H - 2 complex, however, the high polymorphism increases the probability of heterozygosity, and the muItiple nucleotide differences within a single gene create a situation in which gene conversion could be relatively frequent. Could the postulated high frequency of H - 2 gene conversion be the explanation for the high H - 2 mutation rate? If H - 2 K and H - 2 D were single genes (rather than gene families, see below), then the question would have to be answered in the negative. The occurrence of gene conversion requires gene heterozygosity, and this condition was not fulfilled by the animals in which the known H - 2 mutations arose. Even though some H - 2 mutations were detected in F, hybrids, the mutational event must have occurred in the homozygous parents. If, on the other hand, the H - 2 K (H-2D) locus were a cluster of closely related but nonidentical genes, pairing between nonhomologous genes might occur and lead to gene conversion even in an H - 2 homozygote. However, as discussed below, the evidence for H - 2 K ( H - 2 D ) gene cluster is virtually nonexistent, and hence the explanation of H - 2 mutations by gene conversion is not a very likely one.

B . REGULATOR-GENE MUTATION The regulator-gene hypothesis postulates that H - 2 mutations are the result of changes in regulatory, rather than structural genes. One specific form of the hypothesis is this: H - 2 K and H - 2 D are assumed to be clusters of closely related genes that arose from a common ancestor by duplication. Subsequent mutational diversification resulted in gene families in which individual members were like alleles except that they occupied different positions (loci) on the chromosome. T h e hypothesis further states that always in each family only one gene is expressed, and that the expression is controlled by a common regulator gene. The multi-gene-family hypothesis of H - 2 was first explicitly formulated b y Bodmer (1972, 1973) and later popularized primarily by Hood (Hood and Silver, 1977). H - 2 mutations are explained by the hypothesis as the result of a genetic change in the regulator gene, a change leading to an expression of a structural gene different from that expressed in the standard-type animal. The arguments in favor of the regulator gene hypothesis are these.

102

JAN KLEIN

First, sequence data indicate that products of two H-2K ( H - 2 D ) alleles differ in as many as 30-40% of their amino acids, while two allelic products of most other gene systems differ in only one amino acid (for review and references, see Hood and Silver, 1977). Two H - 2 K ( H - 2 D ) alleles thus behave like genes at two different loci. Second, some tumors have been reported to express H-2 antigens absent in the strain of origin but present in other, unrelated strains (Parmiani and Invernizzi, 1975).The expression of the new antigens is thought to be caused by the activation in tumor cells of previously silent H - 2 K or H - 2 D genes. Third, the high mutation rate of H-2K and H - 2 D genes could be explained by a mutation at the regulatory locus and activation of new genes in the H-2K and H - 2 D families. Fourth, gene families may exist in other systems, for example, in genes coding for the constant region of immunoglobulin molecules. In the rabbit, for instance, allotypes have been found that behave as though controlled by a single Mendelian gene but show certain irregularities in their inheritance (Strosberg, 1977). The irregularities can be interpreted as evidence for a family of genes of which only one is expressed. The products of these complex allotypes differ-as do allelic H-2 molecules-by more than one amino acid. Let us now have a closer look at these arguments to evaluate their worth. The first argument is not at all compelling. An equally plausible explanation of the multiple amino acid differences found between allelic H-2 products is that the rapid diversification of alleles is needed to generate high polymorphism, and high polymorphism is needed for the filnction of the H - 2 K and H - 2 D loci. The multiple amino acid substitutions could thus accumulate because they are favored by selection. The polymorphism of H - 2 loci is unique, and this uniqueness is reflected in the primary structure of the H-2K and H-2D molecules. One can, of course, turn this argument around and contend that because of the need for high polymorphism, a special systemthe gene family system-evolved for generating the polymorphism. But the important fact is that to explain the multiple amino acid differences between H - 2 K or H - 2 D alleles, one does not need to evoke gene families. The second argument would be compelling, if it were based on convincing experimental data. Evidence for the expression of foreign H-2 antigens in tumor cells is based on two sets of observations: histogenetic and serologicaI (Parmiani and Invernizzi, 1975; Invernizzi and Parmiani, 1975; Garrido et al., 1976). In the histogenetic studies, mice of strain A were presensitized with various tissues of strain B and then challenged with strain-A tumors. The observed slowdown of tumor

H-2 MUTATIONS

103

growth in so-challenged recipients was attributed to immunity against histocompatibility antigens shared by strain-B normal cells and strainA tumors. Although the H antigens could not be identified genetically, they were presumed to be controlled by the H-2 complex. The latter presumption was, however, contradicted by at least some studies (see, for example, Invernizzi et al., 1977) in which the strain combinations were such as to exclude H-2 participation. In the serological studies, H-2 antisera made against strain-B H-2 antigens were found to react with strain-A tumor cells, but not with strain-A normal cells. The interpretation offered by the authors of these studies is that the anomalous histogenetic and serological reactions are directed against tumorspecific transplantation antigens (TSTA) and that TSTA are H-2 antigens, expressed because of derepression of normally silent H - 2 genes. A major criticism of these studies is that, in both the histogenetic and serological experiments, evidence is lacking for the observed reactions being directed against products ofH-2 loci. In the histogenetic studies one would wish to see the experiments repeated in congenic strain combinations, in which contributions by the non-H-2 loci have been excluded. As the data stand now, they can, in fact, be interpreted as indicating that the reaction is not against H-2 antigens. This conclusion is supported by the notorious weakness of the TSTA with respect to their ability to induce either cellular or humoral immunity (for a review, see Klein, 1969). The published serological data purporting to show the expression of new H-2 antigens on tumor cells are even less convincing. Anyone who has worked with H-2 antisera for some time knows how easy it is to fall into the trap of overinterpretation. To do serological studies properly, one would want to use antisera that have been carefully analyzed by extensive panel tests and then rendered monospecific by carefully planned absorptions. Instead, the authors of these studies used sera that were produced in other laboratories, did not bother to panel-test them by the same technique used for tumor typing, made no attempt to restrict the specificity of the antisera, and relied solely on the H-2 chart in drawing conclusions about the antibodies present. No wonder they found all kinds of anomalous reactions. Until careful serological comparison of normal and neoplastic tissues is performed with defined antisera, the anomalous reactions must be considered as artifacts. However, let us assume for a while that a carefully controlled study would show anomalous serological reactions with tumor cells. Would that constitute a proof of foreign H-2 antigen expression of tumor cells? Certainly not, since alternative-and more probable-explana-

104

JAN KLEIN

tions would still be available. For example, the anomalous reactions could be caused b y previously undetected contaminating antibodies (evidence is accumulating that even the most monospecific antisera can contain such unexpected antibodies), or they could represent cross-reactions with various viral antigens, TSTA, or the complex of these antigens and H-2. We have proposed that the complex of H-2 and TSTA may, in certain circumstances (e.g., during associative recognition by T lymphocytes), resemble an H-2 alloantigen (Zaleski and Klein, 1977a). If so, the observed cross-reactivity could be caused by this resemblance rather than derepression of genes in the purported H - 2 cluster. And even if evidence for genetic association between TSTA and H - 2 could be obtained, it still would be more likely that the TSTA would be H-2 variants derived by point mutations, rather than derepression of genes in a cluster. The notion of TSTAs being controlled by the H - 2 complex is further contradicted by the finding that TSTA genes are not linked to H - 2 (Klein and Klein, 1975),and that TSTA and H-2 are physically separable molecules (Yefenof and Klein, 1974; Davies et al., 1974), although claims to the contrary have also been published (0th et al., 1975). Considering the difficulties with histogenetic, serological, and genetic analysis of tumor cells, the only convincing proof of H-2-TSTA identity would be a biochemical one. If someone were to isolate TSTA from uncontaminated strain-A tumor and show that it has the composition of strain-B H-2 antigens, the derepression hypothesis would have to be seriously entertained. Until then, however, such a hypothesis must be considered as one of those fashionable ideas that catch on rapidly and a few years later are as rapidly forgotten. The third argument in favor of the regulator-gene hypothesis is the high mutation rate of H - 2 genes. Hood and Silver (1977) argue that the mutation rate can be explained by postulating each new mutation to be an expression of a different gene in the presumed H - 2 K or H - 2 D gene cluster. If their argument were correct, one would expect each new mutation to differ from the standard type as much as two alleles of independent origin differ from each other. (The various genes in the H - 2 K or H - 2 D clusters-although not expressed-would have the same opportunity to mutate and to differentiate from one another as have genes carried by different chromosomes.) This expectation, however, is not borne out b y the facts. The preliminary biochemical analysis of the H - 2 mutants (see Section VII) shows that, with one exception, mutant H-2 molecules differ from standard-type molecules in one or two peptides, whereas H-2 molecules controlled by two independent alleles differ in some 30-40% of their peptides. The one ex-

H-2

MUTATIONS

105

ception is the M504 mutant which, as discussed in Section VII, might have a different origin from the other known mutations. The H-2 mutation data, therefore, do not support the gene cluster hypothesis. The fourth argument in favor of the regulatory-gene hypothesis is the existence of complex allotypes in the rabbit immunoglobulin system. If multigene families exist in the Ig systems-so the argument goes-they may also exist in the MHC system, since the two systems might be evolutionarily related (Klein, 1977). The fact is, however, that there is no compelling evidence that the complex allotypes are the result of gene derepression in a multigene family. Here, too, alternative explanations exist. Thus, the argument is circular: complex allotypes are used as per analogiam evidence for the existence of multigene families in MHC, and MHC is used as per analogiam evidence for the existence of multigene families controlling complex Ig allotypes. The value of such an argument is self-evident. The existence of multiple amino acid differences between products of two allelic genes is not without a precedent, although, admittedly, one must climb deep down the evolutionary scale to find the precedent. I n the tobacco mosaic virus, two strains differ in about 30 positions of their coat protein, which constitutes about a 20% difference between two alleles (Anderer et al., 1965). If allelic genes of a virus could differ in so many nucleotide pairs, why could a similar situation not exist in the mouse? In summary, none of the four arguments cited in favor of regulatory H - 2 gene mutations is in any way compelling. At present, there is simply no evidence that H-2K and H - 2 D are large clusters of genes of which always only one is expressed. C. POINTMUTATIONS By far the simplest explanation of the H - 2 mutations’ origin is by substitution of one nucleotide pair with another in the DNA sequence of the H - 2 cistron. This point mutation then leads to a substitution of one amino acid with another in the polypeptide chain of the H-2 molecule. The point-mutation hypothesis is supported by preliminary biochemical data (cf. Section VII) and is compatible with all that is known about H - 2 mutations. However, considering the high H - 2 mutation rate, it is likely that point mutations accumulate rapidly in H-2 genes, so that at the time when the mutants are analyzed, more than one mutation might already be present. An alternative explanation, that the H-2 mutations are the result of more complex genetic changes, such as deletions or duplications of nucleotides within H - 2 , is unlikely. These changes would result in alteration of the reading frame

106

JAN KLEIN

and thus in considerable changes of the H-2 molecule’s primary structure. So far there is no evidence for such changes. In summary, among the various explanations of the H - 2 mutations’ origin, point mutation is the most likely one. At any rate, the answer to the question of how H - 2 mutations arise should be known before long -as soon as the amino acid sequence of the altered peptides in the mutant molecule is determined. IX. Mutations and Polymorphism

Polymorphism is defined as the presence in a population of two or more alleles at a single locus, with frequencies exceeding those that can b e explained by concurrent mutations. One point needs to be emphasized in this definition, namely that it is not the number of alleles but the frequency of a given allele that makes a locus polymorphic. Thus, the human hemoglobin locus has some 300 alleles (Vogel, 1969), yet the locus is not considered highly polymorphic because most of the alleles are extremely rare and can be, in each case, explained as being the result of a recent mutation. The human ABO system, on the other hand, is polymorphic because the three alleles coding for the A, B, and 0 traits occur (in northern European population) with frequencies 0.3, 0.1, and 0.6, respectively (Race and Sanger, 1968). Hence the fact that, among inbred mouse strains, there are some 10 alleles at the H-2K locus and 10 at the H - 2 D locus does not constitute a proof that H - 2 loci are polymorphic. The alleles might be the result of mutations preserved by the special mating system used for the estabhshment of inbred strains. To determine the polymorphism of these two loci, one must determine the frequencies of the individual alleles in wild mouse populations. We are in the midst of such a determination. The preliminary results obtained so far (W. Duncan, E. Wakeland, Z. Zaleska-Rutczynska, and J. Klein, unpublished data) indicate that multiple alleles at H - 2 loci do indeed occur in wild populations. But the H - 2 polymorphism is different from polymorphisms at other loci. When one tests wild mice for an isozyme locus, for example Es-1, or esterase-1 (Selander and Yang, 1969), one finds, among mice from a given geographical area, one allele with very high frequency (say 0.8), and one or two additional alleles with much lower frequencies. Not so with the H - 2 loci. Here, in a local population (a single barn), oneH-2 allele predominates over a few others, but when one looks at a population in a larger geographical area, one does not find any high-frequency alleles. Instead, one finds a large number of alleles, each present with a rela-

H-2

MUTATIONS

107

tively low frequency. How low the frequencies are and how many alleles exist among wild mice remains to be determined. However, if the present trend continues, we may run into a situation in which the system might be so polymorphic that, by definition, it will not be polymorphic any more. This statement may sound absurd whereas, in fact, it is quite appropriate. The present definition of polymorphism requires an allele to b e present in a population with a frequency of 1% or more. Imagine, however, a situation in which there would be, say, 200 alleles present in a population, each with a frequency of 0.5%.By definition, such a system would not b e polymorphic despite the large number of alleles present in the population. In other words, at a certain point the mutation rate might be so high and the population so favorably structured that there would be a new kind of polymorphism in which new alleles would be constantly generated, spread through local populations, and relatively rapidly eliminated when these populations became extinct. The H - 2 mutation rates seem to be high enough to provide sufficient material for such a dynamic polymorphism. Further studies on H-2 population genetics should reveal how close the real situation among wild mice comes to the hypothetical one. It is interesting to note that recent studies on wild rats (Shonnard et al., 1976) do not show any extensive polymorphism at the major histocompatibility complex of this species. This finding is surprising, since the rat and the mouse are closely related species with similar geographical distributions. There are several possible explanations for the difference in MHC polymorphism between the rat and the mouse. The first possibility is that the finding of Shonnard and his co-workers is fortuitous, and that when other rat populations are tested extensive polymorphisms will be found. Isozyme studies suggest that there might be considerable genetic differences among individual local rat populations. Thus, Serov (1972) tested feral Siberian rats and found no evidence of any polymorphism at the approximately 20 investigated loci. On the other hand, Eriksson and his co-workers (1976) found rats in the Helsinki area to be polymorphic at several loci, with the degree of polymorphism approaching that of the mouse. The second possibility is that rats, because of periodic decimation b y poisoning, may b e undergoing population “bottlenecks” more frequently than mice. The founder effect would then be more pronounced in the former than in the latter species. The third possibility is that the difference in the degree of MHC polymorphism between rat and mouse is related to a difference in the

108

JAN KLEIN

social structures of the two species. Subdivision into demes seems to be much less pronounced in rats than in mice, and rats also seem to migrate more than mice (Calhoun, 1963). Minor differences in social behavior may have profound effect on genetic polymorphism of the MHC loci. In view of the differences in polymorphism, regardless of whether the differences are general or only local, it should be interesting to find out whether the mutation rate of the rat MHC differs appreciably from that of the mouse. X. Effect of H-2 Mutations on immune Functions

The genetic and biochemical studies discussed in the preceding sections suggest that most H-2 mutations are alterations in a single gene, either H-2K or H-2D, and that many of them are point mutations or possibly accumulations of point mutations. This finding provides a unique opportunity to determine which traits are controlled by the H - 2 K and H-2D loci. Realizing this opportunity, we initiated some time ago a systematic study of the effect that H-2 mutations have on immune functions. The mutations used were ba, bd, da, fa, and kaall but one (ba)produced and kindly supplied to us by Dr. Igor K. Egorov; the ba mutation was obtained through the courtesy of Dr. Donald W. Bailey. As comparison of our work with more limited work done by other investigators on other mutations shows, these mutations are fairly representative of the entire group. The immune functions we studied were: allograft reaction, cell-mediated lymphocytotoxicity (CML), mixed lymphocyte reaction (MLR), graft-versus-host reaction (GVHR), graft-versus-host disease (GVHD), generation of allogeneic supernatants, and immune response. In each study, we compared mutant combinations (presumably a single gene difference) with H-2 recombinant combinations differing in the entire K or D regions (potentially a multigene difference). These comparisons are summarized in Tables IX through XXIII, in which mutant data are always shown in the upper part and recombinant data in the lower part of each table. A summary of the results appears in Table XXIV.

A. ALLOGRAFT REACTION AND IMMUNOLOGICAL TOLERANCE Since skin grafting was used for the detection of the available H - 2 mutants, it is not surprising that in all mutant-standard strain combinations (with two exceptions), exchanged skin grafts were reciprocally rejected (Bailey and Kohn, 1965; Bailey et al., 1971; Apt et al., 1975; Melvold and Kohn, 1976; Egorov, 1967; Blandova et al., 1975; Egorov

H-2

MUTATIONS

109

and Blandova, 1972; Mobraaten and Bailey, 1973; Klein et ul., 1974a, 1975, 1976a; Forman and Klein, 1975a). The two exceptions were H-2*" and H-2*', in which graft rejection occurred in the direction d + db or d + dc, but not in the direction of db + d or dc + d (McKenzie et al., 1977a; H. Kohn, personal communication). Histogenetically, the H - 2 d b and H-2dc mutations, therefore, behave as though of the loss type (H antigens present in H - 2 d were lost by H-2** and H-2*', and no new H antigen was gained). Grafts exchanged between strains carrying mutations in the same gene (for example, among the various H - 2 b mutant strains) are rejected, as a rule. For example, H-2ba rejects H-2** grafts, and vice versa. However, there are three exceptions to this rule, since grafts exchanged among strains carrying mutations H-2bQ',H-26Q2,and H-2*Q3are not rejected (Melvold and Kohn, 1976).The observation that three independently arisen mutations apparently have the same antigenic composition can be explained in one of two ways. First, the H - 2 K b gene could contain a super-hot spot and most, if not all, of the H-2* mutations could be alterations in this spot. If the spot were to be small, it would have only a limited number of options when it would mutate. Consequently, repeated mutations would frequently be obtained. If this interpretation is correct, b g l , b g 2 , and b g 3 should be biochemically indistinguishable. The second explanation is that the entire haplotype has only a limited number of options when it mutates. According to this hypothesis, amino acid substitutions in different parts of the H-2K* molecule might lead to similar changes in the molecule's tertiary structure. Consequently, the different mutations might look the same to a T cell, which might be recognizing primarily differences in tertiary, rather than primary, structure. This hypothesis predicts that the H-2*Q1,H-2*@, and H - P s 3 mutations will differ in the primary structure of the H-2K molecule. In most of the mutant combinations tested by us (Klein et al., 1975, 1976a; Forman and Klein, 1975a; J. Klein, unpublished data), the median survival time (MST) of tail-to-dorsum skin grafts was between 15 and 21 days, which was comparable to the MST observed in H - 2 recombinant combinations (Table IX).There were, however, three instances in which grafts in the mutant combinations survived longer than in the recombinant combinations: H - 2 d a grafts placed on H-2" recipients were usually rejected within 3 weeks after grafting, but occasionally some survived for prolonged periods of time (more than 70 days); and in the H-2" + H-2ka and H-2ku + H-2" combinations, the MSTs were 39 and 25 days, respectively. The finding with the H-2*a grafts is a paradoxical one, since this strain seems to have accumulated

110

JAN KLEIN

TABLE IX SKIN-GRAFT REJECTIONS IN H-2 MUTANT AND H-2 RECOMBINANTCOMBINATIONF Donor

Recipient

Target antigens

MSTb

Range

16.5 15.1 21.8 18.5 16.2 17.0' 20.5 21.3 39.6 25.9

12-22 12-17 16-25 13-23 14- 19 13-22 16-34 17-27 16-55 13-30

16.5 16.7 15.5 14.6

14- 18 13-18 12-19 11-15

H - 2 mutants Hzl B6 M505 B6 M504 B10.D2 M506 A.CA M523 CBA

B6 Hzl B6 M505 B10.D2 M504 A.CA M506 CBA M523

Kha Kb Kbd

Kh

Dda Dd

K'" K' Kka Kk

H-2 recombinants B1O.A B 10.AQR B1O.G BlO.T(GR)

B 10.AQR B1O.A BIO.T (6R) B1O.G

Kk Kq

Dq

D*

Based on Klein et al. (1974a, 1975, 1976a),Forman and Klein (1975a), and J. Klein (unpublished data). Median survival time. Two grafts survived for more than 70 days.

more mutational differences than any other of the known H-2 mutants. One can speculate that the prolonged survival of M504 grafts is related to the fact that H-Zda antibodies are often produced in this strain combination, and that these antibodies have an enhancing effect on the graA (Egorov, 1967; Klein, 1975).Long MTSs were also observed by Melvold and Kohn (1976) in combinations b 4 b g l and b --.* b g 2 . To determine whether antigens responsible for the rejection of tissues other than skin were also affected by H-2 mutations, we performed a series of heart muscle transplants (Klein et al., 1976b). In these studies, hearts of newborn mice were extirpated and cut into small fragments, and individual fragments were inserted under the ear skin of an adult recipient. Graft survival was monitored by visual inspection, and cessation of the heart mucle pulsation was taken as a sign of rejection. We observed (Table X) that between 50% and 80% of the grafts (depending on the allelic combination) were rejected within 6 weeks after grafting. The frequencies of rejection in the mutant and recombinant combinations were comparable. Thus, the effect of H-2 mutations is not restricted to antigens of the skin, but expresses itself in other tissues as well.

H-2

111

MUTATIONS

TABLE X HEART-GRAFT REJECTIONSIN H - 2 MUTANTAND H - 2 RECOMBINANTCOMEIINATIONS~ Donor

M50S B6 M504 B 10.D2 B 10.AQR B1O.A B10.AKM B1O.BR

Target antigen

Recipient

H - 2 mutants Kdb

B6 MSOS B 10.D2 MSO4 BI0.A B1O.AQR R1O.BR B 1O.AKM

Percent rejected

70

Kb

so

Dda

80

LY

70

H - 2 recombinants Kq Kk Dq Dk

Rejection time (weeks)

100 SO 100

80

Data from Klein et al. (1976b).

The mutant grafting data shed new light on the old problem of antigenic strength. Before, one could argue that the strength of an H molecule was merely a function of the number of antigenic determinants it carried. According to this view, alleles at minor H loci differ by a single mutation, their products carry only a small number of determinants, and so the antigens stimulate only weak immunity. By the same token, unrelated alleles at H-2 loci differ by multiple mutations and their allelic products carry a large number of determinants, which together induce strong immune response to an allograft. The hypothesis predicts that determinants generated by a single H-2 mutation would be weak, as are the determinants controlled b y minor H loci. The observation that, in fact, the majority of mutant H-2 determinants act as strong inducers of allograft reaction indicates that the hypothesis is probably incorrect. Apparently, antigenic strength is not the result of a simple summation of determinants; more likely, it is determined by some qualitative properties of the H molecules. Parenthetically, one must add that some summation of effects apparently occurs between antigens controlled by different loci, since the MSTs in combinations differing in the entire H - 2 haplotype are usually shorter than MSTs in monoregional strain differences. While the number of determinants seems to be of no consequence for the induction of immunity, it does seem to play some role in the elicitation of immunological tolerance. We injected intravenously 15 x lo6 spleen and bone marrow F, hybrid cells into newborn mice of the parental strain, and 7-9 weeks later we measured tolerance by skin grafting (Streilein and Klein, 1977). As Table XI shows, we found

112

JAN KLEIN

TABLE XI NEONATALTOLERANCE INDUCTION ACROSS H - 2 REGIONS ACCORDING TO SOURCE OF DISPARITY: RECOMBINANT VS MUTANT"

H - 2 region ~

K D

"

Source of disparity

No. tolerant/No. tested

Recom1)inant Mutant Recombinant Mutant

0117

Percent tolerant

~

6/26 19/30

17/20

0 23 63 85

From Streilein and Klein (1977).

it somewhat easier to induce tolerance in mutant (closely related), than in recombinant (unrelated), strain combinations. Perhaps this difference in tolerance induction has something to do with the number of T-cell clones one has to tolerize in the two types of strain combinations. B. CML TO ALLOANTIGENS The CML reactivity of H - 2 mutant strains correlates with their graft rejection ability: strain combinations in which grafts are rejected give positive CML, whereas those in which graft rejection does not occur are also negative in the CML assay (Brondz, 1972; Berke and Amos, 1973; Klein et al., 1974a, 1975, 1976a; Forman and Klein, 1975a,b; Nabholz et al., 1975; Chauvenet and Amos, 1975; Melief et aE., 1975, 1976, 1977). The strength of the CML reaction in terms of net 51Crrelease falls within the same range in all mutant combinations tested, and is comparable to that observed in H - 2 recombinant combinations (Table XII). The availability of several mutations affecting the same gene allows one to test CML cross-reactivity among closely related genetic variants. An example of such a test is shown in Table XII. In this experiment, H-2bacells sensitized to H-2b stimulators lysed not only H-2b but also H-2bdtarget cells; similarly, H-2ba cells sensitized to H-2bd stimulators lysed both H-2bd and H-2b targets (Forman and Klein, 1975a,b).These results can be interpreted in terms of shared antigenic determinants as follows. One can assume that the ba anti-b effector cells react with both b and bd, because the two haplotypes share an antigenic determinant that is absent in ba. We assign the symbol H-2K.1 to this determinant. Furthermore, since the b anti-ba effectors lyse ba, but not bd targets, one can assume that ba has a determinant (H-2K.3) that is absent in both b and ba. Similarly, since b anti-bd effector cells react with bd but not with ba, one can conclude that bd

H-2

113

MUTATIONS

T A B L E XI1 CELL-MEDIATED LYMPHOCYTOTOXICITY (CM L) WITH H-2 MUTANT AND RECOMBINANTSTRAINS' Net "Cr-release Responder

Stimulator (target)

B6 Hzl B6 M505 B10.D2 M504 A.CA M506 CBA M 523

Hzl B6 M505 B6 M504 B10.D2 M506 A.CA M523 CBA

BIO.A B1O.AQR B1O.AKM B1O.BR

Target antigens

H-2 mutants K ha Kb K hd K h

D*a D* K 'a K' Kk= Kk H-2 recombinants B 1O.AQR K" B1O.A Kk B10.BR Dk B1O.AKM DQ

Meanb

Range

35 32 41 33 29 32 25 28 29 31

28-37 15-34 29-48 31-35 18-35 25-36 20-28 14 -35 12-38 19-33

35

31-37 30-42 20-35 15-4 1

37 29 34

' Data from Klein et al. (1974a, l975,1976a), Forman a n d Klein (1975a,b), a n d J. Klein (unpublished data). ' Mean of several experiments.

carries a determinant (H-2K.4) that is lacking in both b and ba. Additional determinants can be defined by stimulating an unrelated H - 2 haplotype (e.g., H-2h4,which shares the D region with H-2* and differs in the K region) with a mutant or standard strain (Table XIII). The h4 anti-b(ba, bd) effectors react with all three haplotypes ( b , ba, bd), indicating that the haplotypes have at least one determinant (H-2K.2) in common. From these data a chart can b e constructed that very much resembles the H-2 chart defined by serological methods (Table XIV). (In this chart, the existence of H-2K.5 was deduced from the absorption experiment described below.) Using additional H - 2 b mutants and their F, hybrids, Melief and his co-workers (1977) have recently expanded the H-2 chart by the definition of several additional determinants (Table XV). (Unfortunately, these authors introduced a new numbering system, so that the entries in Tables XIV and XV do not correspond to each other. Another nomenclatorial mess is thus in the making.) Melief s chart contains 19 determinants, of which only 7 are defined by strong CML reactions. At

114

JAN KLEIN

TABLE XI11 CELL-MEDIATED LYMPHOCYTOTOXICITY (CML) BETWEEN H-2b, H-2h, AND H-2bd a CROSS-REACTNITY Mean net 51Cr-release Responder

Stimulator

Target

(%)

96 96 96 B6 Hzl Hzl Hzl Hz 1

M505 M505 Hz 1 Hzl 96 96 M505 M505 96 96 Hzl Hzl

M505 Hzl Hzl M 505 B6 M 505 M505 B6 B6 Hzl Hzl 96

20.9 6.0 27.8 5.5 24.6 21.6 27.4 20.6 26.8 6.5 25.2 14.3

M505 M505 M505 M505 (I

Based on data of Forman and Klein (1975a).

least some of the weak determinants (shown in parentheses) could be artifacts of the in vitro method. The presence of artifacts in the CML system used by Melief and his co-workers is seen from the fact that some of their F, hybrid cells reacted with the H-2 haplotype present in the responder. On the other hand, an artifactual nature of determinants defined by strong CML reactions is unlikely, since the existence of such determinants is also supported by in vivo data. Thus, Apt and his co-workers (1975)reported that H-2bamice presensitized by H-2b spleen rejected H-2bd grafts in an accelerated fashion. The authors TABLE XIV CHARTOF H-2K ANTIGENS RECOGNIZEDI N ALLOGENEIC CELL-MEDIATED LYMPHOCYTOTOXICITY (CML) WITH H - 2 b MUTANTS" ~

~

Antigens

C57BLl6 BG.C(Hz1) 96.M505

H-2 haplotype

1

2

3

4

5

b ba bd

1 1

2 2 2

3 -

-

5 -

4

-

Based on Forman and Klein (1975a), Klein and Forman (1976). and Geib et a!. (1977). (I

ANTIGENIC DETERMINANTS DEFINED BY

THE

T A B L E XV REACTIONS OF FIVEH - 2 MUTANTS IN CELL-MEDIATED LYMPHOCYTOTOXICITY(CML)" Antigenic determinant

H-2 haplotype

b ba bd bg bh

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

lb (1) 1 1

(2) 2 (2) (2)

(3) (3) 3 (3)

4 (4) (4) -

5

8 8 8

10 10 10 -

11 (11) 11

13 -

(14) 14 14 (14)

-

(5)

9 (9) (9) 9

(12) (12)

(4)

(7) (7) (7) 7

-

(5) -

(6) 6 (6) (6)

-

_

5

(8)

_

-

12

-

(15) 15 (15) (15)

(16) (16) (16) 16

17

_ 17 -

18

(18) (18) 18

19

2

&

E

-

19

" Based on Melief et al. (1977). Numbers in parentheses indicate determinants defined b y weak reactions; all other numbers stand for determinants defined b y strong reactions; dash (-) = absence of a given determinant.

5k =I 0

5

116

JAN KLEIN

also observed accelerated rejection of H-2b grafts placed on H-2ba mice presensitized to H-2bdtissues, and failed to find any evidence of second-set rejection of H-2bdgrafts placed on H-2b mice presensitized to H-2ba.These data are in keeping with the cross-reactions observed in the in vitro CML assay. (Apt and his co-workers did not, however, observe any cross-reactivity with H-2b of H-2bd mice presensitized to H-2ba-a result in disagreement with the in vitro data. Whether this discrepancy is real or merely the result of inadequate cell dosage used for presensitization remains to be determined.) In our experiments, we observed cross-reactivity only when the mutant strain was the responder, and never when effector cells were derived from the standard strain. This finding is as expected, since cross-reactivity is the consequence of one mutant strain’s losing an antigenic determinant that the other mutant still shares with the standard strain. Cross-reactivity between two mutant strains (with the standard strain as a responder) would mean that two independent mutations effected a gain of the same antigenic determinant-an event that probably occurs only infrequently. It must be emphasized that the charts in Tables XIV and XV are based on cellular (T-cell) rather than humoral (antibody) reactions. Since it is very likely that T cells and humoral antibodies “see” different portions of the H-2 molecule, no comparisons can be made between the T-cell defined and the serologically defined H-2 charts. Since the H-2bamolecule differs from both H-2band H-2bdin at least two antigenic determinants, the question of whether the two determinants are recognized by one or two different T-cell clones can be raised. We attempted to answer this question by an experiment in which we removed reactive clones by adsorption on a spleen monolayer derived from one mutant strain and then tested the nonadherent cells on target cells derived from another mutant strain (Geib et al., 1977). We observed (Fig. 11) that ba anti-b effector cells can be adsorbed on bd targets so that virtually all the bd-specific reactivity of the nonadherent cells is removed; but these cells still react with b cells. We interpret this finding to indicate that when ba anti-b effector cells are generated, at least two T-cell clones are stimulated-one specific forb cells and the other reacting with a determinant shared by b and bd. The CML mutant studies thus reveal previously unsuspected complexity at both the target-antigen and the T-cell level. The target antigens apparently are complex mosaics of determinants, many of which can activate different T-cell clones. Although the data on the T-cell-defined H-2 determinants are still limited, they can be used to draw certain conclusions about the topo-

H-2

MUTATIONS

117

E A Ratio

FIG. 11. Adsorption of Hzl (H-2bfl)anti-B6 ( H - 2 b ) effector (E) cells on Hzl, M505 ( H - 2 9 , and B6 monolayers, and testing of the nonadherent cells on B6 targets (T) in the CML assay. [From Geib et al. (1977), reprinted with the permission of Williams and Wilkins Co., Baltimore, Maryland.]

logical properties and evolution of H-2 molecules. The first conclusion is that a single H-2 molecule carries multiple, T-cell-defined determinants. The evidence for this conclusion was discussed above. The second conclusion follows from the observation that H-2h4cells recognize a determinant (H-2.2) common to different H-26 mutants. Since this determinant has not been affected by any of the mutations, it must be located in a region (site) distinct from the mutant site. Thus, there are at least two-probably nonoverlapping-regions (sites)-in the H-2 molecule: one in which mostH-26 mutations occurred, and another which none of the mutations has altered. Whether determinants H-2K.1,3,4, and 5 all reside in the same site or whether they form two or more sites cannot b e resolved at the present time. However, at least some do seem to occupy the same general area, and one can, therefore, draw a third conclusion-that one region probably can carry more than one antigenic determinant. The fourth conclusion is that mutants in the same gene show different degrees of relatedness among themselves and toward the standard allele. Thus, for example, the fact that ba anti-b effectors cross-react with bd, that ba anti-bd effectors cross-react with b, and that no cross-reactivity occurs in any

118

JAN KLEI N

other combination of these three haplotypes (Forman and Klein, 1975a) can be taken to mean that bd is more cIosely related to b that is ba. When one adds to these findings the data of Melief and his coworkers (1976,1977), one can arrange the H-2b mutants into a series in the order of decreasing relatedness to b: b > bg > (bd > bh) > ba. In this series, bg is the most closely related haplotype to b , and ba is the most unrelated. The closeness of bg to b is also supported by the slow rejection of skin grafts exchanged between strains bearing the two haplotypes (Melvold and Kohn, 1976) and by the weakness of the b t* bg MLR (Melief et al., 1976). The relative unrelatedness of ba to b is supported by the complete lack of cross-reactivity between the two haplotypes in CML against virus-infected cells (see Section X,C). A hypothetical scheme of how two mutants might have originated from the standard H-2b haplotype is shown in Fig. 12. One can extend this kind of speculation and suggest a general scheme of how the evolution of an H-2 haplotype might proceed (Fig. 13). Since unrelated H-2 haplotypes show very little cross-reactivity in the CML assay: the evolution can be envisioned as a progressive loss of shared determinants and generation of unshared ones. C . CML TO ASSOCIATIVE ANTIGENS Certain antigens, when attached to or incorporated in the cell membrane, are recognized by T lymphocytes in association with MHC

haw 1

lost, 5-3/

\4

b

d

w

FIG. 12. Changes in antigenic determinants during the derivation of H-2ba and H-2W mutants from H-2*. Ovals represent the H-2K molecule; circles, antigenic sites; numerals, antigenic determinants; and letters, H- 2K alleles. This conclusion was first reached by Brondz (1964)and has since been confirmed by numerous other investigators-too numerous, in fact, to quote them all here. The fact that occasionally some investigators d o find CML cross-reactivity among at least some unrelated H - 2 haplotypes (e.g., Phillips et al., 1973) suggests that some sharing of antigenic determinants must exist, even among haplotypes that are not demonstrably related. The reason that most investigators d o not find such cross-reactivity is probably the insensitivity of their CML assays.

H-2

MUTATIONS

119

FIG. 13. A hypothetical scheme of H-2 haplotype evolution through multiple changes in antigenic determinants. Letters indicate H-2 haplotypes; numbers, antigenic determinants.

molecules present on the same cell. These associative antigens can either be generated by viral infections (Zinkernagel and Doherty, 1975) and chemical modification of membrane components (Shearer et al., 1975; Forman, 1975), or they can be present in the cell as natural constituents of the membrane in the form of minor H antigens (Bevan, 1975; Gordon et al., 1975). T lymphocytes stimulated by an associative antigen develop into effector cells capable of lysing appropriate target cells in the standard CML assay. However, the CML reaction occurs only when target cells carry not only the same associative antigen as stimulators, but also similar MHC molecules. The H-2 mutants provide an opportunity to determine how similar the MHC molecules of the stimulator and the target must be for cell lysis to occur. One can, for example, generate effector cells to associative antigens in the context of the H-2Kb molecules and then ask the question whether these cells will lyse targets carrying the same antigen in the context of, say, H-2Kbamolecules. When such experiments were carried out by Zinkernagel(1976),Zinkernagel and Klein (1977), Forman and Klein (1977), Blanden et al. (1976, 1977), and J. Klein and C. L. Chiang (unpublished data), some interesting findings emerged (summary in Table XVI). First, it seems that different types of associative antigen have different requirements as regards the MHC similarity of the stimulator and target. Thus, for example, H-2Kb effectors stimulated by trinitrophenyl (TNP)-modified H-2Kbcells will lyse not only TNP-modified H-2Kb but also H-2Kba cells (Forman and Klein, 1977).At the same time, however, H-2Kb effectors stimulated by lymphocytic choriomeningitis (LCM) virus-infected H - 2 K b cells will lyse LCM-infected H-2Kb but not H-2Kbatargets (Zinkernagel, 1976).In fact, a complete cross-reactivity has been observed among all the mutants so far tested when TNP is used as the associative antigen (Forman and Klein, 1977), while only limited cross-reactivity is seen when the antigen is provided by a virus (Zinkernagel, 1976; Zinkernagel and Klein, 1977; Blanden et al., 1977). Whether this result is caused by some basic difference in the mechanism of the response to the two types of antigen,

120

JAN KLEIN

TABLE XVI BETWEEN H - 2 MUTANTAND CROSS-REACTIVITY STANDARD STRAINSIN ASSOCIATIVE CELL-MEDIATED LYMPHOCYTOTOXICITY (CML)O Cross-reactivityb of cells sensitized to

H-2 haplotype

Lymphocytic choriomeningitis virus

Ectromelia virus

Vaccinia virus

ba bd bf bg bh

NT NT

+

da

db fa ku

*

+

+

2

'.

NT NT NT

NT -

+

Trinitrophenyl

Minor H antigens

+ + +

-? NT

NT NT NT

NT

+ +

NT

2

+ f

+

NT NT

" Based on Zinkemagel (1976), Blanden et al. (1976, 1977), Zinkernagel and Klein (1977), Forman and Klein (1977), R. Zinkernagel (personal communication), and J. Klein and C. Chiang (unpublished data). +, Cross-reactivity present between standard and mutant strains; -, cross-reactivity absent; ?, weak cross-reactivity; NT, not tested.

or whether it merely reflects differences in the quantity or manner of antigen presentation, is not clear at this time. It is interesting to note, however, that the pattern of response to minor H antigens (J. Klein and C. Chiang, unpublished data) seems to follow, with one exception, the virus-response pattern. The exception is that b and ba haplotypes strongly cross-react in the minor H but not in the virus systems. The cross-reactivity in the minor H system is totally unexpected, since b and ba appear to be the most unrelated haplotypes among the b mutants. One possible explanation of the cross-reactivity is this. To generate effector cells against, say, H-3 antigens, we stimulated B1O.LPa cells by BlO.A(5R) and then tested the effector cells on C57BL/6 and the various mutant targets. The combination is such that, in addition to anti-H-3 response, anti-H-2Dd response is also stimulated. Since the H-2b and H-2b-mutant cells lack H-2Dd, only the anti-H-3 response should be detected. One can speculate, however, that the H-2ba mutation altered the H-2Kb molecule in such a way as to resemble the H-2Ddmolecule. If so, then the observed cross-reactivity of i5 anti-b cells would not be directed against H-3 but against H-2 antigens. Preliminary genetic studies seem to support this contention. Second, the response within the same group of antigens is strikingly

H-2

MUTATIONS

121

similar. Thus, the response pattern to the three viruses tested (LCM, vaccinia, and ectromelia) is the same, despite the fact that LCM, on the one hand, and vaccinia and ectromelia, on the other, are quite different types of virus. In general, the presence or the absence of CML cross-reactivity to viral antigens reflects the degree of relatedness of the dif‘ferent mutants, as determined b y allogeneic CML. No crossreactivity is seen between b and ba, the two most unrelated haplotypes among the H-2* mutants, whereas some cross-reactivity is observed between b , on the one hand, and bd, ha, and bh, on the other, where the haplotypes seem to be more closely related. Third, in all combinations and with all antigens tested, the presence or the absence of cross-reactivity is always reciprocal. For example, H-2Kb effectors stimulated by LCM-infected H-2Kb cells lyse H-2Kbd targets; and, reciprocally, H-2Kb effectors stimulated by H-2Kbdcells lyse H-2Kb targets (Zinkernagel and Klein, 1977). A similar observation has been made with all other types of antigen and the various mutant strains.

1 . MLR

In 1965, when Dutton first described the MLR-stimulatory role of H-2 antigens, very little was known about the functional complexities of the H-2 system. Therefore, it seemed logical to assume that the stiniulatory antigens were identical to those defined b y serological methods, and controlled by the K and D regions (Klein and Shreffler, 1971). The first indication that this kind of logic might not apply to the H-2 system came from the study of Rychlikovh and her co-workers (1970), who reported that the K end seemed far more important for MLR stimulation that the D end. In fact, the D-end antigens did not seemed to stimulate at all. Spurred b y this finding, Bach and his coworkers (1972) and, practically at the same time, Meo et al. (1973)initiated a study of the then available H-2 recombinants and soon were forced to the unexpected conclusion that the main stimulus for MLR was coming not from the K or D regions, but from the center of the complex, in which no genes for serologically defined antigens were known at that time. As a result of this observation, the opinion immediately swung to the other extreme, and the participation of MLR of the K and D regions was denied completely by many investigators. The low degree of MLR stimulation that was observed in some K or D region-disparate combinations (Klein et al., 1972) was dismissed as being caused by “contaminating” 1-region loci that recombinations

122

JAN KLEIN

failed to separate from H-2K or H - 2 D . The labors of the H - 2 complex thus seemed to be neatly divided among different regions, with the center primarily responsible for MLR and the periphery inciting CML and like reactions. This view of the MHC was also supported by human studies in which no MLR stimulatory role could be found for the H L A - A and -73 loci (for a review, see Svejgaard et al., 1975). However, when the first H - 2 mutant was tested in MLR (Rychlikova, and Ivanyi, 1969; (Rychlikovh et al., 1972), the most surprising result was obtained. Although the mutation (H-2da)mapped in the H - 2 D locus, the mutant cells clearly stimulated the corresponding standard-strain lymphocytes in a mixed-lymphocyte culture. One would think that this finding would raise some doubts in the minds of those immunologists who believed that there was no MLR outside of the Z region. Not so! Instead, these investigators suggested that the H-2da haplotype carried two mutations, one in the H - 2 D locus and another somewhere in the Z region. And when mapping studies proved that there was no MLR mutation in the Z region 0 f H - 2 ~ "(Forman and Klein, 1975b), they postulated an Z-like locus in the D region, close to H - 2 D . They were not deterred in their conviction, even when later studies showed that one mutation after another was affecting MLR. Apparently, according to these investigators, each one of the dozen or so mutant haplotypes carried two mutations, one in H - 2 K or H - 2 D 7 or another in an unidentifiedz-like locus. It did not help much when further mapping studies with other H - 2 mutants again excluded the participation of the Z region, as currently defined, in the mutant MLR (Klein et al., 1975), since the existence of hypothetical I-like loci closely linked to H - 2 K and H - 2 D could not be normally excluded. In an attempt to obtain a direct proof of H-2K and H - 2 D loci participation in MLR, we took advantage of the recent finding by Nagy and his co-workers (1976).These authors observed that the MLR is accompanied by the release of stimulatory antigens from the stimulating cells and binding of these antigens on the responding cells in the process of their transformation into blasts. The blast-bound antigens could then be demonstrated by the fluorescent antibody techniques. Since, in two strain combinations (M523 anti-CBA and M504 antiB10.D2), we were able to produce antisera distinguishing the mutant and the standard strains, we used the antisera to determine whether the antibodies they contain bind to blasts generated in M523 antiCBA and M504 anti-BlO.D2 mixed lymphocyte cultures. The results (Table XVIII; cf. Geib and Klein, 1978)indicate that they do. Since we have shown previously (Klein et al., 1976c)that the antisera react with antigens that cocap with H-2K (M523 anti-CBA) or H-2D (M504

H-2

123

MUTATIONS

TABLE XVII MIXED LYMPHOCYTE REACTION (MLR) AND GRAFT-VERSUS-HOST REACTION (GVHR) WITH H - 2 MUTANTS AND H - 2 RECOMBINANT STIWNS" ~~

Activity index" Responder (donor)

Stimulator (host)

R6 Hzl B6 M505 B10.D2 h4 504 A.CA h4 506 CBA h1523

Hz 1 B6 M505 B6 M504 B10.D2 M506 A.CA M523 CBA

A.TH A.TL

A.TL A.TH

Stimulating anti gen s

MLR

GVHR

5.8 5.9 8.7 7.7 9.2 2.5 2.3 1.7 2.3 3.1

2.1 1.5 1.5 1.9 1.9 1.4 1.4 2.1 1.6 1.7

9.2 12.5

2.1 1.9

H - 2 mutants Kha Kb Kbd

Kh

D *a Dd

K'a

K' Kka

Kk H - 2 recomhinutits Ik

I'

Data from Klein (1976a, 1977), K l e i n et u1. (1974a, 1975, 1976a), Forman and Klein (1975a,h),Klein and Eyorov (1973), and J . Klein (unpublished data). Stimulation index for MLR and spleen index for GVHR; means from several experiments.

anti-BlO.D2) antigens, we conclude that the antigens binding to MLR blasts are controlled by the H-2K and H - 2 D loci. In summary, the thoroughly studied H - 2 mutations that have been shown to affect either H - 2 K or H - 2 D loci all affect MLR; in mutations in which genetic mapping was possible the loci controlling MLR-stimulating antigens were localized in the K or D regions; and when MLR occurs in the mutant-standard strain combinations, H-2K or H-2D antigens of the stimulator are found on the responding cells. If all these data do not convince one that H-2K and H-2D antigens are able to stimulate MLR, nothing will! Although significant MLR occurs in all mutant-standard strain combinations tested, the degree of stimulation varies, depending on the specific combination (Rychlikova et al., 1972; Widmer e t al., 1973; Forman and Klein, 1975a,b; Klein et ul., 1974a, 1975, 1976a; Melief et al., 1975). The reaction is weak in some combinations [stimulation index (SI) between 1.3 and 2.01, moderate in others (SI between 2.0 and 5.0), and strong in still others (SI above 5.0, cf. Table XVII). In the

124

JAN KLEIN

strongly stimulating combinations, the mutant MLR is comparable to that obtained in the recombinant strains differing at the Z region. However, when one compares the MLR in mutants and recombinants differing at the K or D regions, the former usually show a higher degree of stimulation than the latter. This finding is paradoxical: the more closely related H-2K or H-2D alleles of the mutants stimulate more than the distantly related alleles of H-2 recombinants. The paradox is probably real (not, for example, the consequence of a small sample size and interallelic variation), since in the human H L A system, where only unrelated haplotypes are tested, the K - and D-region homologs do not seem to stimulate at all (Yunis and Amos, 1971). How can this paradox be explained? We would like to suggest that the high mutant MLR is a consequence of some internal property of the T-cell receptor for the I(- and D-region alloantigens. We have speculated (Zaleski and Klein, 1977a) that part of the diversity of the T-cell receptors for KID-region antigens is generated by the H-2WH2D antigens themselves, and that the diversification proceeds from self to nonself antigens in a radial fashion. As a consequence of this diversification, an H-2a mouse contains the highest numbers of T-cell clones with receptors for H-2a-like molecules, an H-2bmouse contains a high frequency of T cells with receptors for H-2b-like antigens, etc. Since H-2 mutations alter antigens only slightly, the altered antigens encounter a high frequency of T cells with receptors for them, and a relatively high MLR stimulation ensues.

2 . GVHR The response of the H-2 mutants in the graft-versus-host reaction (GVHR), whether measured by splenomegaly or lymph node enlargement, closely resembles their response in MLR. This observation thus supports the notion that the two reactions may operate on the same principle and against the same antigens. Although GVHR has been observed in all standard-mutant combinations tested, the strength of the reaction varies (Klein and Egorov, 1973; Klein et al., 1974a, 1975, 1976a; Mnatsakanyan and Egorov, 1975; Forman and Klein, 1975a, Egorov et al., 1977). Not infrequently, the reaction is quite weak in one direction and strong in the opposite direction (Table XVII). An extreme example of such asymmetrical behavior of two alleles at the same locus has been reported by Mnatsakanyan and Egorov (1975), who observed strong reaction of bd grafts to ba hosts, but no reaction of ba cells injected into bd recipients, Surprisingly, bidirectional GVHR occurs between b and ba, as well as between b and bd. The nonresponsiveness of ba to bd cannot be explained by the as-

TABLE XVIII H-2K AND H-2D ANTIGENS BY MIXED-LYMPHOCYTE REACTION (MLR) BLASTSGENERATED I N STANDARD-MUTANTSTRAIN COMBINATIONS" ,\CQUlSITION OF

Responder

Stimulator

Stimulation index

Antiserum"

No. of cells with diffuse staining

No. of cells with patchlike staining

Total No. of cells counted

L M523

M504

CBA

B10.D2

2.5

3.5

Anti-H-2.60 Anti-H-2.23 NMS' Anti-H-2.40 Anti-H-2.4 NMS

17 71 2

5 73 8

63 5 3 86 19 7

100 100 100

102 100

100

Data from Geib and Klein (1978).

* First treatment with alloantiserum;

second treatment with goat antimouse Ig antiserum conjugated with fluorescein isothiocyanate. Both treatments were carried out in noncapping conditions. T h e relevant antigens of the strains used are: M523(H-2.23+,H-2.60-), CBA(H-2.23+, H-2.60+),M504(H-2.4+, H-2.40-), and B10.D2(H-2.4+, H-2.40'). Patchy staining is an indication of an antigen acquisition by the cell; diffuse staining is indicative of cells' own antigens. NMS, normal mouse serum.

sz 9 4

5z v1

126

JAN KLEIN

sumption that ba has all the antigenic determinants present in bd because the defect is complemented by b (blba hybrids respond normally to bd), as well as by several unrelated H - 2 haplotypes (Egorov et al., 1977). Egorov and his co-workers discussed several possible explanations of the ba anti-bd nonresponsiveness and concluded that probably the H-2 molecules themselves are involved in the recognition of other H-2 antigens and that the recognition in ba is distorted by the mutation. However, it is important to note that ba responds poorly-both in GVHR and MLR-not only to bd but also to b and all other b mutants tested (Egorov et al., 1977; Melief et aZ., 1976). For example, in the b t* ba combination, the spleen index of b responding to ba is 2.1, whereas in the reciprocal direction the SI is only 1.5 (Table XVII). The ba strain, therefore, seems to be a generally poor responder to other H-2Kb alleles, and one can argue that the lack of GVHR in the ba anti-bd combination is merely an extreme expression of this general defect. (The ba haplotype responds normally to unrelated haplotypes, but this is not surprising since such a response involves genes outside the K region.) The poor responsiveness itself can be explained by the relatively large distance (in terms of relatedness) between ba and the other K b alleles, and by the assumption that the distance is larger in the direction from ba to other b alleles than in the opposite direction. The lengthening of the distance might result in weakening of the GVHR and MLR, as discussed in the preceding section. E. PRODUCTION OF AN ALLOGENEICSUPERNATANT T lymphocytes cultured with allogeneic stimulating cells produce a soluble factor that nonspecifically helps B cells to mature into plasma cells (Schimpl and Wecker, 1972). The maturation can be measured by the production of antibodies against an antigen [e.g., sheep red blood cells (SRBC)] added to the culture. Until recently, the production of this allogeneic factor (or supernatant) was thought to be controlled exclusively by the I region of the H-2 complex. However, experiments with H - 2 mutants (Kettman et al., 1977) indicate that the K region (more specifically the H - 2 K locus) can perform this function as well. For example, a mixture of H-2kand H-2kacells produces an allogeneic factor capable of aiding anti-SRBC response (as measured by the number of plaque-forming cells) to the same degree as a factor produced in a mixture of cells differing in the entire H - 2 complex (H-2k + H-2d, cf. Table XIX). The MLR occurring in the culture of standard and mutant cells thus has the same characteristics as MLR in Z region-disparate strain combinations.

H-2

TO

127

MUTATIONS

TABLE XIX ABILITY OF MIXTURES OF MUTANT CELLS GENERATE AN ALLOCENEIC SUPERNATANT (AS)" Anti-SRBC PFClwellb (SEM)

Cell mixtures

Response at 50% AS

CBA

M523 BALBlc CBA + M523 M523 iBALBIc

Response at 90% AS

N Db 28 (3) 37 ( 3 ) 175 (17) 124 (13)

From Kettman et al. (1977).

* SRBC, sheep red blood cells; PFC, plaque-forniing cells. ND, not done. F. GRAFT-VERSUS-HOST DISEASE Injection of parental lymphocytes into an F, hybrid leads to a host's response referred to as the graft-versus-host disease (GVHD). The severity of the GVHD depends on the number of cells injected, the genetic disparity between the donor and the recipient, the age of the host, and many other factors (for review, see Elkins, 1971). To study the effects ofH-2 mutations on GVHD, we injected 5 X lo' parental spleen cells into young adult, sublethally irradiated (400 R) F, hybrids and recorded recipients' death as the ulterior sign of the disease. Using this system, we observed (Table XX) that the GVHD mortality in the mutant-standard strain combinations varied between 50% and 10070, depending on the particular allelic combination (Klein, 1976a, 1977a; Klein and Chiang 1976; J. Klein, unpublished data). Most recipients died within 4 weeks after injection, but in some instances the time of death was delayed until several months postinjection. Paradoxically, GVHD in the mutant combinations was more severe than that in the recombinant strains (Table XXI). The reason for this paradoxical behavior is not known, but perhaps the phenomenon is somehow related to the strength of the proliferative response in the two groups of animals. As discussed earlier, both MLR and GVHR are stronger in the mutant than in the corresponding recombinant combinations, and perhaps the strong proliferative reaction then leads to a strong cytotoxic GVHD response.

G. SEROLOCYOFH-2 MUTANTS Serological differences between H-2 mutant and a standard strains can, theoretically, be detected in two ways: first, by testing the mutant

128

JAN KLEIN

TABLE XX GRAFT-VERSUS-HOST DISEASE(GVHD) WITH H - 2 MUTANTSAND H - 2 RECOMBINANTS" Donor

Target antigens

Recipient

B6 M505 M504 BlO.D2 CBA M523 BIO.A B1O.AQR BIO.T (6R) Bl0.Q

Percent dead

Time of death (days)

53 44 100 85 67 66

24-72 15-55 16-63 12-75 10-92 11-30

10 20 13 20

15 17-19 21-42 17-21

H - 2 mutunts (B6 x M505)F, Kbd (B6 x M505)F1 Kb (B10.D2 X M504)F1 Dd (B10.D2 x M504)F1 Dda (CBA x M523)F1 Kka (CBA X M523)F, Kk H - 2 recombinants (B1O.A X BlO.AQR)F, Kq (B1O.A x BIO.AQR)F1 Kk LB1O.T (6R) x BIO.QIFI Dq Dd IB1O.T (6R) x BIO.QIFI

Based on Klein (1976a, 19771, Klein and Chiang (1976), and J. Klein (unpublished data).

strain with a battery of antisera directed against known antigens (and thus establishing whether any of the known H-2 antigens have been altered, lost, or gained); and second, by cross-immunizing the two strains (and thus producing antisera specific for either the mutant or the standard strain). Using the former approach, serological differences from the standard strain have been found in all H - 2 mutations that have been

TABLE XXI CONTRIBUTION OF MUTANT AND RECOMBINANT H - 2 ALLELES TO THE INDUCTION OF (GVHD)" Difference H - 2 region

Type

No. of combinations tested

Mortality Total inoculated mice ~

K K D D

Mutant Recombinant Mutant Recombinant

4 2 2 2

70

MSTb

62 15 100 17

36 16 27 29

~~

73 20 26 23

Based on Klein (1976a, 1977), Klein and Chiang (1976), and J. Klein (unpublished data). * Median survival time (days).

H-2

MUTATIONS

129

thoroughly studied. That is all with the exception of the Kohn-Melvold H-2” mutants, which have so far been studied serologically only by McKenzie and his co-workers (1977a). These authors failed to find any serological difference between H-2” and H-2bg1,H-2*g2, and H-2bh. However, as discussed below, McKenzie et a l . also failed to find serological difference between H-2” and H-2”-two haplotypes that were successfully differentiated b y other authors. The Kohn-Melvold mutants therefore deserve a second look. Using the latter approach, antisera were produced by cross-immunization between standard strains and mutants H-2da (Dishkant et al., 1973), H-2db (McKenzie et al., 1977a), H-2’” (Egorov, 1976),H-2’b (L. Mobraaten, quoted by McKenzie et al., 1977b), and H-2ku (Klein et al., 1977). No antiserum has as yet been produced against any of the H-2” mutants. We tried various immunization “tricks,” such as inoculation of killed cells (to avoid GVHR), immunization with different doses of cells, injection of cells to which “carrier” molecules (LPS) had been attached, immunization of F, hybrids carrying different background genes, and immunization across an H-2 barrier added on top of the mutant difference. But we failed invariably in all these attempts (J. Klein, unpublished data), as have other investigators using the same or different approaches (Bailey et nl., 1971; McKenzie et al.,

1977b). Do all these failures mean that there are no serological differences between the H-2” mutants and the standard C57BL/6 strain? Certainly not! As discussed below, serological differences between H-2” and its mutant haplotypes have been found by several investigators, using antisera against known H-2 antigens. Why, then, the failures? Anyone who has tried to produce a monospecific H-2 antiserum in a strain combination with restricted antigenic difference knows how difficult such a task is. In general, the more restricted the difference between two strains, the more difficult it is to obtain an antiserum by crossimmunization of these strains. So the failures to produce antibodies by cross-immunization of strains that are absolutely identical except for a little bump on their H-2 molecules are not unexpected. In fact, it is surprising that one can produce antibodies in some of the mutantstandard strain combinations. A summary of the H - 2 mutant serology appears in Table XXII. Serological data pertaining to the individual mutations are discussed below. H-2*”, H-26ff,H - 2 b f .Evidence for serological differences among H-2”, H-2””, and H-2*d haplotypes was obtained independently by Klein et al. (197413, 1975) and Apt et al. (1975). The latter authors observed

TABLE XXII ANTIGENICCOMPOSITION OF MUTANT AND STANDARD H - 2 HAPLOTYPES

H-2 haplotype

b ba bd d do

f

fa

k

ka (I

H-2 antigens 1

2

3

5

4

-

2 - 5 2 - 5" - 2 - 5b - _ 3 4 - - 3 4" -

- - - - 1 1

-

3 3

-

5 5

6

6 6 6 6 6 6 6

7

8

8 - 8 7 ? 7 ? 7 8 7 8

9

11

13 13 13

23

25

26

31

26 26"

-

Quantity of antigen reduced in comparison with the standard strain. Quantity of antigen increased in comparison with the standard strain.

31 31

32

33

34

37

39

33 33" 33"

- - - -

39 39 39

3 4 - 3 4 - - 37 39 - 37" 39

40

49

49 49

49 49

50

60

62

H-2 MUTATIONS

131

that absorption of an H-2dlH-2a anti-H-2i5 serum (anti-KbZb)by H-2ba cells removed all of the antiserum’s reactivity against H-2ba,but did not remove reactivity against H-2b and H-2bd; absorption by H-2bd cells, on the other hand, removed all the antiserum’s reactivity against H-2bdcells, but left some reactivity against H-2b and H-2bacells. Apt and his colleagues interpreted these results as evidence for the presence in H-2b and H-2bdof an antigen that is lacking in H-2ba,and of another antigen shared by H-2b and H-2babut absent in H-2bd.Essentially similar results were obtained by Klein et al. (1974b) with an H-2d/H-2aanti-H-2i serum. These authors, however, were able to absorb out the anti-H-2* reactivity by large doeses of either H-2baor HZbd cells. They concluded, therefore, that the serological differences among H-24 H-2bn,and H-2bdhaplotypes were of a quantitative nature. McKenzie and his colleagues (1976) repeated the absorption experiment with an antiserum similar to that of Apt et al. (1975), and reported that they could not find any serological difference between H-2b and H-2ba.How could this discrepancy be explained? First, it is important to realize that there is no principal difference between the data of Apt et al. (1975) and Klein et al. (1975). In a quantitative absorption experiment, one can take a point on the curve where the absorption by one strain is complete and say that the remaining activity against a second strain is either caused by the same antibody, which the first strain for some reason failed to remove, or by a different antibody that does not react with the first strain. To distinguish between the two possibilities, one attempts to absorb out the reactivity against the second strain by increased doses of cells from the first strain. If the absorption is successful, one says that the difference between the two strains is quantitative; otherwise one concludes that the two strains differ qualitatively. This approach works well when one is dealing with unrelated antigens and when the second antibody is drastically different from the first one. However, when one works with a population of closely related antibody molecules-and such are known to be present in most H-2 antisera, the distinction between quantitative and qualitative becomes a matter of semantics. All one can say in such a case is that the two strains differ serologically; whether the difference is labeled “quantitative” or “qualitative” is irrelevant. We shall return to this point later. The second important point in considering the negative data of McKenzie and his colleagues is that these authors are the only ones who failed to find serological differences among H-2bmutants. In contrast, three laboratories did find such a difference. In addition to the two studies already mentioned, a more recent study by David and his

132

JAN KLEIN

co-workers (1977) confirmed (without properly acknowledging it) the earlier work on the antigenic difference between H-2b and H-2bd.The latter authors found that an antiserum similar to the one used by Klein et al. (1975) retained anti-H-2b activity after complete absorption by H-2bdcells. They named the antigen defined by the residual activity 62 and identified it by biochemical analysis as H-2 (rather than Ia). Further support for serological differences among H-2b mutations comes from a study by Mnatsakanyan et al. (1977), who reported that three of their antisera (d antii7, d anti&, and dlf anti-s) failed to react with H-2baand H-2bf erythrocytes in the PVP hemagglutination test. The same antisera reacted strongly with H-2b and H-2bderythrocytes (more strongly with the latter than with the former). Since the only antibody that all three antisera have in common is anti-H-2.5, the authors concluded that the H-2.5 antigen has either been lost or drastically changed by the H-2baand H-2bfmutations. Absorption data with appropriate erythrocytes support this conclusion. Thus, the score between those who do and who do not find serological differences among the H-2bmutants is, at present, 4 : 1. In a soccer game between two good teams such would probably be the winning score. Furthermore, the serological differences among the H-2b mutants cannot be dismissed as “trivial” or “only minor.” A loss of one of the strongest, most widely shared, and most stable antigens (H-2.5) is anything but trivial. H-2b0’,H-2b02,and H-2h. McKenzie et al. (1977a) found no serological difference between these mutations and the standard strain. The mutations have not been tested serologically in any other laboratory. H-2da.The mutation was studied in detail by Dishkant et al. (1973) and Klein et al. (1974a, 1976~). The former investigators came to the conclusion that the H-2d and H-2da strains differ in antigens H-2.40, 49, and 50, defined by H-2d anti-H-2da,and H-2daanti-H-2d sera. Antigen H-2.40 resembles H-2.4 in that the two have the same strain distribution, with one exception: the H-2.40 antigen is absent in H-2da, which is considered to be H-2.4-positive (but see below). Antigen H2.49 resembles H-2.3, with H-2daagain being H-2.49-negative, H-2.3positive. Antigen H-2.50 is specific for the mutant H-2da haplotype. Klein et al. (1976~) confirmed these results and found, in addition, that anti-H-2.4 sera reacted less strongly with H-2dathan with H-2d cells, an observation in agreement with the biochemical data of Brown et al. (1978). Clearly H-2.4 has been altered by the mutation, but whether the change should be considered quantitative or qualitative is debatable. In most antisera, the H-2.40 antibody reacts in both the hemagglu-

H-2

MUTATIONS

133

tination and the cytotoxic tests; the anti-H-2.50 is usually only a hemagglutinating antibody. H-2.50, however, is present on lymphocytes because these cells absorb the H-2.50 antibody. Perhaps the inefficacy of anti-H-2.50 in the cytotoxic test has something to do with the carbohydrate changes in the H-2Dda molecule reported by Brown et al. (1978). Cocapping exaperiments have established that H-2.40 and H-2.4 are carried by the same molecule encoded by the H - 2 D locus (Klein et aZ., 1977). Therefore, at least one mutation in the H-2da haplotype must have occurred in this locus. H-2db. McKenzie et al. (1977b) reported that H-2dbcells, in contrast to H-2d cells, do not react with antisera detecting antigen H-2.28, and that the H-2db anti-H-2d serum contains H-2.28 antibodies. Since the H-2.28 antigen is presumably encoded by the H - 2 L locus (see Section 11), they concluded that the mutation occurred at that locus. This conclusion is supported by the biochemical data of Hansen et al. (1977), who demonstrated the absence of H-2.28 and presence of H-2.4 peaks in the H-2dbstrain. Surprisingly, skin-grafting experiments have established that H-2db does not complement the defect in the H - 2 d a mutation [H-2d grafts were rejected by H-2da/H-2db F, hybrids, cf. McKenzie et al. (1977b)I. This result would normally mean that H-2da and H-2db mutations occurred in the same locus, but in this particular instance some other explanations must apply, since the serology indicates that the mutations occurred at different loci. One possibility is that in one of the strains, two mutations occurred, one in a locus controlling the serologically detectable antigens, and the other in a locus responsible for the H antigens. In fact, since da and db were detected in different standard strains, one of the two mutations might have been present long before the second mutation was detected. Another possibility is that a single mutation in either H-2da or H-2db affected two loci ( H - 2 D and H - 2 L ) . One can, for example, visualize a deletion encompassing the terminal portion of the H - 2 D gene and the entire H - 2 L gene. Either of these two hypotheses wouId explain the discrepancy that currently exists between the serological data and the results of the complementation study. H-2". H-2facells do not react in a direct cytotoxic test with our antiH-2.26 antisera (H-2.26 is the private antigen of the H - 2 K f allele), but One do absorb these antisera in an in uitro test (Klein et al., 1976~). thus faces the same problem as with the H - 2 b or H-2dn mutants: should the change be considered quantitative or qualitative? All one can say is that the H-2faantigens are sufficiently changed so that at least some H-2.26 antibodies react with them very poorly. In addition to H-2.26,

134

JAN KLEIN

antibodies to antigens H-2.37 and H-2.39 also showed reduced reactivity with the H-2facells (in comparison to H-2'cells). Reactivity of antibodies to other H-2f antigens was not changed. A successful production of an antiserum by cross-immunization of H-2' and H-2fa strains has been reported by Egorov (1976). H-2*. L. Mobraaten (quoted by McKenzie et al., 1977b) obtained an H-2* anti-H-af serum that reacts with H-2f erythrocytes in the PVP hemagglutination test. H-2ka.We obtained an H-2 kn anti-H-2 serum defining antigen HThe anti2.60 (present in H-2 k, absent in H-2ku;cf. Klein et al., 1976~). gen cocaps with H-2.23, an antigen controlled by the H-2K locus. No other H-2k antigen detected by our battery of antisera seems to have been changed by the H-2kamutation. In addition to the tests for H-2K and H-2D antigens, we also tested H-2 mutants ba, bd, da,f a , and ka for their reactivity with anti-la sera (Klein et al., 1976~).In all instances, we found the mutant and the standard cells to be serologically indistinguishable, indicating that the serologically detectable Ia antigens have not been altered by these mutations. The foregoing discussion leads to the conclusion that all thoroughly studied mutations display some alteration in their serologically detectable sites. The degree of the alteration, however, varies among the different mutations. The mutations can be arranged into a continuous spectrum, with those carrying relatively small serologically detectable alterations at one end of the spectrum and mutations with differences easily detectable b y cross-immunization at the other. Clearly, there is no basis for a division of the H-2 mutations into two classes, one with only histogenetic and no serological changes, and another with both histogenetic and serological alterations (McKenzie et al., 1977b). So far there is only one class of H-2 mutations with both serological and histogenetic changes; if the second class exists, it has not been found. At any event, such a classification is grossly misleading, since it is based on data that cannot be confirmed by other laboratories. The available serological data provide important information on the topology of the serologically detectable H-2 antigens (Klein et al., 1976~). Thus, the fact that the H-2fumutation changed the expression not only of H-2.26 (the private antigen of H-2Kf), but also of H-2.37 and 39 (two public antigens), argues that these antigens are carried by the same molecule. This finding extends our previous observation (Hauptfeld and Klein, 1975) that public and private H-2 antigens are controlled by the same locus. The observation that a single mutation (H-2"")apparently changed

H-2

MUTATIONS

135

the expression of three different determinants (H-2.26, 37, and 39) suggests that these antigens may all be part of the same site and that, therefore, one site may carry several H-2 determinants. However, this situation apparently is not a general rule, since, in the case of the H 2kamutation, an antigen (H-2.60) was lost, while others (H-2.1, 11,25, and 23) apparently were not affected. Similarly, the H-2*a mutation resulted in a loss of two antigens (H-2.40 and 49) and a change of another antigen (H-2.4), while other antigens coded for by the H - 2 d haplotype were apparently not affected at all. These two examples (H2ka and H-2du) argue that an H-2 molecule may carry more than one serologically detectable site. One can therefore envision the H-2 molecule as a complex structure carrying several antigenic sites, with individual sites carrying several antigenic determinants. The finding that several mutations (H-2fu,H-2da, H-2ba, and H-2bd) affected the serologically detectable antigens quantitatively can be used to argue that conformational features of the H-2 molecule play an important role in determining the antigenicity of this molecule (Klein, 1976b). In this way, a mutation occurring outside an antigenic site could affect this site indirectly if it were also to generate, in addition to a change in the primary structure, a change in the tertiary structure of the H-2 molecule. The existence of a complex H-2 antigen topology is also indicated by the fact that in none of the H-2 mutants have all the antigens controlled by a particular allele been changed to a degree that they would become unrecognizable by the corresponding monospecific antisera. Apparently, the serotype of a given H-2 molecule is the result of several amino acid substitutions at different positions along the polypeptide chain, so that the generation of a completely different serotype requires not one but many changes in the corresponding gene. This hypothesis explains why all attempts have failed to detect H - 2 mutations by serological methods. The methods would have detected only mutations that had the entire serotype changed, but such a change would require multiple mutational steps. Perhaps it would be more productive to use, for instance, the ka anti-k serum to look for the reoccurrence of the H - 2 k a mutation. In such an experiment, one would presumably be looking for an alteration caused by a single mutational step, but even then, the probability of detecting the ka + k reversion might not be very high. The observation that none of the H - 2 mutants has completely lost the private antigen could also be of significance. It suggests that probably even this most restricted serological specificity is generated by more than one amino acid substitution in the H-2 polypeptide chain.

136

JAN KLEIN

On a more general note, the mutant studies should serve as a warning for those who still take serological symbolism quite literally. In the mind of some immunologists, the assignment of a number to an antigen transforms this antigen into a rigid structure-a little pyramid or groove on the H-2 molecule-to which the corresponding antibody fits like a key into a lock. A complex antigen, such as H-2, is nothing of the sort. The relationship between the antigen and antibody is far more delicate than this grossly mechanistic visualization can express. Take, for example, the effect of the H-2dQmutation on the H-2.4 antigen. Has the antigen been changed by the mutation? The answer to this question apparently depends on what antibody one uses to look at the antigen. For some anti-H-2.4 antibodies, the antigen looks the same; for others, it has changed so drastically that they fail to react with it. And in between is an array of antibody populations for which the antigen has changed to some extent. And all these antibodies are present in a strictly monospecific antiserum to a single private antigen. Depending on the conditions during the production of anti-H-2.4 serum, some of the antibody populations might prevail over others, and antisera made in the same strain combination using more or less the same immunization protocol might, nevertheless, be different. To a serologist, all this is obvious; to an investigator who uses H-2 antisera merely as a tool, it often is not. There are still many immunologists who think of antigens in terms of little pyramids and of antibodies as sophisticated keys, and who view the H-2 chart as a tally in which numbers can be added up or subtracted according to rigid rules of arithmetic. Hopefully, H - 2 mutations will shatter once and for all this naive and distorted view of H-2 serology. H. GENETICCONTROL OF IMMUNE RESPONSE Since immune response to many antigens is controlled by the I region, and since no Z-region mutation has yet been detected, one would not expect any of the known H - 2 mutations to affect I r gene functions. Indeed, most mutant strains behave the same as the standard strains when tested for their response to a variety of synthetic polypeptides and serum proteins (C. Merryman, P. Maurer, and J. Klein, unpublished data). Occasionally, a few individuals of a mutant strain respond in a nonconforming fashion, but this exceptional behavior has generally been difficult to reproduce. However, there is at least one response that is clearly affected by an H - 2 mutation-the response to the Thy-1 antigen (Zaleski and Klein, 1977b).When AKR (H-2", Thy1") thymocytes are injected into CBA (H-2k, T h y - l b )mice, one finds a large number of plaque-forming cells with specificity for the Thy-1.1

H-2

137

MUTATIONS

TABLE XXIII

EFFECTOF H-2kfl MUTATION ON THE Strain CBAlLacStoY (CRA x C57BL/6)Fi (CBA x DBA/2)Fi

M523 (M523 X C57BL/6)F, (M523 x DBA/2)Fi

IMMUNE

RESPONSE TO THE THY-1.1ANTIGEN"

H - 2 haplotype

PFCfspI een

klk klb kld kulku ku l b kald

10,254 ? 1,144 2,167 ? 296 2,300 ? 291 sffi -t_ 87 178 5 37 125 ? 45

'I Plaque-forming-cell (PFC) response of indicated strains to a single intravenous injection of 4 x lo7AKR thymocytes. (Rased on Zaleski and Klein, 1977b.)

antigen in the spleens of these recipients. When M523 (H-2ka,Thy-1 *) mice are used as recipients instead of CBA, the number of plaqueforming cells is markedly reduced (Table XXIII): a mutation (H-2k") has changed a high-responder strain (CBA) into a low-responder strain (M523). The implications of this finding are discussed in detail in another review (Zaleski and Klein, 1977a). Here it suffices to say that the interpretation we favor is based on two assumptions: first, that an antigen (Thy-1) is recognized by T cells in conjunction with H-2 antigens of the immunizing cell; and second, that a mouse of a given haplotype has a high frequency of T cells with receptors for antigens presented in the context of its own H-2 molecules. XI. W h a t Have H-2 Mutations Contributed to Immunology?

Although the first H - 2 mutation was described 12 years ago (Bailey and Kohn, 1965), only recently has the usefulness of mutations for immunological research been generally recognized. In the few years of extensive research, H - 2 mutations have made many important contributions to immunology. Some of these contributions are enumerated below. 1. The first, and most important, conclusion drawn from the mutant studies concerns H - 2 pleiotropism. The studies have clearly demonstrated that a single locus controls a variety of immunological phenomena. The H - 2 K and H - 2 D loci control antigens that can cause allograft rejection, CML, MLR, GVHR, and GVHD, that can lead to the production of nonspecific allogeneic supernatants, and that can be detected serologically. Among these functions, two deserve special emphasis. There has been a long debate (cf. Klein, 1975, for references) over whether serologically and histogenetically detectable antigens of

0 cu

TABLE XXIV SUMMARY O F H-2 MUTANT EFFECTSON IMMUNE F U N ~ O N S Production

H-2 mutation

ba bb bd bf bd b& bg3 bh bi bj bk bm bn bo M513 da db dc dd fa

fb ka

+, positive

Graft rejection +a

+ + + + + + + + + + +

+ + + +

+ + + + + +

of allogeneic CML

MLR

GVHR

GVHD

supernatant

Immune response

+ +

NT NT NT NT NT NT NT NT NT NT NT NT NT NT NT NT NT NT NT NT NT

-? NT -? NT NT NT NT NT NT NT NT NT NT NT NT -? NT NT NT -? NT

+

+

+

NT

NT

NT

NT

NT NT NT NT NT NT NT

NT NT NT NT NT NT NT NT NT NT NT

NT NT NT NT NT NT NT NT NT NT NT NT

+ + + + + +

NT NT NT NT NT NT NT

+ +

NT NT

+ + +

+ + + + + +

+ +

NT NT

+ + +

+ +

+ +

NT NT

+ + +

+

NT NT NT NT NT

+

+

reaction between standard and mutant strain; -, negative reaction; NT, not tested.

+

Serological difference

+ + +

NT -? -? -? NT NT NT NT NT NT NT

+ +

NT NT

+ -t

t

R

r

E Z

H-2

MUTATIONS

139

the K and D regions are controlled by the same or different loci. The mutant studies have established that they are controlled by the same loci (H-2K or H - 2 D ) . There has also been a long controversy over whether the H - 2 K and H - 2 D locus products can stimulate MLR. The mutants have, once and for all, established that they can. 2. The mutant studies have cast doubt on die simplistic model of K(CD)-I region cooperation in the execution of CML (Bach et al., 1977). They prove that strong CML reaction can occur against H-2K and H-2D antigens in the absence of a proliferative response to Z-region antigens. The cell cooperation model, therefore, must be revised to take this finding into account. 3 . The mutant studies have, for the first time, revealed that histogenetically detectable antigens (for example, those detected by the CML or in vivo allograft reactions) are as complex as antigens detected by serological methods. 4. The studies have suggested that the complexity at the target antigen level has an equivalent at the T-cell receptor level. Heterogeneity exists among T-cell receptors for individual determinants of a single antigenic molecule. 5. The mutants have provided a tool for topological mapping of serologically detectable determinants on the H-2K and H-2D molecules. They have confirmed the presence of multiple determinants, private and public, on a single molecule. 6. The mutant studies have suggested, though not proved, that T and B cells recognize different regions of the H-2 molecule. 7. Biochemical analysis of H - 2 mutations has revealed that a small change in the H-2 molecule, perhaps a single amino acid substitution, has drastic effects on the phenotype: the change can result in the death of a cell, or even of an animal. 8. The observation of a high mutation rate has helped to understand how H - 2 polymorphism is generated. 9. The fact that H - 2 K ( H - 2 D ) and Z-region loci are involved in similar functions (e.g., MLR) has suggested the existence of functional relatedness among the different H - 2 regions, and helped to explain why the complex has remained a unit. XII. Perspectives

A few years of research by a handful of investigators, of course could not have exhausted all that H - 2 mutations can offer to immunology. Much still remains to be done. Undoubtedly, more extensive use of H - 2 mutants will further advance our understanding of the relation-

140

JAN KLEIN

ship between genes and immune functions and produce new insights into the H-2 complex, whose role in the cell physiology remains a great puzzle. Biochemical analysis, in particular, should prove to be rewarding. When, for instance, the nature of the difference among the H-2b8mutations becomes known, we will learn much about the mutational process itself. Biochemistry will provide answers to such questions as: Are there hypermutable regions in the H-2 genes? Does the antigenicity of the H-2 molecule reside in a restricted area? Are there different areas in the molecule for different functions? The functional approach also has much to offer. If, for example, a mutation in the I region were to be found, many of the questions with which immunologists have been struggling for years would probably b e answered. Among these, the one question that wouid get a top priority would concern the relationship between the Ia and Ir genes. To make mutant research more efficient, it might be necessary to develop new techniques to speed up the selection and characterization ofH-2 mutants. If, for instance, a way could be found of introducing in vitro selected mutant cells into a developing embryo, the whole area of mutant research would become wide open to approaches rivaling those currently available only to microbial geneticists. The current situation in immunology very much resembles that shown in Fig. 14. It is my conviction that H-2 mutations will even-

FIG.14. On the situation in immunological sciences according to Lucas Cranach the Elder.

H-2

MUTATIONS

141

tually become one of the important tools for bringing order into the Age of Chaos. ACKNOWLEDGMENTS I thank Ms. Jeanne Lively for secretarial help and Dr. Henry I. Kohn for information concerning his mutants. T h e experimental work cited in this review was supported by Grants Nos. CA 17225, A1 12589, A1 11650, and A1 11879.

REFERENCES Anderer, F. A., Wittmann-Liebolt, B., and Wittmann, H. G. (1965).Z. Naturforsch., Teil B 20,1203. Apt, A. S., Blandova, Z., Dishkant, I., Shumova, T., Vedernikov, A. A., and Egorov, I. K. (1975). Zinmunogenetics 1,444. Arrzt, K., Hamburger, L., and Flaherty, L. (1977).Zmmunogenetics 5,477. Auerbach, C. (1976). “Mutation Research. Problems, Results and Perspectives.” Chapman & Hall, London. Bach, F. H., Widmer, M. B., Segall, M., Bach, M. L., and Klein, J. (1972).Science 176, 1024. Bach, F. H., Grillot-Courvalin, C., Kuperman, 0.J., Sollinger, H. W., Hayes, C., Sondel, P. M., Alter, B. J., and Bach, M. L. (1977). Zmmunol. Reu. 35, 76. Bailey, D. W. (1966). Transplantation 4,482. Bailey, D. W. (1968).In “Advances in Transplantation” (J. Dausset, J. Hamburger, and G. Mathe, eds.), p. 317. Munksgaard, Copenhagen. Bailey, D. W., and Cherry, M. (1975).“Forty-sixth Annual Report,” p. 73. Jackson Laboratory, Bar Harbor, Maine. Bailey, D. W., and Kohn, H. I. (1965).Genet. Res. 6, 330. Bailey, D. W., and Usania, B. (1960).Transplant. Bull. 7, 424. Bailey, D. W., Snell, G. D., and Cherry, M. (1971).In “Immunogenetics of the H-2 System” (A. Lengerova and M. VojtiSkovi, eds.), p. 155.Karger, Basel. Batchelor, A., Phillips, R., and Searle, A. (1969).Br. J . Radiol. 42,448. Bennett, D. (1975). Cell 6,441. Benzer, S. (1961). Proc. Natl. Accid. Sci. U . S. A . 47,403. Berke, G., and Amos, D. B. (1973).Nature (London),New B i d . 242, 237. Bevan, M. J. (1975).J . Exp. Med. 142, 1349. Blanden, R. V., Dunlop, M. B. C., Doherty, P. C., Kohn, H. I., and McKenzie, I. F. C. (1976). Ztnmunogenetics 3, 541. Blanden, R. V., McKenzie, I.F.C., Kees, U., Melvold, R. W., and Kohn, H. I. (1977).J. Exp. Med. 146,869. Blandova, Z. K., Shumova, T. F., Kryshkina, V. P., and Egorov, I. K. (1973).In “Biology ofthe Laboratory Animals” (V. A. Dushkin, ed.), Vol. 2, p. 53. Acad. Med. Sci. USSR, Moscow. Blandova, Z., Mnatsakanyan, Y. A,, and Egorov, I. K. (1975). Zmmuriogenetics 2,291. Bodmer, W. F. (1972). Nature (London) 237, 139. Bodmer, W. F. (1973).Transplant. Proc. 5, 1471. Bonner, J. J., and Slavkin, H. C. (1975). Zrninuirogeiietics 2,213. Boubelik, M., Lengerova, A,, Bailey, D. W., and MatouSek, V. (1975).Deo. Biol. 47,206. Boyse, E. A. (1959). Zmmunology 2, 170. Brondz, B. D. (1964). Folici B i d . (Prague) 10, 164. Brondz, B. D. (1972). Transplant, Reu. 10, 112. Brown, J. L., and Nathenson, S. (1977)./. Immunol. 118, 98.

142

JAN KLEIN

Brown, J. L., Naire, R., and Nathenson, S. G. (1978).J . Zmmunol. 120, 726. Brown, J. L., Kato, K., Silver, J., and Nathenson, S. G. (1974). Biochemistry 13,3174. Calhoun, J. B. (1963). U.S . , Public Health Sew., Publ. 1008. Chauvenet, P. H., and Amos, D. B. (1975). Cell. Zmmunol. 17,477. Cherry, M., and Eicher, E. M. (1976). Mouse News Lett. 54,41. Cudkowicz, G., and Stimpfling, J. H. (1964).Immunology 7,291. Czarnomska, A,, and DBmant, P. (1975).Folia Biol. (Prague)21,419. David, C. S., ShrefHer, D. C., and Frelinger, J. A. (1973). Proc. Natl. Acad. Sci. U.S . A . 70,2509. David, C. S., Neely, B. C., and Cullen, S. E. (1977).Zmmunogenetics 5, 143. Davies, D. A. L., Baugh, V. S. G., Buckham, S., and Manstone, A. J. (1974). Eur. J . Cancer 10,781. DeGiorgi, L., Biasi, G., and Festenstein, H. (1976). Folia B i d . (Prague)2,437. Demant, P., Cherry, M., and Snell, G. D. (1971a). Transplantation 11,238. Demant, P., Snell, G. D., and Cherry, M. (1971b). Transplantation 11,242. Demant, P., capkova, J., Hinzovi, E., and VoriEova, B. (1973). Proc. Natl. Acad. S c i . U.S. A. 70,863. Demant, P., Snell, G. D., Hess, M., Lemonnier, F., Neauport-Sautes, C., and Kourilsky, F. (1975).J. Zmmunogenet. 2,263. Dishkant, I. P., Vedernikov, A. A., and Egorov, I. K. (1973). Genetika (Moskva) 9,82. Dizik, M., and Elliott, R. (1977). Mouse News Lett. 56, 58. Donner, M., and Wioland, M. (1976). Folia Biol. (Prague)22,51. Donner, M., Vaillier, D., and Burg, C. (1973).Eur. J . Zmmunol. 3,424. Dutton, R. W. (1965).J . E x p . Med. 122, 759. Dux, A., Corduwener, D., and Muhlbock, 0. (1971). I n “Immunogenetics of the H-2 System” (A. Lengerovi and M. Vojtiikova, eds.), p. 163. Karger, Basel. Egorov, I. K. (1967). Genetika (Moskua)9, 136. Egorov, I. K. (1974). Zmmunogenetics 1, 97. Egorov, I. K., and Blandova, Z. K. (1968).Genetika (Moskoa) 12,63. Egorov, I. K., and Blandova, Z. K. (1972). Genet. Res. 19,33. Egorov, I. K., Mnatsakanyan, Y.A., and Pospelov, L. E. (1977).Zmmunogenetics 5,65. Ehling, U. H. (1966). Genetics 54, 1381. Ehling, U. H., Cumming, R. B., and Malling, H. V. (1968). Mutat. Res. 5,417. Elkins, W. L. (1971).Prog. Allergy 15,78. Erb, P., and Feldmann, M. (1975).J . Exp. Med. 142,460. Eriksson, K., Halkka, O., Kokki, J., and Saura, A. (1976). Heredity 37,341. Ferreira, A., and Nussenzweig, V. (1976).J.Zmmunol. 117,771. Festenstein, H., Abbasi, K., and Demant, P. (1974).J . Zmmunogenet. 1,47. Festing, M. (1973). Genet. Res. 21, 121. Flaherty, L. (1975). lmmunogenetics 2, 325. Flaherty, L. (1976). Zmmunogenetics 3,533. Flaherty, L., and Wachtel, S. S. (1975).Zmmunogenetics 2,81. Forejt, J. (1975).Nature (London)26, 143. Forman, J. (1975).J. Exp. Med. 142,403. Forman, J., and Klein, J. (1975a). Zmmunogenetics 1,469. Forman, J., and Klein, J. (1975b).J.Zmmunol. 115, 711. Forman, J., and Klein, J. (1977). Zmmunogenetics 4, 183. Garrido, F., Schirrmacher, V., and Festenstein, H. (1976).Nature (London)259, 228. Geib, R., and Klein, J. (1978).J.Zmmunol. (submitted for publication). Geib, R., Chiang, C. L., and Klein, J. (1977).J . Zmmunol. 120,340.

H-2

MUTATIONS

143

Gelfand, M. C., Sachs, D. H., Liebeman, R., and Paul, W. E. (1974).J. E x p . Med. 139, 1142. Gleichmann, H., Gleichrnann, E., Andre-Schwartz, J., and Schwartz, R. S . (1972).]. E x p . Med. 135, 516. Goldrnan, A. S., Katsumata, M., Yaffe, S. J., and Gasser, D. L. (1977).Nature (London) 265, 643. Gornard, E., Duprez, V., Henin, Y., and Levy, J. P. (1976).Nature (London) 260, 707. Gordon, R. D., Sirnpson, E., and Samelson, L. E. (1975).J. E x p . Med. 142, 1108. Corer, P. A. (1936).Br. J. E x p . Pathol. 17,42. Corer, P. A. (1937).J. Pathol. Bacteriol. 44,691. Gotze, D., ed. (1977). “The Major Histocompatibility System in Man and Animals.” Springer-Verlag, Berlin and New York. Green, M. M. (1976).In “The Genetics and Biology of Drosophila” (M. Ashburner and E. Novitski, eds.), Vol. 16, p. 929. Academic Press, New York. Gregorova, S., Ivanyi, P., Sirnonova, D., and Mickova, M. (1977). Immunogenetics 4, 301. Hampl, R., Ivanyi, P., and Starka, L. (1971).Steroidologia 2, 113. Hansen, T. H., Cullen, S. E., and Sachs, D. H. (1977).J.E x p . Med. 145,438. Hauptfeld, V., and Klein, J. (1975).J. E x p . Med. 142,288. Hauptfeld, V., Klein, D., and Klein, J . (1973). Science 181, 167. Hellstrom, K. E., Hellstrorn, I., and Haughton, G. (1964). Nature (London) 204, 661. Hirst, J., and Dutton, R. W. (1970).Cell. Immunol. 1, 190. Hoffman, H. A,, and Grieshaber, C. K. (1977). Mouse News Lett. 56,51. Hood, L., and Silver, J. (1977).Contemp. Top. M o l . Immunol. 5,35. Invernizzi, G., and Parmiani, G. (1975).Nature (London) 254,713. Invernizzi, G., Carbone, G., Meschini, A., and Parmiani, G. (1977).J. Immunogenet. 4, 97. Ivanyi, P., Hampl, R., Starka, L., and Mickova, M. (1972).Nature (London),New Biol. 238,280. Kapp, J . A., Pierce, C. W., and Benacerraf, B. (1974).J. E x p . Med. 142, 50. Kettmnn, J., Klein, J., and Forrnan, J . (1977).1. Immunol. 119, 1189. Kindred, B., and Shreffler, D. C. (1972).J. Immunol. 109,940. Klein, G. (1969).Fed. Proc., Fed. Am. Soc. E x p . B i d . 28, 1739. Klein, G., and Klein, E. (1975).Int. J. Cancer 15,879. Klein, J. (1975). “Biology of the Mouse Histocornpatibility-2 Complex.” Springer-Verlag, Berlin and New York. Klein, J. (1976a).Transplant. Proc. 8,335. Klein, J . (1976b). Contemp. Top. Immunobiol. 5, 297. Klein, J. (1977). Transplant. Proc. 9,847. Klein, J., and Chiang, C. L. (1976).J. Immunol. 117,736. Klein, J., and Egorov, I. K. (1973).J. Immunol. 111,976. Klein, J., and Forman, J. (1976). In “Leukocyte Membrane Determinants Regulating Immune Reactivity” (V. P. Eijsvoogel, D. Roose, and W. P. Zeijlemaker, eds.), p. 443. Academic Press, New York. Klein, J., and Shreffler, D. C. (1971). Transplant. Reo. 6, 3. Klein, J., and Zaleska-Rutczynska, Z. (1977).J. Zmmunol. 119, 1903. Klein, J., Widmer, M.‘B., Segall, M., and Bach, F. H. (1972). Cell. Immunol. 4,442. Klein, J., Hauptfeld, V., and Hauptfeld, M. (1974a). Prog. Immunol., Int. Congr. Immunol., 2nd, 1974 p. 197. Klein, J., Hauptfeld, M., and Hauptfeld, V. (197413).J. E x p . Med. 140, 1127.

144

JAN KLEIN

Klein, J., Forman, J., Hauptfeld, V., and Egorov, I. K. (1975).J.Immunol. 115, 716. Klein, J., Egorov, I. K., Mnatsakanyan, Y. A,, and Hauptfeld, V. (1976a). Scand. J . Immunol. 5,521. Klein, J., Chiang, C. L., Lofgren, J., and Steinmuhr, D. (1976b). Transplantation 22, 384. Klein, J., Hauptfeld, M., Geib, R., and Hammerlxrg, C. (1976~). Transplantation 22, 572. Kohn, H. I. (1973).Mutat. Res. 20,235. Kohn, H. 1. (1976).Nature (London) 263,766. Kohn, H. I., and Melvold, R. W. (1974).Mutat. Res. 24, 163. Kohn, H. I., and Melvold, R. W. (1976). Nature (London) 259,209. Kohn, H. I., Melvold, R. W., and Dunn, G. R. (1976). Mutat. Res. 37, 237. Kralova, J., and Demant, P. (1976).lmmunogenetics 3,583. Lemonnier, P., Neauport-Sautes, C., Kourilsky, F. M., and Demant, P. (1975).lmmunogenetics 2,517. Lilly, F., Boyse, E. A., and OId, L. J. (1964).Lancet 2, 1207. Luning, K. G. (1971). Mutat. Res. 11, 133. Luning, K. G., and Searle, A. G. (1971). Mutat. Res. 12,291. Lyon, M. F. (1977).Mouse News Lett. 56,37. Lyon, M. F., and Morris, T. (1966). Genet. Res. 7, 12. Lyon, M. F., Phillips, R. J. S., and Bailey, H. (1972). Mutat. Res. 15, 185. McClintock, B. (1950). Proc. Nrttt. Acad. Sci. U . S. A. 36,344. McDevitt, H. O., and Chinitz, A. (1969).Science 163, 1207. McDevitt, H. O., Deak, B. D., Shreffler, D. C., Klein, J., Stimpfling, J. H., and Snell, G. D. (1972).J.E r p . Med. 135, 1259. McKenzie, I. F. C., Morgan, G. M., Melvold, R. W., and Kohn, H. I. (1976). Immunogenetics 3, 241. McKenzie, I . F. C., Morgan, G. M., Melvold, R. W., and Kohn, H. I. (19774. Immunogenetics 4,333. McKenzie, I. F. C., Pang, T., and Blanden, R. V. (1977b). Imnwnol. Reu. 35, 181. Melief, C. J. M., Schwartz, R. S., Kohn, H. I., and Melvold, R. W. (1975).Immunogenetics 2,337. Melief, C. M. M., Schwartz, R. S., Kohn, H. I., Melvold, R. W., and Dux, A. (1976). In “Leucocyte Membrane Determinants Regulating Immune Reactivity” (V. P. Eijsvoogel, D. Roose, and W. P. Zeijlemaker, eds.), p. 453. Academic Press, New York. Melief, C. J. M., van der Meulen, M., and Postma, P. (1977).Immunogenetics 5,43. Melvold, R. W., and Kohn, H. I. (1975). Mutat. Res. 27,415. Melvold, R. W., and Kohn, H. 1. (1976). Immunogenetics 3, 185. Meo, T., David, C. S., Nabholz, M., Miggiano, V., and Shreffler, D. C. (1973). Transplant. Proc. 5, 1507. Meruelo, D., and Edidin, M. (1975). Proc. Natl. Acad. Sci. U . S. A . 72, 2644. Mickovi, M., and Ivanyi, P. (1973). Transplant. Proc. 5, 1421. Mickova, M., and Ivanyi, P. (1974).J . Hered. 65,369. Mickovi, M., and Ivinyi, P. (1975).Folia Biol. (Prague) 21,435. Miller, J. F. A. P., Vada, M. A., Whitelaw, A., and Gamble, J. (1976).Proc. Natl. Acad. Sci. U . S. A. 73,2486. Mnatsakanyan, Y. A., and Egorov, I. K. (1975). Genetika (Moskva) 11,30. Mnatsakanyan, Y.A., Pospelov, L. E., and Egorov, I. K. (1977).Genetikn (Moskva) 13,64. Mobraaten, L. E., and Bailey, D. W. (1973).Mouse News Lett. 48, 17. Moller, G., and Moller, E. (1965).Nature (London) 208,260.

H-2

MUTATIONS

145

Mukai, T., and Cockerham, C. C. (1977).Proc. Natl. Acad. Sci. U . S . A. 74,2514. Nabholz, M., Young, H., Meo, T., Miggiano, V., Rijnbeek, A., and Shreffler, D. C. (1975). lmniunogenetics 1,457. Nagy, Z., Elliott, B. E., Nabholz, M., Gammer, P. H., and Pernis, B. (197q.J.Erp. Med. 143, 648. Neauport-Sautes, C., and Dbmant, P. (1977).Mouse News Lett. 57,8. Old, L. J., Stockert, E., Boyse, E. A,, and Kim, J. H. (1963).J.Natl. Cancer Inst. 31,977. Oth, D., Sabolovic, D., and Garrec, Y. (1974). Folia Biol. (Prague)20,20. Oth, D., Berebbi, M., and Meyer, G. (1975).J . Natl. Cancer Znst. 55,903. Parmiani, G . , and Invernizzi, G . (1975). Znt. J. Cancer 16,756. Passmore, H. C., and Shreffler, D. C. (1970). Biochem. Genet. 4,351. Petrinyi, G., Kiessling, R., Povey, S., Klein, G., Herzenberg, L., and Wigzell, H. (1976). Zmmunogenetics 3, 15. Phillips, S. M., Carpenter, C. B., and Strom, T. B. (1973). Transplant. Proc. 5, 1669. Pla, M., Zakany, J., and Fachet, J. (1976).Folia Biol. (Prague) 22,49. Plate, J. M. D. (1976).Nature (London) 260,329. Race, R. R., and Sanger, R. (1968). “Blood Groups in Man,” 5th ed. Davis, Philadelphia, Pennsylvania. Rich, S. S., and Rich, R. R. (1976).J.E x p . Med. 143,672. Rohrborn, G . (1970).In “Chemical Mutagenesis in Mammals and Man” (F. Vogel and G . Rohrborn, eds.), p. 148. Springer-Verlag, Berlin and New York. Russell, W. L. (1951). Cold Spring Harbor Symp. Quant. Biol. 16,327. Rychlikova, M., and Ivanyi, P. (1974).Folia Biol. (Prague) 20,68. Rychlikova, M., and Ivanyi, P. (1969).Folia Biol. (Prague) 15, 126. Rychlikova, M., Demant, P., and Ivanyi, P. (1970).Folia Biol. (Prague) 16, 218. Rychlikova, M., Demant, P., and Egorov, I. K. (1972).Folia Biol. (Prague) 18,360. Sabolovic, D., Oth, D., and Burg, C. (1971).Immunology 20, 341. Schimpl, A,, and Wecker, E. (1972). Nature (London),New Biol. 237, 15. Schlager, G., and Dickie, M. M. (1971). Mutat. Res. 11,89. Schwartz, R. H., and Paul, W. E. (1976).J . E x p . Med. 143,529. Searle, A. G. (1974).Ado. Radial. Biol. 4, 131. Selander, R. K., and Yang, S. Y. (1969). Genetics 63, 653. Selby, P. B., and Selby, P. R. (1977). Mutat. Res. 43, 357. Serov, 0. L. (1972). Dokl. Akad. Nauk SSSR 204, 978. Shearer, G. M., Rehn, T. G., and Garbino, C. A. (1975).J. E x p . Med. 141, 1348. Shonnard, J. W., Cramer, D. V., Poloskey, P. E., Kunz, H. W., and Gill, T. J., 111. (1976). lmmunogenetics 3, 193. Shreffler, D. C., and Owen, R. D. (1963).Genetics 48, 9. Shumova, T. E., Kryshkina, V. P., and Egorov, I. K. (1972).Cenetika (Moskoa) 8, 171. Simonsen, M., and Jensen, E. (1959).In “Biological Problems of Grafting” ( F . Albert and G. Lejeune-Ledant, eds.), p. 214. Thomas, Springfield, Illinois. Snell, G. D. (1958).J.Natl. Cancer Inst. 21,843. Snell, G. D. (1974). “Catalog of Mouse Alloantisera,” 1974 Suppl. Natl. Inst. Health, Bethesda, Maryland. Snell, G . D., Demant, P., and Cherry, M. (1971a). Transplantation 11, 210. Snell, G. D., Graff, R. J., and Cherry, M. (1971b). Transplantation 11,525. Snell, G . D., Cherry, M., and Demant, P. (1974).Folia Biol. (Prague)20, 145. Snell, G . D., Daussett, J., and Nathenson, S. G. (1976).“Histocompatibility.” Academic Press, New York. Staines, N. A., Guy, K., and Davies, D. A. L. (1974).Transplantation 18, 192.

146

JAN KLEIN

Stanton, T. H., and Boyse, E. A. (1976).Immunogenetics 3,525. Stimpfling, J . H., and Pizarro, 0. (1961). Transplant. Bull. 28, 102. Streilein, J. W., and Klein, J. (1977).J.Zmmunol. 119,2147. Strosberg, A. D. (1977). Immunogenetics 4,499. Suslov, A. P., Egorova, S. G., and Brondz, B. D. (1976). Bull. E x p . Biol. Med. 80,86. Svejgaard, A., Hauge, M., Jersild, C., Platz, P., Ryder, L. P., Staub-Nielson, L., and Thomsen, M. (1975). “The HLA System.” Karger, Basel. Tada, T. (1974).In Immunological Tolerance: Mechanisms and Potential Therapeutic Applications” (D. H. Katz and B. Benacerraf, eds.), p. 471. Academic Press, New York. Taussig, M. J., and Munro, A. J. (1974).Nature (London)251,63. Van deBerg, J., and Klein, J. (1978).J. E x p . Zool. 203, 319. Vedernikov, A. A., and Egorov, I. K. (1973).Genetika (Moskoa) 9,60. Viklick9, V., PolGkovP, M., VortiEova, B., and Matousek, V. (1976). Folia Biol. (Prague) 22,402. Vogel, F. (1969).Humangenetik 8, 1. Widmer, M. B., Alter, B. J., Bach, F. H., Bach, M. L., and Bailey, D. W. (1973).Nature (London),New Biol. 242,239. Womack, J. E., and Eicher, E. M. (1976). Mouse News Lett. 54,41. Yamazaki, K., Boyse, E. A., Mike, V., Thaler, H. T., Mathieson, B. J., Abbott, J., Boyse, J., Zayas, Z. A., and Thomas, L. (1976).J.E x p . Med. 144, 1324. Yefenof, E., and Klein, G. (1974). E x p . Cell Res. 88,217. Yunis, E. J., and Amos, D. B. (1971).Proc. Natl. Acad. Sci. U.S. A. 68,3031. Zaleski, M., and Klein, J. (1977a). Zmrnunol. Rev. 38, 120. Zaleski, M., and Klein, J. (197713).J. Exp. Med. 145, 1602. Zinkernagel, R. M. (1974). Nature (London) 251,230. Zinkernagel, R. M. (1976).J. E x p . Med. 143,437. Zinkernagel, R. M., and Doherty, P. C. (1974). Nature (London) 251,547. Zinkernagel, R. M., and Doherty, P. C. (1975).J.E r p . Med. 141, 1427. Zinkemagel, R. M., and Klein, J. (1977). Zmmunogenetics 4,581.

ADVANCES IN IMMUNOLOGY, VOL. 26

The Protein Products of the Murine 17th Chromosome: Genetics a n d Structure ELLEN S. VITETTA A N D J. DONALD CAPRA Department of Microbiology, University of Texas Southwestern Med ical School, Dollar, Texas

I. Introduction .........................................................................................

148 148 149 150 C. Distribution 150 151 154 154 B. Structure. 156 157 IV. The K and D ...................... ................................................................. A. Genetics ......................................................................................... 158 .................. 158 C. Distribution ................................................................................. 159 .................. 160 D. Structure .................. 167 167 168 169 170 171 177 .................... 177 .................. 178 178 B. Function ........................................................................................................ 179 C. Distribution ................................................................................................... 179 D. Structure of the Ss Protein 181 ................ 181 F. Conclusions 181 .................. 182 B. Qa-1 ...................................... .................. 182 182 C. H-2.6, 27, 28, 29 ........................................................................... 182 VIII. TLO ...................................................................................................... ...... 182 A. Definition and Genetics .......... ......... ............................. ,....., B. Tissue Distribution ...................................................................................... 183 183 C. Antigenic Modulation of TL ....................................................................... 184 D. Biochemistry of TL Antigens ..................................................................... .................................

147

Copyrightm 1978 by Academic Press, Inc All rights of reproduction In any form reserved.

ISBN 0-12-0224267

148

ELLEN S. VITETTA AND J. DONALD CAl'RA

E. Conclusions .................................................................................................. IX. Perspectives ....................................................................................................... References

.......................................................................................................... I.

185 186 188

Introduction

Over the past two decades, the genes and gene products specified by chromosome 17 of the mouse have become increasingly important to immunologists, developmental biologists, and geneticists. Thus, many of the gene complexes that map between the centromere and TLa (Fig. 1) appear to play a fundamental role in the control of embryogenesis, immune responsiveness, and differentiation. With the possible exception of the Ss protein, molecules encoded by these genes are expressed on the surface of cells. At least four of these molecules (H-2D7H-2K7TL, and Qa-2) are associated in the cell membrane with µglobulin, and five of them (T/t, H-2K7 H-2D7 Qa-2, and TL) have major subunits of molecular weight ( M , ) 44,000 (Fig. 2). All the molecules that have been studied biochemically are glycoproteins. Despite rapid advances in the elucidation of the genetic map of chromosome 17 and the biochemistry of its products, relatively little is known about the function of most of these molecules. Nevertheless, there are numerous studies suggesting that some of these products (Tlt, H-2K7 Ia, H-2D7and TL) are involved in cell-cell interactions leading to normal differentiation or immune function. The S region, which appears to control the production of one or more complement components, it also intimately involved with the immune response. Several excellent and comprehensive reviews of individual gene complexes of chromosome 17 have recently appeared. The purpose of this review is to emphasize common features of these genes and their products and their interrelationships and evolutionary origins. This will be approached by analyzing the current state of knowledge of the biochemistry, including the primary structure, of these molecules. I I . &Microglobulin

&-Microglobulin was initially isolated by Berggard and Bearn

(1968) from the urine of patients with Wilson's disease and chronic 0

1 1 1 1 1

A B J E C

T/t

K

7S

G

D

I

Qo-i Qo-2 TLa

FIG. 1. A portion of the murine 17th chromosome (not drawn to scale).

T H E MUFUNE

1

7

CHROMOSOME ~ ~

149

2 10

20

30 40 50 FRACTIONS (TLa)

60

70

FIG.2. Molecules encoded by the H-2, TLu, F9, and Qa-2 regions of chromosome 17. Iodinated cells were lysed, and the lysates were treated with the relevant alloantiserum. In the case of spleen and lymph node cells, Ig was depleted from the lysate prior to treatment with alloantiserum. The complexes were precipitated with goat antimouse Ig or were absorbed to Stuph~Jococcusuureus (Cowan I strain). Precipitates or bacterial pellets were washed, the radioactivity was eluted with sodium dodecyl sulfate (SDS), reduced, and the eluates electrophoresed along with 3H-labeled p and L chains on SDS gels. Gels were counted, and the markers were aligned. The major subunit of each molecule has a molecular weight of 44,000. H-2 and F9 also have subunits ofM, 22,000. All four molecules have small subunits ofM, 12,000, which, in at least three cases (H-2, TL, and Qa-2), have been immunologically identified as p,-microglobulin. The small subunit of F9 is difficult to obtain with some batches of sera and does not precipitate with anti-j3,-microglobulin serum. The sera used to precipitate these molecules have been described by Vitetta et al. (1976a,b) and Flaherty (1976).

cadmium poisoning. In man, there is good evidence that the gene coding for p,-microglobulin is carried on chromosome 15, not on chromosome 6 along with H L A , (Goodfellow et al., 1975; Smith et al., 1976). Thus, while there is no evidence that p,-microglobulin is a product of the murine 17th chromosome, it will be discussed here because of its association with at least four of the products specified by the genes on chromosome 17 (Vitetta et al., 1975a, 1976b; Peterson et al., 1975; Geib et al., 1976; Michaelson et al., 1977).

A. GENETICS Despite numerous attempts to find genetic markers, none has been detected in the &-microglobulin of any species. Indeed, all the struc-

150

ELLEN S. VITETTA AND J. DONALD CAPRA

tural data (see below) indicate homogeneity-even from many animals.

in material pooled

B. FUNCTION

The association of p,-microglobulin with several cell surface molecules must provide a clue to its function, but no comprehensive explanation for its role has been forthcoming. The most likely possibility is that it acts as a “light” chain to form a portion of the putative “receptor” site. Another possibility includes stabilization of the cell surface glycoproteins to render them less susceptible to proteolytic degradation (Bismuth et al., 1974). P,-Microglobulin has many properties in common with immunoglobulin domains (Painter et a1 ., 1974). For example, &-microglobulin can fix complement. Whether or not these functions are related to the physiological function of the µglobulin complex on the cell surface, remains to be determined. More likely, these functional properties relate to the common evolutionary origin of &-microglobulin and immunoglobulin. C. DISTFUBUTION While abundantly present in serum, µglobulin is also present on the plasma membranes of several types of cells. Bernier and Fanger (1972) reported that short-term lymphocyte cultures from normal human donors released Pz-microglobulin into the medium. They also demonstrated that the rate of &-microglobulin synthesis and secretion was increased severalfold after phytohemagglutinin stimulation. Hutteroth et al. (1973)confirmed and extended these findings to long-term lymphoid cell lines. Its synthesis seemed independent of immunoglobulin synthesis. Both B and T lymphocytes have membrane &-microglobulin, but p,-microglobulin is not associated just with lymphoid cells. As expected, because of its association with H-2, nonlymphoid cells, including nonneoplastic glial cells, fibroblasts, and various cell lines, contain membrane-associated &-microglobulin (Berggard and Bearn, 1968); µglobulin can also be found on sperm (Fellous et a t , 1976a). I n ontogenetic studies it has been shown that fetal liver, kidney, thymus, and testis are also capable of active de novo synthesis of &-microglobulin. On the cell surface, /&-microglobulin can be found in free form (Tanigaki et al. 1977), in association with histocompatibility antigens (Vitetta et al., 1975a; Silver & Hood, 1974; Ostberg et al., 1975), non H-2 antigens (Anundi et al., 1975; Michaelson et al.,

THE MUFUNE

1

7

CHROMOSOME ~ ~

151

1977; Ostberg et al., 1975; Vitetta et al., 1975a), and possibly tumor antigens (Tada et al., 1978). D. STRUCTURE

1 . Isolation As mentioned previously, @,-microglobulin is present in serum, and large amounts are found in the urine when the renal tubules are damaged. In man, urine is usually obtained from patients with chronic cadmium poisoning. Deliberate renal tubular damage has been used experimentally in many species for the preparation of&-microglobulin (Berggard, 1974; Gordon and Kindt, 1976). Recently, p,-microglobulin has been isolated from the serum of chickens (Winkler and Sanders,

1977). Mouse p,-microglobulin has been extracted from liver cell membranes b y sodium thiocyanate and subjected to standard biochemical purification techniques (Natori et al., 1974, 1975; Appella et al., 1976a).Others have done structural work on mouse µglobulin by pulse labeling techniques identical to those used and described below for H-2 (Silver and Hood, 1974).

2. Molecular Size and Shape In their original description of p,-microglobulin, Berggard and Bearn (1968) reported a molecular weight of 11,600 and a sedimentation coefficient of 1.65 S as determined b y velocity ultracentrifugation. Their analysis revealed that &-microglobulin consisted of 100 amino acid residues including two half-cysteines involved in disulfide linkage. Human &-microglobulin has a Stoke’s radius of 16 A and a low frictional ratio, indicating a spherical shape (Karlsson, 1974). Circular dichroism and optical rotatory dispersion analyses suggest the presence of small amounts of /3 structure. These are common properties of immunoglobulin domains, particularly the absence of significant a-helix.

3 . Sequence Presently, the complete sequence of only human p,-microglobulin is available, but partial sequences are known for dog, rabbit, rat, mouse, guinea pig, and chicken &-microglobulin (Smithies and Poulik, 1972; Poulik, 1976; Poulik et al., 1977; Cunningham and Berggard, 1974, 1975; Cunningham et al., 1976; Appellaet al., 1976a,b).As shown in Fig. 3 , there is a remarkable structural preservation, which

5

m IM PRO LYS I EGNI

~ l y w

ILE GIN

G . PIG

VAL LEU HIS

ALA-

VK

1')

15

m

SER ARG HIS PRO MA

20

ru

*p1

w

LVS

SER AS RL LEU

25

30

35

ASN CIS TYR VAL SER GLY RL HIS PRO YR ASP ILE

GLX V K

FIG. 3. Amino acid sequence comparison of the µglobulins from different mammals. The amino-terminal one-third of each molecule is shown. A11 are compared to the human sequence, which is the only completed one. A straight line indicates sequence identity with the human sequence. Original references are noted in the text. The guinea pig sequence is complete only to position 18 and is from Cunningham et al. (1976).

1

THE MURINE

7

CHROMOSOME ~ ~

153

explains the widespread immunological cross-reactivity. Virtually all substitutions between these proteins are conservative replacements. Most of the differences are in the N-terminal seven residues. The remarkable homology of µglobulin with immunoglobulin-constant regions was first pointed out by Smithies and Poulik (1972) and Peterson et ul. (1972). The complete sequence of human p,-microglobulin and the homology with immunoglobulin are shown in Fig. 4. Although this figure has been reproduced from the Peterson

P2-MICROUOBULIN EU CL

(RESIDUES 109-2141

EU C H l (RESIDUES 119-2201

EU CH2 IRESIDUES 234-3411 EU CH3 IRESIOUES 342-446)

~

-

~

-

TnR

THR L E U h€T

. -

-

20 ASXIGCY]LVS SER ASX P H C W A S N

L E U L Y S SER CLY TnR ALA SER VAL VAL

SERIGLYJGLVTnR ALA

A L A ~ G L V I L E SER ARC THR PRO GLU VAL THR

WT THR L Y S ASN GLN VAL S E R m T H R

W S E R CLX L E U SER PHE SER L Y S ASN

-

60 I T R P

SER GLN GLU SER VAL THR GLU CLN ASP SER L Y S ASP SER TWR SER GLY

-

V A L m T H R PHE PRO ALA VAL L E U GLN SER ALA L V S TnR L Y S PRO A R C GLU GLN GLN T Y R

NR L Y S THR THR PRO PRO VAL L E U ASP SER

- W G L Y

-

ASP SER

ASP GLY SER PHE PHE[LEUI

FIG.4. Comparison of the amino acid sequence of human µglobulin with the constant region of human K chains and the three “domains” of human y chains. Reproduced from Peterson et al. (1972) with permission of the Proceedings of the National Academy ofSciences, USA. Identical residues are enclosed in boxes. Numbering is for µglobulin.

154

ELLEN S. VITETTA AND J. DONALD CAPRA

et ul. (1972)paper in several reviews, it seems worthwhile to reproduce it again in order to point out several features of the homology that will also apply to most arguments of this type, which will follow later in this review. The overall homology between /3,-microglobulin and immunoglobulin is not particularly impressive-even the 28% homology with the CH, domain of IgG is only marginal at best at the statistical level, especially since several gaps have been introduced in both sequences to obtain a number as high as 28%. The convincing data are those showing the extraordinary identity of those residues, which are identical among immunoglobulin domains and also with &-microglobulin. Of 11 positions where all immunoglobulin-constant domains have an identical residue (positions 5,9,14,25,37,39,67,69,81,83, and 92 in Fig. 4), 10 are identical in p,-microglobulin (all those listed above except position 39,the Leu-Trp interchange). By itself this is an extraordinary degree of homology and convincingly demonstrates the close evolutionary relationship of these proteins. However, even with this, it was not until the H-2/~,-rnicroglobulin-HLA//3,-microglobulin relationship was uncovered and the immunoglobulin domain-like functions of µglobulin were reported that this “homology” became widely accepted as significant. Ill. The T/t Region

The Tlt region is 14 centimorgans to the left ofH-2 and is characterized by a class of semidominant T (Brachyury) mutations, which, when heterozygous, cause a short tail and, when homozygous, are lethal (Chesley, 1935;Bennett, 1975).A large number of different mutants have been detected, and some are viable. These lethal mutations exist as natural polymorphisms in mice and are maintained by a distortion of their rate of transmission through males (which can be as high as 90% instead of the expected 50%) (Dunn and Bennett, 1964). A. GENETICS Genes in the Tlt region are of critical importance in early stages of embryonic development. Thus, as indicated in Fig. 5, homozygotes for any one of eleven different mutant alleles die at characteristic stages during embryogenesis (Bennett, 1964; Spiegelman and Bennett, 1974).Morphological studies of affected homozygous embryos suggest failure of cell-cell interactions (Bennett, 1964;Bennett et ul., 1972;Spiegelman and Bennett, 1974).These results have been interpreted as indicating that the Tlt region encodes cell surface molecules that are necessary for cell interactions leading to normal embryonic differentiation.

THE MUFUNE

1

7

Extra-embryonic cell mass

+

Homozygote:

Oiesat day:

t"lt"

4

tole

t"'1t"'

6- 7

7-10

CHROMOSOME ~ ~

155

+

Y TIT,

Primitive streak

+

Residual efnbrvonic &I1 Illas

t9/t9

9-11

Ectoderm

{ +

Brain stem

ectoderm Forebrain

t"'W"'

9 . birth

tW2/tw2 birth

In an attempt to generate antibodies against such cell surface structures, adult mice were injected with syngeneic nullipotential primitive teratocarcinoma (F9) cells that lacked H-2 antigens. It was reasoned that these cells could contain T/t antigens, and that these molecules would be immunogenic in an adult mouse. The antisera produced by this procedure (anti-F9)were not cytotoxic for adult cells or for a wide variety of tumor cells (Artzt et nl., 1973).However, antiF9 did react with sperm and normal cleavage embryos (up to 8 cells) (Artzt et al., 1973). Moreover, using another teratocarcinoma cell line that could differentiate in vitro, it was demonstrated that, as cells differentiated, the expression of F9 antigen decreased (Nicolas et al., 1975). In contrast, cells initially H-2 negative became H-2 positive. These results suggest a reciprocity in the expression of H-2 and F9. In addition, on normal embryos, F9 disappeared by day 10 and H-2 appeared (Jacob, 1977), further suggesting that the expression of F9 and H-2 is reciprocal (Snell, 1968). The experiments described above support the notion that F9 is a product of the + allele of t" since t" homozygotes often die at the morula stage. This possibility was further tested b y taking advantage of the fact that the expression of T/t antigens on sperm is codominant. Thus, if F9 were specified by an allele of tI2,sperm from T/tI2 (or +/t'l) males should carry 50% as much F9 antigen as sperm from T/+ (or +/+) males. In such an experiment, b y absorbing with graded numbers of the two types of sperm, this prediction proved correct (Artzt et al., 1974). Experiments using the anti-F9 serum have also shown that there are extensive cross-reactions between anti-F9 serum and molecules on rat, rabbit, bull, and human sperm (Buc-Caron et al., 1974;

156

ELLEN S. VITETTA AND J. DONALD CAPRA

Jacob, 1977) These results demonstrate that some structural features of F9 are preserved among many mammals, suggesting a fundamental role for F9 antigens on spermatocytes.

B. STRUCTURE The structure of the F9 antigen(s) was first described b y Vitetta et al. (1975b). Using either enzymic radioiodination of sperm or teratocarcinoma cells, or internal IabeIing of teratocarcinoma cells followed by immunoprecipitation and gel electrophoresis of the precipitates, the major form of the antigen appeared to be a disulfide-bonded dimer of two M , 44,000 subunits bound to a M , 12,000 protein. However, monomers and larger polymers of the M,44,000 subunits, as well as an additional molecule of M , 22,000, were frequently observed. Coelectrophoresis of the M , 44,000, 22,000, and 12,000 subunits of F9 and H-2 (isolated in a similar manner from spleen) indicated a remarkable structural similarity between the two antigens (Vitetta et al., 1975b). Moreover, when the 44,000 subunits were coelectrophoresed on gels of several different concentrations of acrylamide, the similarity of the molecular weights was confirmed [Vitetta et al., 1977b (Fig. 6)l. Other studies have shown that mild treatment of F 9 cells with periodate results in a large reduction in fluorescent anti-F9 staining, suggesting that the antigenic determinants of F9 may reside in carbohydrate (Jacob, 1977). Recent studies suggest that the M , 12,000 subunit of F9, unlike that

P

6 L

7.5x

4 (81

2 (41 0

M

30

40

50 20 FRACTIONS

30

40

50

FIG.6. Coelectrophoresis of H-2 and F9 antigens on gels of four different polyacrylamide concentrations. From Vitetta et al. (1977a).

THE MUFUNE

1

7 CHROMOSOME ~ ~

157

TABLE I DEPLETEF9 (T/t) FROM LYSATES OF TERATOMACELLS"

FAILURE OF ANTI-&M

TO

Initial precipitation

(% depletion of F9)O

Teratoma

Spleen (% depletion of H-2)

Mouse anti-F9 Rabbit anti-P,M Mouse anti-H-2

77 (2)' 1.6 (4) -11 (3)

-22 (5) 73 (3) 95 (4)

Teratoma and spleen cell lysates were treated with rabhit or mouse serum and goat anti-mouse Ig or goat anti-rabbit Ig. Supernatants were then precipitated with syngeneic anti-F9 (teratoma cells) or anti-H-2 (splenocytes) and goat anti-mouse Ig. Detemiined by calculating the area under the 44,000 peak ohtained after sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Number of experiments is indicated in parentheses.

found associated with H-2, is not conventional &-microglobulin (Dubois et al., 1976;Vitettaet al., 1977a) (Table I). Thus, despite similarities in molecular weight and subunit structure, the smaller subunits of F9 and H-2 are not immunologically related. Recent studies in our laboratories (R. Cook, E. S. Vitetta, and J . W. Uhr, unpublished observations) demonstrated that (1)the M , 12,000 subunit can be absent from gels of anti-F9 immunoprecipitates depending upon the batch of anti-F9 sera used, and (2) the M , 44,000 molecules of F9 and H-2 do not cross-react, using rabbit sera raised against each of the two molecules (Vitetta et al., 1977b). I n addition, F9 can be biosynthetically labeled in suspensions of mouse testicular cells (R. Cook, E. S. Vitetta, J. W. Uhr, and J. D. Capra, unpublished observations). The F9 molecule(s) which are precipitated from lysates of testicular cells have molecular weights of 44,000 and 22,000. At the present time, therefore, the question of homology between H-2 and F9 is unresolved and awaits peptide map comparisons and the determination of the amino acid sequence of an F9 molecule. I n addition, the roles of the M , 22,000 and 12,000 subunits are obscure. IV. The K and D Regions

The K and D regions are located on opposite ends of the H-2 complex. They are known to code for antigens capable of stimulating alloantibody production (Corer, 1936) and graft rejection (Corer, 1937).

158

ELLEN S. VITETTA AND J. DONALD CAPRA

The H-2K and H-2D antigens are serologically complex (Amos et d., 1955) and genetically extremely polymorphic (Klein, 1973).They are found on almost all tissues.

A. GENETICS The H-2D and H-2K gene products exhibit extensive genetic polymorphism. The genetic control of the classical H-2 antigens is best explained by postulating two loci, H - 2 K in the K region, and H-2D in the D region. Klein (1975)has stressed that there is no compelling reason at present to postulate more than one locus in each of these two regions. At each locus there are several alleles. Among the standard inbred mouse strains, approximately 11alleles of both the K and D loci have been defined serologically. There is an even more extensive polymorphism when wild mice from different localities are examined. The antigenic reactivity of the H-2 glycoproteins is determined by their protein structure. Carbohydrate appears to play little if any role in the serological specificities (Nathenson and Cullen, 1974). Most K and D gene products possess at least one unique (private) specificity and multiple common (public) antigenic specificities (Shreffler and David, 1975; Klein, 1975). Shreffler (1971) has suggested that the cross-reactivity of public antigenic specificities among the K and D alleles indicates that they diverged from a common ancestral gene. Most of the available biochemical data support this view.

B. FUNCTION Despite extensive analysis, the function of these molecules has remained elusive. Perhaps the most promising clues come from a study of lymphocyte activation. When T lymphocytes interact with a variety of different substances, they are “activated.” Myriads of materials will activate T cells by binding to cell surface molecules. The H-2 glycoproteins belong to this group of cell surface molecules and are potent activators of T lymphocytes. In fact, it has been shown that the products of the H-2K and H-2D loci alone not only can activate lymphocytes in the mixed lymphocyte reaction, but also stimulate killer T cell differentiation (reviewed in Shreffler and David, 1975). However, these data do not provide even a hint as to how H-2 functions. It has been suggested, however, that these molecules function in the discrimination of altered self-antigens, e.g., in distinguishing normal cells from virally infected or transformed cells (Doherty et al., 1976; or chemically modified cells Shearer et al., 1975; Forman,

1975).

THE MURINE

1

7

CHROMOSOME ~ ~

159

C. DISTRIBUTION There are two general methods for determining the presence of H-2 antigens-transplantation and serologic. The former is more sensitive, but the latter is more quantitative. Figure 7 shows the quantity of H-2 antigens in various tissues. Spleen has been arbitrarily set at 1.0. Every tissue in the body has a detectable amount of H-2, but in some it is extremely low (e.g., red cells, skeletal muscle). Of the tissues shown, only spleen, lymph node, and liver have been routinely used to isolate H-2 antigens for structural analyses. Cell lines have also been useful sources of H-2 glycoproteins, and their H-2 phenotypes have been shown to have remarkable stability. Several of the biochemical studies mentioned below were performed on molecules isolated from these cell lines. Finally, transplantable tumors have H-2 antigens that, in general, have been stable. Since these can be passaged in inbred mouse lines, they too have been rich sources of H-2 glycoproteins for biochemical studies (Thomas et at., 1976).

spleen lymph node llM1

thymus

lung adrenal gut

kidney YllYIry gl8nd

erythrocytes he811 brain testlS

skeletal muscle

h 0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

H-2activity

FIG.7 . Quantity of H-2 antigens in various tissues. Spleen has been arbitrarily set at 1.0. Taken from Klein (1975) with permission from Springer-Verlag.

160

ELLEN S. VITETTA AND J. DONALD CAPRA

D. STRUCTURE 1. Molecular Weight, Subunits, Chemical Composition

There is no general agreement as to the exact nature of the H-2 glycoprotein complex in the cell membrane, especially as to the arrangement of the constituent chains. There are at least two chains-a heavy chain of approximately M, 44,000 and a light chain of M, 12,000. The light chain, which is devoid of carbohydrate and has been shown conclusively to be µglobulin (Silver and Hood, 1974; Ostberg et al., 1975; Vitettaet al., 1976b),is noncovalently associated with H-2 in the plasma membrane. Antisera directed to H-2 specificities coprecipitate &-microglobulin (Vitetta et al., 1976b). The heavy chain, which contains the H-2 specificities, contains carbohydrate. The smallest glycopeptide isolated from H-2 has a molecular weight of 3500 (Muramatsu et al., 1973). It has been convincingly demonstrated that the carbohydrate moiety is not involved in the serologically determined H-2 specificity (Nathenson and Cullen, 1974). By amino acid sequence analysis, it has been shown that the H-2 molecule is inserted into the cell membrane via its C terminus (Ewenstein et al., 1976; Henning et al., 1976)and that the whole molecule is soluble only in detergent (Nathenson and Cullen, 1974). However, when treated with papain, a C-terminal M, 5000-10,000 piece is removed from the H-2 molecule, which renders the remainder soluble. The soluble portion is associated with µglobulin. There is controversy as to the stoichiometry of the H-2/jlz-microglobulin complex on the membrane. There is no firm evidence that the ratio of H-2/~,-microglobulinis 1: 1, but this is generally assumed. Additionally, it is difficult to be dogmatic about the covalent and/or noncovalent forces between the H-2//3z-microglobulin monomers. Some evidence suggests that two H-2 monomers can be disulfide linked near their C-termini to form an M, 116,000 structure as shown schematically in Fig. 8, but half-molecules are also found in cell lysates. The resemblance of the four-chain structure to immunoglobulin is apparent. Structural data continue to mount, attesting to the real but distant evolutionary relationships between histocompatibility antigens and immunoglobulins (Peterson et al., 1975; Capra et al., 1976). 2. Peptide Mapping Two general methods have been utilized to study the peptide structure of H-2: (1) classical peptide mapping utilizing paper electrophoresis in one direction and chromatography in the second; (2) column peptide mapping utilizing various ion-exchange resins. Both

THE MUFUNE

1

7

CHROMOSOME ~ ~

161

FIG.8. Subunit structure and orientation of H-2 antigens in solution and on the cell surface. Detergent-solubilized H-2 antigens contain two heavy and two light chains with at least one disulfide bond linking the two heavy chains. Papain treatment of intact cells or detergent extraction generates H-2 heavy-chain fragments of identical size. H = H-2 heavy chain; Ltel, = H-2 light chain (p,-microglobulin); Fs = water-soluble fragment (FHplus L) obtained after papain treatment of cell surfaces or detergent extracts; FH= fragments of the H-2 heavy chain obtained after papain treatment; F, = portion of the H-2 heavy chain cleaved from the bulk of the molecule by papain and apparently associated with the cell membrane. Molecular weights are given in parentheses. Reproduced from Cunningham et al. (1976)with permission from the Cold Spring Harbor Symposia on Quantitative Biology.

procedures are based on the principal that identical peptides will coelute or comigrate, and coincidence in one of these techniques is tantamount to proving peptide identity. While this is generally true, it is important to appreciate that these techniques magnify differences. Thus, when tryptic mapping techniques are used on a molecule of this size, 20 strategically placed amino acid interchanges in the entire sequence of350 amino acids could result in maps that are so totally different that no similarity would be noted, despite overwhelming sequence homology. An example from immunoglobulin structure will b e illustrative: if one did peptide maps of K and X chains, no peptides would “overlap,” yet we know from sequence and tertiary structure analyses that they are closely related. Using these techniques, Nathenson and his colleagues have provided us with the major insights into H-2 structure (Nathenson, 1970; Yamane et al., 1972; Schwartz et al., 1973; Brown et al., 1974; Nathenson and Cullen, 1974; Cullen et al., 1974). Their general procedure

162

ELLEN S. VITETTA AND J. DONALD CAPRA

involves use of Nonidet PA0 (NP40) lysates from lymphocytes pulsed with radioactive arginine and/or lysine (Schwartz and Nathenson, 1971). H-2 molecules are immunoprecipitated, digested with trypsin, and then column-mapped. By choosing appropriate radioisotopes and mixing 3H- + I4C-labeled immunoprecipitates, Nathenson and associates have compared alleles at a single locus-H-2Kk vs H-2Kb and H-2Dk vs H-2Db; and gene products from different loci-H-2Kk vs H-2Dk and H-2Kd vs H-2Dd. The general conclusions from the work of Nathenson and his colleagues and of others are as follows: (1)the different loci (Kvs D ) have about 25-50% of their peptides in common, depending on whether the comparison was Kb vs Db or Kd vs Dd; (2) comparisons of alleles showed about 30% peptide identities. Thus, by these techniques, the H-2 gene products were thought to be markedly different structurally, and, surprisingly, the allelic differences appear to be as great or greater than the regional differences (see Figs. 9 and 10). 3 . Sequence Studies The first partial sequence analysis of H-2 molecules was obtained by Silver and Hood (1975). While these studies contained only a few amino acid assignments, their impact has been impressive. Shortly thereafter several groups presented partial N-terminal sequence data on several H-2 glycoproteins (Capra et al., 1976; Cook et al., 1978;

FRACTION NUMBER

FIG. 9. Comparison of the tryptic peptides of a K gene product, [14C]argininelabeled H-2.31 (-), and of a D gene product, [3H]arginine-labeled H-2.4 (---). [3H]Arginine-labeled H-2.4 and [14C]arginine-labeledH-2.31 were mixed and digested with trypsin. The supernatant containing the soluble peptides was separated on an ion-exchange peptide column. Reproduced from Yamane et al. (1972) with permission from Biochemistry. H-2.4 is the D gene product of B10.D2 (H-2d),and H-2.31 is the K gene product of

B10.D2.

THE MURINE

m

M

a

m

1

7

CHROMOSOME ~ ~

IW 110 F R A C l l O N NUMBER

I)O

160

163

IW

?a,

FIG.10. Comparison of the tryptic peptides of two K gene products: [“Clarginine-laand 3H arginine-labeled H-2.31 (----). Preparation of the sample for the peptide column and its elution were the same as in Fig. 9. H-2.31 is the K gene product of B10.D2 (H-2“) and H-2.33 is the K gene product of B10 (H-2b).Reproduced from Yamane et u2. (1972) with permission from Biochemistry.

beled H-2.33 (-1

Ewenstein et al., 1976; Henning et al., 1976; Silver and Hood, 1976a,b; Vitetta et al., 1976a). Conclusions originally reached by Silver and Hood have been confirmed and extended with the accumulation of approximately 100 assignments at the N-termini of several molecules (Fig. 11). 1. The K and D gene products are homologous to each other. Sequence data now confirm the peptide mapping conclusions that there are major structural similarities in the products of the K and D region. Presently, greater than 85% of the comparable positions are identical when one focuses on positions that are not seen to vary among alleles. Certain positions that appear to be invariant among regions and alleles (6,7,8, 11, 12, 14,15,27) presumably form the basis of the major peptide similarities between all K and D gene products. These may also explain the long standing known serologic (“public”) cross-reactions (Hauptfeld and Klein, 1975). 2. There are modest differences between the products of allelic genes. Although the data are far from complete, no high degree of variability has been shown in the amino-terminal region of the H-2 alloantigens; only two differences were detected in the three H-2K allelic products (Kk,Kq, and Kb) examined in this laboratory. The two amino acid interchanges (positions 9 and 22) observed in these three H-2K molecules may be attributed to classical allelic variability. The data for other H-2 antigens (Kd, Dd,and Db)is much less complete, yet only modest variation other than the His-Val and Phe-Tyr interchanges at positions 9 and 22 is evident. The most likely interpretation is that the

164

ELLEN S. VITETTA AND J. DONALD CAPRA 5

Kk

fi

** PRO HIS

(GLY)PRO HIS SER LEU ARG TYR

8

**

Kd

bET

lP Dd

P W HIS SER LEU

HIS PRO

MtT

0

FW HIS THR

25

PHE *’

LEU

LYS PRO ARG

PfiE HIS TWI ALA VAL SER ARG PRO

LEU

LYS PRO

LEU

LYS PRO ARG TYR

ARG TYR FW VAL THR ALA VAL ARG

20

ALA VAL SER ARG PRO

T(R

*ARGTYRPHE+ HIS

15

10

S ! 3 LEU ARG TYR

SER

ARG PRO

(m) * ALAVAL

0

ARG

L E U A K TYR PHEVAL(THR)ALAVAL(THR)ARG PRO

LEU

ARG PHE

TYR

**

(PR0)ARG PHE

PRO ARG PRO

TYR

T(R

ARG TYR

FIG. 11. Partial amino acid sequences of murine H-2 molecules. Data were taken from the original work by Capra et al. (1976), Ewenstein et al. (1976), Henning et al. (1976). Silver and Hood (1975, 1976a,b), and Vitetta et al. (1976a). Additional assignments are from a summary compiled by Silver (1976) as well as more recent data from our laboratory. A blank indicates that no assignment was made at the indicated position; *, absence of the residue reported in another molecule; +, valine not found, histidine not tested; ’, both valine and leucine have been reported for these positions; **, previously assigned, not reproduced in more recent studies. Parentheses indicate a tentative assignment. The AldPhe in positions 21-22 of the Kk molecule were erroneously transposed in a previous publication.

mutations at positions 9 or 22 were independent events in the K and D regions which occurred after duplication of a primordial H-2 gene into modem day K and D ; the subsequent mutation at the other position might then have been selected because of conformational constraints. An alternative interpretation is that the NH,-terminal region of murine histocompatibility antigens consists of a limited number of “ groups”-one containing His at 9 and Phe at 22 and another having Val and Tyr at 9 and 22, respectively, and the entire molecule results from the fusion of two “genes.” Thus, one gene shared b y the K and D regions (selected from a small pool of group specijc genes) encodes the NH,-terminus, and the other a region specijic gene encodes the carboxy terminus. This interpretation is consistent with the peptide mapping data, the widespread serological cross-reactivity of H-2 alloantigens between alleles and between the K and D products, as well as the marked differences between allelic products within a region. 3. There is strong homology between H-2 and HLA. Figure 12 compares the sequence of the two most completely sequenced murine histocompatibility antigens with the reported amino terminal sequence of two human transplantation antigens recently described by Terhorst et al., (1976).Additional assignments have been taken from Ballou et

TYR

TYR TYR

THE MURINE

1

7 CHROMOSOME ~ ~

10

5

SE R

VAL

TYR PHE SER SER

15

ALA

165 25

20

(SER)ASX

FIG.12. A comparison of the partial amino acid sequence of two protein products of the K region and a human HLA antigen. See legend to Fig. 10 for original citations for the H-2 sequences. HLA sequences are taken from Ballou et a1. (1976), Terhorst et (11. (1976), and Appella et al. (1976b).

al. (1976). The extraordinary similarity is obvious. For example, between positions three and eight there is only one difference in this six residue stretch. Since glycine has not been labeled as yet, positions 1 6 , and 18 in the murine sequence may also be homologous to the human structures. Thus, the degree of homology between the human and murine transplantation antigens may b e even higher than is evident from the presently available data. 4. There is modest homology between transplantation antigens, p2microglobulin and immunoglobulin. Figure 13 displays the sequences of the H-2K9 and H-2Kb molecules derived in this laboratory, the human HLA transplantation antigens, and the amino-terminal 27 residues of rat p,-rnicroglobulin (Poulik et al., 1977). p,-microglobulins have extraordinarily conserved sequences with the exception of the region around the amino-terminus. The rat sequence is displayed because the sequence of the same portion of the molecule has been determined in the mouse (Silver and Hood 1976a; Appella et al., 1976a,b), and it is typical of µglobulin sequences. There are seven positions in µglobulin in which either the mouse or human transplantation antigens have an identical residue. The con-

5

10

25

I L E GLN LYS THR PRO GLX I t € G L N m T Y R m A R T . H I S PRO

($4

H-ZK~ H-ZK~

HLA-2

(GLY)PRO H I S SER L E U ARG T V R PHE HIS

THR A L A V A L SER ILE

PRO t i i s SER L E U ART. T Y R P H E ~ T H ALA R VAL SER ARG

GLY SER HIS

SER MET ARG TYR PHE PHE T Y R ~ V A SER L ARC

FIG. 13. Amino acid sequence comparison of rat p,-microglobulin and rnurine and human transplantation antigens. Homologous positions have been outlined. *, previously assigned, not reproduced in more recent studies.

TVR TYR

166

ELLEN S. VITETTA AND J. DONALD CAPRA

stituent polypeptide chains of many proteins composed of subunits often bear structural similarities (hemoglobins, immunoglobulins) and the fact that p,-microglobulin is found in physical association with transplantation antigens might suggest that their structures would display some rudimentary structural similarities based on the necessity for their association. However, additional homologies will have to be established between the transplantation antigens and µglobulin, in order to consider seriously that they evolved from a common primordial gene. Figure 14 compares the amino acid sequence of a prototypic murine immunoglobulin VJII heavy chain (Capra et al., 1975a) with the human and murine transplantation antigens. Without introducing insertions or deletions in any of these chains, there are two positions (5 and 19) in which the murine VHIII prototype is identical to both murine H-2K gene products. In two additional positions (16 and 24) the murine VHIII prototype has the same amino acid as both human transplantation antigens. In position 12 virtually all murine VHIIIproteins, murine transplantation antigens, and human transplantation antigens contain valine. When the murine H-2K sequence was subjected to a computer search (National Biomedical Research Foundation) a human VJII heavy chain emerged as the best “fit” of over 800 known sequences in the computer bank. This modest degree of amino acid sequence homology between immunoglobulin V regions and murine and human transplantation antigens can only be definitively established by additional amino acid sequence data. The regions around the cysteine residues should prove informative in this regard. 5. There is a striking sequence homology among the major histocompatibility antigens of mammals and birds. Figure 15 compares the available data on murine H-2, human HLA, guinea pig GPLA and recent work from our laboratories on the chicken B locus (Vitetta et al., 10 GLU VAL GLN LEU

V~III

*

H-2Kk

H-ZK~

HLA-A~

15

20

25

LEU SER CYS A L A ~ S E R ~ P H E

PRO H I S SER

TYR

PRO H I S SER

TYR

G ~ YSER H I S SER

FIG. 14. Amino acid sequence comparison of a prototypic murine VJII heavy chain and murine and human transplantation antigens. Homologous positions have been outlined. *, see legend to Fig. 13.

THE MURINE 5

MllSE H-2

1

7

CHROMOSOME ~ ~ 15

10

(CLY?PraOHIS SER LEU ARG R R FttE VAL IN3 ALA VAL M R

AR(; PllD

GLY SER HIS SER kEET SER

oJINu\

GPIA

PIG

HIS

TIR FHE HIS THR SER VAL SER ARG PRO GLY VAL

nanasERALA

LEUARGT(RR1ETYR

&\VAL

AIA

LYS PRO ARG PHE ILE

LBJ

AR(;

Ltllna

AIA IYR

CLY GLY SER GUI

PRO

25

20

FttE

HIS HMAN HLA

167

ASX FHE

ILE AIA

VAL

GLY TIR

m PRO R(E VAL

na

HIS ARC TIR(R&)AIEG m PRO (m) na FIG. 15. Amino acid sequence comparison of histocompatibility antigens from human, mouse, guinea pig, and chicken. See legends to Figs. 11 and 12 for original citations for H-2 and HLA sequences. The guinea pig and chicken sequences are from B.D. Schwartz et al. (197613) and Vitetta et al. (19771~).

CHICKEV B

1977b). With the exception of position 12, in every tested position, where all three mammalian species had a particular amino acid residue, the same residue was found in the chicken (positions 3, 6, 7, 8, 15,27). Thus, the basic structure of transplantation antigens must be at least 200 million years old and the available data suggests they have varied little in the ensuing years. E. CONCLUSIONS After the careful genetic studies of the past five decades, and the initial chemical analyses of the past five years, the elucidation of the amino acid sequences of the major murine histocompatibility antigens is now at hand. The preliminary data which have been useful in sharply defining some of the arguments concerning the evolution and control of the entire H - 2 complex, and the borderline homologies to immunoglobulin serve as tantalizing morsels which urge us on to complete the difficult task of unravelling the entire structures of the H-2D and H-2K gene products. One need only look at the field of immunochemistry to appreciate the enormous advances that followed from the elucidation of the primary structure of the central molecule of that discipline. V. The/ Region

The Z region is located in the H-2 complex between H - 2 K and Ss (David et al., 1973; Hauptfeld et al., 1973; Sachs and Cone, 1973; Hammerling et al., 1976).In the 4 years since its discovery, the Z region and its products have been vigorously studied in many laboratories, and several important findings have emerged.

168

ELLEN S. VITETTA AND J. DONALD CAPRA

A. GENETICS The Z region is currently divided into 5 subregions termed, from left to right, ZR-ZA, -ZB, -ZJ, -ZE, and -ZC. The ZR-ZA, -ZB, and -ZC subregions are each defined by genes that control the immune responses to a variety of antigens (Shreffler and David, 1975; David, 1976). With defined doses of antigen and routes of immunization, strains of mice can usually be defined as either “high” or “low” responders. In contrast, the ZR-ZE subregion is defined by an alloantigen (David et al., 1976, 1977b), and the -ZJ subregion by suppressor function and suppressor factors (Murphy et al., 1976; Tada and Taniguchi, 1976).In addition, all the subregions, with the possible exception of ZR-ZB, are associated with the expression of specific products called Ia antigens. Each subregion will be discussed briefly in the following sections. ZR-ZA. This subregion determines immune responsiveness to (T, G)-A--L,ovalbumin (McDevitt and Sela, 1967), bovine IgG (Benacerraf and Katz, 1975), hen ovomucoid (Benacerraf and Katz, 1975), IgA (Benacerraf and Katz, 1975), and ragweed pollen (Benacerraf and Katz, 1975). In addition, it encodes cell surface Ia antigens carrying specificities 1, 2, 3, 8, 9, and 20 (Cullen et al., 1976; David, 1976; McDevitt et al., 1976). These assignments have been made by a combination of recombinant typing and sequential precipitation of Ia molecules (see later sections). The ZR-ZA subregion appears to determine the elaboration of helper factors which carry Ia specificities (Armerding et al., 1974; Armerding and Katz, 1974; David et al., 1976; Frelinger et d., 1975; Hammerling et al., 1976; Pierce and Klinman, 1975; McDevitt, 1976; McDevitt et al., 1976). ZR-ZB. The ZR-ZB subregion controls the response to at least 3 antigens-the myeloma protein MOPC- 173 (Lieberman and Humphrey, 1971, 1972), lactic dehydrogenase B (Melchers et al., 1973), and staphylococcal nuclease (Lozner et al., 1974; Sachs et al., 1977). Although there are no agreed-upon Ia specificities that map exclusively in the ZR-ZB subregion, Ia specificities 4, 19, 13, and 15 do map between H-2K and ZR-ZC, placing them in either the ZR-ZA or -ZB subregions (David, 1976). ZR-ZC. This subregion contains immune response genes to the random terpolymer GLT (Merryman and Maurer, 1974). The Ia specificities that map only in the ZR-ZC subregion are 6, 7, and 21 (David, 1976).In addition, Ia antigen specificities 10 and 15 map between H-2K and Ss (David, 1976). ZR-ZE. Recent data have suggested the existence of two distinct loci between IR-ZB and -ZC which have been caled -ZJ and -ZE. The ZR-ZE subregion codes for a B-cell Ia antigen, Ia-22.

THE MUFUNE

1

7

CHROMOSOME ~ ~

169

ZR-ZJ. Recent studies from two laboratories (Murphy et al., 1976; Tada and Taniguchi, 1976) have suggested that the ZR-ZJ subregion (located between ZR-ZA and -ZC, possibly to the left of -ZE) is involved in T-cell suppressor activity. The two systems that have been studied are allotype suppression and antigen-specific suppression. In the second system, the ZR-ZJ determinants have been found on a soluble supressor factor (Tada and Taniguchi, 1976). The detection of ZR-ZJ determinants on concanavalin A (Con A)-induced suppressor cells suggests another association between ZR-ZJ and suppression of immune response (Frelinger et al., 1976). Thus, treatment of T cells with anti-ZJ inhibits Con A-induced elaboration of T cell suppressor substances (J. A. Frelinger, unpublished observations). As mentioned above, -ZJ determinants are found on suppressor factors. In addition, the Ia-4 specificity, previously assigned to the -ZB subregion, has now been reassigned to the -ZJ subregion. Ia-4 appears to be present on a very minor population of peripheral T cells (Murphy et al., 1976). B. FUNCTIONAL RELATIONSHIPS BETWEEN ZR GENESAND Ia ANTIGENS As discussed in the preceding section, the Z region codes for both immune response (ZR) genes and surface molecules on immunocytes known collectively as Ia antigens. In order to examine the relationship between ZR genes and Ia antigens, several laboratories have attempted to block specific immune responses with defined anti-Ia sera. Using mouse lymphocytes, antigen-induced stimulation of T lymphocytes from high-responder strains has been blocked with a specific anti-Ia serum (Schwartz et al., 1976b). In contrast, the anti-Ia sera directed against Ia antigens of the nonresponder haplotype give little or no inhibition of the responder cells. In addition, anti-H-2 sera have no effect on the response. A number of other studies have suggested a functional role for Ia antigens, but these studies do not define their relationship to ZR genes. Thus, anti-Ia sera can block the stimulating cell in the mixed lymphocyte reaction (MLR) ( M e 0 et al., 1975a, 1976), the acceptor site of the B cell for the T helper factor (Taussig et al., 1975), and the binding of aggregated Ig to Fc receptors (Dickler and Sachs, 1974). In addition, as mentioned in Section V,A, Ia antigens have been reported to be associated with both suppressor and helper factors (Armerding et al., 1974; Murphy et al., 1976; Tada and Taniguchi, 1976), but there is some conflicting evidence regarding the presence of Ia on these factors (David et al., 1977b). A number of studies strongly suggest that the Z region determines

170

ELLEN S. VITETTA AND J. DONALD CAPRA

whether IgG antibody production will result from the interaction of B and T cells (Katz and Benacerraf, 1975; Pierce and Klinman, 1975). Thus, although Ia- B cells can develop into cells that secrete IgM antibody, IgG production requires Ia+ B cells cooperating with T cells. This cooperation is restricted by the -ZA subregion. Thus, in a secondary IgG response, the T cells must be confronted with primed B cells of the same ZA haplotype as in the primary response. This restriction has been interpreted as a requirement for the T and/or B cells to recognize both the antigen and the Ia antigen on the cooperating cell in order for a triggering event to occur. The manner in which this recognition is carried out is not yet clear.

c. TISSUEDISTRIBUTIONOF Ia

ANTIGENS Ia antigens have a more limited tissue distribution than H-2 and have not been detected on fibroblasts, erythrocytes, brain, liver, or kidney (Hammerling, 1976). They have been detected on lymphocytes (David et al., 1973; Hauptfeld et al., 1973; Sachs and Cone, 1973; Hammerling, 1976), macrophages (Hammerling, 1976), sperm (Hammerling et al., 1975, 1976; Fellous et al., 1976a,b, Hammerling, 1976), epidermal cells (Hammerling, 1976), and some tumors (Hammerling, 1976). Ia antigens have been detected on fetal liver as early as 10 days of gestation (Hammerling, 1976). Lymphocytes a. B Cells. Approximately 90% of Ig+ B cells express high concentrations of Ia antigens, whereas 10%express little or no Ia. LPS blasts, which are precursors of IgM-secreting plasma cells, generally express large amounts of Ia antigens (Sachs and Cone, 1973; Hammerling et al., 1975; Press et al., 1976). b. T Cells. It has been difficult to detect Ia on thymocytes and peripheral T cells. However, it is now generally agreed that some T cells express low concentrations of (at least some) Ia specificities (Frelinger et al., 1974; Hammerling, 1976). In addition, Con A blasts express large amounts of Ia antigen (Wagner et al., 1975), although phytohemagglutinin (PHA) blasts do not (Hammerling, 1976). With the recent discovery that certain Ia specificities may be restricted to defined subsets of T cells (Murphy et al., 1976; Okumuraet al., 1976), it is obvious that part of the difficulty in detecting Ia antigens on T cells may reside in the specificity of the anti-Ia being used, the expression of particular Ia antigens on minor subsets of T cells, and the relatively small quantities of Ia antigens on T cells. c. Macrophages. Ia antigens are present on 20-50% of peritoneal

THE MUFUNE

1

7

CHROMOSOME ~ ~

171

macrophages (Hammerlinget al., 1975; Hammerling, 1976),and it has been suggested that these determinants play a major role in antigen recognition by T cells of macrophage-bound antigen (Erb and Feldmann, 1975). Numerous studies along these lines have been performed in the guinea pig system (Shevach et al., 1972; B. D. Schwartz et al., 1976a; Thomas and Shevach, 1976; Paul et al., 1977).

D. STRUCTURE 1 . Zsolation Techniques Cells bearing Ia can be treated with detergent or papain, and the antigens can b e isolated by conventional immunologic or biochemical methods (Vitetta et al., 1974; Delovitch and McDevitt, 1975; Cullen et al., 1974, 1976; Hess, 1976).In addition, molecules carrying Ia determinants have been detected in mouse serum and tissue culture medium (Vitetta et al., 1974; Parish et al., 1976a,b; Jackson et al., 1977). Three recent improvements in techniques, described primarily b y Cullen and co-workers (1976), now allow the effective isolation of Ia antigen using immunologic methods. First, Ia antigens adhere to, and can be eluted from, the lectin from Lens culinaris (Cullen et al., 1976). Thus, using Sepharose-lectin, it is possible to separate Ia molecules from >95% of other cellular proteins. Although this step results in losses of up to 50% of the Ia molecules, those molecules that are recovered are highly enriched. Second, the use of fixed S. aureus (Cowan I strain) rather than anti-Ig sera to bind Ia-anti-Ia complexes has eliminated the need for prior removal of Ig from cell lysates (Kessler, 1975).It has also permitted the application of larger amounts of material to SDS gels. Third, the use of a modified Laemmli gel system now facilitates the separation of the Ia subunits (Cullen et al., 1976).As shown in Fig. 16, Ia subunits are difficult to separate on the Maize1 gel system, but can be resolved on the Laemmli system.

2. Za Antigens in Serum Recent studies from two laboratories have demonstrated the presence of Ia antigens in serum. Thus, Callahan et al. (1976)found Ia antigens associated with high-density serum lipoprotein molecules. I n contrast, Parish et al. (1976a) reported that 90% of serum Ia activity resided in a dialyzable glycopeptide and 10% copurified with a2-macroglobulin (Jackson et al., 1977). In addition, Parish et al. (1976a) reported that Ia is secreted by T cells and can be found in the serum in greatly increased amounts after activation of T cells by antigens, mitogens, or graft-versus-host (GVH) reactions. However, in one recent re-

172

ELLEN S. VITETTA AND J. DONALD CAPRA

FRACTION

FIG. 16. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of anti-Ia.7 precipitates from lysates of radioiodinated B1O.AQR splenocytes ( 0 , anti-Ia; 0, normal mouse serum). Precipitates were dissolved and elecbophoresed on either Laemmli or Maize1 gels. Since goat anti-mouse Ig was used to precipitate Ia-anti-Ia complexes, the cell surface Ig is also present in each precipitate. The molecular weight of the Ia molecule was derived from a semilog plot of the apparent molecular weight of p (80,000)and L (22,000) chains on these gels (R. Cook et al., unpublished results).

port (David et al., 1977b), these findings were not reproduced, but instead were attributed to nonspecific inhibition of anti-Ia sera by mouse serum in general. Thus, further studies will be required to determine whether Ia antigens are present in mouse sera and, if so, what is the relationship between these molecules and their cell surface counterparts. 3 . Molecular Weight, Antigenicity, and Subunit Structure Ia antigens isolated both from the cell surface and from lysates of cells labeled with radioactive precursors have two subunits of M , 35,000-38,000 and 25,000-28,000 called (Y and p chains, respectively (Cullen et al., 1976). SQme,but not all, of these subunits are present as disulfide-bonded dimers of M , 58,000 (Cullen et al., 1976). Ia molecules clearly contain carbohydrate, but their role in some or all of Ia antigenicity is not clear. Parish et al. (1976b) reported the loss of Ia

THE MUFUNE

1

7

CHROMOSOME ~ ~

173

antigenicity after periodate oxidation or neuraminidase treatment, but not Pronase digestion. Cullen et al. (1976), however, reported conflicting results. Although minor differences in electrophoretic mobility have been described for Ia molecules from different haplotypes, the origin of these differences is not clear. I n addition, when cell surface Ia molecules (labeled with lZ5I)are compared to total cellular Ia (labeled with L3H1-amino acids), the ratios of radioactivity in alp vary (Schwartz et

al., 1978). One major application of the technique of immunoprecipitation of Ia molecules from cell lysates has been the determination of which antigenic specificities are on the same or different molecules (Cullen et al., 1976). Thus, using sequential immunoprecipitation techniques, Ia-8 and Ia-9 are coprecipitated and are presumably present on the same molecule (Cullen et al., 1976).The same is true of Ia-9 and Ia-20, and also Ia-7, and Ia-21 (Cullen et al., 1976). In contrast, Ia-9 and Ia-11 are on different molecules (Cullen et al., 1976), as are Ia-7 and Ia-22 (David et al., 1977b).Ia-15 coprecipitates with other molecules carrying ZR-ZA subregion specificities, indicating that it probably is not coded for by the -ZE region, as suggested previously.

4 . Papain-Cleaved Za Molecules As opposed to the estimated > 90% yields of Ia molecules obtained by detergent lysis, removal of molecules from cells by papain results in yields of less than 5%. However, after purification of the extracts by ion-exchange and molecular-sieve chromatography, a 40-50-fold purification has been achieved. Molecules purified in this manner have isoelectric points of 4.4-4.6 and are resistant to mild acid treatment and heating at 56°C. In addition, the molecules consist of single subunits of M , 28,000 with occasional fragments of 15,000 (Hess, 1976).

5. Comparative Peptide Mapping of l a Molecules When a and p chains from Za' and lak haplotypes were digested with trypsin and the peptides compared by cation-exchange chromatography, the following features were noted: When the aand p chains from Ia' were compared, 25% of the lysine-labeled peptides coeluted (Cullen et al., 1976) (Fig. 17). This suggests limited structural homology between the two chains and suggests, Imt does not prove, that the smaller p chain is not generated from the a chain by proteolysis. In contrast, when the a chains from Iak and Ia' were compared (Fig. 18), greater than 70% of the peptides coeluted (Cullen et al., 1976). Even though the 1ysine-labeled peptides probably represent approximately

174

ELLEN S. VITETTA AND J. DONALD CAPRA

fMCTlON NUMBER

FIG. 17. Comparison of tryptic fragments of two Ia-polypeptide chains from B10.RIII. The samples were extensively purified as described in the text and were homogeneous when examined by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The samples were digested exhaustively with TPCK-trypsin at pH 8.0. At the completion of the digestion the mixtures were acidified with acetic acid and the insoluble cores were removed by centrifugation. The soluble supernatants were examined by cation-exchange chromatography on a Spherix XX-8-60-0 resin using pyridine/acetic acid gradients (pH 3.12 to pH 5.0)for elution. This figure shows comparative chromatographic patterns for the M , 36,000 (-) and 27,000 (---) Ia-polypeptide chains from [3H]lysine-labeled B10.RIII. See Brown et ul. (1974) for further details of the method. Reproduced from Cullen et al. (1976) with permission from Trunsplantution Reoiews.

50% of the Ia peptides, the data suggest that (Y chains from different haplotypes may be structurally very similar. These findings are reminiscent of the peptide similarities between H-2 molecules from the same haplotypes. 6. Primury Sequence of l a Antigens Using radiolabeled Ia antigens eluted from SDS gels, limited sequence data on the N-terminal portions of both the (Y and p chains from both the IR-ZA and -IC subregions has recently been reported (Silver et al., 1976; Cook et ul., 1977). Figures 19 and 20 compare these results with data presently available on what are thought to be homologous glycoproteins obtained from human (Springer et al., 1976) and guinea pig (B.D. Schwartz et al., 1976b),The available data are extremely limited, but certain tentative conclusions can be drawn.

THE MUFUNE

1

7

175

CHROMOSOME ~ ~

FRACMNWULUYA

FIG.18. Comparison of tryptic fragments from la-polypeptide chains from B1O.BR and B1O.RIII mice. Both samples were precipitated with the same alloantiserum (C3H.OL and anti-C3H) and were purified as described in the text. The digestions with trypsin and subsequent cation-exchange chromatography were the same as in Fig. 5. This figure shows comparative chromatographic patterns for the M , 36,000 polypeptide chains from [3H]lysine-labeled B1O.RIII (-) and ['4C]lysine B1O.BR (---). Reproduced from Cullen et al. (1976) with permission from Transplantution Reoiews.

1

2

A: 4: E/C:

I L E LYS

E/Ck

ILE

Human

034

3

4

5

6

7

8

9

10

11

12

13

14

15

16

17

ILL

ALA

VAL

TYR

( T H R ) VAL

TYR

ILE

ALA

VAL

TYR

VAL

TYR

H I S THR I L E I L E

ALA

PHE

TYR

LEU

LEU

ILE ILL

ALA

PHE

TVR

LEU

LEU

PHE

TYR

LEU

ASN

I L E L Y S GLU GLU ARG VAL I L L I L L GLN

ALA

GLU

18

PRO

TYR

19

20

ARG

ASP

PHE

FIG.19. Partial amino acid sequences of the a polypeptides of Ia molecules. Two IA and two I E/C subregion molecules have been partially sequenced. The two 1A subregion allelic products are identical at the six positions which can be compared. The two I EIC subregion allelic products also appear identical. The a polypeptide chains of mouse IA and I E/C bear no obvious homology to each other. There is a striking homology between the I E/C a chains and the human Ia equivalent, p34. Data are from Cook et al. (1977),R. Cook, E. V. Vitetta, J. Uhr, and J. D. Capra, (unpublished observations), McMillan et al. (1977), Silver et al. ( 1976), Snary et al. (1976), and Springer et al. (1976).

176

ELLEN S. VITETTA AND J. DONALD CAPRA 2

1 Ak

3 5ER

5

6

7

8

9

10

(METI

11

12

TYR

PHE LEU

LYS

GLU

Guinea 4.5 ( Pig

) l L E TYR

PHE

PHE

PHE

PHE

LEU PRO

PHE LEU ARG ARG PHE

16

17

TYR

PHE

18

19

PHE

TYR

20

TYR

)GLY ASP THR PRO GLU ARG PHE L E U

)

15

PHE

Human p29 (

3.5 (

14

VAL TVR

VAL ARG MET SER ARG PRO

PRO

13

PH E

ARG H I S PHE VAL A RG

A:

E/C:

4

GLN

SER

VAL

TYR LYS

TYR

TYR

FIG.20. Partial amino acid sequences of the fl polypeptides of Ia molecules. The p polypeptide chains of mouse IA and I E/C bear no obvious homology to each other. Data on E/C$ are from Cook et al. (1977); A$ is from R. W. Cook, E. V. Vitetta, J. Uhr, and J. D. Capra (unpublished observation); Ab, is from McMillan et al. (1977); for other references, see text or legend to Fig. 19.

a . The a Polypeptide Chains. (1)The a chains derived from different haplotypes of the ZR-ZA subregion are homologous to each other. Thus, of the 10 positions that can be compared, only two are different (8 and 9). This degree of homology is similar to the homology seen in H-2 antigens from different haplotypes. (2) The a chains of the ZR-ZC subregion derived from different haplotypes are also similar to each other. (Note that in Fig. 19 the ZR-ZA subregion haplotypes are different from the ZR-ZC subregion haplotypes.) (3) There is no obvious homology between the primary structures of the products of the ZR-ZA and -ZC subregions. Indeed, there is not a single position from 1 to 17 in which the same amino acids is found in the product of even a single ZR-ZA or -ZC subregion. (4) The murine ZR-ZC subregion and the human “p 34” [which is likely a product of the HLA-D region (Springer et al., 19761 are extremely homologous, with no differences noted. ( 5 ) Neither the ZR-ZA nor -ZC a chains (or the human “p 34” for that matter) show any homology to H-2, µglobulin, or immunoglobulin. b. The /3 Polypeptide Chains. (1) The /3 chains derived from two different haplotypes of the ZR-ZA subregion are homologous (Fig. 20). (2) As in the a chains, the /3 polypeptides of the ZR-ZA and -ZC subregions bear no obvious homology to each other. (3) The a and fl chains are not homologous. (4) While the human “p 29” and the guinea pig sequence are somewhat similar (3 identifies in the 7 positions that can be compared), there is no apparent homology between the murine /3 chains of

THE MUFUNE

1

7

CHROMOSOME ~ ~

177

either subregion and either the human or guinea pig structure. If anything “p 29” is more homologous to ZA beta chains. It should be stressed that the above conclusions are based on only minimal sequence data, but in general, they could have been predicted from the peptide maps that have been discussed. However, since it is not at all clear that both the a and /3 chains are encoded within the major histocompatibility complex (MHC), or, if both are, whether they are the result of proteolytic cleavage of a larger molecule, as may be the case for human and guinea pig C 4 (see Section VI,E), many of these conclusions are tentative until more primary structural data are available. For example, if the a and /3 chains result from the digestion of a larger molecule, a slightly different cleavage point could produce very different N-terminal sequences. Complete structural analysis should resolve this problem. E. CONCLUSIONS At present, although significant data concerning the genetics, function, and structure of the Z region and its products have been obtained, it is not yet clear how the parts of the puzzle fit together. Clearly, there is great heterogeneity in the Z region. Z-region genes and products appear to b e involved in a fundamental way with the regulation of the immune response. Future research will most certainly b e directed toward elucidating the precise manner in which the ZR genes function. This will require not only more extensive immunologic studies, but also further attempts to probe the structural relationships among the Ia antigens on cells, on factors. And in serum. VI. The

S

Region

The S region is situated in the center of the H - 2 complex, which is now known to be intimately associated with complement function. The gene products of this region differ from the other products of the murine H - 2 complex by being primarily serum proteins. The Ss protein was originally described by Shreffler and Owen (1963)as a serum globulin that was present in some inbred mouse strains at a low level and in other strains at a high level. They detected the Ss protein by a rabbit antiserum made against a globulin fraction of normal mouse serum. Closely related to the Ss protein is the Slp, or sex-limited protein, which is detected by an alloantiserum and is thought to define an allotypic variant of the Ss protein (Hansen and Shreffler, 1976).A thorough review of the S region has recently appeared (Shreffler, 1976).

178

ELLEN S. VITETTA AND J. DONALD CAPRA

A. GENETICS Utilizing an appropriately absorbed rabbit antimouse antiserum, Shreffler and Owen (1963) showed a quantitative variation in the serum level of the Ss protein among inbred mouse strains. They classified some strains as Ss-H (those with a “high” level of Ss) and SS-L (those with a “low” level of Ss)and estimated a 20-fold variation between Ss-H and Ss-L. Only F, hybrids from crosses of Ss-H and Ss-L were intermediate (denoted Ss-HL). Shreffler and Owen noted at that time that all Ss-L inbred strains were of the H-2k haplotype, thereby providing the first clue that the Ss protein was linked to the murine MHC. By extensive backcross analysis, as well as from studies on intra-H-2 recombinants, they showed that the Ss protein mapped within the H-2 complex between the K and D regions. Since that time the I region has been placed between S and K and the G region between S and D.

B. FUNCTION While it has been clear for some time that Ss levels in mice varied with age, sex, and hormonal influences, it was a finding by Dkmant’s group of a relationship between Ss and complement levels (Hinzova et al., 1972; Demant et al., 1973) which accelerated interest in the Ss protein. Prior to these studies, the Ss protein was a novel genetic variant that was useful in defining the location of recombinants within the murine MHC. Now the question arose whether Ss was a “regulator” of serum complement levels or a complement component itself? Studies in humans at this time were beginning to show linkage between H L A and complement components of both the classical and alternative pathway. The association at a genetic level of integral membrane molecules (H2, Ia, etc.) and membrane “lytic” molecules seemed more than fortuitous. Hansen et al. (1976) and Goldman and Goldman (1976) have since confirmed and extended Dkmant’s observations. In short order, three papers linked the Ss protein with a specific human complement component (Curman et al., 1975; Lachmann et al., 1975; Meo et al., 1975b). Each used different techniques, and, taken together, these papers strongly suggested that the Ss protein was the murine equivalent of human (2-4. Recently, Wilde et al. (1978)have provided further evidence for this relationship by a series of biochemical and functional experiments. Highly purified and radioiodinated murine Ss material was quantitatively precipitated with antiserum specific for human C-4 and vice versa. The molecular weights and subunit structures of the two proteins were found to be indistinguishable. Thus, the evidence is now compelling that a critical component of

THE MUFUNE

1

7

CHROMOSOME ~ ~

179

the classical complement pathway maps in the central region of the murine H-2 complex. Structural studies on the Ss protein might therefore provide insights into the evolution of the MHC and particularly might indicate a relationship between complement components and alloantigens. C. DISTRIBUTION Unlike the other MHC gene products, the Ss protein is present in serum in relatively high amounts. Thus, both its isolation and characterization have been easier than for other MHC antigens, which specify plasma membrane components. Utilizing immunofluorescence to screen cells, Saunders and Edidin (1974) studied the site of synthesis and cellular distribution of the Ss protein. They concluded that both peritoneal macrophages and liver parenchymal cells synthesize Ss for export to serum. Those cells stained for Ss in their cytoplasm and released Ss into the culture medium. Fibroblasts appeared to have Ss bound to their plasma membrane. It was not proved, however, that Ss had not been adsorbed from the serum. Recently, Wilde and Capra (1978) have been able to confirm and extend most of these findings. By culturing mouse peritoneal macrophages with radioactive amino acids, the Ss protein could be easily isolated from the culture supernatants. Hall and Colton (1977) reported that C-4, synthesized by guinea pig liver cells, does not cleave into subunits upon reduction, suggesting that if the Ss protein is the murine homolog of C-4, it too may be synthesized as a single chain and that the three-chain structure results from postsynthetic cleavage.

D. STRUCTUREOF THE Ss PROTEIN 1 . Isolation Until recently, the major source of Ss protein has been either mouse serum or plasma. Capra et al. (1975b) reported purification of the SS protein from mouse plasma utilizing a series of ion-exchange and gelfiltration procedures. However, the molecular weight and subunit structure have proved to be incorrect, as the commercially obtained murine plasma undoubtedly contained activated complement. T h e complete antigenic identity between this “cleaved” Ss and recently isolated “native” Ss suggests that most if not all Ss antisera are directed against determinants on “cleaved” material. Carroll and Capra (1978) have recently described the isolation of Ss protein from murine ascitic fluid. Utilizing procedures originally described by Tung et al. (1976), 5-10 ml of ascites fluid can be obtained weekly from individual mice. The amount of Ss in ascitic fluid is about one-third of that found in normal plasma. Ascitic fluid has therefore

180

ELLEN S. VITETTA AND J. DONALD CAPRA

become a rich source of pure Ss protein. This material is antigenically and structurally identical to Ss material isolated from serum.

2 . Molecular Weight and Subunit Structure Original reports on the molecular weight of the Ss protein were based on gel filtration analyses in neutral buffers followed by the use of specific antisera to detect Ss and Slp activity in individual column fractions. These studies indicated that two molecular weight species of both proteins existed, with significant differences in the distribution of the larger (one million) and smaller (180,000) protein in males and females and in Ss-H vs Ss-L mice (Hansen et al., 1974).Recently, Meo et al. (1975b) reported similar results. The first report on the molecular weight of the Ss protein in dissociating buffers was M , 120,000 by Capra et al. (197513). However, since this value was obtained on radioiodinated, partially purified Ss derived from commercially available serum, in retrospect, it probably represented a proteolytic fragment of Ss that still retained antigenic activity. Curman et al. (1975)radioiodinated the Ss protein which had been purified by affinity chromatography and obtained a molecular weight of “slightly less than 200,000.” Our most recent results indicate that the molecular weight by SDS-polyacrylamide gel electrophoresis is 210,000 (Wilde et al., 1978). This result has been obtained from both Ss-H and Ss-L strains, and strains that-are Slp positive and Slp negative. Higher molecular weight material has not been ohserved in these analyses. This would suggest that the earlier studies (done in neutral buffers) were detecting noncovalent aggregation. The subunits of Ss were originally reported to be 45,000 and 35,000 by Capra et al. (197513).Curman et al. (1975) reported two subunits of 70,000-80,000 and a third of 23,000. Both groups reported occasionally finding subunits of 13,000. These discrepancies have been clarified as more became known of the relationship of Ss to complement, and the stringent need for precautions during isolation. Wilde et al. (1978) using radiolabeled material found subunit molecular weight of approximately 90,000, 80,000 and 35,000. These subunits were virtually identical to the subunits of human C-4 (Schreiber and MullerEberhard, 1974; Bolotin et al., 1977).

E. RELATIONSHE

BETWEEN THE

MUFUNESS PROTEIN AND

HUMANC-4 Lachmann et al. (1975) by functional analysis, and Curman et al. (1975)and Meo et al. (1975b)by antigenic cross-reactivity reported that the Ss protein was the murine equivalent of human C-4. Carroll and Capra (1978) recently described a simple and reproducible assay for murine functional C4 by using a C4-deficient guinea pig system.

THE MUFUNE

1

7

CHROMOSOME ~ ~

181

5 10 15 20 ASN VAL ASN PHE GLU LYS ALA I L E ASN GLU LYS LEU GLY GLU TYR ALA SER PRO THR ALA LYS

LYS PRO GLY LEU LEU LEU PHE PHE CYS LEU

X

GLU ALA PRO LYS VAL VAL GLU GLU GLN GLU SER ARG VAL HIS TYR THR VAL CYS I L E TRP GLY HIS ALA CY s FIG.21. Partial amino terminal sequences of the three chains of human C 4 . Data are from Bolotin et al. (1977), Gigli et al., 1977), and Andrews and Capra (1978). The y

chain may be polymorphic, as three different positions appear to have alternative amino acids. Each of the above authors has noted this heterogeneity at different positions.

Utilizing this assay, experiments were described which conclusively demonstrated that the murine Ss protein is functional C4 by showing that: 1) pretreatment of mouse plasma with F(ab)’2 anti-Ss completely abrogated the C4 hemolytic titer; and 2) in the isolation of Ss protein, Ss immunochemical activity and C4 functional activity coeluted after each step of a four-step pruification procedure. The final proof will have to await amino acid sequence analysis (Shreffler, 1976). Gigli et al. (1977), Bolotin et al. (1977) and Andrews and Capra (1978) have recently reported the N-terminal sequence of the three chains of human C-4 (Fig. 21). The amino acid sequence of the Ss protein should soon be available for comparison.

F. CONCLUSIONS The evidence is now compelling that the Ss protein represents the murine equivalent of the fourth component of the classical complement pathway. Recent experiments indicate that the entire molecule may be encoded within the H-2 complex at the Ss locus. Further studies will undoubtedly determine more precisely the relationship between Ss and Slp, and possible structural similarities between Ss and other H-2 complex encoded glycoproteins. In man, C2, C4 and factor B of the alternative pathway of complement are all linked to the MHC. These diverse components have in common a “C3 convertase” capacity for the generation of C3b. One might speculate that in man C 3 may also be linked to the MHC; and in the mouse, studies should be directed toward C 2 , Bf and C 3 linkage. Perhaps the generation of C3b is the critical link between H-2 and Z region which has kept this entire “supergene” complex together (see Section IX). VII. The Region between H-2D and Tla

Recent studies have mapped two, perhaps three, genes between H-2D and TLa:Qa-l (Stanton and Boyse, 1976),Qa-2 (Flaherty, 1976),

182

ELLEN S. VITETTA AND J. DONALD CAPRA

and a molecule carrying the public H-2.28 antigen but not the private H-2D specificity, H-2.4 (Hansen et ul., 1977).These genes encode cell surface molecules. A. Qa-2

Using two new C57BL/6 stocks that have weak histocompatibility differences, a recombination in the region between H-2D and TLa was defined (Flaherty, 1976). Cross-immunizations between these mice resulted in the production of an antiserum that defined antigens on lymph node, spleen, thymus, and liver, but not erythrocytes, epidermal cells, brain, or kidney. Using the enzymic radioiodination technique, the Qa-2 antigen was found to contain an M , 44,000 subunit bound noncovalently to &-microglobulin (Michaelson et a1., 1977)(Fig. 2). B. Qa-1 This antigen is present on both thymocytes and lymph node cells and i s defined by an antibody found in conventional anti-TL typing sera. Cytotoxicity tests demonstrate that Qa-1 is present on a subset of peripheral T cells. No structural data on Qa-1 are available at present (Stanton and Boyse, 1976). C. H-2.6, 27, 28, 29 Both cocapping (Lemonnier et al., 1975) and immunoprecipitation (Hansen et al., 1977) studies indicate that two molecules may exist in the H-2D region, one carrying both private and public specificities and one carrying only public specificities. The latter molecule has the same size and subunit structure as H-2 (Hansen et al., 1977) and may represent the product of a second gene closely linked to H-2D. VIII. TLo

A. DEFINITION AND GENETICS

The genetic locus TLa, mapping to the right of H-2D on chromosome 17, encodes a group of cell surface alloantigens known collectively as TL (thymus leukemia) antigens (Boyse et al., 1966). Four specificities have been defined for this locus: TL.l, TL.2, TL.3, and TL.4 (Boyse and Old, 1971). The thymocytes of TL+ mice may be grouped into three phenotypes: TL.l, 2, 3, TL.2, and TL.1, 2 (Boyse and Old, 1971). During the course of leukemogenesis, TL- mice may

THE MUFUNE

1

7 CHROMOSOME ~ ~

183

develop TL+ leukemias and TL+ mice may express additional TL specificities, such as TL.4 (Old et d.,1963). These findings suggest that the TLa complex consists of two loci: structural genes encoding the different TL specificities and regulator genes determining the expression of the structural genes. Apparently all strains carry TLa 1and TLa 2 structural genes, while only some strains carry TLa 3 and TLa 4 . It is also possible that TLa 1 and TLa 2 may be separate loci, and TLa 3 and TLa 4 may be alleles at the same locus. The linkage of TLa with H-2 is supported by substantial genetic evidence, and it has been demonstrated that the two loci are 1.5 map units apart (Klein, 1975). While the TLa region does not appear to be a histocompatibility locus, it is nevertheless closely linked to the major histocompatibility complex.

B. TISSUEDISTRIBUTION TL specificities 1, 2, and 3 have been detected on normal thymocytes in some, but not other, strains of mice. TL+ thymocytes are small, dense lymphocytes that contain high concentrations of Thy-1, GV-1, and Ly-1, 2, 3 but little H-2 (Konda et al., 1972, 1973). The larger, more differentiated cells in the thymus lack TL, as do peripheral T lymphocytes (Kondaet al., 1973). However, several lines of evidence suggest that a population of peripheral TL- cells may become TL+ under the influence of the thymus (Komuro et al., 1973).

c. ANTIGENIC MODULATIONOF TL When TL+ tumors are inoculated into syngeneic mice that have high titers of cytotoxic anti-TL antibody, the tumors become resistant to the antibody and grow normally (Boyse et aZ., 1963; Old et al., 1968). Such tumors can no longer be killed in vitro with anti-TL plus complement (C’). However, after subsequent passage into normal mice, the tumors acquire TL and again become sensitive to treatment with anti-TL + C’. This process has been termed “antigenic modulation” (Boyse et al., 1963, 1967), and it has the following properties: First, it is an active metabolic process (Old et al., 1968).Second, it can occur using Fab fragments as well as whole antibody (Lamm et al., 1968; Esmon and Little, 1976). Third, it is accompanied by capping and endocytosis, but in neither instance is the TL completely cleared from the cell surface (Stackpole et al., 1974; Yu and Cohen, 1974; Esmon and Little, 1976). Moreover, cells become refractory to C’-mediated lysis even prior to extensive capping or loss of surface TL mole-

184

ELLEN S. VITETTA AND J. DONALD CAPRA

cules from the cell (Esmon and Little, 1976). Biochemical studies indicate that TL antigens disappear more rapidly from cells undergoing modulation than from control cells. However, this disappearance is not accompanied by increases in either de nouo synthesis or release of TL-anti-TL complexes into the medium (Yu and Cohen, 1974). Taken together, these studies suggest that the binding of surface TL molecules by either univalent or divalent antibody is sufficient to inhibit subsequent C'-mediated lysis. Fab fragments probably block the binding of C'-fixing intact antibody and, thus, inhibit lysis. When intact antibody is used, sufficient capping and removal of complexes occurs to make subsequent complement fixation, by remaining TL antibody complexes, ineffective. Thus, antigenic modulation of TL appears similar in many respects to capping and modulation of other surface antigens. D. BIOCHEMISTRY OF TL ANTIGENS Three general approaches have been used to extract TL antigens from thymocytes or TL+ leukemias. These include treatment of intact cells or membrane fractions with (1)detergents (NP40 or Triton X-loo), (2) proteolytic enzymes in high concentration (papain), or (3) EDTA. The extracted T L antigens are labeled prior to extraction with 3H-labeled precursors or with lZ5Ior are radioiodinated after extraction in order to monitor further purification. Partial purification has been obtained with molecular exclusion chromatography, binding, and elution from immunoabsorbants or immunoprecipitation. Studies using such techniques have established several features of TL structure. In the case of extraction with NP40, TL appears to be a noncovalently bound dimer with subunits of M , 44,000 and 12,000 (Vitetta et al., 1972; 1975a; Ostberg et al., 1975). Since TL can be depleted from thymocyte lysates by an antiserum prepared against rat &-microglobulin, the small subunit is mouse p,-microglobulin (Anundi et al., 1975; Ostberg et al., 1975; Vitetta et al., 1976b). Biosynthetic studies indicate that the heavy chain of TL contains fucose, making the TL heavy chain a glycoprotein (Muramatsu et al., 1973). When leukemia cells are extracted with EDTA, and the TL-containing fraction is further purified on an immunoabsorbent antihuman Pz-microglobulin, the eluted material has a molecular weight of 120,000, consistent with a tetrameric structure composed of two disulfide-bonded heavy chains linked noncovalently to two &-microglobulin molecules (Anundi et al., 1975). After reduction, the heavy chains have a molecular weight of approximately 50,000, consistent with studies performed using

THE MUIUNE

1

7 CHROMOSOME ~ ~

185

NP40 (Anundi et al., 1975). Although it is not clear at this time why the molecule prepared in EDTA is a tetramer, while that isolated from NP40 lysates is a dimer, it is possible that either the EDTA or NP40containing preparations promote reduction, oxidation, or disulfide exchange. Third, T L heavy chains might aggregate under certain conditions. Extraction of several TL+ leukemia cell lines with papain has yielded fragments of M, 58,000 (Wolcott et at., 1975,38,000 (Anundi et al., 1975; Muramatsu et al., 1973),31,000 (Wolcott et al., 1975),and 19,000 (Stanton et al., 1975).Papain cleavage appears to occur distal to the heavy-chain disulfide bond, since heavy-chain dimers are not obtained with the papain procedure (Anundi et al., 1975).The fragment ofM, 19,000 may represent a T L heavy chain which is further cleaved after release from the membrane. The 58,000 fragment may be a portion of the heavy chain linked to µglobulin since it can be retained on a column of insolubilized anti-p,-microglobulin (Ostberg et al., 1975).T h e exact nature of these fragments awaits fiirther structural studies. Extensive Pronase digestion of papain-released T L results in the recovery of an M,4500 glycopeptide, which is slightly larger than the M, 3500 glycopeptide obtained from H-2 using similar procedures (Muramatsu et al., 1973)

E. CONCLUSIONS Taken together, the above studies suggest that the native structure of T L is composed of either a TL-P,-microglobulin dimer and/or a (TL),-(&-microglobulin), tetramer, which is attached to the plasma membrane, presumably by the C-terminal portion of the H chain. The native T L heavy chain has a molecular weight of 45,000, and the small chain is the murine analog of /I,-microglobulin. The two chains are noncovalently linked. Proteolytic cleavage of T L from the plasma membrane results in breakage of the molecule at several sites, yielding fragments ranging from M , 19,000-58,000. Only the 58,000 fragment is definitively associated with &-microglobulin. Some of the major questions which remain to be clarified concern (1) the dimeric vs tetrameric nature of the native molecule, (2) the attachment site of the molecule to the plasma membrane, (3) the structural relationship between the T L 1, 2, 3, and 4 specificities, i.e., whether they are on the same or different molecules, (4) the antigenic site or sites that establish TL antigenicity (carbohydrate or protein), and (5) the possible primary sequence homology between T L molecules and other products of the 17th chromosome.

186

ELLEN S. VITETTA AND J. DONALD CAPRA

IX. Perspectives

In viewing the recent progress in our understanding of the biochemistry of the products of chromosome 17 in the mouse, one is deeply impressed by the enormous contributions of classical genetics and serology to this field. Thus, the major issues were developed and the hypotheses posed by a relatively small number of geneticists, embryologists, and immunologists who sensed the potential importance of the MHC and the adjacent gene complexes. Current biochemical studies are possible because of years of painstaking genetic analysis, creation of appropriate inbred mouse strains, including congenic lines, and generation of antisera. Indeed, the present situation is reminiscent of that point in time when meaningful structural information concerning immunoglobulins became available. Surely, there will be an even greater surge of new information from the geneticists, immunologists, and developmental biologists as information about structure is disseminated to these biological disciplines, and the converse is also true. Indeed, our understanding of the genetics of chromosome 17 is in a very early stage. Perhaps the most striking feature concerning chromosome 17 at this time is that it appears to act in some ways as a “supergene.” Thus, there is close association between a number of genes and gene complexes on this chromosome. The close linkage between the I , Ss, H-2D, and H-2K genes, all of which play roles in the immune response, suggest that there are strong selective pressures to maintain this linkage. This view is reinforced by a similar linkage of the counterparts of these genes in man. Another example is the linkage disequilibrium between certain T f t and H-2 haplotypes resulting in a remarkably diminished rate of recombination between the H-2 and T/t complexes. In addition, some of these gene products appear to be reciprocally expressed during differentiation such as F9 (a Tlt product) and H-2 on embryos, TL and Qa-2 on T cells, and TL and H-2D on thymocytes. Indeed, the need for this tight regulation may explain, in part, the necessity for maintaining these gene complexes on the same chromosome. The biochemical studies have revealed a relatively monotonous subunit structure for five gene products on the chromosome including two (Tlt and T L ) which are some distance apart: namely, an M , 44,000 major subunit attached noncovalently to a 12,000 subunit, which in most, if not all, instances is encoded by another chromosome. The most likely prediction is that there will be significant structural homology among these gene products of chromosome 17, suggesting their evolution from a common ancestral gene.

THE MURINE

1

7

CHROMOSOME ~ ~

187

With considerable genetic, immunologic, and biochemical data now at hand, it is possible to discuss the possible interrelationships among the gene products in question. To begin with, all these molecules, with the exception of Ss, are expressed on the cell surface. Presumably, such molecules are acting as receptors. Clearly, some of these receptors are concerned with cell-cell interaction and recognition. The most impressive examples would b e T cell-mediated cytotoxicity of chemically modified (or virally infected) target cells in which the two cell types must be in contact, and T cell stimulation by antigenpulsed macrophages in which interactions between these two cell types precedes stimulation of T cells. In both examples, MHC products play crucial roles. In the case of Tlt, histologic evidence has been interpreted to indicate that these gene products markedly affect cell interactions during embryogenesis. Mutations of the Tlt region are associated with changes in normal patterns of cell-cell interactions resulting in abnormal differentiation and/or death of the embryo. It must be emphasized, however, that the role of MHC products may not be limited to cell-cell interactions. Thus, there is already evidence that they may act as receptors for humoral factors or, indeed, function as serum proteins. Thus, Ia-containing molecular complexes have been found in the incubation media of lymphoid cells as well as in serum. Moreover, there is one report of an Z region-determined genetic restriction for accepting a T cell suppressive factor. The Ss protein represents a very clear example of a molecule in which entrance into the circulation is probably an obligatory step for the molecule to exert its varied functions. One of the most provocative questions to emerge from the early primary sequence studies of the H-2 molecules has been the issue of genetic organization of the MHC. Thus, the relatively high proportion of amino acid differences between the presumed alleles of H-2D and H-2K reinforce the possibility posed b y Bodmer (1973) that the MHC of a single mouse strain might possess a bank of tandemly arranged genes, each representing the presumed alleles of the species with a regulatory mechanism that mimics Mendelian inheritance. Further biochemical information will bring new insight into this question, although a definitive answer probably awaits nucleic acid hybridization if appropriate probes can be developed. There is, however, a large array of important questions that can b e answered only by a combination of biochemical and functional information. Examples include the presence of variable (v) and constant (c)portions of 17th chromosome products and the relationship of antigenicity to structure, i.e., the assignment of public and private specificities, the presence and characteristics of active biological sites on the molecule, possible enzymic

188

ELLEN S. VITETTA AND J. DONALD CAPRG

activity, and the role of the carbohydrate portions in antigenicity and function. It is particularly provocative that there is a recent report that the genes for heavy chains of Ig are on chromosome 17, adding impetus to search for v-c portions of other 17th chromosome products. The one prediction that can be made with complete confidence is that the growing interest in the genetics, structure, and function of the products of chromosome 17 will continue to grow at a rapid rate. This chromosome may, indeed, hold the key to understanding embryogenesis, differentiation, and gene regulation in higher organisms.

ACKNOWLEDGMENTS The authors acknowledge the invaluable secretarial assistance of Ms. Bechtel and Ms. Hahn. We extend special appreciation to Drs. Charles E. Wilde 111, James Forman, Richard Cook, and Jonathan W. Uhr for critically reading portions of this manuscript and for many informative discussions. We also thank the many people who have allowed us to reproduce portions of their published work and our many colleagues who have collaborated with us on aspects of our own work presented in this review. The generous grant support of the National Institutes of Health and the National Science Foundation is gratefully acknowledged.

REFERENCES Amos, D. B., Gorer, P. A,, and Mikulska, Z. B. (1955).Proc. R. SOC. London, Ser. B 144, 369. Andrews, D., and Capra, J. D. (1978). Submitted for publication. Anundi, H., Rask, L., Ostberg, L., and Peterson, P. A. (1975). Biochemistry 14,5046. Appella, E., Tanigaki, N., Henriksen, O., Pressman, D., Smith, D. F., and Fairwellin, T. (1976a). Cold Spring Harbor Symp. Quant. Biol. 417,341. Appella, E., Tanigaki, N., Natori, T., and Pressman, D. (1976b).Biochem. Biophys. Res. Commun. 70,425. Armerding, D., and Katz, D. (1974).J . E x p . Med. 140, 19. Armerding, D., Sachs, D. H., and Katz, D. H. (1974).J.E x p . Med. 140,1717. Artzt, K., Dubois, P., Bennett, D., Condamine, H., Babinet, C., and Jacob, F. (1973). Proc. Natl. Acad. S c i . U.S . A . 70, 2988. Artzt, K., Bennett, D., and Jacob, F. (1974).Proc. Natl. Acad. Sci. U.S . A . 71,811. Ballou, B., McKean, D. J., Freedlander, E. F., and Smithies, O., (1976). Proc. Natl. Acad. Sci. U.S . A . 73,448. Benacerraf, B., and Katz, D. (1975).Ado. Cancer Res. 21, 121. Bennett, D. (1964).Science 144,263. Bennett, D. (1975). Genet. Res. 26,95. Bennett, D., Boyse, E. A,, and Old, L. J. (1972).In “Cell Interactions” (L. G. Silvestri, ed.), p. 247. North-Holland Publ., Amsterdam. Berggard, I. (1974). Biochem. Biophys. Res. Commun. 57, 1159. Berggard, I., and Bearn, A. G. (1968).J . Biol. Chem. 243,4095. Bemier, G. M., and Fanger, M. W. (1972).J. Zmmunol. 109,407. Bismuth, A., Neuport-Sautes, C., Kourilsky, F. M., Manuel, Y.,Greenland, T., and Silvestre, D. (1974).J . Immunol. 112,2036. Bodmer, W. (1973). Transplant. Proc. 5, 1471.

T H E MURINE

1

7

CHROMOSOME ~ ~

189

Bolotin, C., Morris, S., Tack, B., and Prahl, J. (1977). Biochemistry 16, 2008. Boyse, E. A., and Old, L. J. (1971). Trunsplantution 11, 561. Boyse, E. A., Old, L. J., and Luell, S. (1963).J.Notl. Cancer Znst. 31, 987. Boyse, E. A., Old, I,. J., and Stockert, E. (1966).Zinmunoputhol., Znt. S!ynip.,4 t h , 1965 p. 23. Boyse, E. A,, Stockert, E., and Old, L. J. (1967). Proc. Nut]. Acud. Sci. U . S . A . 58,954. Brown, J . L., Kato, K., Silver, J., and Nathenson, S. G. (1974).Biochemistry 13,3174. Buc-Caron, M. H . , Gachelin, G., Hodnung, M., and Jacob, F. (1974).Proc. Natl. Acad. Sci. U . S . A . 71, 1730. Callahan, C. N., Ferrone, S., Klein, J . , Poulik, M. D., and Reisfeld, R. A. (1976).Fed. Proc., Fed. A m . Soc. E x p . Biol. 35, 513 (abstr.). Capra, J . D., and Kehoe, J. M. (1975). Adu. Zmmunol. 20, 1. Capra, J. D., Tung, A. S., and Nisonoff, A. (1975a).J.Intmunol. 114, 1548. Capra, J. D., Vitetta, E. S., and Klein, J. (1975b).J.E x p . Med. 142,664. Capra, J. D., Vitetta, E. S., Klapper, D. G., Uhr, J. W., and Klein, J. (1976).Proc. Natl. Acad. Sci. U . S. A . 73,3661. Carroll, M., and Capra, J . D. (1978). Proc. Natl. Acad. Sci. U . S. A . (in press). Chesley, P. (1935).J . E x p . Zool. 70,429. Cook, R., Vitetta, E. S., Capra, J. D., and Uhr, J. W. (1977).Immunogetietics 5,437. Cullen, S. E., David, C. S., Shreffler, D. C., and Nathenson, S. G. (1974).Proc. Nutl. Acad. Sci. U . S . A . 71,648. Cullen, S. E., Freed, J. H., and Nathenson, S.G. (1976). Trunsplunt. Reu. 30,236. Cunningham, B. A., and Berggard, I. (1974).Transplant. Reu. 21,3. Cunningham, B. A., and Berggard, I. (1975). Science 187, 1079. Cunningham, D. A., Henning, R., Milner, R. J., Reske, K., Ziffer, J. A., and Edelman, G . M. (1976).Cold Spring Hnrhor Synip. Quatit. Biol. 41,351. Curman, B., Ostberg, L., Sandberg, L., Malmheden-Eriksson, I., Stalenheim, G., Rask, L., and Peterson, P. A. (1975).Nature (London)258, 243. David, C. S. (1976).Trunsplant. Reo. 30,299. David, C. S., Shreffler, D. C., and Frelinger, J. A. (1973).Proc. Nutl. Acad. Sci. U . S . A . 70, 2509. David, C. S., Neiderhuher, J. E., Frelinger, J. E., Dugan, E., Meo, T., and Shreffler, D. C. (1976).In “1,eukocyte Membrane Determinants Regulating Immune Reactivity” (V. P. Eijsvoogel, D. Roos, and W. P. Zeijlemaker, eds.). Academic Press, New York. David, C. S., McCormick, J. F., and Stimpfling, J. H . (1977a). Third Zr Gene Workshop (in press). David, C. S., Neely, B. C., and Cullen, S. E. (197711).Third Ir Gene Workshop (in press). Delovitch, T. L., and McDevitt, H. 0. (1975). Zmmunogenetics 2,39. Demant, P., Capkovrl, J., Hinzova, E., and Vorai.ova, B. (1973). Proc. Natl. Acad. Sci. U . S . A . 70,863. Dickler, H. B., and Sachs, D. H. (1974).J.E x p . Med. 140, 779. Doherty, P. C., Blanden, R. V., and Zinkernagel, R. M. (1976).Transplant. Reo. 29,89. Dubois, P., Fellous, M., Gachelin, G., Jacob, F., Kemler, R., Pressman, D., and Tanigaki, N. (1976).Transplantation 22, 467. Dunn, L. C., and Bennett, D. (1964).Science 144,260. Erh, P., and Feldmann, M. (1975).J.E x p . Med. 142,460. Esmon, N. L., and Little, J. R. (1976).J.Zmmunol. 117, 919. Ewenstein, B. M., Freed, J. H., Mole, L. E., and Nathenson, S. G. (1976).Proc. Nutl. Acud. Sci. U . S . A . 73, 915.

190

ELLEN S. VITETTA AND J. DONALD CAPRA

Fellous, M., Colle, A., and Tonnelle, C. (1976a).Eur. J . Zmmunol. 6, 21. Fellous, M., Erickson, R. P., Gachelin, G., and Jacob, F. (1976b). Transplantation 22, 440. Flaherty, L. (1976). Zmmunogenetics 3, 533. Forman, J. (1975).J.E x p . Med. 142,403. Frelinger, J. A., Niederhuber, J. E., David, C. S., and Shreffler, D. C. (1974)./. Erp. Med. 140, 1273. Frelinger, J. A., Neiderhuber, J. E., and Shreffler, D. C. (1975). Science 188,268. Frelinger, J. A., Niederhuber, J. E., and Shreffler, D. C. (1976).J.Exp. Med. 144, 1141. Geib, R., Poulik, M. D., Vitetta, E. S., Kearney, J. F., and Klein, J. (1976).J.Zmmunol. 117. 1532. Gigli, L’(1977).Fed. Proc., Fed. Am. Soc. E x p . Biol. (in press). Goldman, M. B., and Goldman, J. N. (1976).J.Zmmunol. 117, 1584. Goodfellow, P. N., Jones, E. A., Van Heynigen, V., Solomon, E., Bobrow, M., Miggiano, V., and Bodmer, W. F. (1975). Nature (London) 253,267. Gordon, S. M., and Kindt, T. J. (1976). Scand. J . Zmmunol. 5,310. Corer, P. A. (1936). Br. J . Exp. Pathol. 17,42. Corer, P. A. (1937).J . Pathol. Bacteriol. 44,691. Hall, C. E., and Colton, H. R. (1977). Proc. Natl. Acad. Sci. U.S . A . 74, 1707. Hammerling, G. J. (1976). Transplant. Reu. 30,4. Hammerling, G . J., Mauve, G., Goldberg, E., and McDevitt, H. 0. (1975).Zmmunogenetics 1,428. Hammerling, G. J., Black, S. J., Segal, S., and Eichmann, K. (1976). In “Leukocyte Membrane Determinants Regulating Immune Reactivity” (V. P. Eijsvougel, D. Roos, and W. P. Zeijlemaker, eds. Academic Press, New York. Hansen, T. H., and Shreffler, D. C. (1976).J . Zmmunol. 117, 1507. Hansen, T. H., Shin, H. S., and Shreffler, D. C. (1976).J.E x p . Med. 151, 1216. Hansen, T. H., Cullen, S., and Sachs, D. H. (1977)./. Exp. Med. 145,438. Hauptfeld, V., and Klein, J. ( 1 9 7 5 ) ~Exp. Med. 142,278. Hauptfeld, V., Klein, D., and Klein, J. (1973). Science 181, 167. Henning, R., Milner, R. J., Reske, K., Cunningham, B. A., and Edelman, G. M. (1976). Proc. Natl. Acad. Sci. U.S . A. 73, 118. Hess, M. (1976). Transplant. Reu. 30, 40. Hinzovi, E., Demant, P., and Ivinyi, P. (1972). Folia Biol. (Prague) 18, 237. Hutteroth, T. H., Cleve, H., Litwin, S . D., and Poulik, M. D. (1973).J.Exp. Med. 137, 838. Jackson, D. C., Parish, C. R., and McKenzie, I. F. C. (1977). Zmmunogenetics 4,267. Jacob, F. (1977).Zmmunol. Reu. 33,3. Karlsson, J. (1974). Zrnmunochemistry 11, 111. Katz, D. H., and Benacerraf, B. (1975). Transplant. Reu. 22,175. Kessler, S. W. (1975).J.Zmmunol. 115, 1617. Klein, J. (1973).In “International Symposium on HL-A Reagents” (R. H. Regamey and J. V. Sparck, eds.), p. 251. Klein, J. (1975). “Biology of the Mouse Histocompatibility-2 Complex.” SpringerVerlag, Berlin and New York. Komuro, K., Boyse, E. A., and Old, L. J. (1973).J.E x p . Med. 137,533. Konda, S., Kakao, Y.,and Smith, R. T. (1972).J.E x p . Med. 136, 1461. Konda, S., Stockert, E., and Smith, R. T. (1973). Cell. Zmmunol. 7,275. Lachmann, P. J., Grennan, D., Martin, A., and Demant, P. (1975).Nature (London) 258, 242.

THE MUFUNE

1

7

CHROMOSOME ~ ~

191

Lamm, M. E., Boyse, E. A., Old, L. J., Lisouska-Bernstein, B., and Stockert, E. (1968).]. lmmunol. 101, 99. Lemonnier, P., Neuport-Sautes, C., Kaurilsky, F. M., and Demant, P. (1975). Zmmunogenetics 2, 517. Lieberman, R., and Humphrey, W. J. (1971).Proc. Natl. Acad. Sci. U . S.A . 68, 2510. Lieberman, R., and Humphrey, W. J . (1972).J . E x p . Med. 136, 1222. Lozner, E. C., Sachs, D. H., Shearer, G. M., and Terry, W. D. (1974).Science 183,757. McDevitt, H . 0.(1976).In “The Role of Products of the Histocompatibility Gene Complex in Immune Responses” (U. H. Katz and B. Benacerraf, eds.). Academic Press, New York. McDevitt, H. O., and Sela, M. (1967).J.E x p . Med. 126,969. McDevitt, H. 0.. Delovitch, T. L., Press, J. L., and Murphy, D. B. (1976).Transplant. Rev. 30, 197. McMillan, M., Cecka, J. M., Murphy, D. B., and McDevitt. H. 0. (1977). Proc. Natl. Acad. Sci. U . S . A . 74,5135. Melchers, I. K., Rajewsky, K., and Shreffler, D. C. (1973).Eur. J. Immunol. 3, 754. Meo, T., David, C. S., Rijnheck, A. M., Nabholz, M., Miggiano, V., and Shreffler, D. C. (1975a).Transplant. Proc. 7, 127. Meo, T., Krasteff, T., and Shreffler, D. C. (19751)).Proc. Natl. Acad. Sci. U . S . A . 72, 4536. Meo, T., David, C. S., and Shreffler, D. C. (1976).Z n “The Role of Products of the Histocompatibility Gene Complex in Immune Responses” (B. Benacerraf and D. Katz, eds.). Academic Press, New York. Merryman, C. F., and Maurer, P. H . (1974).lmmunogenetics 1, 549. Michaelson, J., Flaherty, L., Vitetta, E. S., and Poulik, M. D. (1977).J . E x p . Med. 145, 1066. Muraniatsu, T., Nathenson, S . G., Boyse, E. A,, and Old, L. J. (1973).J.E x p . Med. 137, 1256. Murphy, D. B., Herzenherg, L. A,, Okumura, K., Herzenberg, L. A., and McDevitt, H. 0. (1976).J.E x p . Med. 144,699. Nathenson, S. G. (1970).Annu. Reo. Genet. 4,69. Nathenson, S. G., and Cullen, S . E. (1974).Biochim. Biophys. Acta 344, 1. Natori, T., Katagiri, M., Tanigaki, N., and Pressman, D. (1974).Transplantation 18,550. Natori, T., Tanigaki, N., Appella, E., and Pressman, D. (1975).Biochem. Biophys. Res. Commun. 65,611. Nicolas, J. F., Dubois, P., Jakob, H., Gaillard, J., and Jacob, F. (1975).Ann. Microhiol. (Paris) 126a, 3. Okurnura, K., Herzenherg, L. A., Murphy, D. B., McDevitt, H. O., and Herzenberg, L. A. (1976).J.E x p . Med. 144,685. Old, L. J., Boyse, E. A,, and Stockert, E. (1963).J.Natl. Cancer Znst. 31,977. Old, L. J., Stockert, E., Boyse, E. A,, and Kim, J. H. (1968).J.E x p . Med. 127,523. Ostberg, L., Rask, L., Wigzell, H., and Peterson, P. A. (1975).Nature (London)253,735. Painter, R. H., Yasmeen, D., Assimeth, S., Isenman, D. E., and Poulik, M. D. (1974). lmmunol. Commun. 3, 19. Parish, C. R., Chilcott, A. B., and McKenzie, I. F. C. (1976a).lmmunogerietics 3, 129. Parish, C. R., Jackson, D. C., and McKenzie, I. F. C. (1976b).lmmunogenetics 3,455. Paul, W. E., Shevaeh, E. M., Pickeral, S . , Thomas, D. W., and Rosenthal, A. S. (1977).J. E x p . Med. 145, 618. Peterson, P. A,, Cunningham, B. A., Berggard, I., and Edelman, G. M. (1972).Proc. Natl. Acad. Sci. U . S . A. 71, 1967.

192

ELLEN S. VITETTA AND J. DONALD CAPRA

Peterson, P. A., Rask, L., Sege, K., Klareskog, L., Anundi, H., and Ostberg, L. (1975). Proc. Natl. Acad. Sci. U . S. A. 72, 1612. Pierce, S. K., and Klinman, N. R. (1975).J.E x p . Med. 142, 1165. Poulik, M. D. (1976).In “Progress in Clinical and Biological Research” (G. A. Jamieson and T. J. Greenwalt, eds.), pp. 155-177. Alan R. Liss, Inc., New York. Poulik, M. D., Shinnick, C. J., and Smithies, 0. (1977). In press. Press, J. L., Klinman, N. R., and McDevitt, H. 0. (1976).J.E x p . Med. 144,414. Sachs, D. H., and Cone, J . L. (1973).J.E x p . Med. 138, 1289. Sachs, D. H., Berzofsky, J. A., Fathman, C. G., Pisetsky, D. S., Schechter, A. N., and Schwartz, R. H. (1977). Cold Spring Harbor Symp. Quant. Biol. 41,295. Saunders, D., and Edidin, M. (1974).J.Zmmunol. 112,2210. Schwartz, B. D., and Nathenson, S. G. (1971).J.Zmmunol. 107, 1363. Schwartz, B. D., Kato, K., Cullen, S. E., and Nathenson, S. G. (1973).Biochemistry 12, 2157. Schwartz, B. D., Paul, W. E., and Shevach, E. M. (1976a). Transplant. Reu. 30, 174. Schwartz, B. D., Kask, A. M., and Shevach, E. M. (1976b). Cold Spring Harbor Symp. Quant. Biol. 41,397. Schwartz, B. D., Vitetta, E. S., and Cullen, S. E. (1978).J.Zmmunol. 120,671. Schwartz, R. H., David, C. S., Sachs, D. H., and Paul, W. E. (1976).J.Zmrnunol. 117,531. Shearer, G . M., Rehn, T. G., and Garbarino, C. A. (1975).J.E x p . Med. 141, 1348. Shevach, E., Paul, W. E., and Green, I. (1972).J.E x p . Med. 136, 1207. Shreffler, D. C. (1971). In “Immunogenetics of the H-2 System” (A. Lengerovi and M. VojtiSkova, eds.), p. 138, Karger, Basel. Shreffler, D. C. (1976). Transplant. Reo. 32, 149. Shreffler, D. C., and David, C. S. (1975). Adu. Zmmunol. 20, 125. Shreffler, D. C., and Owen, R. D. (1963).In “Immunogenetics of the H-2 System” (A. Lengerova and M. VojtiSkovii, eds.), p. 58. Karger, Basel. Silver, J. (1976).Cold Spring Harbor Symp. Quant. Biol. 41,406. Silver, J., and Hood, L. (1974).Nature (London)249,764. Silver, J., and Hood, L. (1975). Nature (London) 256,63. Silver, J., and Hood, L. (1976a). In “The Role of Products of the Histocompatibility Gene Complex in Immune Responses” (D. H. Katz and B. Benacerraf, eds.), p. 677. Academic Press, New York. Silver, J., and Hood, L. (1976b).Proc. Natl. Acad. Sci. U . S. A. 73,599. Silver, J., Checka, J. M., McMillan, M., and Hood, L. (1976).Cold Spring Harbor Symp. Quant. 41,369. Smith, M., Gold, P., Shuster, J., Tanigaki, N., and Pressman, D. (1976).J.Immunogenet. 3, 105. Smithies, O., and Poulik, M. D. (1972). Proc. Natl. Acad. Sci. U . S. A. 69,2914. Snary, D., Barnstable, C., Bodmer, W. F., Goodfellow, P., and Crumpton, M. J. (1976). Cold Spring Harbor Symp. Quant. Biol. 41,379. Snell, G. C. (1968).Folia Biol. (Prague) 14,335. Spiegelman, M., and Bennett, D. (1974).J.Embryol. E x p . Morphol. 32,723. Springer, P., Kaufman, J . F., Siddoway, L. A., Gephalt, M., Mann, D. L., Terhorst, C., and Strominger, J. L. (1976). Cold Spring Harbor Symp. Quant. 41,387. Stackpole, C. W., Jacob, J. B., and Lardis, M. P. (1974).]. E x p . Med. 140,939. Stanton, T. H., and Boyse, E. A. (1976).Zmmunogenetics 3,525. Stanton, T. H., Bennett, J. C., and Wolcott, M. J. (1975).J.Zmmunol. 115, 1013. Tada, T., Tanigaki, N., and Pressman, D. (1978).J.Zrnmunol. 123,513. Tada, T., and Taniguchi, J . (1976).In “The Role of Products of the Histocompatibility

THE MURINE

1

7

CHROMOSOME ~ ~

193

Gene Complex in Immune Responsers” (D. H. Katz and B. Benacerraf, eds.), p. 513. Academic Press, New York. Taussig, M. J., Munro, A. J., Campbell, R., David, C. S., and Staines, N. A. (1975).J. E x p . Med. 142,694. Terhorst, C., Parham, P., Mann, D. L., and Strominger, J. L. (1976). Proc. Natl. Acad. Sci. U . S. A. 73,910. Thomas, D. W., and Shevach, E. M. (1976).J. E x p . Med. 144, 1263. Thomas, K., Engers, H. D., Cerottini, J. C., and Brunner, K. T. (1976). E u r . ] . Imrnunol. 6, 257. Tung, A. S., Ju, S.-T., Sato, S., and Nisonoff, A. (1976).]. Zmmunol. 116,676. Vitetta, E. S., Uhr, J. W., and Boyse, E. A. (1972). Cell. Zmmunol. 4, 187. Vitetta, E. S., Klein, J., and Uhr, J. W. (1974). Zrnmunogenetics 1,82. Vitetta, E. S., Uhr, J. W., and Boyse, E. A. (19754. j . Irnrnunol. 114,252. Vitetta, E. S., Artzt, K., Bennett, D., Boyse, E. A., and Jacob, F. (1975b). Proc. Nmtl. Acad. Sci. U . S . A . 72, 3215. Vitetta, E. S., Capra, J. D., Klapper, D. G., Klein, J., and Uhr, J. W. (1976a).Proc. Nntl. Acad. Sci. U . S . A . 73,905. Vitetta, E. S., Poulik, M. D., Klein, J., and Uhr, J. W. (1976b).]. E x p . Med. 144, 179. Vitetta, E. S., Cook, R., Artzt, K., Poulik, M. D., and Uhr, J. W. (1977a).E u r . J . Zmnwnol. 7,826. Vitetta, E., Uhr, J. W., Klein, J., Pazderka, F., Moticka, E., Ruth, R. F., and Capra, J. D. (1977b). Nature (London)270,535. Wagner, H. D., Cotze, D., Ptschelinew, L., and Rollinghoff, M. (1975).J. Ewp. Med. 142, 1477.. Wilde, C. E., and Capra, J. D. (1978). Submitted for publication. Wilde, C. E., Vitetta, E. S., Andrews, D. W., Klein, J., and Capra, J. D. (1978).Submitted for publication. Winkler, M. A,, and Sanders, B. G. (1977). Zmrnunochemistry 14,615. Wolcott, M., Stanton, T. H., Williams, J. L., and Bennett, J. C. (1975). Biochemistry 14, 4792. Yamane, K., Shimada, A,, and Nathenson, S. G. (1972). Biochemistry 11,2393. Yu, A., and Cohen, E. P. (1974).j. Zrnrnunol. 112, 1296.

This Page Intentionally Left Blank

ADVANCES IN IMMUNOLOGY, VOL. 26

Expression a n d Function of Idiotypes on lymphocytes K. EICHMANN lnsfitut fur lmmunologie und Genetik, Deutsches Krebrforrchungrzentrum, Heidelberg,

Federal Republic of Germany

195 I. Introduction ................................................................................... 199 11. Anti-idiotypi .................................... 199 ties ................................. A. Anti-idiotypic Reagents to Idiotypic Sulx 205 B. Preparation of Isogeneic Anti-idiotypic R ...................... ace Receptors ... 209 C. Preparation of Anti-idiotyyic Reagents agains 211 111. Analysis of €3- and T-cell Receptor Idiotypes ................................................... 211 A. Analysis of B-Cell Receptor Idiotypes ................................... 217 B. Analysis of T-cell Receptor Idiotypes ................................... 234 IV. Functional Role of Lymphocyte Receptor Idiotypes ........................................ 235 A. Recognition of‘Idiotypes within the Immune System .......... 24 1 B. Consequences of Recognition of Idiotypes within the Imm 247 V. Summary ............................................................... 248 ........................................................ References

I. Introduction

The interaction ofthe immune system with antigens in the environment occurs via the variable domains of receptor molecules on lymphocytes. In addition to a binding site for antigen, each variable domain carries a particular set of antigenic determinants, the combination of which is now temied idiot!ype (Id) which is recognized b y a particular set of anti-icliotrypic antihodies. Perhaps the first to detect such idiotypic determinants on a human myeloma protein were Lohss et al. (1953), who reported that “in two out of three antisera to myeloma-y-globulin fractions a small proportion of remaining antibodies were demonstrated which couId be saturated neither with normal yglobulin nor with total serum.’’ The antigenic individuality of human myeloma proteins was systematically demonstrated by Slater et a1 . (1955),and the concept was thereafter extended to a series of induced antibodies in human (Kunkel et a1 ., 1963) and in rabbit experimental models (Oudin and Michel, 1963; Oudin, 1966). It became clear even 195 Copyrighto 1978 Iiy Academic Press, Inc. All nghh 01 reproduction in any form reserved. ISBN 0-12.022426-7

196

K. EICHMANN

from this early work that the idiotype of an antibody or myeloma protein is an expression of its unique structure in the variable domain. Over a number of years idiotypes (or “individual antigenic specificities’’) were studied with respect to immunochemical questions such as their association with one or both polypeptide chains, with the variable or constant regions, and their relationship to the antibody combining site. These aspects have been extensively reviewed by Hopper and Nisonoff (1971) and by Capra and Kehoe (1975). It may therefore suffice here to point out that idiotypic specificities have been found exclusively on the Fvpiece (Wells et a1 ., 1973; Inbar et a1 ., 1972), that either of the two V regions may exclusively, preferentially, or equally contribute to the complement of idiotypic determinants (Grey et a1 ., 1965; Hurez et ul., 1968; Seligmann and Mihaesco, 1967; Wang et a1 ., 1970; Eichmann, 1977a; Krawinkel et al., 1977a), and that some but not all idiotypic determinants are intimately associated with the antigen-binding site (Brient and Nisonoff, 1970; Brient et a1 ., 1971; Claflin and Davie, 1974a,b,1975; s h e r and Cohn, 1972a; Sheret a1 ., 1971; Briles and Krause, 1974; Carson and Weigert, 1973). Another series of interesting studies related the expression of idiotypes to the question of antibody diversity. This work utilized antiidiotypic reagents of exquisite specificity to search for the occurrence of immunoglobulin molecules with a given set of idiotypic determinants (i.e., that of a myeloma protein) within normal heterogeneous immunoglobulin (Kunkel, 1970; Pawlak et ul., 1973b) In most instances, such molecules were either not found or incredibly rare. From these studies it has been suggested that, under the assumption that the myeloma proteins in question did not possess exceptionally rare idiotypes and that most idiotypes are expressed in about equal proportions, these proportions reflect that of each single molecular species of immunoglobulin molecules, thus pointing to a repertoire of the order of ten million to thirty million (Kunkel, 1970). In all these studies, idiotypes were essentially considered under the dogma of their “individuality,” such that idiotypic cross-reactions between the antibodies of different individuals, undoubtedly observed, were either not considered or attributed to nonspecificity of reagents. It was not until a number of reports on cross-idiotypic specificity had appeared, that a new interest in idiotypy as a tool for studying the immune system was generated. Cross-idiotypic specificity was observed between human cold agglutinins of different individuals (Williams et a1 ., 1968; Feizi e t a1 ., 1971; Feizi and Kabat, 1972), between a mouse myeloma protein with specificity for phosphorylcholine and antibodies induced in mice with this antigen (Cohn et ul.,

IDIOTYPES ON LYMPHOCYTES

197

1969), among antibodies to streptococcal carbohydrates from different ralhits (Braun and Krause, 1968), and between human anti-y-globulins (Kunkel et al., 1973). These cross-reactions between the antibodies of different individuals opened the possibility of assessing such problems a s the genetics of antibody V regions using idiotypes as probes. Furthermore, the possibility of using an anti-idiotypic reagent for the analysis of immunoglobulin molecules of more than one individual suggested the use of these reagents for the study of lymphocyte receptors for antigen. Idiotypic cross-reactions between immunoglobulin molecules from more than one source also promoted a series of studies on the stnictural requirements for the expression ofidiotypic determinants. These studies have been extensively reviewed by Capra and Kehoe (1975), and it may suffice here to point out that structural identity is not required for two Ig molecules to share idiotypic determinants. Extensive structural similarity is required in one of the two V-region polypeptides and, in particular, in its hypervariable regions (Carson and Weigert, 1973; Capra et d., 1972, 1973, 1974, 1975, 1976; Capra and Kehoe, 1975). Thus, one anti-idiotypic reagent identifies a group of structurally closely related inimunogloliulin molecules. One of the domains of idiotype research in recent years has been the use of anti-idiotypic antibodies as tools for the analysis of the genetic control of antibody V regions, and this field has also been extensively reviewed (Weigert et nl., 1975; Eichmann, 1975b; Makeli and Karjalainen, 1977a). It is in this area that the term “idiotype” is used not only to define a given set of antigenic determinants recognized by a particular antiserum, h i t also to descrilie a population of antibody molecules reactive with such an antiserum (Mikela and Karjalainen, 1977a,li), and, in addition, to describe populations of antibody molecules that share certain specificity characteristics without necessarily being defined by an anti-idiotypic reagent. In the mouse, the expression of idiotypes thus defined is controlled b y genes, most of which have been found in close linkage to the genes that control the constant regions (CH) of the heavy chain (Rlomberg et al., 1972; Eichmann, 1974; 1972, 1973; Eichmann and Berek, 1973, 1974; Eichmann et d., Imanishi and Makela, 1973, 1974, 1975; Lieberman et al., 1974; Pawlak et al., 1973a; Riblet et t r l . , 1975a,h; Berek et a1 ., 1976). These genes are therefore thought to be V,, genes that control the structure of the heavy-chain variable region and the chromosomal region containing the C, and V, genes is now referred to a s Zg-1 complex ( Herzenberg et a l . , 1968; Potter and Lieberman, 1967). At least eleven such VHgenes (Mikela and Karjalainen, 1977a) are known, but their assignment to

198

K. EICHMANN

distinct genetic loci in the Zg-l complex, at first thought to be feasible (Eichmann, 1975b), became increasingly difficult as it was found that different allelic Zg-l complexes carried VH genes controlling antibodies of one specificity at clearly different chromosomal locations (Berek et al., 1976). Nevertheless, much valuable information has come from these studies, such as the formal confirmation of the twogene hypothesis (Eichmann et a1 ., 1974) and of the discontinuous arrangement of V and C genes in the germ line (review in Eichmann, 1975b) In addition, it was observed that some V genes may undergo genetic recombination to the C genes much more frequently than others, a finding whose interpretation with respect to chromosomal distances and with respect to the size of the V-gene cluster is still a matter of much discussion (Riblet et al., 1975a,b; Krawinkel et al., 1977a; Pisetsky et a1 ., 1977; Eichmann et a1 ., 1977a). Most important for this review, the genetic control of idiotype expression has provided a means to define a molecule that is reactive with anti-idiotypic antibody not only by mere serological parameters; the definition by linkage studies of the genes controlling such a molecule raises the significance of the results to a level far beyond that of simple serology. As all these aspects of idiotypy have been reviewed in great detail (Kunkel, 1970; Hopper and Nisonoff, 1971; Oudin, 1974; Capra and Kehoe, 1975; Weigert et a1 ., 1975; Eichmann, 1975b; Makela and Karjalainen, 1977a) the present paper will restrict itself to a very recent area of idiotype research in which the idiotypes of cell surface receptors are studied. This field of research has basically two aspects, one dealing with the molecular nature and identity of receptor molecules, and the other concerning their possible function as the key elements of regulation within a network of immunocytes that are interlocked to one another through idiotype-anti-idiotype interactions. The former will be the main subject of this article, but the latter will also be discussed, particularly since it offers a totally new view of the immune system, that could resolve many of the unanswered questions on its true function and functioning. This article is divided into three major sections (11-IV): Section I1 summarizes a series of'new procedures for the production of anti-idiotypic reagents, avoiding any repetition of the more conventional procedures previously summarized in other reviews (Hopper and Nisonoff, 1971; Kunkel, 1970; Oudin, 1974). Section I11 deals with the serological analysis of antigen receptors on lymphocytes, and Section IV discusses the functional aspects of lymphocyte receptor idiotypes within an immune network. It should be stressed that the area to be covered by this chapter not only is a huge one, but also is rapidly ex-

IDIOTYPES ON LYMPHOCYTES

199

panding. It is therefore impossible to present a totally comprehensive overview, and the collection of data summarized is heavily influenced by the author’s bias. II. Anti-idiotypic Reagents

A host of different new procedures for the production of anti-idiotypic reagents have recently been described in the literature in a way that the subdivision into “heterologous,” “homologous,” and “isologous” reagents, as was agreed upon at the 1st International Congress of Immunology (Potter and Kunkel, 1971) is no longer satisfactory. Most workers in the field appear to agree now on adopting the principles of the genetic nomenclature used in transplantation immunology and differentiate between “xenoantisera” or “heteroantisera,” “alloantisera,” and “isoantisera,” depending on whether the immunization occurred across species or strain barriers or within an inbred strain. For anti-idiotypic reagents, these three categories have to be extended to a total of four b y adding the category “autoantisera,” which are produced, spontaneously or by deliberate autoimmunization, in the same individual. In addition, a number of new techniques have been described recently for the production of anti-idiotypic reagents in all four categories that allow the restriction of the specificity of these reagents to either the combining site, to regions outside the combining site, or to idiotypic determinants contributed b y either the light-chain or the heavy-chain variable region. Such antisera are best referred to in terms denoting their specificity rather than their source. Furthermore, techniques have been developed to produce anti-Id antisera directly to cell-bound receptors. In some cases, reagents produced by such new techniques have already been applied to the analysis of lymphocyte receptors, and in others this will undoubtedly be done in the near future. Moreover, particularly the production of iso- and auto-antiidiotypic reagents is of special interest for the putative regulatory function of cell surface receptor idiotypes. Therefore, and because it has not been done b y others, these new techniques for the production of anti-idiotypic reagents are summarized in this section. A. ANTI-IDIOTYPICREAGENTS TO IDIOTWIC SUBSPECIFICITIES

1 , Combining Site-Specijic Anti-idiotype Antisera raised in rabbits to human or mouse immunoglobulins possess, after proper absorption with pooled Ig, a mixture of anti-idio-

200

K. EICHMANN

typic (anti-Id) antibodies reactive with binding site-associated and nonbinding site-associated determinants (reviewed by Hopper and Nisonoff, 1971).Recently, an elegant technique has been described to isolate from such conventional anti-Id reagents the fraction of antibodies reactive with the combining-site region (Claflin et d.,1974b; Claflin and Davie, 1974a,b, 1975).Rabbits are immunized with mouse myeloma protein HOPC 8, which is one of a series of mouse myelomas reactive with phosphorylcholine (PC) (Potter and Lieberman, 1970; Potter et al., 1973; Potter, 1971). The resulting antiserum contains, in addition to anti-Id antibodies, those recognizing class- and type-specific determinants on the HOPC 8 protein (IgAK).The rabbit antiserum was then passed over an immunoabsorbent column made of Sepharose 4B and the HOPC 8 protein, which retained virtually all antibody present in the antiserum. Subsequently, a lop3M PC solution was passed over the column in order to compete with those antiId antibodies that are sensitive to the occupation of the binding site of the HOPC 8 protein. The authors eluted around 100 p g of antibody protein per milliliter of antiserum, which differed from the anti-Id in an alloantiserum raised in strain A mice to HOPC 8 by (a) complete blockage by PC, (b) inability to bind myeloma protein TEPC 15 after it has been affinity-labeled with PC, (c) by failure to distinguish between a series of PC-reactive myeloma proteins, and (d),by failure to distinguish induced antibodies to PC in a variety of inbred strains that have been claimed to possess distinct idiotypes as detected by alloanti-Id antibody (Sher and Cohn, 197213; Lieberman et al., 1974). In the meantime, reports have appeared that suggest similar idiotypes of anti-PC antibodies in all strains of mice, using even nonfractionated anti-Id (Rudikoff and Claflin, 1976). It should be mentioned in this context that antigenic determinants on immunoglobulin molecules that disappear after combination with antigen are not necessarily idiotypic. Kunkel et a l . (1976)showed that, in addition to idiotypic determinants, subgroup-related antigens were blocked on monovalent F(ab) fragments from anti-Rh antibodies bound to red cells. The authors discussed the possibility that these antigens were also close to the combining site and/or that combination with a large antigen, such as a red blood cell, has effects on the exposure of antigenic determinants that are different from those of binding a haptenic group. Conversely, it is obvious that small haptens may leave some of the site-associated idiotypic determinants unblocked so that they may erroneously appear as non-site associated. Therefore, it has to be borne in mind that the terms “site-associated” or “non-site associated” refer

IDIOTYPES ON LYMPHOCYTES

20 1

to operational definitions valid in special cases but not necessarily in general. It is clear, however, that such idiotypic determinants are of a more defined character than are those identified by conventional antiId reagents. With respect to the analysis of lymphocyte receptors, antisera with this restricted anti-Id specificity should b e extremely useful.

2. Noiiconibining Site-Specijic Anti-idiot!lpe Some conventionally raised anti-Id antisera detect essentially nonbinding site-associated antigenic determinants. T h e most prominent example is the alloantiserum raised in strain A mice to the BALB/c myeloma protein TEPC 15 (T15)which is very poorly inhibited by addition of PC, the relevant hapten, to the reaction (Potter and Lieberman, 1970; Sher and Cohn, 1972a; Claflin and Davie, 1974a,b, 1975). Other such preferentially non-site-associated idiotype recognizing antisera have been obtained in the early work on idiotypy, in which antigen-antibody precipitates or bacterial agglutinates were used for immunization (Kelus and Gell, 1968; Oudin and Michel, 1963, 1969a,b; Oudin, 1966).Such anti-Id reacted with the antibody even in bacterial agglutinates, suggesting that non-site-associated determinants were preferentially detected (Kelus and Gell, 1968). Recently, several systems have been described to deliberately produce anti-Id directed toward non-site-associated idiotypic determinants. In one series of experiments, rabbits were immunized with antibodies to 1,-azobenzoate from other rabbits after incubation with excess rabbit IgG-P-azobenzoate complex. Nearly all anti-Id antisera, in contrast to such antisera formed in response to free antibenzoate antibody, were not inhibited by the benzoate hapten (SpringStewart and Nisonoff, 1973). As a further development in this direction, JBrgenson et al. (1977) produced isoantisera b y immunization of BALBlc mice with the BALB/c myeloma protein MOPC 315, which has specificity for the DNP ligand. One group of animals received MOPC 315 protein after it had been affinity labeled with N-bromacetyl-N-DNP-L ligand (BADL) (JBrgenson et al., 1977). The antiserum thus obtained recognized determinants on the 315 F v fragment (Wells et a1 ., 1973)outside the DNP binding site, which were nevertheless generated b y combination of the variable regions of both heavy ( H ) and light (L) chain (Table I). In contrast, mice immunized with native MOPC 315 protein produced anti-Id, which after short immunization was almost exclusively binding-site directed, and after prolonged immunization still contained 75% binding site-specific antibodies (JBrgenson et at., 1977). These latter data are in agreement with those of Helman et a l .

202

K. EICHMANN

TABLE I REACTIVITY OF BALB/C ANTX-IDIOTYPIC ANTISERARAISEDAGAINST NATIVEOR AFFINITY-LABELED MOPC 315 FOR NATIVE AND AFFINITY-LABELED MOPC 315 AND ITS ISOLATED H AND L CHAINSn Antigens Antiserum

M 315

M 315 (aff.)

Anti M 315* Anti M 315 (aff.)' 1 2 3

73

0

0

n.t.

92 93 84

83 86 79

10 1 12

3 1 0

L 315

H 315

Data from Jbrgenson et al. (1977). Fijpmes represent percent binding of radiolaheled antigens using an antiserum ohtained after short immunization of BALBlc mice with native M 315 protein. Figures represent percent inhihition of the binding of radiolaheled native M 315 to three different antisera against affinity-labeled M 315 (aff.).

(1976) on BALB/c anti-Id antibodies to MOPC 315, which were found to be 60-70% site-specific, whereas rabbit antibodies prepared to the Fv fragment of MOPC 315 were more than 90% non-site specific. It is quite clear from these data that, although in some cases antisera raised by conventional methods are found to possess restricted specificities, in general the production of such antisera requires the use of special methods as described above.

3 . Anti-idiotype with Specijicity f o r V Hund V LDeterminunts A number of studies have been carried out investigating the question whether idiotypic determinants expressed on native immunoglobulin molecules would b e found on isolated H or L chains (Grey et ul., 1965; Brient and Nisonoff, 1970; Hurez et ul., 1968; Seligmann and Mihaescu, 1967; Wang et ul., 1970).These studies have been extensively reviewed by Hopper and Nisonoff (1971). In essence, it became clear from these studies that isolated H and L chains rarely express idiotypic determinants of native molecules, although several A-type human proteins were observed in which isolated chains carried either all or some of the idiotypic determinants (Grey et al., 1965; Hurez et al., 1968; Seligmann and Mihaescu, 1967).In another study in which the isolated chains of a human myeloma protein were recombined to heterologous partners, it was found that some of the idiotypic determinants were maintained in the molecule containing the homologous H chains whereas no activity was found in connection with the L chain (Wang et ul., 1970).

203

IDIOTYPES ON LYMPHOCYTES

In a similar study, anti-Id antisera produced in guinea pigs against

the mouse antibody A5A (Eichmann, 1972, 1973; see Section II1,B) were analyzed for their reactivity with either L- or H-chain variable regions (Krawinkel et nl., 1977a; Eichmann, 1977a). The production of anti-Id antisera in guinea pigs is based on the observation that an intravenous injection of 5 mg of ultracentrifuged pooled normal immunoglobulin renders these animals tolerant to most nonidiotypic determinants of an antibody, 100 p g of which are given subcutaneously in complete Freund's adjuvant at the same time (Eichmann and Kindt, 1971; Henney and Ishisaka, 1969; Eichmann, 1972,1973; Bereket al., 1976; Miikelii and Karjalainen, 1977a,b; Haimovich, 1977; Jack et nl.,

1977). Four different antisera thus obtained against the same antibody (A5A) were analyzed for their reactivity with either L chain- or H chain-associated idiotypic determinants (Krawinkel et a2 ., 1977a). Artificial molecules were reconstructed from the H and L chains from the A5A antibody and from pooled mouse IgG (MIg) such that two homologous and two heterologous recombinant molecules were obtained. The reactivity of the four conventionally raised antisera with native A5A and MIg and with the four types of recombinant molecules is illustrated in Table 11, which shows that two (1and 2) of the antisera preferentially react with A5A H chains whereas the other two (3and 4) TABLE I1 REACTIVITYOF GUINEAPIG ANTI-ASA ANTI-IDIOTYPIC ANTISERAWITH HOMOLOGOUS AND HETEROLOGOUS CHAIN RECOMBINANTS FROM A5A AND POOLED MOUSEIGG (MIG)" Amount (pg/ml) bound to anti-A5Ad A5A

Mk

Native Native

HLh HL L H

H L

1

2

3

4

5

152' <1 48' 2 27 2

135

210 <1 63 3 12 45

175

225 <1 195 <1 171

<1

37 <1 18 <1

26

Percenr' bound to anti-MIg 91 94 58 77 61 69

Data from Krawinkel et a!. (1977a) and Eichmann (1977a). Preparation of chain recoinbiiiants described i n Krawinkel et al. (1977a). ' Native molecules and heterologous a s well a s homologous chain recombinants were radioiodinated, and the binding capacity of the 5 different antisera was determined by radioimmune assay as descril>ed (Eichmann, 1973). Antisera 1-4 were prepared against native A5A; antiserum 5 was prepared against A5AH-MIgL recombined molecule. "

204

K. EICHMANN

preferentially react with A5A L chains. In the cases of antisera 1-4, however, the reactivity is preferential, not exclusive, so that, for example, antiserum 3 reacts four times more strongly with molecules containing the A5A H chain than with those containing the A5A L chain (Krawinkel et a1 ., 1977a) The production of antisera exclusively reacting with the idiotypic determinants on one chain required the use of artificially recombined molecules as immunogens in guinea pigs (Eichmann, 1977a). Using the combination A5A H-MIgL, an antiserum was obtained that was exclusively reactive with A5A V, determinants (5 in Table 11) (Eichmann, 1977a). It is surprising to see that the antiserum did not distinguish between A5AH-A5AL and A5AH-MIgL, indicating that the tolerizing injection of MIg was sufficient to render the animal tolerant to all L-chain determinants in the artificial immunogenic antibody molecule. It is worth noting that none of the heterologous recombinant molecules used in these studies had detectable antigen-binding activity toward group A streptococcal (Strep.A) carbohydrate (K. Eichmann, unpublished), the antigen against which the original A5A antibody is directed (Eichmann, 1972, 1973). Nevertheless, there is appreciable idiotypic activity maintained suggesting that a given idiotypic determinant or set of determinants is not exclusively associated with the antibody population reacting with the original antigen. Observations pointing in a similar direction have been made by Oudin and Cazenave (1971)and Cazenave et al. (1974), who found that immunoglobulin and lymphocyte populations unreactive with antigen reacted with an anti-Id serum raised against an antibody isolated from the same rabbit’s serum, and by Eichmann et a l . (1977b), who found A5A idiotype secreting B cell clones that did not bind to Strep.A particles (see Section IV,B). Another, similar, method for the preparation of anti-Id with exclusive H-chain specificity was employed by Yarmush et a l . (1977). In these experiments, the H chain obtained from a homogeneous rabbit antibody to group C streptococcal carbohydrate (Krause, 1970)was recombined with L chains isolated from a heterogeneous population of antibodies with the same specificity, obtained from another rabbit. The artificial molecule was then injected into the rabbit, which contributed the L chains and, accordingly, recognized only the idiotypic determinants of the homogeneous H chain from the other rabbit, as shown by the inhibition data in Table 111 (Yarmush et d., 1977; Sogn et al., 1977). The same group of authors (Sogn et al., 1976) also produced anti-Id

20s

IDIOTYPES ON LYMPHOCYTES

TABLE 111 HOMOLOGOUS IDIOTYPIC REACTIONBETWEEN NATIVE RECOMBINED RABBIT ANTI-STREPTOCOCCAL ANTIBODY MOLECULES AND THE HEAVYCHAINSPECIFIC ANTI-IDIOTYPIC ANTISERUM

INHIBITION OF THE AND

BY VARIOUS INHIBITORS“

Idiotype H-L

H recomb”

Inhibitor

H -L H recomb” L recombb L H-L L recombb

L

Concentration (mgiml)

Percent inhibition

1.0 1.0 1.0 4.0 1.0 1.0 4.0

95 91 4 2 96 3 4

Data from Yarmush et u1. (1977). Recom1)ined molecules prepared from the hoinoloyous heavy (H) or light (L) chain and heterologous H or L chains, respectively, from allotypically matched rabbit.

antibodies specific for L chains of rabbit antibodies to streptococcal carbohydrate. In this case, a conventional alloantiserum was passed over an immunoadsorbent column made from Sepharose, and the isolated L chain from the antibody was used to produce the antiserum. L-chain-specific anti-Id antibody was then eluted from this column using 3 M NH,SCN (Sogn et al., 1976). In summary, a whole new set of techniques has recently been developed that allows the production of anti-Id reagents possessing specificity toward much more defined areas within the V domain of immunoglobulin molecules. In some cases, such antisera have already been employed in the analysis of lymphocyte receptors (Krawinkel et al., 1977a; Eichmann, 1977a), and it is clear that future work utilizing anti-Id as analytical reagents will largely be based on antisera of this type.

B. PREPARATIONOF ISOGENEIC ANTI-IDIOTYPICREAGENTS The preparation of isogeneic anti-Id reagents is in itself not completely understood, a s it seems to contradict the rules of self-tolerance. It has, however, been demonstrated in all reported instances in which isoimmunization has been attempted that anti-Id antibodies can he obtained in animals of the same inbred strain and even in the very same animal that produced the immunizing immunoglobulin (Sirisinha and Eisen, 1971; Eisen et al., 1975; Sakato and Eisen, 1975; Cosenza, 197s; Cosenza et d.,1977a; Kluskens and Kohler, 1974; McKearn, 1974; McKearn et al., 1974a,b; Rodkey, 1974, 1976; Eich-

206

K. EICHMANN

mann, 1972; Yakulis et al., 1972; Geczy et al., 1976, 1977; Iverson, 1970; Iverson and Dresser, 1970; Janeway and Paul, 1973; Janeway et aZ.,1975a; Anderson et aZ.,1977a). The implications of the successful iso- and autoimmunization with regard to the network concept will be discussed in Section IV; in the present context, only the technical details will be presented and a distinction will be made between isoantiId (raised in the same strain) and autoanti-Id (raised in the same individual).

1 . Isoanti-idiotypic Antibodies Isoanti-Id antibodies against mouse immunoglobulins are prepared

by injecting mouse antibodies or myeloma proteins in animals of the same strain, which are, however, individually distinct from the animal producing the original idiotypic immunoglobulin. The immunizations follow the rules first described for alloimmunization with mouse myeloma proteins (Potter and Lieberman, 1970): a subcutaneous injection of 100 p g of purified myeloma protein in Freund’s complete adjuvant is followed by one or two intraperitoneal injections of the same amount in incomplete Freund’s adjuvant, and a series of intraperitoneal injections of the same amount in saline. The intervals between injections are 2-3 weeks (Potter and Lieberman, 1970; Sirisinha and Eisen, 1971; Yakulis et al., 1972; Sakato and Eisen, 1975; Cosenza, 1975; Cosenza et al., 1977a; Jbrgenson et n l . , 1977; C . Berek and K. Eichmann, unpublished). I n the case of an induced antibody to Strep.A carbohydrate, agglutinates of streptococci and this antibody were repeatedly injected intraperitoneally (Eichmann, 1972). Enhancement of an isoanti-Id response was achieved by coupling allogeneic mouse immunoglobulin to the myeloma protein (Fraker et al., 1974; Seppali and Eichmann, 1978). Methods for the detection and quantitation of isoanti-Id antibody in some cases have to be more sophisticated and sensitive than those employed for hetero- or alloantisera because, on one hand, the usual double precipitation techniques (Hopper et al., 1970; Hopper and Nisonoff, 1971; MacDonald and Nisonoff, 1970; Eichmann, 1973)cannot b e employed and, on the other hand, small amounts of antibody are frequently present in those antisera. Therefore, the test systems most frequently employed have been passive hemagglutination techniques (Lieberman et al., 1974, 1975, 1976; Kunkel, 1970), radioimmunoassays using solid-phase support (Riblet et al., 1975a; Sher and Cohn, 1972a,b; Bosma and Bosma, 1974; Askenase and Leonard, 1970; Berek et al., 1976),or radioimmune assays using a second antibody that had been absorbed on the idiotypic immunoglobulin, which is particularly

207

IDIOTYF'ES ON LYMPHOCYTES

feasible in the case of IgA myeloma proteins (Sakato and Eisen, 1975, Jbrgenson et ul., 1977; C. Berek and K. Eichmann, unpublished; Seppala and Eichmann, 1978). The problem of low-activity antisera has been overcome by the use of an exceedingly sensitive technique also employed for autoanti-Id (Cosenza et (11 ., 1977a), the indirectly enhanced passive hemagglutination technique. In this assay, red cells are coated with the idiotypic immunoglobulin and are then incubated with the putative anti-Id which in itself is unable to agglutinate the coated red cells. Agglutination is achieved by a second antibody to mouse immunoglobulin that had been previously absorbed on the idiotypic immunoglobulin (Cosenza et al., 1977a).

2 . Autoanti-idiot!lpic Antibodies Autoanti-Id antibodies can be obtained either b y deliberate autoimmunization or spontaneously a s a by-product of immunization with antigen, with a n allograft, or possi1,ly also in the course of malignancy. Deliberate autoimmunization has heen done with the antibodies to a variety of antigens including haptens and proteins isolated from single rabbits, which, after stoppage of immunization and a certain time interval, are reinjected into the same rabbit together with Freund's adjuvant (Rodkey, 1974,1976, also personal communication). In every case, anti-Id antibodies had been obtained that Iiound between 17% and 41% of the idiotypic immunoglobulin (Table IV). Interestingly, it was unnecessary to wait until the rabbit had completely ceased to proTABLE 1V REACTIVITY OF AUTOANTI-IdAGAINST THE ANTIBODIESTO R4N" OF 5 DIFFERENT RABBITS WITH THE RADIOLABELEDF(ab'), FRAGMENTS OF THESEANTIBODIES~ F(d)')2fragment from rabbit 1 ~~

~~

NRS" Anti-Id rabbit

1 2 3 4 5

2

~

3

1.1' 17.0 0.9 0.6 0.1 0.2

1.5 2.3 39.3 3.4 0.7 1.6

5

4

~~~

~

2.0 0.6 1.2 40.7 0.3 0.1

2.3 2.3 2.0 1.3 23.3 0.3

1.2 2.7 2.4 2.3 6.7 23.3

p-Amiriophenyl-N-trimethylammoninm chloride.

* Data from Rodkey (1974). Maximum percentage of binding of radioiodinated F(ah'), fragment by each antisenm. NRS, riormal rabbit serum.

208

K. EICHMANN

duce idiotypic antibody following immunization with antigens. Already about 100 days after antigen immunization, when circulating idiotype was still detectable, anti-Id production could be initiated by reinjection of the antibody (L. S. Rodkey, personal communication). Spontaneous production of autoanti-Id antibodies has been reported in BALB/c mice immunized with pneumococci against the produced anti-phosphorylcholine (PC) antibodies (Kluskens and Kohler, 1974; Cosenza, 1975; Cosenza et al., 1977c) and in rats immunized with allogeneic cells against the produced alloantibodies (McKearn, 1974; McKearn et al., 1974a,b). The best demonstrated case is undoubtedly the antibody response to PC in BALB/c mice in which virtually all antibodies possess the idiotype of the PC-binding myeloma TEPC 15 (T15) (Cohn et al., 1969; Potter and Lieberman, 1970; Cosenza and Kohler, 1972a, 1973; Claflin et al., 1974a,b). In addition, conventional BALBlc mice possess the T15 idiotype in association with natural antibody to PC (Lieberman et at., 1974). Upon immunization with PC, Kluskens and Kohler (1972) observed the appearance of antibodies in the serum that agglutinated T15coated red cells and apparently possesses idiotypic specificity. This antibody could be separated from the simultaneously circulating antiPC antibody by immunoadsorbent column techniques, suggesting that much of the two complementary antibodies, although together in one serum, were not complexed to one another. As these surprising results could still be interpreted by circulating free PC antigen, these doubts were eliminated by the subsequent experiments of Cosenza (1975), who demonstrated the appearance of plaque-forming cells secreting antibody to the T15 idiotype in BALB/c mice immunized with PC (Table V). TABLE V PLAQUE-FORMING CELLS (PFC) SECRETINGAUTOANTI-IDIOTYPIC ANTIBODY TO T15 IN BALB/C MICE IMMUNIZED WITH PHOSPHORYLCHOLINE' Indicator cells SRC PnC-SRBC T15-SRBC T15-SRBC Tl5-SRBC

Inhibitor

T15 MOPC 315

PFClspleen" 360 4596 1366

52 146 2 30 366 f 46 1246 112 ?

5

*

Data from Cosenza (1975). Sheep red blood cells (SRBC) coated with pneumococcal C (PnC) polysaccharide. Spleens of 3 immunized BALBic mice were pooled and analyzed. Nonimrnunized BALB/c mice produced no more than 100 PFC per spleen above the SRBC background. a

IDIOTYPES ON LYMPHOCYTES

209

Autoanti-Id was also demonstrated in transplantation immunity. Inbred (Lewis xBN)F, rats hyperimmunized with BN tumor cells develop in their serum an antibody that forms a precipitin line on Ouchterlony analysis with Lewis anti-BN alloantiserum (McKearn et al., 1974a,b). This antibody had the characteristics of an anti-Id antibody, as it did not react with Lewis antisera to other alloantigens and as a similar antibody could be elicited by direct immunization of (Lewis x BN)F, rats with purified Lewis anti-BN alloantibody (McKearn, 1974). In these studies, however, the possibility was not formally ruled out that the precipitin lines formed were results of the interaction of alloantibodies with circulating soluble antigen. c . PREPARATION OF ANTI-IDIOTYPIC REAGENTSAGAINST CELT.-SURFACE RECEPTORS For the analysis of lymphocyte receptors, the preparation of anti-Id reagents directly against cell-bound receptor molecules has proved to be very useful. The principles for the production of such antisera in ar, allogeneic system, first employed b y Ramseier and Lindenmann (1969), consists in the injection of lymphocytes from one parental strain of mice or rats into F, hybrid animals. These F, hybrids are tolerant to all antigens on these cells except for those associated with the alloantigen-recognition site for the histocompatibility antigens of the other parent. The prediction proved indeed to b e correct, and F, animals thus immunized produced anti-Id antibodies specific for recognition sites directed toward the other parent’s histocompatibility antigens. The pioneering work in this system (Ramseier and Lindemann, 1972a, b; Ramseier, 1973, 1974a,b, 1975, 1976; Ramseier et at., 1977) utilized a rather indirect assay, but the principle of this work has now been adopted by several other laboratories, which have confirmed and extended the original observations (Binz and Wigzell, 1975a,b; Andersson et al., 1977a; Krammer and Eichmann, 1978; D. B. Wilson, personal communication). Four different procedures are in use for the production of anti-Id according to these principles. Although not all the procedures utilize intact lymphocytes for immunization, they are summarized here, because all four are based on the same principle.

1 . Zmmunization with Alloantibodies This is done by immunization of F, animals with alloantibodies of one parent against the other, which are either applied as whole serum, using multiple intradermal injections, or, after glutaraldehyde crosslinking of whole serum, the material is subcutaneously applied in

K. EICHMANN

2 10

complete Freund's adjuvant (Binz and Lindenmann, 1972a,b, 1974; Binz and Wigzell, 1975a; Binz et a1 ., 1973, 1974a,b; McKearn, 1974). These immunization procedures frequently lead to production of antiId antibodies with a preferential reactivity toward parental idiotypepositive B cells (Binz and Wigzell, 1977a).

2. Immunization with Normal Lymphocytes Parental lymphocytes can either be injected into F, animals as a mixture of T and B cells or, after fractionation on IgG-complex columns, as highly enriched T cells, but not as normal purified B lymphocytes (Binz and Wigzell, 1975a; Binz, 1975). Purified T lymphocytes are considerably better in the induction of anti-Id then mixtures of T and B cells (Table VI) (Binz, 1975; Binz and Askonas, 1975). Upon repeated injection of intact purified rat T lymphocytes into adult F, hybrid rats, virtually all animals produce anti-Id detectable by passive indirect hemagglutination assays using alloantibody-coated red cells. Only a few animals, however, synthesize sufficient antibody to employ direct radioimmunoassays for the detection of idiotypic receptors on lymphocytes (Binz and Wigzell, 1975a, 1977a,b). Unfortunately, such high-titered antisera have thus far, in a predictable fashion, been obtained only in rats, whereas in mice the incidence of responding animals was exceedingly low (Binz and Wigzell, 1977a; Binz and Askonas , 1975). 3 . lmmunization with Positively Selected Lymphocytes In this procedure, T lymphocytes with reactivity toward the other parental histocompatibility antigens are preselected in a mixedlymphocyte reaction (MLR) and isolated using Ficoll gradient cenTABLE VI PASSIVE HEMAGGLUTINATION TITEROF ANTI-IDIOTYPIC ANTISERA M S E D IN (LEWISx DA)F, RATS BY INJECTIONOF PARENTAL SPLEEN CELLSOR PURIFIED T CELLS"

a

Cells injected

TiteP (log 2)

DA spleen' DA T Lewis spleen Lewis T

2.0 -+ 0 3.0 2 0.4 2.4 2 0.5 7.0 -r 1.1

Data from Binz and Wigzell (1977a).

* Sheep red blood cells coated with the appropriate alloantibody were used for hemagglutination; log 2 titers are given. Cells (2.5 x 10') were injected, and the animals were bled 5 days thereafter.

IDIOTYPES ON LYMPHOCYTES

211

trifugation or unit gravity sedimentation procedures (Andersson et

al., 1977a; Krammer and Eichmann, 1978).T blasts isolated in this way are 60-90% reactive toward the relevant histocompatibility antigen (Nagy et al., 1976; Elliott et al., 1977), and their injection into either F, hybrid animals (Krammer and Eichmann, 1978) or animals of the donor strain (Andersson et al., 1977a) leads to reliable production of anti-Id antibodies that can be detected b y immunofluorescence (Krammer and Eichmann, 1978), by inhibition of MLR and cytotoxic T cells (Krammer and Eichmann, 1978; Andersson et al., 1977a; H. Binz, personal communication), or by the delayed rejection of relevant grafts by the immunized animals (Andersson et al., 1977a; Binz and Wigzell, 1977a,b). The major advantage of this method is that it is successful in mice (Andersson et al., 1977a; Krammer and Eichmann, 1978).

4 . Zmmunizution with Soluble T-cell Receptors Binz and Wigzell(1977a7b,c)have isolated from normal rat serum or urine, using F, anti-Id antibody as immunoadsorbent, a fraction of idiotype-positive, antigen-binding molecules. These have been used to immunize rabbits (Binz et ul., 1975a) or other rats (Binz and Wigzell, 1977b). In both instances, anti-Id antibodies reactive with the relevant alloantigen-reactive lymphocytes have been elicited.

Ill. Analysis of B- and T-cell Receptor ldiotypes

A. ANALYSISOF B-CELLRECEPTORIDIOTYPES

1 . Direct Analysis of Zdiotypes on or in B Cells In an early study it was shown that anti-Id antibodies to a human A-type Bence-Jones protein react, when used in fluorescence microscopy, with a small proportion (- lo7 positive cells per human spleen) of lymphocytes in the spleen, suggesting that lymphocytes and plasma cells show a similar degree of idiotypic heterogeneity as do circulating antibodies (Pernis, 1967). Studies that directly demonstrate idiotypes on normal antigen-binding lymphocytes are scarce. Claflin et (11. (1974a,b) used double-labeling experiments to show that mouse spleen cells binding the phosphorylcholine (PC) hapten react with anti-Id antibody to the PC-binding myeloma protein HOPC 8. It was found that this anti-Id antibody inhibits the binding of PC to lymphocytes from strain BALB/c but not to the lymphocytes from other strains; this is somewhat surprising, as it has since been shown that the HOPC 8R15 idio-

212

K. EICHMANN

type is genetically not restricted to BALB/c antibodies to PC (Rudikoff and Claflin, 1976). Some of the earliest evidence for idiotypic determinants on the Ig of lymphocytes was obtained in the human system in studies on the lymphocytes of patients with Waldenstrom’s macroglobulinemia (Wemet et a1 ., 1972). High percentages of peripheral blood lymphocytes of these individuals showed on their membranes the same idiotype as the serum Ig. This was limited to Ig-bearing B cells. Studies in myeloma (Lindstrom et a1 ., 1973; Mellstedt et a1 ., 1974) also showed the myeloma protein idiotype on a limited percentage of peripheral blood B cells. More recently chronic lymphatic leukemia cells have been widely studied as models of monoclonal B-cell proliferation, and idiotypic determinants have been demonstrated on the membrane Igs (Fu et al., 1974; Kubo et al., 1974). These studies have been extensively reviewed (Natvig and Kunkel, 1973; Preud’homme and Seligmann, 1972), and no attempt will be made here to comprehensively summarize this work. One of the most significant findings to stem from this work has been the demonstration that IgM and IgD, when present on the same B cell, as is frequently the case, possess the same idiotype. This was demonstrated in a variety of ways but was most apparent in capping experiments with the anti-Id antisera (Fu et al., 1974, 1975; Salsano et al., 1974; Natvig et al., 1975).

2 . Analysis of Zdiotypes on B Cells Using Functional Assays The functional approach to the analysis of B-cell receptor idiotypes consisted in the measurement of antibody production of B cells in the presence and in the absence of anti-Id antibody. A variety of studies of this sort have been performed both in vitro and in vivo, using a variety of idiotypic systems. Some of this work has been reviewed previously (Nisonoff and Bangasser, 1975; Nisonoffet ul., 1977; Kohler, 1975), so only the principles are summarized here: anti-Id antibody is prepared against an antibody or myeloma protein whose idiotype is common or cross-reactive among the members of one inbred strain of mice. Such idiotypes include, among others, that of the strain A/J antibody clone A5A with specificity for the group-specific carbohydrate of group A streptococci (Eichmann, 1972, 1973, 1974, 1975a), that of BALB/c myeloma proteins T15 and HOPC 8 with specificity for PC (Cohn et al., 1969; Lieberman et al., 1974; Cosenza and Kohler, 1972~1,1973; Kohler, 1975), and that of antibodies to the azobenzoarsonate hapten (ARS) in strain AIJ (Hart et a1 ., 1972; Pawlak et al., 1974; Bangasser et al., 1975; Nisonoff et al., 1977; Shyr-Te et a l . , 1977). The anti-Id antibodies used are of xenogeneic origin for the A5A and ARS idiotypes

213

IDIOTYPES ON LYMPHOCYTES

and of allogeneic origin for the T15 idiotype. Injection of the anti-Id antibody into mice before or shortly after immunization with antigen results in unresponsiveness of those B cells that produce the antibodies bearing the idiotypic determinants (Table VII) (Hart et al., 1972; Pawlak et ul., 1973b,c, 1974; Nisonoff and Shyr-Te, 1976; Eichmann, 1974, 1975a). Similarly, addition of anti-Id antibody to Michel-Dutton type cultures before or shortly after adding the appropriate antigen results in prevention of the formation of idiotype-secreting TABLE VII SUPPRESSION OF B-CELL RESPONSES BY ANTI-IDIOTYPIC ANTIBODY Antigen, anti-idiotype

Normal response

Suppressed response

pLG/ml antibody (Ab) and idiotype (Id) produced

518 2 79pg/ml 4 2 0.9 pg/ml

Strep.A-CHO, an ti-A5A"

Ah: 927 ? 107 pg/ml Id: 284 ? 56pghnl

Benzoarsonate, an ti - A M b

Percent binding of idiotype in the presence o f X pl of antiarsonate hyperiininune serum 8-10 (1 pl)

Phosphorylcholine (PC), anti-T 15'

91-101 (10 pl)

Anti-PC-PFC/culture (in oitro)

3 160 343 Aliti-PC-PFCispleen ( i n oioo) 139,000

1300

a Data from Eichmann (1974). A/J mice received 10 pg of IBC of guinea pig anti-A5A antiserum or normal guinea pig serum and were immunized with 6 injections of group A streptococcal vaccine 10-15 days later. On day 20, the mice were bled and anti-A-CHO antibody as well as antibody with A5A idiotype were determined by radioiniinune assays (Eichmann, 1972, 1973). Data from Hartet nl. (1972).A/J mice received 2 injections of0.2 ml ofrabbit anti-ARS idiotypic antiserum, spaced 1 day apart. Mice were immunized with keyhole limpet hernocyanin-azobenzoarsonateseveral times beginning 2 days after injection of antiidiotype. Hyperimmune sera from these and from normal mice were used a s inhibitors of the idiotypic binding between the rabbit anti-ARS antiserum and radiolabeled anti-ARS antibody. Data from Cosenza and Kohler (197%). lo7 BALB/c spleen cells were cultured together with 10' R06A pneumococci, and the responses to PC were determined using C polysaccharide coupled to sheep red blood cells. A/J anti-T15 idiotypic antisera or nonnal A/J serum were added to the cultures. For the in oioo experiment, BALB/c mice were injected with loy R36A pneumococci and 200 pl of A/J anti-T15 antiserum or nonnal A/J serum a s control. PFC, plaque-forming cells.

214

K. EICHMANN

plaque-forming cells (Table VII) (Cosenza and Kohler, 1972b, 1973); Cosenza et al., 1977a). In the case of the A5A and ARS idiotypes, which are associated with only a certain fraction (25-30%) of the antibodies induced by immunization with antigen, the magnitude of the total antibody is not dramatically reduced by the elimination of the idiotype-positive portion (Hart et al., 1972; Eichmann, 1974). In contrast, in the case of the T15 idiotype, which is present on virtually all the antibodies to PC, the entire antibody response may be abrogated (Cosenza and Kohler, 1972a; Kohler et al., 1974; Cosenza et al.,

1977b). Common to all these studies is that the suppression requires relatively large quantities of anti-Id antibody and that the suppression is transient unless injection of either anti-Id or antigen is continued (Bangasser et al., 1975; Pawlak et al., 1973~).These observations point to the notion that the major mechanism is a transient functional of physical elimination of idiotype-bearing precursor B cells, and it has been concluded from these studies that B precursor cells bear receptors possessing the same idiotypes as the antibodies that are eventually secreted by their descendant plasma cells (Nisonoff and Bangasser, 1975; Kohler, 1975; Eichmann, 1974). It should be pointed out in this connection that idiotype suppression can also be achieved by injection of newborn mice or of adult mice with small doses of anti-Id antibody (Eichmann, 1975a; Augustin and Cosenza, 1976; Cosenza et nl., 1977~). This suppression differs from the suppression with large doses by a longer duration (Eichmann, 1975a; Cosenza et d., 1977c) and b y the fact that it is T-cell mediated (Eichmann, 1975a). Thus, as suppression b y high doses of anti-Id gives information as to the idiotypic nature of the B cell receptor for antigen, low-dose suppression is more complicated and may be mediated via accessory cell types. This latter type of suppression will be discussed in Section IV below. Taken together, the analysis of the idiotypes of the antigen-receptor on B cells has fully supported and confirmed the clonal selection theory (Burnet, 1959; Jerne, 1955) in revealing (a) that binding sites are expressed prior to antigen exposure, (b) that they are clonally distributed and clonally restricted among B lymphocytes, and (c)that, even in the case of intraclonal class heterogeneity, there is only one binding site (idiotype) per clone of lymphocytes.

3 . Precursor Frequencies of Zdiotypically Defined B Cells A number of laboratories have reported data on the B-cell precursor frequency for certain antigen-binding specificities (Klinman et al., 1974; Andersson et al., 1977b,c; Quintans and Levkovitz, 1973,

IDIOTYPES ON LYMPHOCYTES

215

1974a,b, 1976).Here we limit the discussion to those reports that deal with the idiotypic analyses of B-cell precursors, and summarize data on the T15 idiotype (Cosenza et ul., 1975; Quintans and Cosenza, 1976) and on the A5A idiotype (Eichmann et al., 1977b). The frequency of B cells for the T15 idiotype in the spleens of BALB/c mice was determined in Michel-Dutton cultures of spleen cells stimulated with Pneumococcus pneumoniae strain R36A, using concentrations of spleen cells that limit the number of precursor cells in each culture to an average of one. The detection system was either the appearance of plaque-forming cells or the appearance of soluble antibody in the supernatant, using PC-coupled sheep red blood cells (SRBC) as indicator cells (Cosenza et al., 1975). The fraction of T15 idiotype-producing cultures was determined by adding anti-T15 idiotypic antibody and monitoring the reduction in the number of positive cultures. Analysis of the data according to Poisson’s distribution yielded a frequency for PC-specific B precursor cells of 1 in 4 x lo4to 1 in lo5 spleen cells, 88% of which expressed T15 idiotypic determinants. These results are in agreement with those using fragment cultures (Gearhart et al., 1975; Sigal et al., 1975) in suggesting a pronounced but not complete association of the PC-binding specificity with the T15 idiotype at the precursor cell level. The data suffer from the fact that the cell population analyzed was limited to that responding to PC under the in vitro culture conditions used. Thus, a possible nonresponding cell population carrying the T15 idiotype may have been missed, and therefore the precursor frequency may have been underestimated. A fundamental result borne out by these studies was that the expression of the T15 idiotype was the same on the B cells involved in Tdependent as well as in T-independent responses to PC, which was presented either as strain R36A pneumococci for a T-independent response, or as hapten coupled to keyhole limpit hemocyanin (KLH) for a T-dependent response (Quintans and Cosenza, 1976). Thus, with respect to receptor idiotype, the B-cell population involved in Tdependent and T-independent responses appears to be the same. Recently, the B-precursor frequency for the A5A idiotype was determined in cultures of spleen cells stimulated with lipopolysaccharide (LPS), thus analyzing lymphocytes independent of their reactivity to antigen (Eichmann et al., 1977b). In this experimental system, the fraction of B cells responding to LPS can be monitored independently by plaque formation on SRBCs coated with protein A (Gronowicz et al., 1976; Anderson et al., 1977b,c), and the proportion of idiotypepositive cells within all splenic B cells can be corrected by this figure

216

K. EICHMANN

TABLE VIII PRECURSOR FREQUENCIES OF IDIOTWE-POSITIVEPRECURSOR B CELLS FOR THE A5A AND T15 IDIOTYPES I N SPLEEN CELLS OF NORMAL MICE Sample Uncorrected Corrected for 50% B cells Corrected for 30% LPS-reactive cells

A5A"

PCb (% T15+)

1:1.53 x 104 1:7.6 x 103 1:2.5 x 103

1:5.4 x lo4 (88%) NR' NR

Data from Eichmann et al. (1977b). Graded numbers of spleen cells were cultured together with normal thymocytes as filler cells and 50 pg of lipopolysaccharide (LPS) per milliliter in 200-p1 cultures. A5A-positive cultures were determined by measuring the amount of A5A idiotype in the supernatants (Eichmann, 1973). The data were analyzed according to Poisson's distribution. Data from Cosenza et al. (1975); the value is a mean of 4 reported experiments. Graded numbers of spleen cells were cultured together with R36A pneumococci, and positive cultures were determined by spot tests of the supernatants using complement lysis of phosphorylcholine (PC)-coupled sheep red blood cells. The data were analyzed according to Poisson's distribution. ' NR, not reported.

(usually 1 in 6 spleen cells or 1in 3 B cells). Table VIII shows the A5A precursor frequencies determined in limiting-dilution experiments in which the average number of A5A-positive B precursor cells was around 1per culture. The detection system used in these experiments was a radioimmune assay that determines the appearance of A5Abearing immunoglobulin in the culture-supernatants (Eichmann, 1973; Eichmann et al., 1977b). It is surprising to see the about 3- to 4-fold greater frequency for A5A-secreting precursor B cells as compared to that for T15-secreting precursor B cells. This discrepancy can be partially explained by (a) the greater efficiency of LPS stimulation as compared to antigenic stimulation, (b) the correction factors used to calculate the actual A5A precursor frequency, and (c) the independence of antigen stimulation (Eichmann et al., 197713).Indeed, a certain fraction of the B cells secreted antibody with A5A idiotype that was unreactive with group A streptococci, the antigen fitting the original A5A antibody (Eichmann, 1972, 1973). This antigen-negative, idiotype-positive fraction would not have been detected in experiments using antigenic stimulation of lymphocyte cultures, as was the case in the experiments performed to determine the precursor frequency for T15 idiotype-positive B cells (Cosenza et al., 1975). This apparent discrepancy will be further discussed in the context of the network hypothesis in Section IV.

IDIOTYPES ON LYMPHOCYTES

217

B. ANALYSISOF T-CELL RECEPTORIDIOTYPES 1 . ldiotypes of Helper T Cells

The discovery of antibody idiotypes on helper T cells came originally from the notion that under certain circumstances the injection of anti-Id antibody into mice may be stimulating rather than suppressive (Eichmann, 1974; Trenker and Riblet, 1975) and that stimulation was seen in the T-cell compartment in addition to stimulation in the B-cell compartment (Eichmann and Rajewsky, 1975). These experiments were done primarily using anti-Id antibodies to the ASA idiotype that represents the major idiotype of strain A/J antibodies to group A streptococcal (Strep. A) carbohydrate (Eichmann, 1972, 1973,1974,197Sa). The expression of the A5A idiotype in association with anti-A-CHO antibodies is controlled by a VHgene, termedA5A+, which is linked to the ZgZ' allotype locus of strain A/J (review in Eichmann, 1975b). Helper-cell stimulation experiments have also been done using anti-Id antibodies to BALB/c myeloma protein S 117 which has specificity for A-CHO (Vicari et al ., 1970) and whose idiotype is associated with induced antibodies to A-CHO of strain BALB/c and of other strains carrying the Zg-1 a allotype, to which it is genetically linked (Berek et al., 1976, review in Eichmann, 1975b). Furthermore, an idiotype associated with strain AKR antibodies to A-CHO has been utilized in helper-cell stimulation experiments (Krammer and Eichmann, 1978). Helper-cell stimulation is achieved with the IgG, fractions obtained from guinea pig antisera to the antibodies described above (Eichmann and Kindt, 1971; Eichmann, 1972,1973,1974; Benacerraf et al., 1963). The principles of the induction of anti-Id antibodies in guinea pigs to mouse antibodies have been described in Section II,A. The preparation of IgG, fractions from these antisera is easily achieved by agarose block electrophoresis (Braun and Krause, 1968; Eichmann and Greenblatt, 1971, Eichmann, 1974), monitoring the idiotype binding capacity (IBC) and the passive hemolysis of idiotype-coated SRBCs of each fraction. IgG, fractions are calibrated according to their IBC, centrifuged at 100,000 g for 30 minutes, and 0.1 pg of IBC is injected intraperitoneally or intravenously into mice of the appropriate strains. The demonstration of helper activity in these systems was done using in uiuo adoptive transfers (Rajewsky, 1971; Eichmann and Rajewsky, 1975; Rajewsky and Eichmann, 1977; Rajewsky et al., 1976) as well as in vitro microcultures (Black et a1 ,, 1975, 1976a,b; Hammerling et al., 1976a; Eichmann, 1977a; Seppiilii and Eichmann, 1978; Krammer and Eichmann, 1978).The principle of both types of experi-

218

K. EICHMANN

ments is the demonstration of helper T cells with specificity to the Strep. A carrier by the generation of hapten-specific antibodies or plaque-forming cells in response to challenge with a hapten4trep.A conjugate. Table IX summarizes the results of in vitro microculture experiments using spleen cells from A/J mice, BALB/c mice, and AKR mice that have been presensitized with Strep.A or with the IgG, fractions of anti-Id antisera to A5A, to S117, or to AKR anti-A-CHO antibodies, respectively. It is clear from the data that in all three strains helper activity is successfully activated by injections of the appropriate anti-Id antibody. The helper activity is exerted by T cells, since it can be abolished by treatment of the spleen cells with anti-Thy-1 and complement (Eichmann and Rajewksy, 1975; Black et nl., 1976a) and since more than 98% pure lymph node T cells can be used (K. Eichmann, unpublished). It is specific for Strep.A, as TNP coupled to heterologous carriers, such as SRBCs, chicken Ig, or guinea pig Ig, did not elicit a response (Eichmann and Rajewsky, 1975; K. Eichmann, unpublished). The data thus obtained were interpreted to suggest a direct interaction of anti-Id antibodies with T helper-cell precursors resulting in proliferation and differentiation to functional helper cells. Further evidence for this was obtained in studies that (a) further corroborated the identical antigen-binding specificity of T cells with that of B cells TABLE IX RESPONSIVENESS OF STRAINSA/J, AKR, AND BALB/c TO STRAIN-SPECIFIC ANTI-IDIOTYPIC REAGENTS WITH THE DEVELOPMENT OF HELPER ACTIVITY SPECIFIC FOR STREP.AAS CARRIER" Strain

MJ AKR Balb/c

Immunized with Strep.A Anti-A5AC Strep.A Anti(AKRiA-CHO)'

-

Strep.A Anti61 17'

Anti-TNP PFCb per lo6cultured cells 4

167 112 12 222 89 9 189 134

K. Eichmann, P. Krammer, and I. Seppala, unpublished results. Spleen cells (1 x 1Oa)cultured for 4 days in 100 pl of Click's Medium (Click et al., 1972) with 1 x 10' streptococcal group A (Strep.A) particles coupled with TNP. Plaqueforming cells (PFC) were developed on sheep red blood cells coupled with TNP (Rittenberg and Pratt, 1969; Black et al., 1976a; Eichmann, 1977a). IgC, fractions of the respective guinea pig anti-idiotypic antisera; 0.1 pg of idiotype binding capacity injected (Eichmann and Rajewsky, 1975).

IDIOTYF'ES ON LYMPHOCYTES

219

and antibodies reacting with the same anti-Id antibody, (11) showed the idiotypic uniformity of T helper cells induced by anti-Id antibody as opposed to the heterogeneity with respect to both idiotype and specificity of T helper cells induced b y immunization with Strep.A (Black et a1 ., 197.5, 1976a,b), and (c) demonstrated that the genes controlling T helper cell responsiveness to anti-Id antibody reside in the same linkage group as those controlling the expression of antibodyassociated and B cell-associated idiotypes, namely the Zg-l complex that contains the genes encoding the mouse Ig H chain (Hammerling et aZ., 1976a; Krawinkel et aZ., 1977a; Eichmann et al., 1977a). These data have recently been comprehensively discussed in another article (Rajewsky and Eichmann, 1977) and are therefore merely summarized in Table X. Recently, antibodies to idiotypic suhpecificities of the ASA antibody have been used to further characterize idiotype presentation on helper T cells. Antibodies with highly preferential reactivity toward idiotypic determinants on either the VH or the V, region (see Section II,A) of the ASA antibody have been used to stimulate helper cells in uico and to inhibit helper activity in uitro (Eichmann, 1977a). The data are presented in Fig. 1 and Table XI. With respect to helper-cell stimulation i n viuo it is apparent that antisera with preferential reactivity toward Vcregion determinants are 10- to 20-fold less effective than are antisera with preferential VH reactivity. There is essentially no difference in the effectiveness of the two kinds of antisera with respect to B-cell stimulation. The difference between the two types of

001

01

I 0 0.01 pg H 1d iniecled

01

10

FIG.1 . Differential capacities of anti-A5A idiotypic antibodies (IgC,) with preferential V w reactivity ( @ ) and with preferential VL reactivity (0)to sensitize T helper cells (left panel) and B precursor cells (right panel). Abscissa: micrograms of idiotype binding capacity (IBC) injected per mouse; left ordinate: TNP-plaque-forming cells as percent of the response of cells sensitized with Group A streptococci (Strep.A) as a measure of T helper-cell activity; right ordinate: A-CHO-plaque-forming cells as a measure of B-precursor cell activity. Upper dotted lines: responses of cells primed with Strep.A; lower dotted lines: responses of unprimed cells. All responses were determined in microcultures, stimulated with Strep.A T N P or Strep.A. For experimental details see Black et (11. (1976a). The two sets of curves i n the left panel represent two different experiments.

K. EICHMANN

220

TABLE X EXPERIMENTAL EVIDENCE THAT T AND B CELLS RESPONSIVE TO THE SAME ANTI-IDIOTYPE POSSESS THE SAME ANTIGEN-BINDING SPECIFICITY AND A SIMILAR DEGREEOF IDIOTYPIC UNIFORMITY, AND THATT-CELLIDIOTWES ARE CONTROLLED BY VH GENES" B cell response to

T cell response to

Strep.A

Anti-A5A

Strep.A

Anti-A5A

30% 10-50%

>80% >90%

10-3070 0-10%

>80% >90%

+ + + +

+ + +

+ + + +

+ +

_ _ _ _ _ _ _ ~

Inhibition with A-CHO* Inhibition with anti-ASA* Linkage to genetic markers'

V" A5A+ A5A+ A5AA5A" a

Ig-1 e e a,b,d,f c

H-2 a S

a,d,k,s,b d,k

-

+

Reviewed in Eichmann and Rajewsky (1977).

* Data from Black et al. (19764. Responses were determined in in oitro microcultures using Strep.A as antigen and A-CHO sheep red blood cells (SRC) as indicator cells for the B-cell response, and TNP4trep.A as antigen and TNP-SRC as indicator cells for the T helper-cell response. Inhibition of the B-cell response by A-CHO or anti-A5A idiotype was done by including the inhibitors in the plaque test. Inhibition of the T helper-cell response was done by including the inhibitors into the culture medium. Data from Hammerling et al. (1976a) and from Krawinkel et al. (1977a). A panel of inbred strains including congenic strains were tested for their ability to produce T helper cells in response to anti-idiotype, as analyzed by their i n oitro helper effect in the anti-TNP response to TNP4trep.A. Similar data were obtained for the S117 idiotype (Hammerling et al., 1976a) and, recently, for an idiotype associated with strain AKR antibodies to A-CHO (Gammer and Eichmann, 1978).

antisera is even more pronounced in the in vitro inhibition experiments. VL reactive antisera show minimal helper-cell inhibition at concentrations at which VH reactive antibodies are more than 80% inhibitory. These data have been interpreted to suggest that T helper cells present VH-associatedidiotypic determinants, whereas B precursor cells present both VH- and VL-associated idiotypic determinants (Eichmann, 1977a). The reason for this difference may be either a complete absence of VLregions on T helper cells, or a surface composition that results in inaccessibility of the VL region to anti-Id antibody. Furthermore, it is possible that the VL regions that are commonly associated with a given VH region in an antibody are different from the VL regions associated with the same VH region on the T-cell receptor. At present there is no strong evidence for or against any one of the three alternatives. Taken together, the present evidence suggests that T helper cells display receptors that share with antibody

IDIOTYPES ON LYMPHOCYTES

22 1

TABLE XI OF HELPERACTIVITY in Vitro by ANTI-IDIOTYPIC INHIBITION ANTISERA WITH v, OR vt REACTIVITY" Idiotype binding capacity per culture (ng)

Anti-TNP PFC/cu 1tiire*

Percent inhibition

-

7 165

0

-

119 147 147 140 64 21 83 46 23

28 11 11 15 61 87 50 69 86

Nonpriined cells Anti-A5A-primed cells Inhibitors used NGPS 1 :500' iMIg(M1g abs) 1 : 500 Anti-A5A L > H Anti-A5A H > L Anti-A5A H

10 30 10 30 1 3 10

" Data from Eichmann (1977a).

* Plaque-forming cell (PFC) response to TNP3trep.A of lo6 spleen cells in a 4-day microculture; mean of 4-6 identical cultures. Nornial guinea pig serum and MIg-absorbed anti-MIg serum in concentration corresponding to the highest concentration of anti-idiotype. All inhibitors were present during entire culture period. molecules those idiotypic determinants that are V, associated, but not the full complement of idiotypic determinants. Idiotypic determinants on helper T cells have also been demonstrated using anti-Id antibodies to the phosphorylcholine (PC)-binding myeloma protein TEPC 15 (T15)(Cosenza et ul., 1977b,c; Julius et ul., 1977).As was pointed out above, the T15 idiotype is expressed in association with induced anti-PC antibodies in BALB/c mice and in other strains (Cohn et d.,1969; Lieberman et al., 1974), and under normal conditions virtually all the antibody is T15 positive (Cosenza and Kohler, 1972a). Demonstration of T15-positive helper cells employed three approaches, one consisting in the inhibition of antigen-induced PCspecific helper cells, the second consisting in the induction of PC-specific helper cells b y anti-T15 idiotype, and the third demonstrating that suppression of the T15-positive antibody response is accompanied by suppression of the T15 positive helper-cell population (Julius et a1 ., 1977; Cosenza et d., 1 9 7 7 ~ )PC-specific . helpers are induced in BALB/c mice using PC-conjugated BALB/c myeloma proteins. These helper cells can be revealed in adoptive transfer experiments (Mitchi-

222

K. EICHMANN

son, 1971; Mitchison et al., 1970) using PC-conjugated BSA as antigen and BSA-primed spleen cells as a B-cell source (Table XII). The helper effect could be specifically inhibited (>90%) by injecting antiT15 idiotypic antibody together with the transferred cells and the antigen into the recipient mice. In the second type of experiment, Julius et al. (1977) showed that splenic T cells from mice pretreated with anti-T15 idiotypic antibody carried helper activity in adoptive transfer experiments using BSAprimed spleen cells and PC-BSA as the challenging antigen. As in the experiments in which Strep.A-specific helper cells are induced by anti-Id antibodies (see Table IX), 0.1 pg of IBC was optimal for helper-cell induction to PC. Together with experiments in which it was shown that PC-specific helper cells that arose by immunization of T15 suppressed mice were not inhibited by anti-T15 antibodies, the data were interpreted to suggest that T cells as well as antibodies to PC possess similar or identical idiotypes (Cosenza et a l . , 1977b,c). TABLE XI1 INHIBITION BY ANTI-TIS IDIOTYPE OF PHOSPHORYLCHOLINE(PC)-SPECIFIC HELPERCELLS, AND STIMULATIONOF PC-SPECIFIC HELPERCELLS BY ANTI-T15 IDIOTYPE, SHOWN IN ADOPTIVE TRANSFEREXPERIMENTS" _

~

B cells BSA-primed

T cells PC-MOPC 315 primed

+ + +

+ + +

_

_

_

_

_

_

~

~

~ ~

~

Anti-T 15 serum

Anti-BSA PFCIspleen

+

1400 200 29700 1400

Anti-T 15 primed (CLg IBC) 10 1

-

+ + +

0.1 10 1 0.1

230 70 250 6000

9800 12.800

Data from Cosenza et 01. (1977~)and Julius et ol. (1977). Pooled results from several experiments; some controls were omitted for the sake of simplicity. B cells were anti-Thy-1-treated spleen cells; T cells were nylon wool-filtered spleen cells. Anti-T15 serum was given together with the injection of the cells, and all recipient mice received PC-bovine serum albumin (BSA) as antigenic challenge. BSA-sheep red blood cells were used as indicator cells in plaque test. Priming with AIJ anti-TI5 serum was done 4-6 weeks before the experiment by intraperitoneal injection of appropriate amounts of whole serum.

~

~

IDIOTYF’ES ON LYMPHOCYTES

223

In this system, experiments using anti-Id antibodies with specificity for VHor VLdeterminants have not been performed. It is, however, interesting to note that the commonly used alloantisera to T15 detect idiotypic determinants that are almost exclusively not associated with the T15 binding site (Claflin et nl., 1974a,b; Claflin and Davie, 1975) and therefore detect so-called “framework” idiotypes. Thus, the demonstration of such idiotypic determinants on helper T cells with specificity for PC allows the conclusion that both hypervariable as well as framework sequences are shared between T-cell receptors and antibodies. The demonstration of the chain association of the idiotypic determinants in the T15PC system still needs to he done. Another point worth mentioning here is the surprising similarity of the degree of idiotypic heterogeneity observed between PC-specific helper cells and antibodies of the same specificity. As do anti-PC antibodies, virtually all PC-specific helper T cells carry T15 idiotype (Table XII). This result is in agreement with the similar heterogeneity observed between T and B cells reactive with A-CHO and carrying A5A idiotypic determinants (Table X). This result alone is sufficiently strong evidence that the combining sites of both T-cell and B-cell receptors are of similar molecular construction and of similar genetic origin.

2 . Zdiotypes of Suppressor T Cells There is considerably less information on the receptors of suppressor T cells than on the receptors of helper T cells. Using the IgG, fraction of anti-Id antibodies to A5A, suppressor T cells are induced in A/J mice that specifically suppress the production of A5A idiotypepositive antibodies (Eichmann, 1974, 1975a) when transferred into slightly irradiated recipient mice. The T cells responsible for this suppression possess the I-J subregion antigen commonly found on suppressor cells and, therefore, belong to this class of T cell (Hammerling and Eichmann, 1976, 1977; Hammerling et al., 1976b,c; Murphy et al., 1976; Tada et nl., 1976). In analogy to the induction of helper T cells by IgG, anti-A5A-idiotypic antibodies, the induction of suppressor T cells by IgG, anti-A5A-idiotypic antibodies has been interpreted to suggest the presence of the A5A idiotype on suppressor T cells (Eichmann, 1975a; Rajewsky and Eichmann, 1977). The different effector functions of the induced T-cell types have been attributed to the different Ig classes of the sensitizing antibodies (Eichmann, 1975a). Attempts to further analyze suppressor T-cell idiotypes have not been very successful. Using guinea pig IgG, antibodies to idiotypes associated with anti-A-CHO and other antibody responses in strains other than AfJ, no suppressor T-cell induction has been observed (C.

224

K. EICHMANN

Berek, unpublished; R. S. Jack and T. Imanishi, personal communication). Thus it may be that a genetic polymorphism exists that governs the induction of suppressor T cells by guinea pig anti-Id antibodies. Furthermore, attempts to use anti-Id antibodies to lyse suppressor T cells in the presence of complement have occasionally been successful, but inconsistently so (G. C. Hammerling and K. Eichmann, unpublished). Thus, at present no clear picture exists of the presentation of idiotypes on the receptors of suppressor T cells.

3. Zdiotypes of DTH-Reactive T Cells Recently, experiments have been reported on the induction of DTH-reactive T cells by injection of anti-T15 idiotypic antibody into BALB/c mice (Cosenza et al., 1977b). DTH was elicited by injection of pneumococci (strain R36A) into the ears of mice that had received anti-T15 antibody, and the data showed that at doses optimal for helper-cell induction (see above) there was also optimal induction DTH reactivity. This observation is particularly interesting as unrelated work on DTH-reactive T cells has revealed H-2 restriction of antigen recognition by this type of T cell (Miller and Vadas, 1977). Thus, this T-cell subclass is similar to the cytotoxic T cell that recognizes foreign antigens in relation to H-2 coded antigenic determinants on target cells (Zinkernagel, 1976; Shearer et al., 1976). Two alternative possibilities have been considered to explain this double requirement of antigen recognition: alteration of the H-2 antigen by the introduced haptenic determinant (“altered self”) leads to the recognition of the target cell by one receptor, or “double recognition” of an H-2 antigen, and an introduced haptenic group requires recognition by two receptors. Many investigators feel that the double requirement for the antigen-recognition by cytotoxic T cells or DTH-reactive T cells may provide a handle for solving the T-cell receptor puzzle, and that anti-Id reagents are to be used in this approach. The observation that DTH-reactive T cells with PC specificity are induced by anti-Id antibody to T15 may be a first indication that the recognition sites on DTH cells for the anti-T15-Id antibody are similar to the PC-binding site of T15 and are therefore distinct from binding sites recognizing H-2-coded antigens that may be on the same or another cell. Many more experiments need to be done in this context, particularly using cytotoxic T cells, before any firm conclusions can be drawn. 4 . Zdiotypes on Rabbit Zg-Negative ( T ) Cells The study of rabbit T cells by anti-Id antibodies offers the advantage that with the a locus allotypes (Kindt, 1975) there are independent markers for the VHregion of rabbit immunoglobulins of all classes.

IDIOTYPES ON LYMPHOCYTES

22s

Cazenave et a l . (1977) have exploited this advantage in their study of peripheral blood lymphocytes carrying specific binding sites for ribonuclease (RN) and for anti-Id antibodies to anti-RN antibodies of the same rabbit. These authors separated peripheral Ig-negative and Ig-positive lymphocytes from rabbit immunized with RN, using Ficoll gradient separations of anti-Ig rosetted from nonrosetted cells. In the fractions thus obtained, RN-binding lymphocytes were determined b y rosetting, and the rosetted cells were analyzed using peroxidaselabeled or iodine-labeled Fall fragments from antibodies to the idiotype of each individual rabbit's anti-RN antibody, from antibodies to the appropriate allotypic specificities of the a (H chain) and h (L chain) series, and from a goat anti-VHantibody. Ig-negative cells from immunized rabbits bound RN antigen as well as labeled Fall fragments from anti-Id antibodies, anti-cl locus allotype antibodies, and anti-VHantibodies from a goat. Not bound to Ig-negative cells were anti-b locus allotype antibodies and common anti-Ig antibodies. It was interesting to note that apparently only those Ig-negative cells that bound anti-Id also bound anti-ci locus antibodies, suggesting that, in the Ig-negative population only those cells that are presently committed to antigen recognition express receptors. In this respect, as well as with regard to the functional role of Ig-negative, Id-positive cells, many questions are still open in the rabbit system, particularly as other investigators have not found detectable amounts ofa-locus allotypic determinants on rabbit Ig-negative cells (Jensenius et a1 ., 1977). It should be mentioned in this context, however, that the data are mutually supportive with the findings of Krawinkel et ul. (1977a,b,c) and of Cramer et (12. (1977) that antigen binding receptor material isolated from rabbit Ig-negative cells carries a-locus allotype determinants. 5 . ldiotypes on Allocintijien-Reactiue T Cells

The production of anti-Id antibodies to receptors (antibodies or Tcell receptors) with alloantigen specificity has been discussed in Section II,C. Here we discuss the use of these antibodies to analyze the receptors of lymphocytes with reactivity toward alloantigen, particularly the receptors of T cells. This type ofanalysis, including the comparison of alloantigen receptors of T and B cells, is particularly important, since it is not quite certain whether alloantigen recognition is mediated by a principally distinct receptor system, at least as far as T cells are concerned, or b y the same receptor system that is used to recognize conventional foreign antigens. Evidence supporting a distinctness of T-cell-alloantigen recognition primarily comes from reports on the high proportions of alloantigen-reactive T cells, on the double requirements for antigen-recognition by cytotoxic T cells, and on the

226

K. EICHMANN

possible double role of the same T cells both in alloantigen recognition and as effector cells in humoral immunity (Heber-Katz and Wilson, 1976; Howard and Wilson, 1974; Binz and Wigzell, 1977a,b; Fischer-Lindahl and Wilson, 1977; Zinkernagel, 1976; Shearer et al., 1976). Experiments comparing the idiotypes of alloantigen-reactive B and T cells have been done using antisera produced against alloantibodies as well as antisera produced against T cells (Binz et al., 1974a,b, 1975a,b; Binz and Wigzell, 1975a,b, 1977a,b; McKearn, 1974; McKearn et al., 1974a,b). (Lewis xDa)F, rats immunized with Lewis anti-DA alloantibodies produce anti-Id antibodies that react with Lewis spleen cells as well as with purified T cells from that strain, using assay systems such as the uptake of radiolabeled anti-rat Ig antibodies or staphylococcal protein A onto lymphocytes that were previously incubated with the anti-Id antiserum. The reaction was specific because DA cells did not show any uptake of anti-Id antibody (Table XIII) (Binz and Wigzell, 1977a). Similarly (CBA xC57B16)Fl mice immunized with CBA anti-C57B 16 alloantibody yielded antisera that stained spleen cells of CBA mice but not of C57B16 mice, using 1251labeling of anti-Id antibodies (Binz and Lindenmann, 1972a). The exact enumeration of the number of idiotype-positive cells in normal populations employed light microscopy-autoradiography using lZ5Ilabeled anti-Ig antibodies (Binz and Wigzell, 1975a), fluorescein-labeled anti-Ig antibodies in double-sandwich techniques (Table XIV), or triple-layer sandwich techniques with ferritin-labeled TABLE XI11

F, ANTI(LEWIS ANTI-DA)ANTIBODY PRODUCED AGAINST LEWISANTI-DA ALLOANTIBODIES REACTS WITH LEWISSPLEEN CELLS AS WELL AS WITH PURIFIED LEWIST CELLS'

Cellsb Lewis spleen Lewis T DA spleen DA T Lewis spleen Lewis T DA spleen DA T

Incubation with Anti(Lewis anti-DA)

F, normal serum

Uptake of I z 5 I protein A (cP4 4528 k 132 3103 rt 290 835 t 96 822 2 177 822 2 110 810 rt 112

8082 94 9 4 3 % 24

Data from Binz and Wigzell (1977a). 1 x 10' cells analyzed. T = anti-immunoglobulin columns purified T cells. Cells were incubated with anti-idiotypic antibody or normal F, serum, washed, and then incubated with radioiodinated protein A. a

227

IDIOTYPES ON LYMPHOCYTES

antibodies and electron microscopic enumeration (Binz et al., 1975a,b). In all these experiments, about 1% of'the relevant B cells and 5-7% of the relevant T cells were positive for reactivity with antiId antibodies to parental antibodies against a totally different set of histocompatibility determinants (Binz et (11 ., 1975a,b; Binz and Wigzell, 1975a, 1977a,b). Similar proportions of idiotype-positive cells were observed using F, anti-Id antisera produced against parental T cells. Thus, anti-Id antibodies against B cell-derived antibodies as well as those against T-cell receptors detect high proportions of idiotype-positive lymphocytes in the T-cell Compartment and somewhat lower proportions in the B-cell compartment. Even in the B-cell compartment, the proportion of positive cells is significantly above that expected for any conventional antigen-binding specificity, so that the proportion of B cells bearing a particular alloantigen reactivity-associated idiotype is surprisingly high. The proportion of Id-positive T cells is dramatically above that observed for antigen-binding T cells (Hammerling and McDevitt, 1974), and a proportion of about 5% for a particular alloantigen reactivity-associated idiotype has been concluded (Binz and Wigzell, 1975a, 1977a,b). The association of idiotype expression with alloantigen reactivity in these experiments has been established in a number of experiments that showed elimination of the entire alloantigen-reactive lymphocyte population by elimination of the idiotype-positive population (see below). This shows that the alloantigen-reactive population is contained within the idiotype-positive population but does not reveal the identity of both populations. The evidence that all the idiotype-posiTABLE XIV PERCENTAGE OF LEWIST AND B LYMPHOCYTES STAINED WITH FITCLABELEDF,-ANTI(LEWIS ANTI-DA)ANTI-IDIOTYPIC ANTIBODIES~ Cellsb

Anti-idiotype

Lewis T DA T Lewis B DA B Lewis T DA T Lewis B DA' B

Anti(Lewis anti-DA)

Percent positive cells'

6.23 0 1.03 0

0.47 0.75 98.21

Anti-rat Ig

98.10 ~~

" Data from Binz and Wigzell (1975b). T cells were prepared by anti-immunoglobulin (Ig) columns filtration; B cells were prepared by treatment of spleen cells with rabbit anti-rat T-cell antiserum and complement. Between 200 and 3200 cells were counted.

228

K. EICHMANN

tive T cells are reactive toward the relevant alloantigen is as yet rather indirect and relies on the linear relationship between the degree of graft-versus-host (GVH) reactivity and the proportion of idiotype-positive T cells injected. Therefore, it is not formally ruled out at present that the proportion of 5% idiotype-positive T cells for reactivity to one set of alloantigens may b e an overestimate resulting from extensive cross-reactivity of some of the anti-Id reagents within the T-cell compartment. In another system using anti-Id antibodies of (AKR x C57B16)F1mice obtained by immunization with AKR T blast positively selected in mixed lymphocyte culture (MLC) for reactivity with C57B16 alloantigen, splenic concanavalin A (Con A)-stimulated T blasts did not stain above background with the anti-Id antibody and FITC-labeled anti-Ig, whereas generally more than 60% of T blast selected in the relevant MLR were stained (Krammer, 1978).These latter experiments suggest a smaller proportion of idiotype-bearing T cells for each set of alloantigens. Taken together, the question of the idiotypic diversity of T-cell receptors for alloantigen needs further evaluation. Anti-Id antibodies to alloantibodies and to T-cell receptors with reactivity toward the same set of alloantigens have been shown to differ with respect to the number of idiotypic determinants detected. Whereas absorption with spleen cells or insolubilized alloantibodies removes all activity from both kinds of antisera, absorption with purified T cells removes all activity from anti-Id antisera toward T-cell receptors but leaves some anti-B cell activity in anti-B-Id antisera (Table XV) (Binz and Wigzell, 1977a). This has been interpreted to suggest that T cells express or present part of the idiotypic determinants of B cells and, whereas B cells also possess unique idiotypic determinants, T cells possess only the common ones. This result is in perfect agreement with the data on idiotype presentation of T helper cells which express the VHidiotypic determinants in common with B cells but lack their VL determinants and possibly also their VH-VL combination determinants (Eichmann, 1977a; Krawinkel et a2 ., 1977a). The VH nature of alloantigen-specific T-cell receptors is also suggested b y experiments that showed genetic linkage of T-cell idiotype expression in the rat to the IgA allotype locus (Binz et al., 1977; Binz and Wigzell, 197713). F2 animals were produced between strains Lewis and DA that differ both in their H-1 (Ag-B) major histocompatibility complex (MHC) as well as in their IgA allotype. Typing of each individual for H-1 and for uptake of F,-anti (Lewis anti-DA) anti-Id antibody by its lymphocytes revealed nonlinkage of idiotype expression to H-1. Allotype linkage was shown by typing of eight individuals that were homozygous for the Lewis allele of H-1. Three animals showed

IDIOTYPES ON LYMPHOCYTES

229

uptake of anti-Id as well as the Lewis IgA allotype whereas the remaining 5 animals lacked this allotype and did not show uptake of the anti-Id antibody. Although obtained with a small number of animals, these results are indicative of the linkage of genes that control T-cell idiotype expression to the H-chain linkage group of the rat, and are therefore in perfect accord with the data on the genetic control of T helper-cell idiotypes in the mouse (see above). Nonetheless, the control of T cell receptor idiotypes by VH genes only was in a way unsatisfactory as it did not explain the many specificity differences observed between T and B cell receptors as well as the profound influences of the H-2 complex on T cell antigen recognition and on the mature T cell repertoire. Krammer and Eichmann (1977, 1978)have therefore reinvestigated the question of the genetic control of T cell idiotypes in a system similar to that of Binz and Wigzell, using an (AKR x C57B1/6)Fl anti-(AKR anti-C57B1/6) T cell blast antiserum (Krammer, 1978). As shown in Table XVI, this antiserum reacts with T cell blasts recovered from MLRs of AKR T cells stimulated by C57B16 lymphocytes, but not with T cell blasts recovered from MLRs in which SJL T cells were used as responders. The difference in H-2 as well as Ig-1 haplotypes between strains AKR and SJL facilitated the study of the genetic control of T cell idiotype expression in a (AKR x SJL)Fl x SJL backcross experiment. The data in Table XVI clearly show that only those backcross mice express the idiotype that possess both the H-2 haplotype and the Ig-1 allele of the Id-positive parental strain AKR, suggesting that at least two genes, one in the MHC and the other in the heavy chain linkage group, control the idiotypes of T cells. These data are not in contrast with the single gene (V,) control of the expression of the A5A idiotype on T helper cells (Hammerling et al., 1976) since in the latter study antisera produced against T cells were used whereas the former utilized antisera produced against antibodies. Rather, the data strongly suggest that T cells possess, in addition to those shared with antibodies, unique idiotypic determinants that are carried either by products of the MHC or by other gene products upon which the MHC exerts a regulatory influence (Krammer and Eichmann, 1978). Anti-Id antibodies to allo-reactive T cells were also tested in a variety of functional T-cell assays. All of those data reveal the reactivity of the anti-Id antibody with the functionally defined T-cell populations involved in the recognition of and the reactivity to the relevant set of alloantigens. The functional assays used and the results obtained have recently been quite comprehensively reviewed (Binz and Wigzell, 1977a) and are therefore summarized in Table XVII (Binz and Askonas, 1975; Binz and Wigzell, 1975a,b, 1977a,b; Binz, 1975; McKearn et al., 1974a,b; Krammer, 1977). The interesting aspect of this work

TABLE XV

T CELLSABSORBONLYPARTOF THE ACTIVITY OF F, ANTI-IDIOTYPE TOWARDB CELLS BUT ABSORB ALL ACTIVITY TOWARD T CELLS; B CELLSOR ALLOANTIBODIES ABSORB ALL ACTIVITY OF F, ANTI-IDIOTYPE TOWARDB AND T CELLS Lewis cells treated with

Absorbed with

Experiment and results Cpm:uptake of 12sI-labeled protein A*

F, anti (L anti-DA Ah)

Lewis T Lewis spleen

Normal F, serum

Lewis spleen

Lewis T

4528 2221 1136 822

3 103 1050 1010 8 10

Cpm:uptake of 1251-labeled anti-rat Ig on Lewis T cellsC (L anti-DA) IgG -

F, anti (L anti-DA T) Normal F, serum

70075 24505 22376 MLC:Lewis anti& x DA)FLd Cpm: PH]TdR incorporation

F, anti (L anti-Da T) + C’ Normal F, serum Untreated

+ C’

(L anti-DA) IgG -

1938 30929 32112 32933 Local GVH reaction; Lewis or DA cells in (L X DA)F1‘ Weight (mg) of

F, anti (L anti-DA T) + C’ Normal F, serum ~

+ C’

_

_

_

(L anti-DA) IgG -

“Lewis” node

“DA” node

6.1 27.7 28.5

32.6 25.0 26.4

_ _ _ _ _ _ ~ ___

Data from Binz and Wigzell (1977a). Data were pooled from several experiments; many controls were omitted for the sake of simplicity. Lewis spleen or Lewis purified splenic T cells were incubated with anti-idiotype to alloantibody before and after the indicated absorptions. Cells were washed and thereafter incubated with radioiodinated protein A. Purified Lewis T cells were incubated with anti-idiotypic antiserum to T cells before and after absorption on insolubilized IgG from alloantiserum. Cells were washed and then incubated with radioiodinated rabbit anti-rat Ig. Mixed-lymphocyte culture performed with responder spleen cells with and without treatment with the anti-idiotypic antibody and complement. Graft-versus-host reaction performed by injecting Lewis spleen cells into left foot (“Lewis” node) and DA spleen cells into right foot (“DA” node) of F, animal. Cells were treated with anti-idiotype and complement before injection. Regional lymph node weight was taken as measurement of GVH reaction. a

231

IDIOTYF'ES ON LYMPHOCYTES

with regard to the analysis of T-cell receptors for alloantigen is that all or nearly all of the functionally distinct effector cells reactive toward one set of alloantigens appear to be recognized by one anti-Id reagent, although it is known that, for instance, MLR-reactive cells react to determinants that are largely distinct from those that are detected by cytotoxic T cells (Shreffler and David, 1975). Thus, parental T cells injected into an F, animal appear to give rise to various subpopulations of T effector cells, the sum of which expresses the total or almost total complement of T-cell combining sites for one haplotype, whereas each of the subpopulations expresses only those idiotypes that are characteristic for recognition of its particular MHC subspecificity.

6 . Idiotypes of Soluble T-cell Products Anti-Id antibodies with alloantigen reactivity have been used to demonstrate and isolate receptor material in normal rat serum or urine, or culture supernatants. Insolubilized F,-anti (Lewis-anti-DA) idiotypic antiserum absorbed, from large quantities of normal Lewis serum or urine, a population of idiotype bearing alloantigen-reactive molecules that fractionated on Sephadex G-200 into two fractions, one cochromatographing with IgG, and the other with albumin (Binz and Wigzell, 1977a,b). Sodium dodecyl sulfate (SDS) polyacrylamide electrophoresis revealed four components , one of which corresponded to 7 S IgM whereas the remaining three had molecular weights ( M J of TABLE XVI CONTROL OF T CELL RECEPTOR IDIOTWE EXPRESSION BY GENES IN THE HEAVYCHAIN LINKAGE GROUPAND THE MAJOR HISTOCOMPATIBILITY COMPLEX^ T cell blasts from MLR" Responder

AKR (H-2kIg-ld) SJL (H-2s Ig-lb) Group 1 (H-2d8Ig-lb'b)D Group 2 (H-2dsIg-lb'd)D Group 3 (H-2s'kIg-lb'b)b Group 4 (H-2"kIg-lb'd)b

Stimulator

Positive responder T cell blasts with F l a (AKRaB6) anti-idiotypic serum (%)'

B6 B6 B6 B6 B6 B6

35.1 5.0 6.2 6.6 6.3 40.6

a Details of the preparation of T cells, MLR conditions and blast isolation from Krammer (1978). Group 1 to 4 are mice from a backcross of SJL mice with (AKR x SJL) F1 hybrid mice. The conditions to detect Id+T cell blasts in indirect double immunofluorescence have been described in Krammer (1978). Data from Krammer and Eichmann (1977).

232

K. EICHMANN

TABLE XVII DEMONSTRATIONOF IDIOTYPESON T CELLS USING F, ANTI-IDIOTYPIC ANTISERA IN FUNCTIONAL ASSAYS"

Functional assay Inhibition of mixed lymphocyte reacaction (MLR)

Inhibition of Graft-versus-Host (GVH) reaction

Species, strain (responderlstimulator)

Treatment with anti-idiotype

+

Rat: LewisIDNA Mouse: CBAIC57B16

Pretreated with anti-Id complement

Mouse: AKRIC57B16

Addition of anti-Id to cultureb

Lewisl(Lx DA)F, DA/(LX DA)F, CBA/(CBAx C57B16)F1 Lewis/(Lx DA)F,

Preimmunization of F, animal with parental T cells

CBAI(CBA xC57B 16)F, Rat: DAIBN LewislBN LewislDA Rat: LewislAu LewislDA LewislBN

Injection of F, animal with anti-Id Transfer of immune F, T cells into F, host Treatment of parental responder cells with anti-Id + complement Fractionation of parental responder cells on affinity column made from insolubilized anti-Id

Accumulation of specific cells in lymph node draining skin graft

LewislDA

Injection of radioiodinated anti-Id and subsequent autoradiography of regional lymph node

Delay of graft rejection

LewislDA

Preimmunization of recipient with positively selected syngeneic T blast' Preimmunization of recipient with idiotype bearing receptor material isolated from normal serum

LewislDA

Data extensively reviewed in Binz and Wigzell (1977a). P. Krammer, personal communication. The effect of this immunization has also been deomonstrated by unresponsiveness in MLR and inability to develop cytotoxic T cells in MLR and GVH reactions (Binz and Wigzell, 1977c; also personal communication).

IDIOTWES O N LYMPHOCYTES

233

150,000, 70,000, and between 30,000 and 40,000, respectively. Upon reduction and alkylation, the 150,000 M , peak disappeared whereas the 70,000 and 35,000 peaks remained. The latter three components were found to b e T cell-derived receptor material as shown by their presence in supernatants from purified T cells, by their absorption to DA lymphocytes, and by inhibition of the uptake of F, anti-Id onto Lewis lymphocytes (Binz and Wigzell, 1977aJi). These data are in agreement with the molecular weight detemiination of T-cell receptor material recovered from haptenated nylon (see below) (Krawinkel et nl., 1977a,b; Cramer et al., 1977) in suggesting that T cells produce a dimeric receptor molecule with two subunits of 70,000 M,, a breakdown product of which has an M , of about 35,000. In partial accord with this interpretation are recent results on the molecular weight of molecules carrying ASA idiotype (see above) recovered from the supernatants of Con A-stimulated purified T cells of strain A/J mice presensitized with anti-A5A IgG, idiotype. SDS-acrylamide gels revealed a major component of -36,000 M , together with a very small peak of -70,000, both of which were T-cell specific. In addition, IgG as well as IgM peaks were observed that were most likely derived from costimulated contaminating B cells (Eichmann and Melchers,

1978). Receptor material has also been isolated and idiotypically analyzed from mouse spleen cells and enriched T- and B-cell populations, using antigen-coupled nylon nets (Krawinkel and Rajewsky, 1976; Krawinkel et ul., 1977a,b,c; Cramer et d., 1977; Reth et d . , 1977). Nylon nets coupled with hapten retain a population of hapten-primed B and T cells at 4”C, which can be released by shifting the temperature to 25°C. Subsequent treatment of the nets with acid buffer or hapten yields soluble hapten-binding material that can b e determined in haptenated-phage-inhibition assays. This material has been characterized with respect to its biochemical and immunochemical properties and cellular origin. Shortly, material recovered from B cells carries all the immunochemical properties of Ig whereas material recovered from T cell has a molecular weight (- 150,000) similar to that of IgG but does not carry constant-region determinants of Ig. However, variable-region (V,) determinants are observed on T cell-derived as well as on B cell-derived material, including u locus allotypes in the rabbit and idiotypic determinants of antibodies to the hapten NP in mouse strain C57B16. The idiotype of primary anti-NP antibodies in this strain is associated with a particular fine-specificity behavior, socalled “heteroclicity,” which consists in a greater affinity of these antibodies to the heterologous hapten NIP (Imanishi and Makela, 1973, 1974; Makela and Karjalainen, 1977a,b; Reth et d . , 1977). Heteroclicity as well as idiotype are genetically polymorphic characters that are

K. EICHMANN

234

TABLE XVIII ABSORPTION OF HAPTEN-BINDING,IG-NEGATIVERECEPTOR MATERIALTO INSOLUBILIZED ANTI-NP-IDIOTYPIC ANTIBODY (ANTI-Id)”

Hapten-binding material

Percent phage-inactivating activity removed by anti-Id

C57BL/6 NP specific receptor (spleen) C57BW6 NP specific receptor (T cells) CBNJ NP specific receptor (spleen) C57BL DNP specific receptor (spleen)

54 85 6 7

a Data from Reth et al. (1977) and Krawinkel et al. (1977b). Anti-NP idiotypic antibody was prepared by injecting isolated primary anti-NP antibody from C57BU6 mice into IgC-tolerant guinea pigs (Jack et al., 1977). Receptor material was passed first through an anti-immunoglobulin (Ig) column to remove Ig-positive receptor material. Ig-negative receptor material was then passed through anti-Id column.

controlled by a VHgene (Makela and Karjalainen, 1977a). Ig-negative receptor material recovered from NP-primed C57B16T cells possess both heteroclicity and idiotype in common with antibodies (Table XVIII suggesting a V, participation in the receptor structure (Krawinkel et a1 ., 1977a; Rajewsky and Eichmann, 1977; Cramer et a1 ., 1977). A series of recent experiments have dealt with soluble T-cell products that appear to be gene products of the MHC (H-2) complex of the mouse since they carry I-region determinants (Taussig et al., 1975; Munro and Taussig, 1975; Tada et al., 1975; Takemori and Tada, 1975).These molecules are demonstrable as activities in supernatants or lysates of antigen-primed T cells that possess either T helper cellreplacing or T suppressor cell-replacing activities. In one instance, a TGAL-specific helper factor has been analyzed for the presence or the absence of idiotypic determinants using an anti-Id antiserum prepared against anti-TGAL antibodies (Haimovich, 1977; Moses and Haimovich, 1977), and it was found that the helper activity was abrogated after passage through insolubilized anti-Id columns. Since this factor is also removed on anti-Ia columns (Munro and Taussig, 1975), the experiments may be a first indication that gene products of the MHC as well as VHgenes contribute to this factor. This has been suggested as one of the possibilities for a unifying hypothesis on the structure of the T-cell receptor (Rajewsky and Eichmann, 1977), but many more experiments need to be done in order to clarify a possible association between products of the MHC and immunoglobulin VH genes.

IDIOTYPES ON LYMPHOCYTES

235

IV. Functional Role of lymphocyte Receptor ldiotypes

Jerne (1974, 1976) has proposed a concept of the immune system as a dynamic multispecies system in which the various species are the variable domains of the lymphocyte receptors. Each combining site interacts with a number of others that carry complementary idiotypic determinants, such that by the sum of all interactions a stable network is formed. This concept has been refined, discussed, and extended in a number of recent publications (Richter, 1975; Hoffmann, 1975; Raff, 1977), so that a brief outline may suffice here: the concept is based on the large diversity of both combining sites and idiotypes and on the degeneracy of the combining specificity. Taken together, these properties imply that each combining site, in addition to binding a number of foreign antigenic determinants, necessarily also combines with a number of idiotypic determinants within the same immune system. This leads to the formation of a large number of idiotype-anti-Id bonds involving receptors on lymphocytes as well as soluble antigenrecognizing molecules, thus establishing a “formal” network. The concept is further based on the assumption that a given idiotypic determinant may be associated with a number of different combining sites, not all of which bind the same antigen, and that combining sites binding the same antigen may be associated with a number of different idiotypic determinants. These assumptions lead to the postulate of a “functional” network: the formation of a first set of combining sites (Abl) is stimulated by antigen. This first set affects the number of combining sites in a second, idiotypically complementary set (Ab2) either in a suppressive or stimulatory fashion, and so on until certain thresholds are reached and the chain reaction stops. The nonidentity of idiotypically defined sets of combining sites with those sharing the same antigen-binding specificity implies that each successive Ab set consists of two operationally different populations, one complementary to the site-associated idiotypic determinants of the Ab(n-1) set and resembling the shape of the Ab(n-2) set, and the other complementary to the nonsite-associated idiotypic determinants of the Ab(n1) set and not resembling the Ab(n-2) set. In the Ab2 set these two populations are referred to as “internal image” of the antigen, and as “parallel set,” respectively. The existence of the two sets ensures the eventual termination of the chain reaction because of the increase in heterogeneity with each successive step. Taken together, the network concept leads to a view of the immune system as a self-regulatory antigen-independent entity, which stabilizes itself using its own elements. Introduction of an antigen leads to the temporary disturbance of this stability until all receptors participating in the reaction have reached a new state of equilibrium. Be-

236

K. EICHMANN

cause much of the evidence on the idiotypic properties of lymphocyte receptors has been summarized in this article, it seems to be an appropriate place to critically evaluate this experimental evidence together with the functional aspects of lymphocyte receptor idiotypes predicted in a network model. A functional network requires the mutual recognition of idiotypes on and by B and T lymphocytes, and quantitative consequences of this recognition.

A. RECOGNITIONOF IDIOTYPES WITHIN THE IMMUNE SYSTEM 1 . Recognition of B- and T-cell ldiotypes by B Cells In principle, B-cell recognition of B-cell or antibody idiotypes is shown by the ability of an immune system to produce anti-Id antibodies to other antibodies or to T-cell receptors. For a network, it is important to evaluate the production of anti-Id antibodies against autologous idiotypes, and most of the experiments on this subject have been summarized in Sections II,B and II,C. To reiterate, anti-Id production in mice has been demonstrated to autogenous antibodies to phosphorylcholine (Kluskens and Kohler, 1974; Cosenza, 1975; Cosenza et al., 1977c), in rats to H-2-coded antigens (McKeam et al., 1974a,b), and in rabbits against autogenous antibodies to p-aminophenyl-N-trimethylammonium-chlorideand several other haptens, such as DNP and TNP (Rodkey, 1974,1976, also personal communication). Since these antigens have no particular property in common that would distinguish them from other conventional antigens, it is not unreasonable to propose that each individual’s immune system is capable of recognizing its own antibody idiotypes. Anti-Id production to isogeneic immunoglobulins has been demonstrated using BALB/c myeloma proteins with and without known antigen-binding specificity (Sakato and Eisen, 1975; Sakato et ul., 1977; Eisen et a1 ., 1975; Sirisinha and Eisen, 1971; Hannestad et a1 ., 1972; Cosenza et al., 1977c) and the A5A antibody to A-CHO in A/J mice (Eichmann, 1972). In these experiments it was shown that an antibody response could be elicited to any immunoglobulin that has been tried, although differences in immunogenicity were observed (Sakato et al. 1977). In many respects, however, the isogeneic situation may not be strictly comparable to an autogenous one, as it is not clear whether idiotype expression is under strict genetic control in general or only in certain cases (Eichmann, 1975b; Weigertet al., 1975; Makela and Karjalainen, 1977a). In the latter case, many idiotypes could be analogous to alloantigens although immunization is done in isogeneic combinations. It should be mentioned, however, that even with those idiotypes that are expressed in close to 100% of the members of one inbred strain of mice, isoanti-Id antibodies are obtained (Eichmann,

IDIOTYF’ES ON LYMPHOCYTES

237

1972; Cosenza, 1975). I n these cases, which include the A5A idiotype and the T15 idiotype, isogeneic and autogeneic anti-Id responses may be analogous to one another, both suggesting that antibody idiotypes are exceptions to the law of self-nonself discrimination. Recognition of T-cell idiotypes by B cells has been shown in all those experiments that employ anti-Id antisera produced in F, heterozygotes against the T cells of one parent that are specific for MHC antigens of the other parent (Ramseier et al., 1977; Krammer and Eichmann, 1978; Binz and Wigzell, 1977a). These experiments have been summarized in Sections II,C and III,B. The relevance of these experiments for the network hypothesis depends on the interpretation of the immunogenetic situation in the parent-F, combination in which the F, animal is thought to lack the parent’s receptors for the other parent’s MHC antigens. If this were the basis for immunogenicity of parental idiotypes in F, animals, the model would be irrelevant for the network concept. In some recent experiments, however, it was shown that T-cell idiotypes are in fact immunogenic in animals of the same strain, i.e., in an isogeneic situation. T-lymphocyte blasts recovered from an MLR (CBA anti C57B16) were reinjected into CBA mice together with Freund’s complete adjuvant. This treatment rendered the mice unresponsive to C57B16 MHC antigens, but not to third-party MHC antigens, suggesting that anti-Id immunity has developed (Andersson et al., 1977a). Furthermore, Lewis rats injected with T-cell receptor material isolated from normal Lewis serum using F, anti-(Lewis-antiDA) anti-Id antibody became partially tolerant to DA grafts, but not to third-party grafts (Binz and Wigzell, 1977b,c). In the latter experiments, it was not investigated whether the induced tolerance is d u e to B- or T-cell autoimmunity or to both. Thus, recognition of T-cell idiotypes by B cells has not been clearly demonstrated thus far within a tnie autogeneic situation. As in the previous example, however, it has been shown that the idiotypes of T cells involved in alloantigen recognition are expressed in almost all animals of an inbred strain, such that isogeneic and autogeneic situations may be analogous in these experiments. In any case, experiments clearly demonstrating the recognition of T-cell idiotypes by B cells of the same immune system still needs to be done.

2 . Recognition c$T- and B-Cell Zcliotypes by T Cells T cells of BALB/c mice immunized with BALBlc myeloma proteins exert helper function for B cells sensitized to haptens that are coupled to these myeloma proteins (Janeway et ul., 1975b; Jbrgenson et at., 1977). In the experiments of Janeway et a l . (1975b), BALB/c mice were primed with several myeloma proteins, and their spleen cells

238

K. EICHMANN

TABLE XIX IDIOTYPE-SPECIFIC HELPER ACTIVITY OF BALB/C SPLEEN CELLSPRIMED WITH MYELOMAPROTEINS FOR BALB/C B CELLSPRIMED WITH DNP" IN ADOPTIVE TRANSFER EXPERIMENTS DNP-primed B cells plus cells primed with

MOPC 167 MOPC 603 ~

DNP-M167

Host boosted with DNP-M603

DNP-M460

4.5b 6 18 4.5

0.6 8.4 35.4

0.36 6.3 0.12

~

~~

~

~~

" Data froin Janeway et al.

(1975b). * Geometric means of reciprocal serum volume required to bind 3.3 x DNP-lysine at 3.3 pmol/rnl DNP-lysine.

pmol of

were transferred together with purified B cells from BALB/c mice primed with DNP-KLH. The adoptive hosts were boosted with the myeloma proteins to which DNP has been coupled and idiotype-specific helper activity was observed (Table XIX). These experiments clearly show that T helper cells recognize isogeneic idiotypic determinants on B-cell products. In another series of very recent experiments, many of which are unpublished, it could be shown that T helper cells recognize idiotypic determinants on the B cells with which they cooperate or which they suppress. Cantor (1977), injecting rabbit anti-Id antibodies to strain A/J anti-arsonate (ARS) antibodies (Shyr-Te et al., 1977; Nisonoff et al., 1977) into A/j mice before they were carrier-primed with KLH, showed that the KLH-specific helper population recovered from these mice lacked the ability to cooperate with arsonate-specific B cells producing ARS-idiotype-positive antibodies but retained the ability to cooperate with idiotype-negative arsonate-specific B cells (Cantor, 1977). Conversely, T helper cells sensitized with anti-Id antibody to the A5A idiotype, when offered B cells sensitized with Strep.A that consist of A5A-positive and A5A-negative A-CHO specific B cells, cooperate almost exclusively with the A5A-positive B cell population (Table XX) (Eichmann, 197713) whereas T helper cells sensitized with Strep.A cooperate both with ASA-positive and A5A-negative ACHO specific B cells. It is difficult to explain these results without the assumption that Bcell idiotypes are recognized by T helper cells. T helper cells, however, are also specific for antigen and, in addition, recognize I-region determinants on B cells with which they cooperate (Katz et al., 1973, 1975; Sprent and von Boehmer, 1976). Without a network concept, one would have to assume specific receptors for all these structures on one type of helper T lymphocyte. In contrast, within a strict network

239

IDIOTYPES O N LYMPHOCYTES

model, one T cell recognizing the idiotype of one B cell is sufficient to explain all phenomena. This T cell cooperates with idiotype-bearing B cells because of its anti-Id receptor. In addition, it acquires specificity for antigen by uptake of soluble idiotype-positive receptor molecules. When offered idiotype-positive and idiotype-negative B cells, both with antigen-specificity, idiotype-positive B cells will be more frequently cooperated with because of the additional recognition of the B-cell receptor idiotype. Alternatively, anti-Id T helper cells could function as a cooperative cell type in addition to antigen-specific T helper cells. This latter model does not need the assumption of uptake of receptors by the T helper cell but requires two cooperative T cells for one B cell. It should be mentioned that except for the experiments described above, which clearly demonstrate the recognition of B-cell idiotypes by helper T cells, none of the above-mentioned speculations have been corroborated by direct experimentation. On the other hand, the network concept offers a useful approach to dissect the observed phenomenology into a number of testable working hypotheses. Recognition of B-cell idiotypes by T suppressor cells has been demonstrated in A/J mice that are suppressed for the ARS idiotype (ShyrTe et al., 1977; Nisonoff et al., 1977).When these mice are injected repeatedly with high doses of antigen, they develop suppressor T cells whose specificity can be demonstrated by rosetting with mouse TABLE XX SELECTIVE COOPERATION OF T CELLSSENSITIZED BY ANTI-IDIOTYPE B CELLS DEVELOPING INTO PFC SECRETING ANTIBODY WITH THE SAME IDIOTYPE" ~

~~

Cells

T'

B*

Unprimed 1" Strep.A 1" anti-A5A Unprimed 1" Strep.A I" anti-A5A U npri in ed 1" Strep.A 1" anti-A5A

Unprimed U nprimed Unprimed 1" Anti-A5A 1" Anti-A5A 1" Anti-A5A 1" Strep.A 1" Strep.A 1" Strep.A

WITH

PFCb anti-A-CHO 4

17 33 33

154 200 16

2.53 199

PFC after inhibition with anti-A5A

8 17 9 17 9

17 28 153 29

Percent A5A-positive PFC

95 92 39 86

Data from Eichrnann (1977b). cells (PFC) per 106viable cells recovered from culture; mean of 4-6 identical cultures, standard error between 1.1 and 1.5. Nylon wool-purified lymph node T cells, >95% immunoglobulin (Ig) negative. Anti-Thy-1.2 treated spleen, >85% Ig positive. a

* Plaque-forming

240

K. EICHMANN

red blood cells coated with Fab fragments of the idiotypic antibody (Owen et a1 ., 1977) The rosetting phenomenon is idiotype-specific and can be used to deplete or enrich for the suppressor cells. Recognition of T cell idiotypes by suppressor T cells is demonstrated in A/J mice pretreated with low doses of guinea pig IgG-2 anti-A5A-idiotypic antibody. These mice develop suppressor T cells that prevent the formation A5A-positive antibodies upon subsequent immunization with Strep.A vaccine (Eichmann, 1974, 1975a). In recent experiments it has been shown that the targets of these suppressor T cells are the precursors of the A5A-positive T helper cells, suggesting that the idiotypes of these T helper cells are recognized by the suppressor T cells (Hetzelberger and Eichmann, in preparation). Recognition of T-cell idiotypes by other T cells is suggested by experiments of Binz and Wigzell (1977a) in which purified T cells from F, mice immunized with parental T cells were transferred into irradiated F1recipients. The recipients became resistant to the GVH reaction of the relevant parental lymphocytes, but not to that of the other parent, suggesting that the transferred anti-Id F, T cells suppress the reactivity of the GVH reactive parental T cells via idiotypic recognition.

3. Different ldiotypic Determinants Recognized by T and B Cells In a recent study, Jbrgenson and Hannestad (1977) compared the idiotypic determinants detected by T helper cells and by antibodies of BALB/c mice immunized with the isogeneic myeloma protein MOPC 315 (M315). Immunizations were performed with the intact myeloma protein and with several fragments including Fv as well as free L chains. Anti-Id antibodies induced with free M315 L chain reacted specifically with free M315 L chain, but only minimally with intact M315 molecules. In contrast, helper cells induced with L chains of M315 augmented the response of B cells to hapten conjugated M315 equally well as helper cells induced with total M315 protein. These data have been interpreted to suggest that helper cells recognize determinants that are present on isolated chains as well as on the total M315 molecule and which are therefore not associated with the combining site. In contrast, B cells react primarily with binding site-related idiotypic determinants as shown also by other authors (see Section I1,A) (Helman et a1 ., 1976). These data are extremely interesting, as they suggest, if generalized, a mechanism for drawing an idiotypic borderline between T- and B-cell compartments within a network: T cells may not distinguish between the incomplete VH idiotype of T cells and the complete VH-VL idiotype on the B cell. In contrast, some B cells may recognize only T-cell idiotypes, and other B cells may rec-

IDIOTYPES ON LYMPHOCYTES

24 1

ognize only B-cell idiotypes. Without any further experimental clue, it is not useful to construct a special meaning of this hypothetical difference for a functional network. It may, however, be of fundamental importance for the distinction between the T- and B-cell compartments. OF RECOGNITIONOF IDIOTYPESWITHIN B. CONSEQUENCES THE IMMUNE SYSTEM

1 . Suppression of Zdiotype Expression Suppression of idiotype expression by large doses of anti-Id antibody in adult mice has been discussed in Section II1,A as a means to analyze idiotype-bearing B-cell populations. Within a network, this model may not reflect any naturally occurring situation, as the concentrations of injected anti-Id antibody may rarely be reached in vivo. Also, the effects are transient, so that normal response patterns are reached as soon as 2-3 weeks after the injection of anti-Id. Suppression phenomena that may indeed reflect network effects have been observed in adult mice using small doses of anti-Id, in adult mice using large doses of anti-Id together with repeated administration of antigen, and in newborn mice. Adult A/J mice injected with small doses (0.01-0.1 p g of IBC) of guinea pig anti-ASA idiotype antibody or of its IgG2 fraction become suppressed gradually within a period of about 6 weeks for their ability day: i l t i r iId

FIG. 2. Antibodies with A5A idiotype elicited by immunization with Group A streptococci (Strep.A) in slightly irradiated (200 r) A/J mice that had received 1 x 1oB spleen cells from A/J mice treated at various times before transfer with IgG2 anti-ASA idiotypic antibody (0.1 pg of idiotype binding capacity) (0)or with equivalent amounts of normal guinea pig IgG (0).For experimental details see Eichmann (1975a).

242

K. E I C H M A "

to produce antibodies with A5A idiotype upon immunization with Strep.A (Fig. 2). The suppressed state is maintained for the lifetime of the mouse and can be transferred b y splenic T cells in serial consecutive transfers into 200 r-irradiated syngeneic recipients (Eichmann, 1974, 1975a).Without the injection of antigen, suppressor cell activity decreases with time, but with immunization of each consecutive host, suppressor cells have been transferred up to 10 times over a period of 3 years (K. Eichmann, unpublished). Unlike other suppressor cells, 46 weeks are required to achieve complete suppression of the host, and sometimes transfer of less than lo5 spleen cells is sufficient (Eichmann, 1975a).The suppression is essentially due to unresponsiveness of the A5A-positive T helper cells to subsequent immunization with the Strep.A antigen (Hetzelberger and Eichmann, unpublished), although a slight reduction of the A5A-positive B precursor cells has been observed (Eichmann et al., 1977).These data imply that the response of the A5A-positive B cells depends on the responsiveness of A5A-positive T helper cells, and are another indication for the recognition of idiotypes in T-B cooperation (see also Table XX) In another system, adult A/J mice can be maintained in an idiotypesuppressed state by injecting large quantities of rabbit-anti-Id antibody directed to a cross-reactive idiotype (ARS) associated with the antibody response to the arsonate hapten, when repeated antigen injections follow the injection of the anti-Id antibody (Shyr-Te et al., 1977; Pawlak et a1 ., 1974; Nisonoff et al., 1977). Suppression can be transferred using peritoneal exudate cells containing a large proportion (- 10%) that form rosettes with idiotype-coated red cells. Suppression can also be transferred with splenic lymphocytes as well as with purified T and B cells. This suppression was reported at least in part to be mediated by competition for antigen between idiotype-positive B cells of the host and an excess of idiotype-negative transferred B cells. Idiotype-suppressor T cells, however, also play a role (Nisonoff et ul., 1977; Shyr-Te et al., 1977). In the third system it was shown that newborn BALB/c mice, when injected with allogeneic anti-Id antibody to the T15 idiotype, become suppressed for the entire response to phosphorylcholine (PC) for a period of several months (Cosenza et al., 1977c; Kohler, 1975; Kohler et al., 1974). Thereafter, responsiveness to PC is gradually regained, but the response in both the B-cell and the T-cell compartments is essentially T15 idiotype-negative (Cosenza et al., 1977c; Augustin and Cosenza, 1976). The duration of suppression is dose-dependent and increases with increasing amounts of anti-Id injected. Suppression can b e measured by antibody production, plaque-forming cells as well

IDIOTYPES ON LYMPHOCYTES

243

TABLE XXI PRECURSOR FREQUENCY OF PHOSPHORYLCHOLINE (PC)-SPECIFIC B CELLS RESPONSIVETO R36A PNEUMOCOCCI IN BALB/c MICE INJECTED AS NEWBORNS WITH ANTI-T1SIDIOTYPE" Precursors/lO" spleen cells at 1.5 weeks of age Anti-T 15 Spleen cells

( P a IBCY

T15+

T15-

Normal BALB/c Neonatal suppressed

1.0 0.1 0.01 0.001

106 3 4 0 22

1s 1s 23 30

8

" Data from Cosenza et al. (19774. Graded numbers of spleen cells were cultured together with R36A pneumococci and antibody or plaque-forming cells to PC were detected on PC-coated sheep red blood cells. Positive and negative cultures were counted, and the data were analyzed according to Poisson's distribution. IBC, idiotype binding capacity.

as by enumeration of B-cell precursors (Table XXI) (Cosenza et al., 1977~). During the suppressed state, the mice respond with high titers of autogenous anti-Id when injected with T15 myeloma protein. This autogenous anti-Id is suppressive for other newborn mice, but less so than allogeneic anti-Id (Cosenza et al., 1977a). Suppression of BALB/c mice for the expression of T15 idiotype can also be achieved by irradiation and reconstitution with normal bone marrow or fetal liver (Augustin et al., 1977). This has been interpreted as suggestive of compartmentalization of T15 precursor cells, but experimental proof for this is still missing. I n all these experiments, suppressor cells have not been found in spite of an extensive search (H. Cosenza, personal communication). Taken together, from the suppression experiments summarized above it seems that administration of small quantities of anti-Id antibody has profound and long-lasting consequences in a functional immune system. The amounts administered to achieve chronic suppression on adult A/J mice and in newborn BALB/c mice are equivalent to 1/106to lo7 circulating immunoglobulin molecules. Administration of such quantities of soluble anti-Id antibody may result in permanent dramatic alterations of the lymphocyte compartment, which became evident as (a) amplification of suppressor T cells, (b) a shift in the ratio of idiotype-positive to idiotype-negative precursor B cells, (c) a shift in the number of idiotype-positive to idiotype-negative T helper cells

244

K. EICHMANN

(see Section IV,A) and (d) an increase in the responsiveness of B cells bearing anti-Id receptors for the idiotype in question.

2 . Enhancement of Zdiotype Expression Enhancement of idiotype expression by anti-Id antibody was first observed in experiments using separated guinea pig IgGl and IgG2 anti-Id antibodies to the A5A idiotype. Whereas IgG2 fractions were suppressive, no suppressive effect, but rather an enhancing effect, on the A5A-positive antibody response was observed with the IgGl fractions of the same antisera (Eichmann, 1974). Subsequently, enhancement by anti-Id antibody was seen also in the idiotype-positive T helper-cell compartment (Eichmann and Rajewsky, 1975; Black et a1 ., 1975, 1976a,b; Hammerling et al., 1976a; Rajewsky et al., 1976; Eichmann, 1977a; Cosenza et al., 1977c) in the suppressor T-cell compartment (Eichmann, 1975a), and in the DTH-reactive T-cell compartment (Cosenza et al., 1977b). These experiments have been discussed in Section II1,B. Common to all these experiments is that low doses of anti-Id antibody are required for optimal sensitization in the B- as well as in the T-cell compartments, whereas high doses are commonly suppressive. An exception to this rule appears to be the suppressive effect of low doses of guinea pig IgG2 anti-Id antibodies, which, however, is based on the stimulating effect on suppressor T cells. Another exception appears to be the suppressive effect even of low doses of anti-Id antibodies for the T15-positive response of BALB/c mice treated as newborns with anti-T15-Id antibody. In this case, although suppressor cells have not been found, it is difficult to explain the long-lasting effect without the induction of a cellular suppression mechanism of the host. In the experiments discussed above, enhancement is always demonstrated by administration of antigen in addition to anti-Id antibody. In another series of experiments, enhancing functions of anti-Id antibodies or anti-Id cells have been demonstrated without the administration of antigen. Trenkner and Riblet (1975) observed an in vitro PFC response of mouse spleen cells to PC when rabbit-anti-S103 (a PC-binding myeloma with a T1S-like idiotype) was added to the cultures. No PFC response was observed with mouse anti-S103 antibody, and the response to rabbit anti-S103 antibody could be augmented by addition of spleen cells from mice sensitized to rabbit Ig, suggesting that the rabbit anti-Id fully replaced a T cell-dependent antigen whose haptenic groups are the anti-Id binding sites and whose Fc-

IDIOTYPES ON LYMPHOCYTES

245

portion determinants are recognized as carrier determinants by helper T cells. Antigen-independent demonstration of enhancement of idiotype expression has also been demonstrated in limiting-dilution experiments performed to enumerate idiotype-positive precursor B cells in A/J mice pretreated with anti-Id antibody to the A5A idiotype. In cultures of normal strain A/J spleen cells stimulated with LPS, A5A idiotype-bearing molecules are secreted into the supernatants. A titration of the proportions of A5A-secreting cultures in relation to the numbers of spleen cells suggest a frequency for A5A-positive precursor B cells of about 1: 2500 for nonimmunized A/J mice and of about 1:200 for A/J mice preimmunized with IgGl anti-A5A antibody (Eichmann et al., 1977b). Thus, without addition of antigen, a more than 10-fold increase in the idiotype-bearing B-cell population occurred after administration of anti-Id. An interesting aspect of these studies was that only about one-half of the A5A-idiotype-secreting antibody clones detected by LPS stimulation were in fact antigen (Strep.A)-binding clones. The number of these non-antigen-binding clones was enhanced in parallel with the enhancement of the antigen-binding, idiotype-positive clones, not only in animals pretreated with anti-Id, but also in animals immunized with Strep.A. This suggests that, upon administration of antigen, entire idiotypically defined B-cell populations undergo quantitative changes even if only a certain portion of them detectably interacts with the antigen (Eichmann et ul., 1977b; Oudin and Cazenave, 1971). Such quantitative changes are most readily explained in a network system in which idiotypically similar lymphocyte compartments are under similar control and undergo parallel quantitative changes (Eichmann et ul., 197713). A series of recent experiments also suggest that idiotype-anti-Id interactions at the cellular level may lead to antibody production independent of antigen (Eichmann, 1977b). Spleen cells from A/J mice presensitized with anti-A5A Id antibody, cultured together with cells from A/J mice hyperimmunized with A5A antibody in F.A.c., produce P F C to A-CHO in the absence of antigen (Table XXII). T h e cells from mice immunized with A5A may be purified lymph node T cells, whereas the B cells producing the A-CHO-specific PFC are clearly derived from the anti-A5A sensitized mice. If the B cells are from a mouse immunized with Strep.A., primarily A5A-positive PFC are produced, suggesting a selective, antigen-independent cooperation between anti-Id T helper cells and idiotype-positive B cells. It is partic-

246

K. EICHMANN

TABLE XXII ANTIGEN-INDEPENDENTin Vitro COOPERATION BETWEEN A5A-IDIOTWEPRODUCINGB CELLSAND T HELPERCELLS AUTOIMMUNE TO THE A5A IDIOTWE' Response

Cells" Normal cellsd NC + 1" A5A NC + 1" A5A Tr 1" anti-A5A + NC 1" anti-A5A + 1" A5A 1" anti-A5A + 1" A5A Tf 1" Strep.A + NC 1" Strep.A + 1" A5A a

Antigen: Strep.A.' % A5AP PFC: A-CHO*

8 17 42 195

225 -

None A-CHO

% A5Ae

8 24

37 16 125 162 4 89

93%

-

85%

Data from Eichmann (1977b).

* Plaque-forming cells (PFC) developed on A-CHO sheep red blood cells; numbers are per 106 viable cells recovered from culture, means of 4-6 cultures, standard errors between 1.1 and 1.6. 3 X 106 Streptococcal Group A (Strep.A) organisms per culture of 100 PI. Immunization with anti-A5A: 0.1 pg of idiotype binding capacity (IBC) of IgC, fraction i.p. Immunization with A5A: 100 pg of purified A5A antibody with F.A.C. at several subcutaneous sites. Immunization with Strep.A: lo8 organisms i.p. Determined by inhibition of PFC by anti-A5A antiserum. Nylon wool-purified lymph node cells, >98% immunoglobulin-negative. " 1 x lo6 or 2 x 5 x lo5 cells/culture.

ularly this latter experiment that suggests autostimulatory events within the immune system. Taken together, the data on enhancement of idiotype expression by anti-Id reactions suggest a two-step response. Soluble anti-Id antibody may primarily result in the augmentation of precursor cells that may be the result of cell proliferation. This has been shown by the presence of antigen-specific memory in both the T- and B-cell compartments and the absence of circulating antibody (Eichmann and Rajewsky, 1975; Eichmann, 1977b). Participation of lymphocytes in the anti-Id reaction, in contrast, may lead to B-cell differentiation and secretion of antibody, as shown by the T-cell dependent anti-PC PFC response to anti-S103 anti-Id (Trenkner and Riblet, 1975) and by the antigen-independent PFC response to A-CHO in the presence of antiId T cells (Eichmann, 1977b).

IDIOTYPES ON LYMPHOCYTES

247

V. Summary

Idiotypic determinants of the antigen-specific receptors of lymphocytes are discussed in this chapter with respect to their structural and serological definition as well as to their possible functional role in immune regulation. The serological basis of idiotypes on lymphocytes has been established with anti-idiotypic antisera to total “idiotypes” as well as to idiotypic subspecificities such as VH-and VL-associated idiotypic determinants, and binding site- or framework-associated idiotypic determinants. Idiotype research on B-cell receptors has largely confirmed and supported the clonal selection theory. Idiotype research on T cells is still a rapidly moving field, but a series of functionally defined T-cell subpopulations have been demonstrated to possess idiotypes that are cross-reactive with those of B cells of the corresponding antigen-binding specificity. In those cases in which idiotypic specificities of T cells have been investigated and in which genetic linkage studies have been performed, T cells have been found to display VHregions rather than total antibody V domains. Lymphocyte receptor idiotypes are likely to play a fundamental role in immune regulation. Within one individual’s immune system, spontaneous autoanti-idiotypic antibodies can be observed. In deliberate immunization experiments, it has been shown that both B and T cells recognize B- as well as T-cell idiotypes. Idiotype recognition of B cells may be fundamentally different from that of T cells, such that T cells do not distinguish between T and B cells whereas B cells do, thus providing an idiotypic borderline for network interactions between the T- and B-cell compartments. The difference in discriminatory power between T and B cells for idiotypes may b e intimately related to the difference in idiotype expression between two types of lymphocytes. Idiotype recognition within a functional immune system is not without consequence. Low-dose administration of anti-idiotypes results in enhancement of the immune response to antigen in all cellular compartments thus far tested. The net result of enhancement may be an augmentation or a suppression of the idiotype response, depending on whether helper T cells or suppressor T cells have been stimulated. Interaction of a target lymphocyte with soluble anti-idiotypes in the absence of antigen primarily leads to an enlargement of the precursor cell pool for a particular idiotype. In contrast, the interaction of a target lymphocyte with an anti-idiotype-bearing lymphocyte appears to result in differentiation and, in the case of the interaction of a B cell with an anti-idiotypic helper T cell, antibody production.

248

K. EICHMANN

ACKNOWLEDGMENTS I thank my friends and colleagues who have made this work possible by their generous communication of thoughts, doubts, and data. Their names have been extensively quoted in the text and will not be repeated here because they are too many.

REFERENCES Andersson, L. C., Binz, H., and Wigzell, H. (1977a). Nature (London) 264, 778. Andersson, J., Coutinho, A., and Melchers, F. (1977b).J. E x p . Med. 145, 1511. Andersson, J., Coutinho, A., and Melchers, F. (1977c).J. E x p . Med. 145, 1520. Askenase, P. h.,and Leonard, E, J. (1970).Immunochemistry 7, 29. Augustin, A., and Cosenza, H. (1976). Eur. J . Zrnmunol. 6,497. Augustin, A., Julius, M. H., and Cosenza, H. (1977).J. Supramol. Struct., Suppl. 1,939 (abstr.). Bangasser, S. A., Kapsalis, A. A., Fraker, P. J., and Nisonoff, A. (1975).J. Immunol. 114, 610. Benacerraf, B., Ovary, Z., Bloch, K. J., and Franklin, E. C. (1963).J. E x p . Med. 117,937. Berek, C., Taylor, B., and Eichmann, K. (1976).J. E x p . Med. 144, 1164. Binz, H. (1975). Scand. J. Irnrnunol. 4,79. Binz, H., and Askonas, B. A. (1975). Eur. J. Immunol. 5,618. Binz, H., and Lindenmann, J. (1972a).J. E x p . Med. 136,872. Binz, H., and Lindenmann, J. (1972b). Scand. J. Zmmunol. 1,339. Binz, H., and Lindenmann, J . (1974). Cell. Immunol. 10,260. Binz, H., and Wigzell, H. (1975a).J. E x p . Med. 142, 197. Binz, H., and Wigzell, H. (1975b).J. E x p . Med. 142, 1231. Binz, H., and Wigzell, H. (1977a). Contemp. Top. Immunobiol. 7 , 113. Binz, H., and Wigzell, H. (1977b). Cold Spring Harbor Symp. Quant. Biol. 41 (in press). J . E x p . Med. 147, 63. Binz, H., and Wigzell, H. (1977~). Binz, H., Lindenmann, J., and Wigzell, H. (1973).Nature (London)246, 146. Binz, H., Lindenmann, J., and Wigzell, H. (1974a).J. E x p . Med. 139,877. Binz, H., Lindenmann, J., and Wigzell, H. (1974b).J. E x p . Med. 140,731. Binz, H., Kimura, A., and Wigzell, H. (1975a). Scand. J . Irnrnunol. 4,413. Binz, H., Bachi, T., Wigzell, H., Ramseier, H., and Lindenmann, J. (1975b). Proc. Natl. Acad. Sci. U . S. A. 72,3210. Binz, H., Wigzell, H., and Bazin, H. (1977). Nature (London)264,639. Black, S. J., Eichmann, K., Hiimmerling, G. J., and Rajewsky, K. (1975). I n “Membrane Receptors of Lymphocytes” (M. Seligman, J. L. Preud’homme, and F. M. Kourilsky, eds.), p. 117. North-Holland Publ., Amsterdam. Black, S. J., Hammerling, G. J , , Berek, C., Rajewsky, K., and Eichmann, K. (1976a).J. Exp. Med. 143,846. Black, S. J., Hammerling, G. J., Eichmann, K., and Rajewsky, K. (1976b). In “Leukocyte Membrane Determinants Regulating Immune Reactivity” (V. P. Eijsvoogel, D. Roos, and W. P. Zeijiemaker, eds.), p. 149. Academic Press, New York. Blomberg, B., Geckeler, W. R., and Weigert, M. (1972). Science 177, 178. Bosma, M., and Bosma, G. (1974).J . Exp. Med. 139,521. Braun, D. G., and Krause, R. M. (1968).J . Exp. Med. 128,969. Brient, B. W., and Nisonoff, A. (1970).J. E x p . Med. 132,951. Brient, B. W., Haimovich, J., and Nisonoff, A. (1971).Proc. Natl. Acad. Sci. U . S. A. 68, 3 139.

IDIOTYPES ON LYMPHOCYTES

249

Briles, D. E., and Krause, R. M. (1974).J . Zmmunol. 113,522. Burnet, F. M. (1959). I n “The Clonal Selection Theory of Acquired Immunity.” Oxford Univ. Press, London and New York. Cancro, M. P., SigaI, N. H., and Klinmann, N. R. (1978).j. Exp. Med. 147, 1. Cantor, H. (1977).J . Supramol. Struct., S u p p l . 1, 908. Capra, J. D., and Kehoe, J. M. (1975). Ado. Imrnunol. 20, 1. Capra, J. D., Kehoe, J. M., Williams, R. C., Feizi, T., and Kunkel, H. G. (1972). Proc. Natl. Acad. Sci. U . S . A. 69,40. Capra, J. D., Wasserman, R. W., and Kehoe, J. M. (1973).J. Exp. Med. 138,410. Capra, J. D., Tung, A. S., and Nisonoff, A. (1974).J. Immunol. 114, 1548. Capra, J. D., Tung, A. S., and Nisonoff, A. (1975).J. Immunol. 115, 414. Capra, J. D., Berek, C., and Eichmann, K. (1976).J. Immunol. 117, 7. Carson, D., and Weigert, M. (1973). Proc. Natl. Accid. Sci. U . S. A. 70, 235. Cazenave, P. A., Ternynck, T., and Avrameas, S. (1974). Proc. Natl. Acrid. Sci. U . S . A . 71,4500. Cazenave, P. A., Cavaillon, J. M., and Bona, C. (1977). Immunol. Rev. 34,34. Claflin, J . L., and Davie, J. M. (1974a).J. Exp. Med. 140, 673. ClaHin, J. L., and Davie, J . M. (197413).J. Z~nmunol.113, 1678. Claflin, J. L., and Davie, J. M. (1975).J. Intmunol. 114,70. ClaHin, J. L., Liebennan, R., and Davie, J. M. (1974a).J. Exp. Med. 139,58. ClaHin, J. L., Liehemian, R., and Davie, J. M. (1974b).J. Immunol. 112, 1747. Click, R. E., Benck, L., and Alto, B. S. (1972).Cell. Zmmunol. 3, 264. Cohn, M., Notani, G., and Rice, S. (1969). Zmmunochemistry 6, 111. Cosenza, H. (1975). Eur. J. Immunol. 6, 114. Cosenza, H., and Kohler, H. (1972a). Proc. Natl. Acad. Sci. U . S . A. 69,2701. Cosenza, H., and Kohler, H. (1972b). Science 176, 1027. Cosenza, H., and Kohler, H. (1973).Speciftc Recept. Antibodtes, Antigens Cells, Int. Convocution Immunol. [Proc.],3rd, 1972 p. 330. Cosenza, H., Q u i n t h s , J., and Lefkovits, I. (1975). Eur. J . Zmmunol. 5, 343. Cosenza, H., Augustin, A,, and Julius, M. H. (1977a). Eur. J. Zmmunol. (in press). Cosenza, H., Julius, M., and Augustin, A. (1977b). I n “The Immune System, Genes and the Cells in which they Function” (E. Sercarz, L. Herzenberg, and F. Fox, eds.). Academic Press, New York (in press). Cosenza, H., Julius, M., and Augustin, A. (19774. Transplant. Reo. 34,3. Cramer, M., Krawinkel, U., Hammerling, G., Black, S. J., Berek, C., Eichmann, K., and Rajewsky, K. (1977).I n “Proceedings of the Third Ia-Workshop” ( H . 0 . McDevitt, W. E. Paul, and B. Benacerraf, eds.). Academic Press, New York (in press). Eichmann, K. (1972). Eur. J. Immunol. 2,301. Eichmann, K. (1973).J. Exp. Med. 137, 603. Eichmann, K. (1974).Eur. J. Zmmunol. 4,296. Eichmann, K. (197%). Eur. J . Zmmunol. 5, 511. Eichmann, K. (1975b).Immunogenetics 2,491. Eichmann, K. (1977a). In “The Immune System: Genetics and Regulation” (E. Sercarz, L. Herzenberg, and F. Fox, eds.). Academic Press, New York (in press). Eichmann, K. (1977b). Z. Zmmunituetsforsch. 153,297 (Abstr.). Eichmann, K., and Berek, C. (1973).Eur. J . Zmmunol. 3, 599. Eichmann, K., and Berek, C. (1974). Ann. Zmmunol. (Paris) 125c, 359. Eichmann, K., and Melchers, F. (1978). In preparation. Eichmann, K., and Greenblatt, J. (1971)./. Exp. Med. 133,424. Eichmann, K., and Kindt, T. J. (1971).J. Exp. Med. 134,532.

250

K. EICHMANN

Eichmann, K., and Rajewsky, K. (1975).Eur. J. Zmmunol. 5, 661. Eichmann, K., Tung, A. S., and Nisonoff, A. (1974).Nature (London)250,511. Eichmann, K., Berek, C., Hammerling, G., Black, S. J., and Rajewsky, K. (1977a). Colloq. Geo. Biol. Chem. 27,65. Eichmann, K., Coutinho, A., and Melchers, F. (1977b).J.Exp. Med. 146, 1436. Eisen, H. N., Sakato, N., and Hall, S. J. (1975). Transplant. Proc. 7, 209. Elliott, B. E., Nagy, Z., Nabholz, M., and Pernis, B. (1977). Eur. J. Zmmunol. 7 , 287. Feizi, T., and Kabat, E: A. (1972).]. Exp. Med. 135, 1247. Feizi, T., Kabat, E. A., Vicari, G., Anderson, B., and Marsh, W. L. (1971).J. Zmmunol. 106, 1578. Fischer-Lindahl, K., and Wilson, D. B. (1977).J. E x p . Med. 145,500. Fraker, P. J., Cicurel, L., and Nisonoff, A. (1974).J. Zmmunol. 113, 791. Fu, S. M., Winchester, R. J., Feizi, T., Walzer, P. D., and Kunkel, H. G. (1974). Proc. Natl. Acad. Sci. U . S.A. 71,4487. Fu, S. M., Winchester, R. J., and Kunkel, H. G. (1975).J.Zmmunol. 114,250. Gearhart, P. J., Sigal, N. H., and Klinman, N. R. (1975).J.Exp. Med. 141,56. Geczy, A. F., Geczy, C. L., and de Weck, A. L. (1976).J. E x p . Med. 144,226. Geczy, A. F., Geczy, C. L., and de Weck, A. L. (1977). In “The Immune System: Genetics and Regulation” (E. Sercarz, L. Herzenberg, and F. Fox, eds.), p. 209. Academic Press, New York. Grey, H. M., Mannik, M., and Kunkel, H. G. (1965).]. Erp. Med. 121, 561. Gronowicz, E., Coutinho, A., and Melchers, F. (1976). Eur. J . Immunol. 6,588. Haimovich, J. (1977).Eur. J. Zmmunol. (in press). Hammerling, G. J., and Eichmann, K. (1976).Eur. J. Zmmunol. 6, 565. Hammerling, G. J., and Eichmann, K. (1977).In “Proceedings of the Third Ia-Workshop” (H. 0. McDevitt, W. E. Paul, and B. Benacerraf, eds.). Academic Press, New York (in press). Hammerling, G. J., and McDevitt, H. 0. (1974).J.Zmmunol. 112, 1726. Hammerling, G. J., Black, C., Berek, C., Eichmann, K., and Rajewsky, K. (1976a).J.Exp. Med. 143,861. Hammerling, G. J., Eichmann, K., and Sorg, C. (197%). In “The Role of Products of the Histocompatibility Gene Complex in Immune Responses” (D. Katz, and B. Benacerraf, eds.), p. 417. Academic Press, New York. Hammerling, G. J., Black, S. J., Segal, S., and Eichmann, K. (1976~).In “Leukocyte Membrane Determinants Regulating Immune Reactivity” (V. P. Eijsvoogel, D. Roos, and W. P. Zeijiemaker, eds.), p. 367. Academic Press, New York. Hannestad, K., Kao, M. S., and Eisen, H. (1972).Proc. Nutl. Acad. Sci. U . S.A. 69,2295. Hart, D. A., Wang, A. L., Pawlak, L. L., and Nisonoff, A. (1972).]. E x p . Med. 135,1293. Heber-Katz, E., and Wilson, D. B. (1976).J.Exp. Med. 143, 701. Helman, M., Shreier, I., and Givol, D. (1976).J.Zmmunol. 117, 1933. Henney, C. S., and Ishisaka, K. (1969).J.Zmmunol. 103,56. Herzenberg, L. A., McDevitt, H. O., and Herzenberg, L. A. (1968).Annu.Rev. Genet. 2, 209. Hoffmann, G. W. (1975).Eur. J. Zmmunol. 5,638. Hopper, J. E., and Nisonoff, A. (1971). Adu. Immunol. 13,57. Hopper, J. E., MacDonald, A. B., and Nisonoff, A. (1970).J. E x p . Med. 131,41. Howard, J. C., and Wilson, D. B. (1974).J.Exp. Med. 140,660. Hurez, D., Meshaka, G., Mihaesco, C.. and Seligmann, M. (1968).J.Immunol. 100,69. Imanishi, T., and Makela, 0. (1973).Eur. J. Zmmunol. 2, 323. Imanishi, T., and Mikela, 0. (1974).J.Exp. Med. 140, 1498.

IDIOTYPES ON LYMPHOCYTES

251

Imanishi, T., and Mikeli, 0. (1975).J.E x p . Med. 141,840. Inbar, D., Hochman, J., and Givol, D. (1972).Proc. Natl. Acad. Sci. U . S. A. 69,2659. Iverson, G. M. (1970). Nature (London)227,273. Iverson, G. M., and Dresser, D. W. (1970).Nature (London)227, 274. Jack, R. S., Imanishi-Kari, T., and Rajewsky, K. (1977). Eur. J . Zmmunol. 7,559. Janeway, C . A., and Paul, W. E. (1973).Eur. J . Zrnrnunol. 3,340. Janeway, C. A., Koven, H. S., and Paul, W. E. (1975a).Eur. J. Zmmunol. 5, 17. Janeway, C. A., Sakato, N., and Eisen, H. (1975b). Proc. Natl. Acad. Sci. U . S. A. 72, 2357. Jensenius, J. C., Williams, A. F., and Mole, L. E. (1977).Eur. J . Imrnunol. 7, 104. Jeme, N. K. (1955).Proc. Natl. Acad. Sci. U . S. A . 41,849. Jerne, N. K. (1974).Ann. Zmmunol. (Paris) 125c, 373. Jeme, N. K. (1976).Haruey Lect., 70, 93. Jorgensen, T., and Hannestad, K. (1977). Eur. J. Zmmunol. 7,426. Jorgensen, T., Gaudernack, G., and Hannestad, K. (1977).Scand. J. Zmmunol. 6, 311. Julius, M. H., Cosenza, H., and Augustin, A. (1977).Eur. J. Zmmunol. 7,273. Katz, D. H., Hamoaka, I., Dorf, M. E., Maurer, P. H., and Benacerraf, B. (1973).J.E x p . Med. 138,734. Katz, D. H., Graves, M., Dorf, M. E., Dimuzio, H., and Benacerraf, €3. (1975).J.E x p . Med. 141,263. Kelus, A. S., and Gel, P. G. H. (1968).J.E x p . Med. 127,215. Kindt, T. J. (1975). Adu. Zmmunol. 21,35. Klinman, N. R., Press, L. J., Pickard, A. R., Woodland, R. T., and Dewley, A. F. (1974). In “The Immune System: Genes, Receptors, Signals” (E. Sercarz, A. R. Williamson, and F. Fox, eds.), p. 357. Academic Press, New York. Kluskens, L., and Kohler, H. (1974). Proc. Natl. Acad. Sci. U . S. A. 71,5083. Kohler, H. (1975).Transplant. Rev. 29,54. Kohler, H., Strayer, D. S., and Kaplan, D. R. (1974).Science 186,643. Krammer, P., and Eichmann, K. (1977).Nature (London)270,733. Krammer, P., and Eichmann, K. (1978).Behring Znst. Mitteilungen 62, (in press). Krause, R. M. (1970). Ado. Zmrnunol. 12, 1. Krawinkel, U . , and Rajewsky, K. (1976).Eur. J . Zrnmunol. 6, 529. Krawinkel, U.,Cramer, M., Berek, C., Himmerling, G., Black, S. J., Rajewsky, K., and Eichmann, K. (1977a). Cold Spring Harbor Symp. Quant. Biol. 41,285. Krawinkel, U., Cramer, M., Imanishi-Kari, T., Jack, R. S., and Rajewsky, K. (1977b). Eur .]. Zmmunol. 7, 566. Krawinkel, U.,Cramer, M., Mage, R., Kelus, A., and Rajewsky, K. (1977c).J.E x p . Med. 146,792. Kubo, R. T., Grey, H. M., and Pirofsky, B. (1974).]. Zmmunol. 112, 1952. Kunkel, H. G. (1970).Fed. Proc., Fed. Am. Soc. E x p . Biol. 9,55. Kunkel, H. G., Mannik, M., and Williams, R. C. (1963). Science 140, 1218. Kunkel, H. G., Agnello, V., Joslin, I. G., Winchester, R. J., and Capra, J. D. (1973).J . E x p . Med. 137,331. Kunkel, H. G., Joslin, F., and Hurley, J. (1976).J.Zmmunol. 116, 1532. Lieberman, R., Potter, M., Mushinski, E. B., Humphrey, J. R. W., and Rudikoff, S. (1974).J.E x p . Med. 139,983. Lieberman, R., Potter, M., Humphrey, J. R. W., Mushinski, E. B., and Vrana, M. (1975). J . E x p . Med. 142, 106. Lieberman, R., Potter, M., Humphrey, W., and Chien, C. C. (1976).]. Zmmunol. 117, 2105.

252

K. EICHMANN

Lindstrom, T. D., Hardy, W. R., Eberle, B. J., and Williams, R. C. (1973).Ann.Intern. Med. 78,837. Lohss, F., Weiler, E., and Hillmann, G . (1953).Z. Naturforsch., Teil B 8,625. MacDonald, A. B., and Nisonoff, A. (1970j.J. Erp. Med. 131, 583. McKearn, T. J . (1974).Science 183,94. McKearn, T. J., Stuart, F. P., and Fitch, F. W. (1974a).J.Immunol. 113, 1876. McKearn, T. J., Hamada, Y., Stuart, F. P., and Fitch, F. W. (197413).Nature (London) 251,648. Makela, O., and Kajalainen, K. (1977a).Immunol. Reu. 34, 119. Makela, O., and Kajalainen, K. (1977b). Cold Spring Horbor Symp. Quant. Biol. 41, 735. Mellstedt, H., Hamrnarstrom, S., and Holm, G. (1974). Clin. E x p . Immunol. 17,371. Miller, J. A. T. P., and Vadas, M. A., (1977). Cold Spring Harbor S y m p . Quant. B i d . 41, 579. Mitchison, N. A. (1971).In “Cell Interactions and Receptor Antibodies in Immune Responses” (0.Makela, A. Cross, and T. U. Kosunen, eds.), p. 249. Academic Press, New York. Mitchison, N. A., Rajewsky, K., and Talor, R. B. (1970). In “Developmental Aspects of Antibody Formation and Structure” (J. Sterzl and I. Riha, eds.), Vol. 2, p. 547. Academic Press, New York. Moses, E. and Haimovich, J. (1977).In “Proceedings of the Third Ia-Workshop” (H. 0. McDevitt, W. E. Paul, and B. Benacerraf, eds.). Academic Press, New York (in press). Munro, A. J., and Taussig, M. (1975). Nature (London) 256, 103. Nagy, Z., Elliott, B. E., Nabholz, M., Krammer, P., and Pernis, B. (1976).J.Exp. Med. 143,648. Natvig, J. B., and Kunkel, H. G. (1973). Ado. Immunol. 16, 1. Natvig, J. B., Salsano, F., Frbland, S. S., and Stavem, P. (1975). In “Membrane Receptors of Lymphocytes” (M. Seligmann, J. L. Preiid’homme, and F. M. Kourilsky, eds.), p. 13. North-Holland Publ., Amsterdam. Nisonoff, A., and Bangasser, S. A. (1975).Transplant. Reu. 27, 100. Nisonoff, A,, and Shyr-Te, J. (1976).Ann. Immunol. (Paris) 127c, 347. Nisonoff, A,, Shyr-Te, J., and Owen, F. L. (1977).Transplant. Reo. 34,89. Oudin, J . (1966).Proc. Soc. London, Ser. B 166,207. Oudin, J . (1974).I n “The Antigens” (M. Sela, ed.), Vol. 2, p. 278. Academic Press, New York. C . Acad. Sci. U . S . A. 68,2616. Oudin, J., and Cazenave, P. A. (1971). P ~ C J Natl. Oudin, J., and Michel, M. (1963). C . R. Hebd. Seances Acad. Sci. 257,805. Oudin, J., and Michel, M. (1969a).J . E r p . Med. 130,595. Oudin, J., and Michel, M. (1969b).J.Erp. Med. 130,619. Owen, F., Ju, S. T., and Nisonoff, A. (1977).J.E x p . Med. 145, 1559. Pawlak, L. L., Mushinski, E. B.. Nisonoff, A., and Potter, M. (1973a).j.E x p . Med. 137, 22. Pawlak, L. L., Hart, D. A,, and Nisonoff, A. (1973b).J.Immunol. 110,587. Pawlak, L. L., Hart, D. A., and Nisonoff, A. (1973c).J.E x p . Med. 137, 1442. Pawlak, L. L., Hart, D. A,, and Nisonoff, A. (1974). Eur.1. Immunol. 4, 10. Pernis, B. (1967).Cold Spring Harbor Symp. Quant. Biol. 32,333. Pisetsky, D., Garrison, C., and Sachs, D. (1977). Zmmunogenetics (in press). Potter, M. (1971).Ann. N.Y. Acad. Sci. 190,306. Potter, M., and Kunkel, H. C. (1971). Prog. Immunot. 1, 1361.

IDIOTYPES ON LYMPHOCYTES

253

Potter, M., and Liebeman, R. (1967).Ado. Immunol. 7,91. Potter, M., and Liebeman, R. (1970).J. E x p . Med. 132,737. Potter, M., Mushinski, E. B., Rudikoff, S., and Apella, E. (1973).Specific Recept. Antibodies, Antigens, Cells, Int. Convocation Immunol. [Proc.],3rd, 1972 p. 270. Preud’homme, J . L., and Seligmann, M. (1972).Blood 40, 777. Quintans, J., and Cosenza, H. (1976).Eur. J. Immunol. 6,393. Quintans, J . , and Lefkovitz, I. (1973).Eur. J . Immunol. 3,392. Quintans, J., and Lefkovitz, I. (1974a).]. Immunol. 113, 1373. Quintans, J., and Lefkovitz, I. (197411).Proc. Int. Workshop Nude Mice, I s t , 1973 p. 89. Quintans, J., and Lefkovitz, I. (1976).I n “Immune Responses of Lymphocytes” (M. Feldnian and A. Globeson, eds.), p. 101. Plenum, New York. Raff, M. (1977).Nature (London) 265, 205. Rajewsky, K. (1971).Proc. R . Soc. London, Ser. B 176,385. Rajewsky, K., and Eichmann, K. (1977). Contemp. Immunobiol. 7, 67. Rajewsky, K., Hammerling, G. J., Black, S . J., Berek, C., and Eichmann, K. (1976).I n “The Role of Products of the Histocompatibility Gene Complex in Immune Responses” (D. H. Katz and B. Benacerraf, eds.), p. 417. Academic Press, New York. Raniseier, H. (1973).Curr. Top. Microbiol. Immunol. 60,31. Raniseier, H . (1974a).J.Exp. Med. 140,603. Raniseier, H . (1974b). Cell. Immunol. 8, 177. Raniseier, H. (1975).Eur. J . Immunol. 5 , 23. Raniseier, H. (1976).Imniunol. Commun. 5, 827. Ramseier, H., and Lindenmann, J. (1969).Pathol. Microbiol. 34,379. Raniseier, H., and Lindenmann, J. (1972a). Tran.~plunt. Rev. 10, 57. Raniseier, H., and Lindenmann, J. (19721)).Eur. J . Immunol. 2, 109. Raniseier, H., Aguet, M., and Lindenmann, J . (1977).Immunol. Reo. 34, 50. Reth, M., Imanishi-Kari, T., Jack, R. S., Cramer, M., Krawinkel, U., Hammerling, C. J., and Rajewsky, K. (1977). I n “The Immune System: Genetics and Regulation” (E. Sercarz, L. A. Herzenberg, and F. Fox, eds.), p. 139. Academic Press, New York. Riblet, R., Weigert, M., and Mikela, 0. (1975a).Eur. J . Immunot. 5, 778. Riblet, R., Cohn, M., and Weigert, M. (197%). Inmunogenetics 525,526. Richter, P . H. (1975).Eur. J . Immunol. 5,330. Rittenberg, M. B., and Pratt, K. I. (1969).Proc. Soc. E x p . Riol. M e d . 132, 575. Rodkey, L. S. (1974).J.Exp. Med. 139, 712. Rodkey, L. S. (1976).]. ~mm11710~. 117, 986. Rudikoff, S., and Claflin, J. L. (1976).J.Erp. Med. 144, 1294. Sakato, N., and Eisen, H. N. (1975).J.E x p . Med. 141, 1411. Sakato, N., Janeway, C. A,, and Eisen, H. N. (1977).Cold Spring Hnrbor Symp. Quant. B i d . 4 1 (in press). Salsano, F., Frdand, S . S., Natvig, J. B., and Michelson, T. E. (1974).Scand. ]. Immunol. 3,841. Seligmann, M., and Mihaesco, C. (1967).Gammu Globulins, Proc. Nobel Symp., 3rd, I967 p. 169. Seligmann, M., Meshaka, G., Hurez, D., and Mihaesco, C. (1966).Immunoputhol.,Int. Srymp., 4 t h , 1965 p. 229. Seppalii, I., and Eichmann, K. (1978).In preparation. Shearer, G. M., Rehn, T. G., and Schmitt-Verhulst, A. M. (1976).Transp/ant. Reo. 29, 222. Sher, A,, and Cohn, M. (1972a).J.Immunol. 109, 176. Sher, A., and Cohn, M. (197211).Eu r. J. Immunol. 2, 319.

254

K. EICHMANN

Sher, A., Lord, E., and Cohn, M. (1971).J.lrnrnunol. 107, 1226. Shreffler, D. C., and David, C. S. (1975).Ado. lmmunol. 20, 125. Shyr-Te, J., Owen, F. L., and Nisonoff, A. (1977).Cold S p r i n g Harbor Syrnp. Q u a n t . BioZ. 41, 699. Sigal, N. H., Gearhard, P. J., and Klinman, N. R. (1975).J. Zrnrnunol. 68, 1354. Sirisinha, S., and Eisen, H. N. (1971).Proc. N a t l . Acud. Sci. U . S. A. 68,3130. Slater, R. J., Ward, S. M., and Kunkel, H. G. (1955).J. E x p . Med. 101, 85. Sogn, J. A., Yarmush, M. L., and Kindt, T. J. (1976).Ann. Irnrnunol. (Paris) 127c, 397. Sogn, J. A., Coligan, J. E., and Kindt, T. J. (1977).Fed.Proc., Fed. Am. SOC. E x p . B i d . 36, 214. Sprent, J., and von Boehmer, H. (1976).J. E x p . Med. 144,617. Spring-Stewart, S., and Nisonoff, A. (1973).J. lmrnunol. 110,679. Tada, T., Taniguchi, M., and Takemori, T. (1975).Transplunt. Reu. 26, 106. Tada, T., Taniguchi, M., and David, C. G . (1976).J. E x p . Med. 144,713. Takemori, T., and Tada, T. (1975).J. E x p . Med. 142, 1241. Taussig, M. J., Munro, A. J., CampheII, R., David, C. S., and Stained, N. A. (1975).J. E x p . Med. 142,694. Trenkner, E., and Riblet, R. (1975).J. E x p . Med. 142, 1121. Vicari, G., Sher, A,, Cohn, M., and Kahat, E. (1970).lrnrnunochernistry 7,829. Wang, A. C., Wilson, S. K., Hopper, J. E., Fudenherg, H. H., and Nisonoff, A. (1970). Proc. N a t l . Acad. Sci. U . S. A. 66,337. Weigert, M., Potter, M., and Sachs, D. (1975).lmrnunogenetics 1, 52. Wells, J. V., Fudenberg, H. H., and Givol, D. (1973).Proc. N a t l . Acad. Sci. U . S. A. 70, 1585. Wernet, P., Feizi, T., and Kunkel, H. G. (1972).J. E x p . Med. 136,650. Williams, R. C., Kunkel, H. G., and Capra, J. D. (1968).Science 161,379. Yakulis, V., Bhoogalam, N., and Heller, P. (1972).J. Irnrnunol. 108, 1119. Yamush, M. L., Sogn, J. A., Mudgett, M., and Kindt, T. J. (1977).J. E x p . Med. 145,916. Zinkemagel, R. M. (1976).Nature (London) 261, 139.

The B-Cell Clonotype Repertoire NOLAN H. SlGAL AND NORMAN R. KLINMAN Department of Pathology, University of Pennsylvania School of M e d i c i n e , P h i l a d e l p h i a , Pennsylvania

....................... .......................

I. Introduction ........................................ 11. The Clonal Selection Hypothesis

A. Introduction ................................... B. Immunoglobulin Nature of the C ............................................. C. Specificity of the Cell’s Receptor .................................................................. D. Precommitment of the Precursor Cell .... ........................... E. Integrity of the B-Cell Clone ........................................................................ F. Contradictions to the Clonal Selection Theory C . Conclusion ........................... .............................................. Methods of C1 tion .......................................... ....... ...................... A. Isoelectric Fo B. Fine-Specifici ............................................................................... C. Idiotype .................... D. Sequence ........................................................................................................ E. Other Probes ................................................................................................... Defining the Adult B-Cell Repertoire ................................................................ A. Antigen-Binding Cell Analysis ...................................................................... B. Techniques to Estimate Precursor Frequencies ...................... C. Estimates of the Adult B-Cell Repertoire ................................. Defining the Neonatal B-Cell Repertoire .......................................................... A. Antigen-Binding Cell Analysis ...................................................................... B. I n Vioo Studies of Clonotype Development ................................................ C. I n Vitro Studies of Clonotype Development Conclusion ............................................................................................................. ....................................................... References I

111.

IV.

V.

VI.

255 256 256 257 258 263 269 271 276 276 277 280 285 291 294 294 295 298 305 3 16 317 319 323 327 328

I. I n t r o d u c t i o n

The emergence of the clonal selection hypothesis (Burnet, 1959) has had a profound impact on the directions of immunological ideas and investigations in recent years. Although uncertainty persists in areas such as the time and stability of cellular commitment and the role of cell-to-cell information transfer in cellular specificity (Adler et al., 1966; Bell and Dray, 1971), the elucidation of immunological mechanisms has shifted from the analysis of antigenic induction to investigation of mechanisms that control repertoire expression and anti255 Copyright4 1978 by Academic Press, lnc. All rights of reproduction in any form reserved. ISBN 012022426-7

256

NOLAN H. SIGAL AND NORMAN R. KLINMAN

genic selection. It would appear that both of these processes represent enormously complex processes that ultimately emanate from an interplay of (a) genetic predetermination of variable region amino acid sequences; (b) mechanisms that allow for expression of this information and enable the combinatorial association of light-chain and heavychain specificities; (c) genetic control of clonotype expression both before and after antigenic stimulation; (d) cell-cell interactions; and (e)positive and negative interactions of precursor cells with their environment, antigens, and factors elaborated by adjacent cells. Because of the recent advent of a variety of methods, such as idiotypic and isoelectric focusing analyses, it has been possible for a number of investigators to begin a detailed analysis of the extensiveness of the B-cell specificity repertoire, particularly in inbred murine strains. These studies have already permitted insights into the mode of repertoire expression. We will review the current status of the clonal selection hypothesis, the methods of evaluating repertoire expression, and the status of our current understanding of both the adult and neonatal B-cell clonotype repertoires. II. The Clonal Selection Hypothesis

A. INTRODUCTION In 1900, Ehrlich postulated that the receptor for antigen on immunocompetent cells was identical to the antibody product that he could isolate from serum and that these receptors existed in a preformed state on the cell membrane. The notion of cellular precommitment was largely ignored for the next 55 years, while the instructional theories of antibody formation were popularized by Mudd (1932), Pauling (1940), and Karush (1962). These hypotheses suggested that antigen, in some manner, could program immunocompetent cells to synthesize antibody molecules complementary to the original antigen. The discovery that a cell’s DNA encodes the information responsible for synthesis of a protein and that the ability of an antibody molecule to bind an antigen is directly dependent on its primary amino acid sequence and consequent tertiary structure made these theories untenable (Haber, 1964; Whitney and Tanford, 1961; Jaton et al.,

1968). The clonal selection hypothesis, proposed by Jerne (1955)and Burnet (1959) and elaborated by Mitchison (1967) as the “receptor hypothesis,” has served as the theoretical framework for understanding antibody specificity during the past 20 years. The postulates of this theory can briefly be stated as follows: (1)the receptor of the antibodyforming cell precursor is identical in specificity to the eventual anti-

THE B-CELL CLONOTYPE REPERTOIRE

257

body product of that cell; (2) each precursor cell is committed to one, and only one, antibody specificity; and ( 3 )each antibody-forming cell precursor is committed to a unique antibody specificity prior to antigenic stimulation and the progeny of that precursor remains committed to that specificity. Since the validity of these postulates are of central concern for the discussion of the B-cell repertoire, the experimental evidence that supports each tenet will be reviewed, as well as some recent findings that challenge these assumptions. B. IMMUNOGLOBULIN NATURE OF THE CELL RECEPTOR

If the receptors for antigen on B cells are immunoglobulin (Ig) molecules, then membrane Ig should be detectable on lymphoid cells using specific anti-Ig reagents. Immunoglobulin has been demonstrated by a large variety of techniques to be present on the surface of lymphocytes from mouse, man, rabbit, guinea pig, amphibians, chicken, and fish. Since this subject has been extensively reviewed b y a number ofworkers (Greaves and Hogg, 1971; Marchalonis, 1974; Vitetta and Uhr, 1973; Warner, 1974), we will discuss the topic only briefly . In 1961, Moller observed that a proportion of viable mouse lymphoid cells reacted with a fluorescein (Fl)-labeled anti-Ig reagent and showed membrane staining. This finding has been confirmed b y a multitude of investigators (see Warner, 1974). In general, techniques for demonstrating Ig on the surface of B cells fall into several broad categories: direct visualization, indirect (sandwich) techniques, and physical methods. By far the greatest number of studies have been done b y directly binding Fl-labeled antibodies or radiolabeling cells with 1251-labeledanti-Igs. Using these techniques, investigators have determined the proportion of membrane-bound Ig-bearing cells in a variety of lymphoid tissues (reviewed by Basten and Howard, 1973), the Ig class expression of cell-bound Ig, and the ontogeny of Ig-bearing cells (Raff, 1977). Raff et a l . (1970) compared the uptake of 1251-labeled anti-Ig to mouse spleen cells with uptake of F1-labeled anti-Ig and demonstrated a marked difference in sensitivity between the two techniques. While autoradiography is far more sensitive than immunofluorescence, binding of aggregated lZ5I-labeledanti-Ig to damaged cell membranes and to macrophages requires that the specificity of this method b e carefully controlled. Indirect systems for detecting surface Ig abound. Several techniques employ the bridging of anti-Ig molecules to specially coated erythrocytes, such as the mixed antiglobulin technique (Coombs et ul., 1970) and reverse immune cytoadherence (Paraskevas et al., 1971). Other techniques, such as the hemagglutination inhibition sys-

258

NOLAN H. SICAL AND NORMAN R. KLINMAN

tem (Klein et al., 1970), binding of horseradish peroxidase-labeled antibody (Avrameas and Guilbert, 1971), and lactoperoxidase-catalyzed surface iodination of cells with lZ5I(Marchalonis and Cone, 1973), have been used to quantitate the average number of membrane-bound Ig molecules per B cell; in general, approximately lo4 to lo5 molecules per cell have been found. Surface protein iodination with lZ5I,followed by disruption and analysis in polyacrylamide gel electrophoresis (reviewed in Marchalonis, 1974; Vitetta et al., 1976) has been used to conclusively demonstrate the presence of Ig on the B-cell membrane. In addition, studies (Vitetta and Uhr, 1973; Marchalonis, 1974) that used a double-label technique showed that lymphocyte synthesizes the Ig, which is subsequently found on its surface. While murine cell surface Ig is predominantly composed of an IgM monomer (8 S) molecule (Marchalonis, 1974), a portion of human lymphocytes have IgD on their surfaces (Rowe et al., 1973; Fu et al., 1974), and mouse lymphocytes can also possess an IgD-like molecule (Vitetta et al., 1976; Abney et al., 1976; Goding et al., 1976).

c. SPECIFICITY OF THE CELL’S RECEPTOR It is clear that on its surface the B cell has Ig, which is synthesized within the cell, but such studies do not establish that Ig is the relevant receptor for antigen on the cell. One general approach to this question has been to examine the ability of B cells to bind antigen and then determine whether this interaction can be inhibited by anti-Ig. Although a number of techniques have been employed to examine the interaction of antigen and lymphoid cells, it has been much more difficult to address the problem of whether the cells detected in the antigen-binding cell (ABC) assay are related to the precursor cells triggered by antigen in in oivo immunization. The subject of antigen binding by B cells has been extensively reviewed (Ada, 1970; Roelants, 1972; Bach, 1973; Davie and Paul, 1974; Warner, 1974; Leflcovits, 1975), and therefore pertinent findings and techniques will merely be summarized here.

1 . Antigen-Binding Cells The first studies to demonstrate antigen binding employed the ability of large particulate antigens, such as bacteria (Makela and Nossal, 1961) or heterologous red cells (Nota et al., 1964; Zaalberg, 1964) to form clusters (rosettes) around lymphoid cells. Erythrocytes can be coated with protein antigens (Bankhurst and Wilson, 1971), polysaccharides (Howard et al., 1969; Sjoberg and Moller, 1970), and hap-

THE B-CELL CLONOTYPE REPERTOIRE

259

tenic determinants (Moller and Sjoberg, 1972) to form rosettes as well. These rosetting techniques appear to detect both B cells and T cells, since the frequency of rosette-forming cells decreases approximately 50% after treatment of an unimmunized mouse spleen cell suspension with anti-Thy 1 and complement (Roelants, 1972). The most widely used technique for detecting ABCs has been the use of isotopically labeled antigens, first studied by Naor and Sulitzeanu (1967)and since then in many other laboratories (see Ada, 1970; Roelants, 1972; Davie and Paul, 1974; Warner, 1974). Although the conditions used in most studies appear to detect only antigen-binding to B cells (Warner, 1974; Davie and Paul, 1974), more recent investigations have used this technique to enumerate the frequency of T cells specific for a variety of antigens (Roelants, 1972; Hammerling and McDevitt, 1974; Lawrence et al., 1973). The actual number of ABCs counted will depend on the concentration of the labeled antigen, the specific activity of the label, and the exposure time for autoradiography (Byrt and Ada, 1969; Dwyer and MacKay, 1972a; reviewed in Roelants, 1972; Warner, 1974). These variables are critical in the consideration of the relevance of ABC to antibody-forming cell precursors and the use of ABC frequency to analyze the B-cell repertoire (see Section IV). Other methods used to enumerate ABCs include detection of binding of the enzyme P-galactosidase or horseradish peroxidase (Miller et al., 1971). By employing the activity of the enzyme to cleave a substrate, the ABCs can be detected by fluorescence under the light microscope. Physical separation of ABCs can be accomplished by a variety of techniques; antigen can be attached to glass beads (Wigzell and Andersson, 1969), polyacrylamide beads (Wofsy et al., 1971), nylon fibers (Edelman et al., 1971), polystyrene tubes (Choi et al., 1974), or collagen (Haas and Layton, 1975). The most sophisticated method of separation has utilized the fluorescence-activated cell sorter (FACS), which enables the isolation of virtually pure populations of viable ABCs with functional activity (Julius and Herzenberg, 1974). These physical methods have their greatest utility in demonstrating the correlation between precursor-cell activity and ABCs.

2 . Anti-lg lnhibition of ABCs In order to demonstrate that Ig is responsible for antigen binding to lymphoid cells, a number of investigators have shown that pretreatment of ABCs with anti-Ig inhibits the uptake of antigen by the cell (reviewed in Warner, 1974). I t was first demonstrated b y Byrt and Ada

260

NOLAN H. SIGAL AND NORMAN R. KLINMAN

(1969) that both polyvalent anti-Ig and anti-p sera inhibit the majority of ABCs of unimmunized mice, but over 50% of DNP-specific ABCs of the guinea pig have y2 on their surface (Davie and Paul, 1974).Anti-Ig treatment also eliminates formation of specific rosettes (Bona et al., 1972; Ferrarini et al., 1973), binding to antigen-coated beads (Walters and Wigzell, 1970) and nylon fibers (Rutishauser et al., 1972). Binding of antilymphocyte serum (Ada, 1970; Walters and Wigzell, 1970) or anti-H-2 serum (Hammerling and McDevitt, 1974) does not inhibit antigen uptake. The mechanism by which anti-Ig inhibits antigen binding is not completely understood, since anti-Ig (particularly anti-constant-region antibody) does not, in general, inhibit antigen binding by antibody (Hart et al., 1972). It is possible that anti-Ig limits antigen binding b y inhibiting receptor mobility or by removing receptors via “ capping.”

3 . Cocapping Experiment The phenomenon of cap formation has been employed to provide one of the most direct demonstrations of Ig as the cell’s receptor. Binding of bivalent antibodies directed against Ig determinants as well as binding of multivalent antigens induces a dramatic redistribution of the receptor molecules (Taylor et al., 1971; Unanue et al., 1972; Loor et al., 1972; reviewed iischreiner and Unanue, 1976). The receptors first aggregate into patches that move and become localized over one pole of the cell, forming a “cap.” Univalent anti-Fab and monosubstituted hapten carrier conjugates fail to induce redistribution and cap formation. Roelants et al. (1973) used this technique to show that Ig and the antigen receptor “cocap” simultaneously on the surface of the cells. When capping was induced with a fluoresceinated polyvalent anti-Ig and the cells were subsequently exposed to IZ5I-labeled antigen under noncapping conditions (4”C),the 1251-labeledsilver grains were found to be distributed in the caps superimposed on the Ig fluorescent caps. Thus, these experiments clearly demonstrate that Ig is the receptor for antigen on B cells. 4 . Antigen Inhibition and Suicide of ABCs The question of whether any or all ABCs have relevance for in uiuo immunocompetent cell triggering by antigen is important in the proof of the clonal selection hypothesis and in the use of ABCs to study the B-cell repertoire. In order to verify that the cell’s receptor has the same specificity as serum antibody, it is only necessary to demonstrate that the antibody-forming cell precursors are contained within the ABC population. The problem of whether all ABCs are pre-

THE B-CELL CLONOTYPE REPERTOIRE

26 1

cursor cells is a more difficult question and will be considered later in this section. Direct evidence that precursor cell activity is contained within the ABC population comes from the aforementioned studies, which separated ABCs from the nonbinding population. Inherent in most of these studies (Wigzell and Anderson, 1969; Wofsy et al., 1971; Edelman et al., 1971; Julius and Herzenberg, 1974) was the demonstration that depleted populations had little precursor activity. Ada and Byrt (1969) examined the significance of ABCs in the primary immune response by incubating spleen cells with 1251-labeledflagellin, causing irradiation-induced suicide in those cells that bound antigen. After transfer of the cell population to irradiated syngeneic recipients followed b y challenge with flagellin or a heterologous antigen, Ada and Byrt found that the antiflagellin antibody response was eliminated whiIe responses to other antigens were maintained. Similar results were obtained by other investigators with normal (Unanue, 1971) and primed (Humphrey et al., 1971) spleen populations, and Basten et al. (1971) and Roelants and Askonas (1971)demonstrated that both B cells and T cells can be specifically inactivated by radioactive antigen (reviewed in Roelants, 1972; Warner, 1974). Another approach to the study of the specificity of the cell’s receptor involves the inhibition of antigen binding with free hapten. In a series of elegant experiments, Nitchison (1967)probed the specificity of the precursor cell’s receptors by demonstrating that in the presence of 4hydroxy-3-iodo-5-nitrophenacetyl-~-amino-n-caproic acid (NIP cap) in uitro, stimulation with 4-hydroxy-3,5-iodophenacetyl -chicken y-globulin (DIP-CGG) yielded antibody relatively poor in NIP cross-reactivity. Later, Davie et al. (1971) showed that the receptors of dinitrophenol (DNP) guinea pig albumin (GPA)-binding cells in immune guinea pigs, like that of specific serum antibodies, are primarily directed toward the hapten. The relative capacity of the monovalent ligand e-DNP-L-lysine to inhibit [ 1251]DNP-GPA binding could be used to determine the relative avidity of the cell-bound receptors, and Davie and Paul (1972) subsequently employed this hapten inhibition technique to show that the avidity of ABCs increases after immunization. Using a similar method to assess plaque-forming cell (PFC) avidity, they demonstrated that the avidities of ABC’s, plaque-forming cells, and serum antibody all increase in parallel after immunization. Similar results were obtained b y Moller et al. (1973)in the mouse and, although indirect, support the contention that ABCs indeed represent the precursors of PFCs. A similar conclusion was reached by Diener et al. (1973).Combining autoradiography with cell separation tech-

262

NOLAN H. SIGAL AND NORMAN R. KLINMAN

niques, they demonstrated that the distribution profile of ABCs from unimmunized mice corresponded entirely with that of the in vivo and in vitro immunological activity. Furthermore, antigenic stimulation induced a significant proportion of ABCs to differentiate into larger cells, and the antibody-forming precursor cell function then followed the distribution of the larger cells.

5. Anti-idiotypic lnhibition The most direct proof that the cell’s receptor is identical to its secreted antibody product has centered on evidence that antibodies directed against variable region determinants can specifically inhibit or stimulate the synthesis of that antibody specificity. Cosenza and Kohler (1972a,b) used the immune response to phosphoryicholine (PC) as a model system. Immunization with a rough strain of Diplococcus pneumoniae (R36A)in viuo or in uitro elicited a large number of PFCs specific for the PC haptenic determinant. Since several PCbinding plasmacytoma proteins were known (Leon and Young, 1971; Sher et al., 1971), Cosenza and Kohler used the protocol of Lieberman and Humphrey (1971) to make antibodies against the variable regions of two of these proteins, TEPC 15 and MOPC 167. They first showed that the anti-idiotypic serum recognized variable-region determinants by demonstrating that anti-TEPC 15 or anti-MOPC 167 could inhibit binding of the homologous protein to PC-coated sheep erythrocytes (SRBC) in a hemagglutination-inhibition assay. Moreover, PC-specific PFCs generated in an in viuo or in uitro immune response to R36A Pneumococcus could be inhibited by adding anti-TEPC 15serum at the time of plaque formation. Adding anti-MOPC 167 serum had no effect on anti-PC PFC formation, and neither anti-idiotypic serum suppressed an anti-SRBC PFC response. In addition to showing idiotypic specificity, these results implied that the anti-PC antibodies elicited in response to R36A share idiotypic determinants with TEPC 15 protein. Cosenza and Kohler then demonstrated that the anti-TEPC 15 antibody could prevent the induction of the primary response to PC in uitro. When anti-TEPC 15 was added to spleen cell cultures at the same time as the antigen R36A, no PFCs were seen, whereas addition of the anti-idiotypic serum 1,2, or 3 days after antigenic stimulation had no effect. Thus, antibody recognizing variable-region determinants close to or within the combining site of a well-characterized myeloma protein could prevent the stimulation of antibody-forming cell precursors synthesizing the idiotype. These results strongly suggest that a cell’s receptor for antigen is identical in its variable regions to the antibody product of that cell.

THE

B-CELL CLONOTYPE REPERTOIRE

263

Using the same anti-idiotypic system, Claflin et al. (1974a,b) extended the observations of Cosenza and Kohler by examining the relationship between cells that bind PC and the idiotypic determinants on those cells. Anti-idiotypic serum to HOPC 8, a plasmacytoma protein that is idiotypically indistinguishable from TEPC 15, completely inhibited the binding of [‘2511PCBSA to BALB/c PC-specific spleen cells, whereas other anti-idiotypic sera did not inhibit binding. In addition, the receptors on the ABC displayed the same relative specificity for PC and two of its analogs, glycerophosphorylcholine and choline, as did the PC-binding myelomas TEPC 15 and HOPC 8. These data again clearly indicate that the receptors on BALB/c PC-binding cells have idiotypic determinants and specificities identical with those of TEPC 15 and suggest that ABC are the immunologically relevant precursor cells. The finding that other idiotypes, i.e., MOPC 167, were not detected on BALB/c PC-binding cells and that anti-HOPC 8 almost totally inhibited binding of [1251]PCBSA suggested that the serological response to PC may be highly restricted. This topic will be considered in greater detail in Section IV. More recent studies have confirmed the ability of anti-idiotypic antibody both to inhibit antigenic stimulation of B or T cells specific for PC or to stimulate B cells to antibody formation (Trenker and Riblet, 1975; Cosenza et al., 1977). In addition, T and B “memory” cells have been generated by anti-idiotypic antibodies (Eichmann, 1977). All these findings are preliminary, but, if confirmed, would strongly support the association of receptor and cell specificity.

D. PRECOMMITMENT OF THE PRECURSOR CELL That an antibody-forming cell precursor is committed to one, and only one, specificity during its lifetime is much more difficult to demonstrate than the specificity of the cell’s receptor. The majority of experimental systems that speak to the precommitment question give only indirect evidence to support this hypothesis. While the large body of data that has been generated (reviewed in Makela and Cross, 1970; Lefkovits, 1975) is overwhelmingly in favor of this element of the clonal selection theory, there are relatively few experiments directly demonstrating its validity. Specifically, two points are left unanswered by most experimental systems: (1) while the majority of cells may be committed to the synthesis of only one antibody specificity, there may be a small subpopulation of cells with the ability to respond with more than one antibody; and (2) the possibility exists that all precursor cells go through a developmental stage, where they are capable of being stimulated by antigen but are not yet locked into the synthesis of a single clonotype, thus implying that immature cells may

264

NOLAN H. SIGAL AND NORMAN R. KLINMAN

be less committed than mature B cells. Nevertheless, some recent experiments do address these questions, and the data remain in favor of B-cell precommitment.

1 . lrnmunization with Two Antigens Among the initial attempts to examine the precommitment of antibody-forming cell precursors were analyses of the antibody product of a single isolated PFC after immunization with two or more antigens (Nossal and Lederberg, 1958; Coons, 1958; Nossal, 1960; Makela, 1964a) or with an antigen having two immunogenic determinants (Hiramoto and Hamlin, 1965; Green et al., 1967; Makela, 1964b; Attardi et al., 1964; Gershon et al., 1968; reviewed in Makela and Cross, 1970). If B cells were multipotential, it would be predicted that a significant number of PFCs making antibody to two distinct antigens should exist. Most investigators showed that the antibody from a single cell was restricted by the criterion of lysis of only one antigencoated RBC, although in a few reports a high frequency of double producers was found. Many of these latter findings were shown to be artifacts of the systems used (see Makela and Cross, 1970), but a number of those factors have been controlled in more recent studies (Liacopoulos et al., 1976; see below). Finding few cells or no cells producing detectable antibody to two or more randomly chosen antigens is not, in itself, a strong argument for unipotentiality. The finding serves only to rule out the possibility that a single cell can be stimulated to produce two chosen species of antibody molecules simultaneously in large quantities, and says nothing about the potential of that cell prior to stimulation, the product of the cell at a later point in time, or the specificity of the progeny of that cell.

2 . Frequency of Antigen-Binding and Precursor Cells If an individual cell were multipotential, then it would be expected that a cell could bind and be stimulated to make antibody against a wide variety of related and unrelated antigenic determinants. Conversely, the precommitment hypothesis implies that the frequency of cells binding a given antigen would be relatively low. Experiments assessing the frequency of ABC for a wide variety of antigens (reviewed in Warner, 1974; Lefkovits, 1975)have, in general, shown that the cells binding SRBC, protein antigens, or some haptens comprise less than 0.1% of all lymphoid cells in a nonimmune animal and that this number increases upon specific antigenic stimulation. In addition, several studies have shown that individual ABCs or rosette-forming cells bound only one antigen (Biozzi et al., 1966; Julius et al.,

THE B-CELL CLONOTYPE REPERTOIRE

265

1976; reviewed in Warner, 1974). On the other hand, the high fequency of B cells binding galactosidase and other protein antigens (DeLuca et al., 1974) has been used as an argument against the clonal section hypothesis. This subject will b e considered in Section II,E, and therefore it is sufficient to note at this point that antigen binding may not, in itself, be a sufficient trigger for stimulating a B cell to make antibody. In order to assess whether B-cell precommitment exists on a functional level as well as on the antigen-binding level, a number of experimental systems have evolved that employ a limiting dilution of immunocompetent cells. The details of these techniques will be considered elsewhere in this review (Section IV), but all of them have in common the stimulation ofa graded number of cells in vivo or in vitro. The important point for this discussion is that all limiting dilution systems reveal a low frequency of antibody-forming cell precursors specific for a given antigen. Most notable in this regard are the experimental systems of Klinman and his collaborators (reviewed in Klinman and Press, 1975a) and of Lefiovits (reviewed in Lefkovits, 1975). Both techniques have the advantage of isolating a small number of immunocompetent cells i n vitro and prior to antigenic contact so that any ambiguities of antigenic competition that might be of concern in in vivo limiting-dilution systems (Kennedy et al., 1965; Playfairet nl., 1965; Sheareret d.,1968; Bosma and Weiler, 1970; Askonas et ul., 1970) were eliminated. Both in vitro techniques demonstrate that only an extremely small proportion of immunocompetent B cells respond to a given antigen, and the frequencies of responding cells in these two systems is remarkably similar (see Section IV). In addition, simultaneous stimulation of cultures with two antigens as closely related as DNP and trinitrophenyl (TNP) gives an additive frequency of precursor cells (Klinman et al., 1973). The analysis of precursor cells of neonatal mice in the Klinman splenic-focus system also speaks to one of the caveats raised at the beginning of this section. Press and Klinman (1973a, 1974) have shown that only a small proportion of isolated neonatal spleen cells responds to a given antigen and that stimulation of neonatal cells is as specific as adult cell triggering, since the precursor cell frequency when two closely related antigens are used for stimulation is additive (Press and Klinman, 1974). Moreover, the ability to tolerize this immature Bcell population and the exquisite specificity of this tolerance induction (Metcalf and Klinman, 1976) strongly suggest that B cells, early in their development, are already committed to a given specificity. Thus, the low frequency of ABCs and precursor cells responsive to a

266

NOLAN H. SIGAL AND NORMAN R. KLINMAN

given antigen suggests that both mature and immature B cells are committed to the synthesis of only one antibody specificity prior to antigenic stimulation. It should be noted that although the low frequencies of specific B cells are a logical requirement of the clonal selection hypothesis, the experiments described do not prove precommitment. The alternative hypothesis is that any lymphocyte is potentially able to respond to any antigen, but the probability of successful binding or triggering is low. If this were the case, however, attempts to remove specific cells would fail, since all cells would have the inherent potential to respond to all antigens. Thus, taken together, limiting-dilution analysis and specific precursor-cell depletion yield strong evidence for precommitment.

3. Physical Separation of Cells

The ability to specifically eliminate a population of ABCs and demonstrate the inability of the remaining cells to respond to that antigen directly supports the unipotentiality of the antibody-forming-cell precursors. Wigzell and Andersson (1969)were the first to specifically deplete a population of lymphoid cells of an antigenic reactivity by passage of the cells through an antigen-coated column. Although the initial experiments were performed with an immune cell population, later experiments applied the column technique to lymphocytes from nonimmune mice (Wigzell and Makela, 1970). A mixture of normal lymph node, spleen, and bone marrow cells was passed through a column containing ovalbumin-coated plastic beads. Column-passed and unpassed cells were transferred to lethally irradiated mice followed by immunization with ovalbumin and bovine serum albumin (BSA). There was no response to the ovalbumin in the column-passed cells, whereas the antibody response to BSA was normal; unpassed cells responded normally to both ovalbumin and BSA. While these experiments could demonstrate specific depletion by antigen-coated columns, Wigzell had little success in eluting the antigen-specific cells and testing their functionality. However, experiments using polyacrylamide beads (Henry et al., 1972) or an enzymically digestible matrix (Schlossman and Hudson, 1973; Phillips and Roitt, 1973) confirmed the results of Wigzell and, in addition, isolated the bound cell population. Henry et al. (1972) used hapten elution to recover a highly enriched cell population specific for azophenyl-P-lactoside, but the functionality of those cells was not demonstrated. Haas and Layton (1975) fractionated normal mouse spleen cells in dishes coated with thin layers of DNP-gelatin or NIP-gelatin, which were insoluble at 4°C. Highly viable cells were recovered from the dishes by melting

THE B-CELL CLONOTYPE REPERTOIRE

267

the gel at 37°C. After removal of cell surface-bound DNP-gelatin by treatment with collagenase, the recovered cells could be demonstrated to be enriched in DNP-binding cells. Haas (1975) investigated the functionality of the cell fractions recovered after hapten-gelatin binding. Using the in vitro PFC response to DNP-polymerized flagellin (POL) and/or NIP-POL, Haas (1975) showed that the bound cells were up to 300-fold enriched in antibody-forming cell precursors while cells not binding to the haptenated gelatin were specifically depleted. Purified DNP-specific cells responded only to DNP-POI, but not to NIP-POL. Nossal et al. (1977) conclusively demonstrated the relevance of the majority of these gelatin-binding cells. After two cycles of purification on NIP-coated gelatin and separation of NIPspecific rosette-forming cells, Nossal et al. (1977) found that one cell of every three was an NIP-specific precursor cell as assayed in a microculture system (see below). Scott (1976) has developed a general method of cell separation that has been used to purify B cells specific for POL and TNP, and alloantigen-reactive cytotoxic T cells. Lymphocyte populations were exposed to fluoresceinated antigens in vivo or in vitro, and the ABCs were retained on antifluorescein affinity columns. Specific cells could be eluted with an unrelated F1-labeled protein and shown to be immunocompetent. Another method of specifically depleting a cell population is the use of velocity sedimentation (Osoba, 1970) or Ficoll gradients (Bach et al., 1971) to remove rosette-forming cells. Recently, separation of rosette-forming cells that bind SRBC has been employed in a two-step gradient method designed to isolate large quantities of functional antigen-specific cells (Kenney, 1977). Lymphocytes from immunized animals were obtained at 50-100% purity, and those from nonimmune mice were 30-40% pure. The hapten-derivatized nylon fibers designed b y Edelman et al. (1971) added an additional dimension to physical separation techniques. Confirming the work of others, Rutishauser et al. (1972, 1973) first showed that the fibers could be used to specifically deplete or enrich nonimmune and immune spleen cell populations. In addition, the investigators enumerated the frequency of cells binding to the fibers and determined the relative affinity of the fiber-binding cells by hapten inhibition, demonstrating that only the high-affinity fiber binding cells increased in frequency upon immunization. Physical separation of ABCs in the fluorescence-activated cell sorter (FACS) provided a final level of sophistication to the demonstration of B-cell unipotentiality (Julius et al., 1972; Julius and Herzenberg, 1974). Cells from unimmunized mice binding DNP were found to be

268

NOLAN H. SIGAL AND NORMAN R . KLINMAN

specifically enriched at least 100-fold by this procedure when tested in an adoptive transfer assay for anti-DNP precursor activity. The fraction depleted of binding cells had little anti-DNP precursor activity but responded as well as unfractionated cells to stimulation by hemocyanin. The FACS was also used to demonstrate a positive correlation between the avidity of DNP-binding cells and the avidity of the antiDNP antibody secreted by their progeny (Julius and Henenberg, 1974). High-avidity cells were stained and fractionated using a low concentration of DNP. Medium and low-avidity DNP-binding cells were isolated using a high concentration of fluoresceinated DNPmouse globulin (MGG) plus free hapten to block staining of high-avidity cells. When these fractions were transferred to irradiated recipients along with excess T cells and antigen and the avidity of the direct PFCs was measured by hapten inhibition, it was found that high-avidity DNP-binding cells gave rise to predominantly high-avidity anti-DNP PFCs. Medium- and low-avidity binding cells produced medium- to low-avidity anti-DNP PFCs.

4 . Capping Experiments Raffet al. (1973) designed an experiment to test directly whether an ABC can be committed to synthesize and display on its surface more than one antibody specificity. Their experimental approach is illustrated in Fig. 1.Briefly, spleen cells from nonimmune mice were incubated with a multivalent antigen, polymerized flagellin (POL),under conditions known to induce cap formation. The cells were transferred to noncapping conditions (sodium azide, 4°C)and incubated with rhodamine-coupled rabbit anti-POL to label the POL-binding cells. Then the cells were incubated with fluoresceinated anti-Ig to label all cells bearing Ig. POL-binding cells were identified by scanning for rhodamine caps, and these and all other cells were examined for anti-Ig staining with fluorescein, specifically asking whether some Ig was left on the surface of the POL-binding cells. Seventeen out of 440,000 B cells had a rhodamine-stained cap, and on 14 of the 17 cells all the fluorescein labeling coincided completely with the cap. The remaining 3 cells had most of their Ig in the cap and a small amount of Ig outside the cap, but further experiments showed that the noncapped Ig also had anti-POL specificity. Raff et al. concluded that individual B cells display receptors of a single antibody specificity. The validity of this conclusion is based on the independence of the capping phenomenon for various surface proteins and the sensitivity of the detection techniques. There is ample evidence that separate surface protein molecules, e.g., histocompatibility antigens (Taylor et d.,1971; Preud’homme et al., 1972; Neuport-Sautes et ul., 1973),concanavalin

THE B-CELL CLONOTYPE REPERTOIRE

269

W Spleen cell suspension

POL- induced cops rhodomine -lobeled onti-POL

-

fluorescein-labeled onti -1g

fluorescein-lobeled Ig-bearing cells

10

POL-binding cops

rhodomine-lobeled POL- binding caps

I

fluorescein stoining

FIG. 1. Design of a n experiment to show that individual B cells are committed to one, and only one, specificity (see text for details). POL, polymerized flagellin. From Lefkovits (1975).

A receptors (Unanue et al., 1972), and Ia molecules (Schreiner and Unanue, 1976), all redistribute independently of Ig. On the latter point, the investigators have demonstrated that their method could detect B cells with 3% or more of the normal amount of surface Ig. Therefore, although the experiments do not rule out the possibility of a small subpopulation of B cells with multiple specificities displayed on their surface, they strongly indicate that the vast majority of antibody-forming cell precursors are committed to one, and only one, specificity.

E. INTEGRITY OF THE B-CELLCLONE The previous discussion has centered on the question of whether an immunocompetent virgin B cell has the potential to respond to more

270

NOLAN H. SIGAL AND NORMAN R. KLINMAN

than one antigen at the time of antigen contact. This section deals with a related problem, namely, whether the clonal progeny of the stimulated precursor cell remains committed to that specificity or whether a clone has the capabiIity to rapidly alter its combining site during the course of an immunization. While somatic mutational events may or may not contribute to antibody diversity, the primary concern in this section is the demonstration that such processes do not occur with significantly greater frequency than in other biological systems. The major challenge to this aspect of the clonal selection hypothesis, i.e., the work of Cunningham and his collaborators (Cunningham and Pilarski, 1974a,b,c; Cunningham, 1976a,b), will be considered in Section II,F. The homogeneity of the immunoglobulin product from neoplastic lymphoid cells of mice or humans has been used as evidence for the integrity of the B cell clone. Two groups of investigators (Birshtein et al., 1974; Milstein et al., 1977) have investigated the potential variability with mouse plasmacytoma cell lines. Although the spontaneous mutation rate in such cell lines is as high as lop3per cell per generation, the vast majority of mutants fail to synthesize or release the heavy or light chain, have a deletion of residues between the variable and constant regions, or appear to be frameshift mutations in the constant region (Morrison et al., 1974; Secher et al., 1974). Another approach to the analysis of the potential antibody product of individual B cells is to isolate the clonal progeny of such cells. Such studies have been done by injecting limiting dilutions of cells into lethally irradiated recipients, immunizing the presumptively isolated B cell either in uivo or in uitro, and examining the antibody product of the single cell’s clonal offspring. Bosma and Weiler (1970) injected small numbers of F1 spleen cells into syngeneic irradiated recipients, immunized the reconstituted recipients with either poly-DL-alanine or polyserine, and looked for “all-or-none” PFC response in the spleen. The frequency of responses was lineally related to the cell dose, and the resulting antibody was found to be restricted on the basis of allotype (Bosma and Weiler, 1970) and binding behavior (Bosma et al., 1972). Askonas et al. (1970, 1972) combined the limiting cell dilution system and isoelectric focusing to elegantly demonstrate the unipotentiality of a B cell clone. When they transferred 1 x 10‘ spleen cells from mice previously immunized to DNP-BGG and later boosted the recipients with the homologous antigen, some of the animals were found to have homogeneous anti-DNP antibody in their sera as shown by isoelectric focusing (IEF).In addition, one of the anti-DNP clones

THE

B-CELL CLONOTYPE REPERTOIRE

27 1

(E9) was transferred and expanded through nine recipient generations before senescence occurred (Askonas and Williamson, 1972a; Williamson and Askonas, 1972); the anti-DNP antibody was homogeneous and identical by isoelectric focusing through all passages (Askonas and Williamson, 1972b). Thus, clonal integrity was maintained through approximately 90 divisions of a single immunocompetent B cell clone. Critics note, however, that the E 9 antibody was produced from a highly expandable secondary B cell, which may not be representative of a typical clonal precursor. The splenic focus system (reviewed in Klinman and Press, 1975a; and described in Section IV) isolates individual B cells in fragment culture, and, after stimulation with antigen, the antibody derived from the progeny of such B cells can be analyzed. If a B-cell clone has restricted potential and each fragment contains, at most, one precursor cell, then the antibody should be homogeneous. Analysis of foci derived from both immune and nonimmune adult and neonatal donors supported the validity of these assumptions. The monofocal antibody produced appeared homogeneous by the criteria of: (1)homogeneous hapten-binding properties by equilibrium dialysis (Klinman, 1969, 1971b,c, 1972); (2) regain of binding activity after recombination of separated heavy (H) and light (L) chains (Klinman, 1971a); (3)restricted I E F spectra (Klinman, 1972; Press and Klinman, 1973a; Klinman and Press, 1975b,c); (4)concordance, on a weight basis, of antibody quantitation measured with anti-Fab antibody or with antiidiotype antibody (Gearhart et d.,1975a,b); and (5) the all-or-none binding characteristics of anti-influenza monofocal antibodies to cross-reacting strains of influenza virus (Gerhard et al., 1975). Using a microculture technique that also isolates a single B cell in vitro and permits the analysis of antibody derived from the clonal progeny, Luzzati et d . (1973b) also demonstrated the homogeneity of the clonal antibody product by high-voltage agar electrophoresis and allotype restriction. It should be noted that neither the splenic-focus system nor the microculture technique could rule out a small degree (5%) of contamination with other antibodies. Nonetheless, the analyses performed with isolated B cells stand as perhaps the best evidence for the integrity of the B-cell clone. F. CONTRADICTIONS TO THE CLONAL SELECTION THEORY The evidence presented in the preceding pages provides overwhelming experimental verification of the clonal selection hypothesis. Nevertheless, there are a number of experimental findings that

272

NOLAN H. SIGAL AND NORMAN R. KLINMAN

appear to be incompatible with the theory, and these are described and analyzed in this section.

1. Rapid Variation within a B-Cell Clone Cunningham and his co-workers (reviewed in Cunningham and Pilarski, 1 9 7 4 ~Cunningham, ; 1976a,b) have performed a series of experiments that appear to indicate that B cells undergo rapid variation in their specificity following antigenic stimulation. The experimental findings are summarized in Table I. The assay used to detect changes TABLE I EXPEFUMENTAL EVIDENCEFOR THE RAPID VARIATION WITHIN A B-CELLCLONE Type of experiment

A. Variation within single clones

Protocol

Results

1. Culture of single plaquefonning cells (PFC); 1 PFC isolated by micromanipulation; grown 2 days in oitro; resulting small clone examined for variants (Cunningham and Fordham, 1974)

Of these cultures, 10% developed into 2 or more

2. Limit-dilution cultures in uitro (Pilarski and Cunningham, 1974) 3. Limit-dilution “cultures” in uiuo: clones mown in spleens of irradiated recipient mice (Pilarski and Cunningham, 1975) B. Rare antibody specificities are absent early in a response, but appear later

1. In uivo; following PFC response of mice injected with sheep erythrocytes (Cunningham and Pilarski, 1974b)

2. In uitro; comparing sizes of mitogen-stimulated clones producing common or rare antibody specificities (Pilarski and Cunningham, 1974)

PFC; in most cases, all PFC in a clone made apparently the same antibody specificity; in 10 out of 93 clones, daughter cells made different antibodies Significant proportion of single clones contained mixtures of different antibody specificies As above

Certain highly cross-reactive antibody specificities, present at the peak of a response, are entirely absent from most mice at early stages PFC of rare specificities occur only in small “clones”

THE B-CELL CLONOTYPE REPERTOIRE

273

in antibody specificity involved a mixture of two closely related antigens in PFC assay, e.g., erythrocytes from two different sheep or horse erythrocytes plus SRBC. Plaques arising from antibody-forming cells can be “clear,” implying high affinity for both antigens; “partial,” suggesting high affinity for only one of the two antigens; or “sombreros,” suggesting differing affinities for the two. The initial observation made by Cunningham and Pilarski (1974a) was that if one immunized a mouse to SRBC, the early PFCs were all partials, and only later in the response did PFCs arise that had a wider range of specificities. In fact, more than half of the early PFCs were so specific that they lysed red cells from the immunizing sheep but not from other sheep. These findings led Cunningham to postulate a novel theory for the generation of antibody diversity (Cunningham and Pilarski, 1974c; Cunningham, 1976a,b): the nonimmune animal contains relatively few specificities in his repertoire, but when stimulated b y antigen, an immunocompetent cell can rapidly generate within the clone a family of related but random specificities from which antigen “selects” those that fit best. T o test the theory, the clonal P F C progeny generated from a single B cell were examined in uitro (Pilarski and Cunningham, 1974) or in uivo (Pilarski and Cunningham, 1975). In addition, the daughter cells of a single micromanipulated PFC were analyzed and revealed changes in specificity (Cunningham and Fordham, 1974). Once again, assay for specificity of the antibody was by means of plaque morphology on a mixture of two related antigens. The validity of Cunningham’s findings rest principally with the use of plaque morphology as a marker of antibody specificity, i.e., changes in the variable-region amino acid sequence. Cunningham and Pilarski (197413) have shown that (1) plaques maintain their morphology as they grow in size; (2) plaque morphology is not affected by different concentrations of complement in the medium; ( 3 )only IgM PFCs are generated under the conditions used; and (4) the PFCs are probably not releasing two different antibody specificities. Nevertheless, the use of plaque morphology to follow V-region structure remains controversial. Aside from questions of technique, Cunningham and Pilarski ( 1 9 7 4 ~calculated ) that one division in 30 leads to the generation of a new antibody specificity, but such a phenomenally high mutation rate does not take into account those V-region mutations that would lead to loss or no change in antibody specificity, which might b e expected to be much more probable events. Thus, until such experiments are repeated with more conventional V-region markers such as anti-idiotype or IEF, these theoretical and experimental observations must be considered preliminary.

274

NOLAN H. SIGAL AND NORMAN R. KLINMAN

2 . Multipotentiality of Antibody-Forming Cells Liacopoulos et al. (1976) have investigated the capacity of immunocompetent cells to produce more than one type of antibody molecule, as assessed by the ability of a single PFC to lyse two unrelated determinants. It was initially observed that immunization of Swiss mice with both sheep and horse erythrocytes or with TNP conjugated to SRBC resulted in the transient appearance of 1-4% of the PFCs simultaneously lysing both antigens (Couderc et al., 1973). Experiments have attempted to demonstrate that the simultaneous lysis was due to production of two different antibody molecules by a single cell using micromanipulation of bispecific PFC to media containing one of the two determinants in soluble form (Couderc et al., 1975; Liacopoulos et al., 1976). Results showed that, in 70-80% of “bispecific” cells, inhibition of one activity by soluble antigen did not interfere with the other activity. Another series of experiments involved the micromanipulation of bispecific PFC to microculture in order to examine the progeny of these cells. The investigators observed that the majority of bispecific PFC taken from mice immunized with TNP-SRBC produced daughter cells making only anti-SRBC or anti-TNP, but some of the monospecific cells removed from the mouse gave rise to bispecific PFCs or PFCs making antibody to the other determinant (i.e., anti-SRBC PFCs yielded anti-TNP or bispecific daughter cells). This transformation required the presence of both antigens in the microcultures. It should be noted that experiments from other laboratories have, in general, not observed the appearance of PFCs producing two different antibodies (see Makela and Cross, 1970). In addition, stimulation of isolated B cells in fragment culture with two or more antigens did not result in a higher than expected frequency of clones making antibody to both antigens ( U n m a n et al., 1973; N. H. Sigal and N. R. Klinman, unpublished results). Although one must again question the validity of antibody-mediated cell lysis as a marker of V-region sequence, the results of Liacopoulos and his co-workers remain consistent with the possibility that a small proportion of B cells have the potential to synthesize more than one antibody specificity.

3. Mu 1tip0 tential i t y of An tigen-B inding Cell s While the observation of a low frequency of cells binding an antigen is not a compelling argument in favor of the clonal selection theory, the finding of a high frequency of ABCs for a variety of antigens would be inconsistent with the specificity of a unipotential B cell and might

THE B-CELL CLONOTYF'E REPERTOIRE

275

be considered evidence against the unipotentiality of the precursor cell. DeLuca et al. (1974, 1975a) have enumerated the frequency of ABCs in unimmunized mice which were specific for P-galactosidase arid a number of other protein antigens. They found that approximately 2% of spleen cells and 5% of bone marrow cells bound these antigens under optimal conditions using glutaraldehyde fixation, and that a significant proportion of cells bound two unrelated antigens, such as P-galactosidase and Hy or P-galactosidase and horse spleen ferritin (DeLuca et al., 1975b). Since antigen binding was inhibited by anti-Ig, and the ABCs can regenerate their receptors in uitro, the phenomenon appears to be specific and not due to cytophilic Ig. When the receptors for these ABCs are capped by one of the antigens, DeLuca et al. (1975~) found that the receptors for the other antigen were also found in the cap, suggesting that only one receptor having multiple or cross-reactive antigen-binding specificities existed on the B cell. Other laboratories (Lawrence et al., 1973; Hammerling and McDevitt, 1974; Dwyer and MacKay, 1972a; Cooper et al., 1972; Rutishauser et al., 1972) have also observed high frequencies of ABCs, and, under optimal conditions, up to 4% of murine splenic B cells can bind DNP (Lawrence et al., 1973; Klinman et al., 1976). However, a study by Julius et al. (1976) using the FACS failed to find one cell in the 13,000 examined which simultaneously bound both DNP and Hy. Although antigen binding is necessary for antigenic stimulation, a large body of experimental evidence indicates that this event is insufficient in itself to cause triggering of the B cell. Klinman et al. l(1976) have shown that, for the hapten DNP, less than 10% of the cells that bind the antigen can be specifically stimulated to antibody production. While the fraction of cells binding antigen increases with higher hapten concentrations, the frequency of antibody-forming cell precursors stimulated in uitro remains constant (Klinman, 1972), implying that antigenic stimulation of primary B cells may be highly affinity dependent and that monovalent interaction between receptor and antigenic determinant must exceed a threshold affinity to permit triggering. Thus, low-affinity cells may bind antigen, but these B cells will not be stimulated. Rutishauser et al. (1973), who obtained a high frequency of cells binding to nylon fibers, concluded that since only the higher affinity fiber-binding cells increase upon immunization, not all cells adhering to the fibers are physiologically triggerable by the antigen. Moreover, it has been shown that binding of antigen with cap formation is insufficient for triggering to occur (Elson et d . , 1973). For some antigens, e.g., phosphorylcholine (PC), there is no discrepancy

276

NOLAN H. SIGAL A N D NORMAN R. KLINMAN

between ABC analysis and precursor cell frequency (Claflin et al., 1974a; Klinman et al., 1976; Sigal et al., 1975), perhaps because PC is not as “sticky” as DNP. It should also be pointed out that under certain circumstances antigen binding alone is sufficient for B-cell triggering to proliferation and/or antibody production, e.g., when the antigen is a B-cell mitogen or a hapten is attached to a mitogen such that it is “concentrated” by the relevant B cells (Coutinho et al., 1974; Moller, 1975). However, the relevance of this process to physiologic B-cell stimulation is unclear. A second possibility to explain the high frequency of ABC is that a subpopulation of B cells exists that expresses a multiplicity of antigen receptors but is not stimulatable or is stimulated to produce very small quantities of antibodies. Since overlaps in binding are so high, fewer than 5-10% of all B cells could account for all such cells. This would be consistent with the report of DeLuca et al. (1975a) that most ABC appear to be large eosinophilic cells. Such a subpopulation of B cells would not appear to have receptors for certain antigens such as PC (Claflin et al., 1974a; Klinman et al., 1976; Sigal et al., 1975) and flagellin (Diener et al., 1974) since few cells bind these antigens and all PC and flagellin-binding cells may be stimulatable.

G. CONCLUSION The validity of the clonal selection hypothesis and its consequent receptor hypothesis seems to be well established by the weight of accumulated evidence. Consequently, it is clear that interpretation of the basis for antibody specificity will require a thorough understanding of the B-cell specificity repertoire. However, experimental evidence contradictory to the basic clonal hypothesis may indicate that there are B-cell populations that fail to show commitment and that our understanding of the processes that control variable region expression on circulating or surface antibodies is too primitive to permit rigid interpretations. Such gaps in our understanding have not precluded important and intensive analyses of adult and neonatal B cell repertoires, which will be addressed in Sections IV and V. 111. Methods of Clonotype Identification

In order to determine the number of antibody specificities an individual can make, one must have a means of identifying a clonotype or group of clonotypes and determine the frequency with which the specificity is represented in the B-cell pool. This portion of the review will focus on methods used to identify unique antibody specificities.

THE B-CELL CLONOTYPE REPERTOIRE

277

Each of the principal techniques, isoelectric focusing (IEF), fine-specificity analysis, and idiotypy, takes advantage of a particular characteristic that distinguishes one clonotype from another. For example: I E F utilizes charge differences resulting from disparities in the amino acid sequence of the variable (or constant) region; fine-specificity analysis identifies peculiar binding properties of a clonotype to the homologous antigen and related determinants; and idiotypy employs antibody specific for variable-region determinants. Although the only definitive method for establishing the identity of the V regions of two antibodies is their complete amino acid sequence, limitations on the quantity of antibody required for sequencing with current techniques precludes its use in most cases. Some of the studies in which sufficient quantities of antibody have been isolated are considered at the end of this section.

A. ISOELECTFUC FOCUSING 1 . S l a b Gel IEF Analytical I E F in thin layers of polyacrylamide gel was shown by Awdeh et al. (1968)to be capable of resolving a specifically purified antibody into a pattern of discrete bands. Carrier ampholytes migrate to a characteristic isoelectric point when placed in an electric field, and thus a mixture of these ampholytes will establish a pH gradient of any desired range. If antibody molecules are placed with the ampholyte mixture in a charged field, they will migrate to the portion of ampholytes of identical isoelectric point, which will vary according to the antibody’s composition with respect to positively and negatively charged amino acids. The technique theoretically has the capability to distinguish between two antibody specificities differing by only one charged amino acid. However, it is also clear that two antibodies with differing sequence can show indistinguishable spectrotypes due either to differences only in neutral amino acids or to the coincidental accumulation of positively and negatively charged amino acids such that two distinct clonotypes have identical net charge. The general method for IEF, as modified by Williamson (1971) for analysis of serum antibody, involves the use of 5% ( w h ) acrylamide slabs containing 2% carrier ampholytes. The pH range of the ampholytes can be varied, but most immunoglobulins focus from pH 4 to 8; narrower gradients will increase resolution. T h e 5% acrylamide excludes IgM antibodies owing to the pore size of the acrylamide. After application o f t h e sample to the surface of the gel by adsorption onto pieces of filter paper, the gels are run at constant voltage for approxi-

278

NOLAN H. SIGAL AND NORMAN R. KLINMAN

mately 20 hours. After measurement of the pH gradient, l3II- or 1251-labeled antigen is applied to the surface of the gel and allowed to combine with the focused antibody; the antibody-antigen complexes are fixed in the gel, the excess label is washed out, and the gels are autoradiographed. The resolving capacity of I E F is tremendously increased by postsynthetic deamidation, which characteristically occurs after release of the antibody molecule from the cell. Focusing of a homogeneous myeloma protein, for example, results in a microheterogeneous isoelectric spectrum consisting of 3-5 closely spaced lines (Awdeh et al., 1970). The number of lines and spacing between the lines tends to be a characteristic property of an individual clonotype, so that clonotype identity requires the matching of multiple antibody bands rather than just one. Williamson et al. (1973) have estimated that the slab gel I E F technique could resolve 5 x lo4different antibody specificities. However, recent studies which focused myeloma proteins known to have multiple amino acid differences revealed that I E F may not have the discriminatory ability once assumed (Pink and Skvaril, 1975). Numerous investigators have employed the slab gel I E F technique to compare one antibody specificity with another. Awdeh et al. (1970) showed that the antibody synthesized by a plasmacytoma in vitro is homogeneous and Askonas et al. (1972) demonstrated the identity of the antibody product of a clone of cells propagated through numerous cell generations. The comparison of clonotypes specific for NIP, DNP, or P-galactosidase by IEF has provided an important tool for defining the repertoire and will be discussed in detail in Section IV. Immunization of rabbits or mice with group A streptococcal vaccine results in electrophoretically restricted antibody. Among BALB/c mice, two spectrotypes occurred in each of three mice and one spectrotype was observed in two mice (Cramer and Braun, 1974). In A/ J mice a single predominant clonotype as defined by IEF is elicited on immunization with streptococcal carbohydrate (A5A clonotype) (Eichmann, 1972), and the inheritance of this spectrotype is linked to the mouse heavychain allotype (Eichmann and Berek, 1973).Another clonotype genetically linked to the murine heavy-chain allotype, the heteroclitic antiNP specificity in C57BL/6 mice (Imanishi and Makela, 1974), has been shown to be associated with a characteristic set of isoelectric spectrotypes (McMichael et al., 1975). Recently, Cramer et al. (1976) have used I E F to demonstrate the similarity in the B-cell repertoires of high- and low-responder strains in their antibody response to TGAL. I E F has also been employed by DuPasquier and Wabl(l976) to investigate the anti-DNP repertoire in amphibian tadpoles and adults (see Section V).

THE

B-CELL CLONOTYPE REPERTOIRE

279

The anti-DNP response is characteristically too heterogeneous to be analyzed by IEF, but the presentation of DNP on unusual carrier moieties can elicit a restricted response. Immunization of strains 2 and 13 guinea pigs with a-DNP-deca-L-lysine gave a restricted response involving an average of fewer than two spectrotypes per animal (Civan et aZ., 1976).Of 51 antibodies elicited, only 19 different spectrotypes could be distinguished, and one was observed in 60% of the guinea pigs studied (Williamson, 1976). A similar pattern is observed in outbred rabbits immunized with DNP-gramicidin S, with most responding individuals maintaining the same one or two spectrotypes throughout the course of immunization (Montgomery et al., 1972, 1975a). The validity of I E F to recognize identical clonotypes was supported by the discovery of two anti-DNP-gramicidin S antibodies with similar p1 that had identical association constants for DNP and maintained their affinity when heavy and light chains from the two clonotypes were recombined with each other (Montgomery et aZ., 1975b). 2 . Other I E F Techniques Sucrose gradient I E F employs the same principles as slab gel I E F but allows high molecular weight immunoglobulins, such as IgM, to be focused (Haglund, 1971; Press and Klinman, 1973a). The disadvantage of this technique is the limited resolution capability, estimated at only 50 different clonotypes. Press and Klinman (1973a) described a microscale sucrose gradient IEF in which as little as 1 ng of homogeneous antibody could be characterized. After focusing, drops from the sucrose gradient are collected into saline, the p H is read, and each of the 100 aliquots can be simultaneously assayed for antibody activity, idiotype or heavy-chain class with sensitive radioimmunoassays. This microtechnique has been used to (1)demonstrate the homogeneity of the antibody derived from splenic fragment cultures (Press and Klinman, 1973a); (2) examine the clonotype repertoire of neonatal BALB/c mice for DNP and TNP (Klinman and Press, 1975b); and (3)demonstrate that antibody from a single clone can express one idiotype in combination with a number of heavy-chain classes (Gearhart et d., 1975b). Preparative-scale I E F has been used by Nisonoff and his co-workers (reviewed in Nisonoff et al., 1977) to dissect the major and minor clonotype repertoires responsive to p-azophenylarsonate (Ars) in A/ J mice. Tung and Nisonoff (1975)utilized two methods to isolate quantities of the predominant idiotype. In one method, I E F was carried out in cylindrical polyacrylamide gels with elution of the protein after focusing. The other method utilized a bed of Sephadex G-75 and elu-

280

NOLAN H. SIGAL AND NORMAN R. KLINMAN

tion with a borate buffer of 0-5-cm slices. The former technique was used by Ju et al. (1977) to isolate private idiotypic specificities from hyperimmune A/ J mice suppressed for the predominant idiotype.

3. Light-Chain Focusing Claflin (1976a,b, 1977)has developed a method to study the isoelectric spectra of isolated light chains in order to examine the repertoire of specificities responsive to phosphorylcholine (PC) in mice. After reduction and alkylation of purified myelotna proteins or anti-PC antibodies, heavy (H) and light (L) chains were separated either on Sephadex G-100 in 1 M propionic acid-4.5 M urea or on disc electrophoresis in sodium dodecyl sulfate. The isolated L chains were focused in 5% acrylamide, 2% ampholytes, and 8 M urea; focused chains were stained with Coomassie blue. H chains can also be focused by this method, but apparently give poor resolution (Claflin et al., 1975). The technique has been used to show that all mouse strains have L chains that are identical to TEPC 15, MOPC 511, and McPc 603 myeloma proteins in their isoelectric spectra (Claflin, 1976a). Claflin (1976b) has also presumably identified a genetic marker in the variable region of mouse K chains based on the variations in anti-PC L chains among various stains. L chains from normal immunoglobulin of immunized mice have been analyzed by Gibson (1976). L chains were separated on Sephadex or by urea-formate gel electrophoresis and focused as described in the preceding paragraph. Examination of focusing patterns, which consisted of over 50 bands, showed that six inbred mouse strains possessed virtually identical spectra of focusing bands and three other strains (RF/J, AKR/ J, and C58/ J) showed differences in several bands. Genetic analysis revealed that the differences in focusing patterns were inherited in a codominant fashion. Gibson (1976)suggested that the banding may have something to do with the inheritance of the Lchain variable region genes, which may be related to a finding of Gottlieb and Durda (1977) that a peptide associated with K L-chains was absent in these strains.

B. FINE-SPECIFICITY ANALYSIS Fine-specificity analysis attempts to characterize an antibody population by its relative affinity for a number of related determinants. In this manner, one can determine more precisely the exact specificity to which the antibody is directed and, most important, define a subset of clonotypes within a more heterogeneous array. Relative affinity can b e examined by hapten inhibition with substances chemically related

28 1

T H E B-CELL CLONOTYPE REPERTOIRE

to the original immunogen or by cross-binding to related (or unrelated) determinants.

1 . Fine Specificity f o r Reluted Haptens Makela et al. (Imanishi and Makela, 1973, 1974,1975; Makela et al., 1976, 1977a,b; Makela and Karjalainen, 1977) have defined a number of clonotype subsets on the basis of an antibody population’s relative affinity for related haptens. The best-studied system has been the antiNP response in C57BLK mice, the antibody of which has a higher relative affinity for NNP and NBrP than for the original immunogen NP (heteroclitic antibody), as assessed by hapten inhibition of serum antibody binding to haptenated phage or by inhibition of PFCs (see Table I1 for compounds) (Imanishi and Makela, 1973). Since all mouse strains other than C57BL, LP, and 101 produce nonheteroclitic antibody, two phenotypes can b e defined. Experiments with congenic, recombinant inbred and backcross mice have shown that the ability to make the heteroclitic clonotype on immunization with NP-BSA is linked to the murine heavy-chain allotype (Imanishi and Makela, 1974). The antibody defined b y heteroclitic fine specificity is a heterogeneous population of molecules as demonstrated b y IEF. Although initial studies indicated that there was only one spectrotype in the C57BL anti-NP response (McMichael et al., 1975), more recent work TABLE I1 RELATED HAPTENSUSED FOR FINE-SPECIFICITY ANALYSIS NP system

Compound NP NNP NBrP NIP NClP DIP

ABA-HOP

system

R,

R,

Compound

R,

R,

R3

H NO, Br I CI 1

NO, NO, NO, NO2 NO, 1

ABA-HOP BOC-ABA-Tyr ABA-NP ABA-MIP ABS-HOP

ASO,H, ASO,H, ASO,H, ASO,H, SO,H,

H H NO, I H

COOH BOC-NHCHCOOH COOH COOH COOH

282

NOLAN H. SIGAL AND NORMAN R. KLINMAN

has identified 3-4 additional spectrotypes (Makela et al., 1977a). An anti-idiotypic antibody has also been made against the purified C57BL antibody, and the anti-idiotypic serum reacts with all 4-5 clonotypes in the population. Thus, fine-specificity analysis has identified a group of clonotypes, and the validity of the method has been confirmed by I E F and idiotypy. Imanishi and Makela (1975) found that the antibody response to NBrP (see Table 11) could be divided into three phenotypes by finespecificity analysis. Anti-NBrP antibodies of all allotype b mice had a high relative affinity for NNP but low for NCIP; another category, which included most tested strains, was characterized by high relative affinity for NCIP but low for NNP; and CBA and CH mice had intermediate fine specificity. Genetic studies demonstrated that the second phenotype (high affinity for NCIP) was allotype linked and dominant over the phenotype found in the allotype b mice. A final system examined by Makela and his co-workers (19764 explored the fine specificity of the antibody response to ABA-HOP (see Table 11)by inhibition of haptenated phage with related compounds. The fine-specificity patterns were more complex than the previous systems described with three or more phenotypes characterized. All C57BL mice had a high relative affinity for ABA-MIP, and all A/J mice had a low relative affinity for ABA-MIP, but fine specificity in individuals of other strains or in heterozygotes and backcrosses indicated that some of the clonotypes were expressed in a pseudoallelic fashion upon immunization. The C57BL antibody also had a characteristic I E F pattern, which was shared by many individuals (Makela et al., 1976). Thus, the experimental system served to identify a unique clonotype subset controlled by a set of germline genes. Claflin and Davie (197413) examined the fine specificity of antiphosphorylcholine (PC) antibodies, utilizing the inhibition of PFC formation with the chemically related compounds glycerophosphorylcholine (GPC), and choline (C). The results demonstrated that the PFCs from all mouse strains have the same relative affinity for these compounds after immunization with PC-Hy or R36A pneumococcus. This finding is related to the observation that all mice share an antibody molecule with the same idiotypically identifiable combining site as TEPC 15 and HOPC 8 plasmacytoma proteins. This fine-specificity analysis has been applied to anti-PC monoclonal antibodies generated in the in vitro splenic focus system (Sigal et al., 1977a). These studies on the fine specificity of homogeneous antibodies revealed that, in combination with sucrose gradient IEF, a large number of clonotypes

THE B-CELL CLONOTWE REPERTOIRE

283

that do not possess the TEPC 15 idiotype could lie defined within the anti-PC antibody popul at’1011.

2. Other Fine-Specijcity Analyses McCarthy and Dutton (1975) used a somewhat different form of fine-specificity analysis in their study of the response of mouse spleen cells to SRBC. It was found that erythrocytes from certain sheep elicited a much larger PFC response than other SRBC, and in some mouse strains the higher response was due to “private” antigens on the SRBC that were not shared among all sheep. Thus, “discriminator” strains of mice made antibody predominantly directed to the private antigens on the SRBC rather than to shared determinants on the erythrocytes; nondiscriminator strains of mice responded only to the common antigens. Since the ability to make antibody against the private SRBC antigens is linked to the H-chain allotype (McCarthy and Dutton, 1975; McCarthy et al., 1976), it is likely that this fine specificity analysis defines a small set of clonotypes, the expression of which is controlled b y a single VHgene. The immune response of C57BL mice to hen egg white lysozyme (HEL) is under the control of a gene located in the I region of the murine major histocompatibility complex (Hill and Sercarz, 1975). While only 10% ofC57BL mice make antibody to HEL, the repertoire of specificities in the responding mice could be examined by looking at their I E F spectrotypes and inhibiting anti-HEL P F C with closely related gallinaceous lysozymes, such as peafowl, Japanese quail, and turkey egg lysozymes. I E F of serum antibody demonstrated that individual C57BL mice made a restricted antibody response, but by I E F criteria and the cross-reactivity with lysozymes differing b y as few as four amino acids, the clonotypes in the individual mice were shown to be different. Since the amino acid sequences of many of the lysozymes is known, fine-specificity analysis in this system has the potential to delineate the precise sequence to which the antibody specificity is directed. This experimental model has also been used to investigate the T-cell specificity repertoire b y inducing T-cell tolerance to H E L and challenging with the related lysozymes (Ceckaet al., 1976). DiPauli (1976)employed a similar system to look at the fine specificity of murine antibodies against lipopolysaccharide (LPS). Mice were immunized with LPS from one strain of Salmonella, and the serum antibody was characterized by its relative avidity for LPS from other Salmonella strains. Two phenotypes were identified by cross-re-

284

NOLAN H. SIGAL AND NORMAN R. KLINMAN

activity patterns against LPS from S. cholera-suis, but each individual was shown to have a heterogeneous anti-LPS repertoire based on fine specificity to two other LPSs. The influenza A viral hemagglutinin (HA) provides another system where the antibody response to a large number of closely related but antigenically distinct proteins can be examined. By an extensive analysis of over 200 monoclonal anti-HA antibodies, derived from secondary B cells in the Klinman splenic-focus system and a panel of 9 immunoadsorbents comprised of PR8 (the homologous stimulating antigen) and 8 other viruses of closely related HA type, Gerhard (1976, 1978) has been able to distinguish 47 different cross-reactivity patterns. These studies established both the enormous variety of determinants identifiable on a single protein as well as the wide array 01 clonotypes present in BALB/c mice capable of recognizing the

PR8 HA. 3 . Cross Binding t o Unrelated Haptens

The experiments discussed in the preceding paragraphs all have in common the use of closely related antigens to assess fine specificity of an antibody population. There is no a priori reason, however, to confine such an analysis to related antigens, since Richards et al. (1974, 1975) have shown that a myeloma protein MOPC 460 can bind two unrelated haptens, DNP and menadione, in apparently different parts of its combining site. The utility of binding unrelated haptens as a method of clonotype identification was illustrated b y the following experiment (Richards e t al., 1974): a rabbit was immunized with uridine-bovine y-globulin (BGG) and its serum was analyzed by IEF; although the IEF had over 300 bands when developed with labeled uridine, one or two clonotypes could be distinguished when labeled DNP was used to overlay the IEF gel; when the rabbit was reimmunized with DNP-BGG, these one or two clonotypes were the first antiDNP antibodies to appear after immunization, illustrating the phenomenon of “original antigenic sin” (Fazekas de St. Groth and Webster, 1966) at the clonotype level. Monoclonal antibodies generated in the splenic-focus system have also been examined for binding to unrelated haptens (Sigal, 1977a).Of 680 anti-DNP antibodies examined for their ability to bind PC, two monoclonal antibodies were identified with these properties, and these two were distinguishable by idiotypic differences (see Section IV).

THE B-CELL CLONOTYPE REPERTOIRE

285

C. IDIOTYPE

The terms “idiotype” (Oudin, 1966)or “individual antigen determinants” (Kunkel et al., 1963)were originally used to designate those antigenic determinants found in a population of antibody molecules that were not obseived in the other immunoglobulins of the donor animal or in other animals of that species. In the ralhit, this definition distinguished idiotypic determinants from variable (V)-region allotypic determinants, since the latter was found in preimmune serum antibody from the donor and were inherited from one generation to the next. However, the discovery of cross-idiotypic specificities among human myeloma proteins (Williams et al., 1968), a large number of murine 1969),and inherited idiotypes intrastrain cross-reactions (Cohn et d., in the mouse (Kuettner et d . , 1972) and rabbit (Eichmann and Kindt, 1971) have made the original definitions obsolete. Indeed, even the distinction between rabbit V-region allotypes and idiotype may be blurring (see helow). Therefore, perhaps a better definition of an idiotypic determinant is simply an antigenic determinant found in the immunoglobulin variable region. The value of idiotypy as a means of clonotype identification is 2-fold: to search for clonotypes that are shared among individuals of a strain and to explore relationships among variable regions from different individuals. Since there have been a number of excellent reviews that cover the historical and genetic aspects of this subject (Nisonoff et nl., 1975; Eichmann, 1975; Makela and Karjalainen, 1977),this review will be confined to those aspects of idiotypy relevant to clonotype identification. 1 . Properties of Anti-idiotypic Antibodies Antisera capable of detecting idiotypic determinants can be raised in an animal ofa different species (xenogeneic anti-idiotype) in an animal of the same species (allogeneic anti-idiotype) or in the same strain (syngeneic anti-idiotype). Tmmrinization with a myeloma protein or a purified antibody of restricted heterogeneity will elicit a population of antibody molecules which will recognize isotypic and allotypic antigens on the inimunogen a s well a s variable region determinants. Therefore, anti-idiotypic sera must be absorbed with appropriate inyeloma proteins or with pooled Ig in order to ensure its specificity. I n the rabbit, allogeneic anti-idiotype can be raised in an allotypically matched partner, negating the necessity for such absorptions (Kindt, 1975).Removal of anti-allotypic and anti-isotypic antibodies does not guarantee that the anti-idiotypic serum will solely recognize the ho-

286

NOLAN H. SIGAL AND NORMAN R. KLINMAN

mologous antibody. For example, allogeneic anti-idiotypic serum made against TEPC 15 plasmacytoma protein cross-reacts with another PC-binding myeloma protein MOPC 511 (Sakato and Eisen, 1975; Lieberman and Potter, 1976), which differs from TEPC 15 in a number of amino acid residues in both H and L chains (Barstad et al., 1974; Hood et nl., 1977). Brient and Nisonoff (1970) studied anti-idiotypic antibodies directed to specifically purified rabbit anti-p-azobenzoate antibody and showed that the binding site of the antibody is part of, or close to, the major idiotypic determinant. The reactions of six different antiidiotypic antibodies with ‘”1-labeled F(ab), fragments of anti-azobenzoate antibody were significantly inhibited by the homologous or closely related haptens. There was a close correlation between the affinity of the hapten for the antibody and the hapten’s ability to inhibit the idiotype-anti-idiotype reaction. Hapten inhibition experiments have demonstrated the relationship between idiotypic determinants and the combining site in a number of anti-idiotypic systems of the mouse (Sirisinha and Eisen, 1971; Carson and Weigert, 1973; Claflin and Davie, 1974a). It should be emphasized that the region of the binding site is not necessarily the only idiotypic determinant on the antibody molecule. For example, the reaction between the allogeneic anti-idiotypic sera made against TEPC 15 and the myeloma protein is not inhibited by PC but is inhibited by PC coupled to a protein carrier (Claflin and Davie, 1974a; Gearhart et al., 1975a). These results indicate either that the anti-idiotype binds to the PC combining site with a very high affinity (Nisonoff and Bangasser, 1975) or that it is directed at variable-region framework determinants outside the combining site (Claflin and Davie, 1974a). Although the majority of anti-idiotypic antibodies are directed against the combining site and are specific for antigenic determinants that are unique to a closely related set of antibody molecules, other “idiotypic” systems have been described in the mouse that more closely resemble rabbit V-region allotypic markers. Claflin and Davie (1975) identified an antigenic determinant, called VH-PC,shared by seven mouse myeloma proteins that all bound PC but had widely differing amino acid sequences. This “idiotype” was present on anti-PC antibody from all mouse strains, but not on antibody or myeloma proteins that lacked specificity for PC. This suggests that the common VH-PCdeterminant is not associated with the combining site per se, since the sequences of the myeloma proteins varied significantly, but rather with the variable region’s capacity to bind PC. Bosma et al. (1977) prepared anti-idiotypic antibody against mouse myeloma

THE B-CELL CLONOTYPE REPERTOIRE

287

MOPC 173 and used the cross-reactivity of this anti-idiotypic serum with another myeloma UPC 10 in order to identify an H-chain marker which may be entirely analogous to the rabbit V-region allotype. The antigen was found in about 1% of total serum Ig and was present in many but not all mouse strains, and its expression was linked to the miirine H-chain allotype. In a number of investigations, H and L chains have been isolated and tested for their reactivity with anti-idiotypic antibody directed against the parent immunoglobulin. I n general, anti-idiotypic sera d o not react as well with the isolated polypeptide chains as with the native protein, but there are exceptions to this rule, and antisera can b e prepared that preferentially recognize one chain more than the other. Grey et al. (1965) originally showed that six human myeloma proteins lost their capacity to react with rabbit anti-idiotypic sera when separated into L and H chains. However, in five other proteins, all with A L-chains, idiotypic determinants could be localized to one of the two chains. Carson and Weigert (1973)demonstrated that the idiotypic determinants on 5558 mouse myeloma protein required both H and L chains and that recombination of the 5558 H chain with an L chain differing by only three amino acids caused the loss of the idiotype. In other murine systems anti-idiotypic sera have been prepared which were predominantly specific for either the H or L chain when the homologous chains were recombined with pooled serum Ig (Eichmann, 1977). Nevertheless, the fact that all idiotypic determinants studied as genetic markers in the mouse and rabbit have been linked to the H-chain allotype indicates that the H chain may have a predominant role in the recognition of the idiotype. Indeed, in order to study the inheritance of L-chain markers, investigators have had to make anti-L-chain idiotypic sera b y immunizing animals with purified L chains (Fraser et al., 1977). If the same idiotypic determinant is detectable in the immune serum of all individuals within an inbred mouse strain or among closely related members of a rabbit family, these idiotypes can be assumed to reflect germline antibody specificities and can be used as genetic markers for the variable region. At the present time, there are 14 such markers in the mouse (Blomberg et al., 1972; Herzenberg, 1973; Pawlak et al., 1973; Eichmann, 1973; Leiberman et al., 1974, 1976; Imanishi and Makela, 1974, 1975; McCarthy and Dutton, 1975; Makela et al., 1977a; Pisetsky et ul., 1977; Berek and Eichmann, 1977), and a number of rabbit idiotypes have also been identified (Kindt et al., 1973; Braun and Kelus, 1973; reviewed in Frazier et al., 1977). Most of the literature on inbred mice concerns instances in

288

NOLAN H. SIGAL AND NORMAN R. KLINMAN

which idiotypic determinants are found in all individuals of the strain, since these idiotypes can be used as genetic markers to investigate immunoglobulin variable-region diversity. Using fine-specificity analysis (Makela and Karjalainen, 1977) or anti-idiotypic sera (Kuettner et al., 1972; Ju et al., 1977), investigators have discovered antibody specificities that appear to be unique to the individual mouse in which they were identified. Such “minor” clonotypes, or private specificities, can be viewed either as somatic variants of germline antibody specificities or as H-L chain pairs that are encoded by germline genetic information, but are expressed at a substantially lower frequency than the major clonotypes. As pointed out in Section IV, while the major clonotypes that have been used as genetic markers are more easily identified and enumerated in the population, they may not be representative of the majority of specificities, and therefore, private idiotypic specificities must be identified as well.

2 . Zdiotype as a Method of Clonotype Zdentijication The previous discussion illustrated the great variety of idiotypic determinants that are recognized by antibodies directed against the variable region. The importance of studies on the properties of anti-idiotypic antibodies for clonotype identification is that any anti-idiotypic serum can be used as a reagent to dissect out a subset of clones, but the properties of the reagent should be known and its use defined within the experimental system under study. In particular, one must know whether the anti-idiotypic serum is recognizing one clonotype (a unique H-L pair), a small set of clonotypes, or a large group of antibodies that all fortuitously share a common determinant. The task of clonotype identification is further compounded by the fact that an anti-idiotypic serum is made up of a heterogeneous mixture of antibodies that may be directed against several different determinants on the same molecule. Thus, the detection of low levels of an idiotype in serum antibody may indicate a diverse population of weakly crossreactive antibodies, none of which are identical to the idiotype in question. Alternatively, the relevant idiotype might be present but represent only a small proportion of serum antibody. The following paragraphs will deal with the question of the number of clonotypes defined b y an anti-idiotypic serum. To summarize, the results are variable, but as a general rule anti-idiotypic antibodies recognize more than one clonotype, e.g., one H-chain sequence, paired with several different L chains or a small group of related specificities. An extreme

THE B-CELL CLONOTYPE REPERTOIRE

289

example of the latter group is the V,-PC marker (Claflin and Davie, 1975) and the idiotypic determinant identified by Bosma et aZ. (1977) referred to above. The allogeneic anti-idiotypic serum specific for TEPC 15 recognizes only one clonotype within the PC-specific repertoire, as demonstrated b y I E F studies (Gearhart et al., 1975b; Kluskens et al., 1975). However, analysis of anti-DNP monoclonal antibodies revealed a rare anti-DNP clonotype, which shared the TEPC 15 idiotype, perhaps owing to similar variable-region framework determinants between the anti-DNP and anti-PC clonotypes (Sigal, 1977a). Thus, within the precursor cells specific for PC, only one clonotype is identified, but it is clear that the same idiotypic determinants can appear on other H-L pairs. A similar situation exists for the rabbit anti-TEPC 15 idiotypic serum: analysis of serum antibody by fine specificity and IEF suggested that the one clonotype recognized by this antiserum was identical to TEPC 15 plasmacytoma protein (Claflin and Davie, 1974b; Claflin, 1976a,b), but examination of monoclonal anti-PC antibodies revealed a small group of clonotypes that possessed the rabbit antiTEPC 15 idiotypic determinants but did not share the allogeneic antiTEPC 15 determinants. Therefore, these clonotypes were not identical to TEPC 15 (Gearhart et al., 1977). Amino acid sequencing of anti-p-azophenylarsonate (Ars) antibody possessing a cross-reactive idiotype has shown that the anti-idiotype reacts with a number of different clonotypes (Capra et al., 1975,1977). The N-terminal50 residues of the H chain consisted of a single homogeneous sequence. On the other hand, while the L chains that possessed the idiotype were more restricted than pooled serum Ig, as many as four alternative amino acids were detected at several positions. The L chains belonged to at least three different variable-region subgroups, but the hypervariable regions of the chains had identical sequences. Thus, this anti-idiotypic serum appears to be specific for the antigen-combining site but the unique L-chain hypervariable regions can be accommodated in a large number of frameworks. IEF analyses of antibodies recognized by other anti-idiotypic sera suggest that these antisera recognize more than one clonotype. The anti-NP idiotype is composed of at least four spectrotypes (Makela and Karjalainen, 1977) and anti-ABA-HOP genetic marker is also characterized b y a unique I E F pattern (Makela et uZ., 1977b).This heterogeneity may b e due to a cluster of H-chain genes or one H chain paired with a number of different L chains. It is likely that other VH gene markers, such as the anti-extra-SRBC antigen response (McCarthy and

290

NOLAN H. SIGAL AND NORMAN R. KLINMAN

Dutton, 1975), and the staphylococcal nuclease idiotype (Pisetsky et

al., 1977) will be composed of more than one clonotype. 3. Molecular Basis of ldiotypy It is important for the purpose of clonotype identification to elucidate the molecular basis of idiotypic determinants, since all clonotype identification must eventually rest on sequence correlation. This subject has been reviewed by Capra and Kehoe (1975) and will be discussed only in its briefest form here. Amino acid sequences of the predominant idiotype associated with AIJ anti-Ars antibody, already mentioned above, revealed a single sequence in all H- and L-chain hypervariable regions and a number of different L-chain frameworks. Capra et al. (1977)have arrived at a similar conclusion in the study of two human IgM anti-y-globulin proteins LAY and POM, which share idiotypic determinants and binding specificity. The H-chain variable regions were remarkably similar, differing by only 8 amino acids among the 120 V-region positions. Of these 8 differences, only one position was in the antigen-binding site hypervariable regions. The Lchain sequences demonstrated identity in two of the three hypervariable regions, despite the fact that the two proteins belong to different V, subgroups. The discovery of myeloma protein CBPC2, a PC-binding protein induced in a CB20 mouse, has permitted sequence correlation in the TEPC 15 idiotype system (Claflin et al., 1975).Since the CB20 mouse strain is congenic to BALB/c but carries the C57BL H-chain allotype, CBPC2 was of the C57BL allotype and did not possess the idiotypic determinants recognized by the allogeneic anti-TEPC 15. This protein did have the binding site-associated idiotypic determinants of the rabbit anti-idiotype. Amino acid sequencing revealed that the CBPC2 L chain was identical to TEPC 15, and the H chain of CBPC2 differed from TEPC 15 in two nonhypervariable-region residues through the first 36 amino acids. Thus, the amino acid sequence correlates with retention of the combining-site idiotype but loss of the frameworkassociated idiotypic determinants. Vrana et al. (1977) have begun to explore the amino acid sequences of 12 BALB/c myeloma proteins that bind inulin and for which both private and cross-reactive idiotypic determinants are known, but the results are incomplete at this time. The recent work of Aasted and Kindt (1976a,b) is an excellent example of the use of idiotypic analysis and sequence to explore the struc-

THE B-CELL CLONOTYPE REPERTOIRE

29 1

bra1 and serological relationships among clonotypes. After hyperimmunization with streptococcal group C vaccine, 10 different clonotypes from one rabbit were isolated by immunoadsorbent chromatography, agarose block electrophoresis, and fractionation on an anti-a' allotype column. Anti-idiotypic sera were prepared against four of these clonotypes, and each antibody could b e characterized by (1) the extent of idiotypic cross-reactivity with these antisera; (2) L-chain N-terminal sequences; and ( 3 ) a' allotype subspecificity (see Table 111). This study provides another observation of idiotypic cross-reactivity among antibodies that differ in their H- and L-chain V-region frameworks, but more important, illustrates the utility of a multifactorial approach to clonotype identification.

D. SEQUENCE Although amino acid sequencing is the only definitive method of clonotype identification, there are few experimental systems where sufficient antibody can be obtained to provide meaningful information. The mouse and human myeloma proteins provide a pool of antibody molecules available in large quantities, and most sequence work to date has been done on these proteins. Obviously, determination of the amino acid sequence of a group of myeloma proteins in itself is not helpful in clonotype identification, but such information can be used to estimate the minimum number of H or L chains necessary to code for antibody diversity. In general, two approaches have been used to arrive at V-gene estimates; since both have been previously reviewed, they will be mentioned here only briefly. One approach (Smith et ul., 1971; Hood et n l . , 1974,1977) is to draw up a genealogic tree of all the sequences in a given family. T h e number of mutations needed to generate each sequence from the most similar common ancestor sequence is then used to deduce, by discounting parallel mutations, whether or not each sequence requires a separate gene. The second method (Cohn, 1971; Cohn et al., 1974) is based on the minimal hypothesis that a variable region can be divided into framework regions and complementarity-determining positions. The number of different framework sequences is taken to represent the minimum number of required germline genes, and the hypervariable region amino acid changes are assumed to be generated by somatic processes. For V, sequences, the latter method yields an estimate of 100-200 germline genes (Cohn et ul., 1974), whereas genealogic analysis suggests that, at the 90% confidence level, the V, pool size ranges between 700 and

TABLE I11 COhWAFUSON OF CHEMICAL, ALLOTYPIC, AND IDIOTYF'IC CHARACTEFUSTICS OF 10 ANTIBODY COMPONENTS PURIFIED FROM ANTI-GROUP C SERUMOF A RABBIT

Antibody 1 2 3 4 5 6 7a

7b 8 9

Immunoadsorbent fraction I I I1

I1

L-chain N-terminal sequence 1

2

3

4

D P V L

-v-vn

Blocked -V-M

% Idiotypic cross-reaction

5

6

7

8

9

T

Q

T

A P-

S

P-

111 111 111

111

IV IV

-V Blocked Blocked Not done Not done

PIA A

a' Subspecificity

2 1 1 2 1 1 1 3 1 1

% Inhibition

Ab2

Ah4

Ab7a

Ab7b

by hapten

100

6

-

-

-

-

-

21 43 20 44 -

-

-

-

12 22 71 29

100 -

-

21 100 125 -

10 28 100

THE B-CELL CLONOTYPE REPERTOIRE

293

10,000 genes (Hood et al., 1974). Similarly disparate estimates have been made for the mouse VH and Vk families. Because of the large amount of protein necessary for analysis, there have been relatively few sequence studies of induced antibodies. In the mouse, amino acid sequencing of anti-PC antibody from A/ J mice revealed a striking degree of similarity to PC-binding myelomas of BALB/c origin (Claflin and Rudikoff, 1976).When the sequence of A/ J H chains was compared to the H chains of TEPC 15, MOPC 511, and McPc 603, both the framework and first hypervariable regions were identical in all cases. Sequence analysis of the L chains through part of the first complementarity region revealed three chains, one similar to each of the myeloma proteins TEPC 15, McPc 603, and MOPC 511. Hyperimmunization of rabbits with pneumococcal vaccine or streptococcal group A carbohydrate can induce large quantities of homogeneous or restricted antibody that can be used for sequence analysis. Haber et al. (1977)have examined the H- and L-chain diversity in an outlired rabbit population by sequencing anti-S3 or -S8 pneumococcal antibodies. Of the 65 L chains sequenced through the first 21 residues, there were 45 unique sequences; among 8 complete V, seqiiences of anti-S3 antibodies, there were no identical sequences. In addition, three anti-S8 antibodies elicited in the same rabbit demonstrate differences in sequence and length in both the framework and the first hypervariable regions. Diversity was also found among VH compleinentarity-determiningsequences, but the framework regions were highly conserved, in contrast to V,. While one must remember that this study was done in an outbred population of rabbits, the data suggest an enormous diversity among V regions and imply that the predominant idiotypes of the mouse are not entirely representative of other clonotypes in the repertoire. In order to reduce the ambiguities present in an analysis of an outbred population, a number of rabbit antistreptococcal antibodies from a large, closely related rabbit family were isolated and sequenced (Braun et al., 1976a). Variability in the complementaritydetermining regions was reduced 2- to 6-fold in comparison with outbred rabbit L chains, with several examples of identical first and third hypervariable regions. Using the minimal hypothesis of Cohn (1971), Braun et nl. (19764 estimated that the rabbit has at least 27 VK germline genes. Significantly less diversity has been found in the sequence analysis of anti-DNP and anti-NIP antibodies in inbred guinea pigs (Cebra, 1977). Although the population of antibodies appeared grossly heterogeneous b y IEF and affinity analysis, even those differing in affinity

294

NOLAN H. SIGAL AND NORMAN R. KLINMAN

shared H-chain sequences through the first two hypervariable regions, differing only in the third VHregion.

E. OTHERPROBES Two other approaches for evaluating clonotype expression should be mentioned here. The first is the use of chain recombinational analysis as a means of assessing polypeptide chain identities. In 1971, it was demonstrated by Bridges and Little (1971) for myeloma proteins and b y Klinman (1971a) for monoclonal antibodies that antigen-binding specificity was dependent on the pairing of homologous H and L chains. Thus, MOPC 315 and MOPC 460 H chains would bind DNP only when recombined with their own L chains, and monoclonal antiDNP antibodies lost much of their binding activity if recombined with their own L chains in the presence of an excess of heterologous L chains. These findings have been confirmed by studies that have assessed the capacity of recombined H and L chains to bind anti-idiotype antibody (Sher et al., 1971) and homogeneous antibodies (Montgomery et al., 1975a). A second probe for the expression of a given H or L chain can be thought of as the biological equivalent of homologous chain pairing. Cohn et al. (1974) demonstrated that the presence of the A, L-chain sequence in antidextran antibody was dependent on the presence of a specific H chain associated with the a' allotype. Conversely, strains expressing low levels of A, had a low level of that H chain in their antidextran response. Thus, the presence of a given H or L chain in the repertoire may act as a probe for the potential to express associated chains. Recently, Nisonoff (1977) has reported that the expression of the cross-reactive anti-Ars idiotype is dependent not only on the presence of the a4 H-H-chain allotype, but on the presence of the L-chain allotype common to most murine strains. Mice expressing PL/J L chains that differ in isofocusing (Gibson, 1976) and peptide mapping (Gottlieb and Durda, 1977) do not express the cross-reactive idiotype even when they contain the appropriate a4 H chain. IV. Defining the Adult B-Cell Repertoire

Having described the methods used for clonotype identification, we will now deal with the application of these techniques to an understanding of the number of different antibody specificities an individual or species can make. At any given time, this would be a direct function of the number of B cells representative of each clonotype (the repeat frequency) and the total number of potentially responsive cells

THE B-CELL CLONOTYPE REPERTOIRE

295

within the population. Determination of the number of clonotypes within the adult repertoire, therefore, may, in part, be derived from an analysis of the frequency of cells specific for a given antigen and identification of the number of clonotypes that respond to the antigen. While the majority of ABC analyses and limiting-dilution studies have accomplished the first task, there are relatively few experiments where both requirements are satisfied. Nevertheless, such investigations are important both in an historical sense, since they provided the initial estimates of repertoire diversity, and because they have supplied a framework around which more definitive studies have been performed. Any analysis of the clonotype repertoire must take into account several fundamental problems: (1) I n nonimmune animals, most clonotypes are represented by only a small proportion of the B-cell population, making identification both difficult and laborious. (2) Repeat frequencies are readily obtained only for “dominant” clonotypes, i.e., those clonotypes represented b y myeloma proteins with antibody specificity, or clonotypes invariably expressed in large amounts in serum as the major antibody responding to a given antigen; while these clonotypes are more amenable to identification, they may not be truly representative of the primary B-cell repertoire. ( 3 )The relationship of the repertoire expressed by an individual at a given point in time to the total potential of the individual (or inbred strain) is unknown; an individual may never express the total potential, or only part of this potential may b e expressed at any point in time and be subject to clonal flux rates and hence varying expression. (4)The effect of antigen contact and environmental influences on the potential extensiveness of the repertoire is unknown. Thus, with these limitations in mind, we review in this section the experiments that have attempted to define the adult B-cell repertoire.

A. ANTIGEN-BINDING CELL ANALYSIS

The first attempts to define the adult B-cell repertoire employed the binding of radiolabeled antigens to lymphocytes in uitro. Beginning with Naor and Sulitzeanu (1967), the frequencies of cells that bind numerous antigens, including autologous antigens such as murine growth hormone (Unanue, 1971), have been enumerated. The frequencies of ABCs and rosette-forming cells for a number of antigens are listed in Tables IV and V, respectively. As noted in Section 11, not all ABCs may be relevant as antibody-forming cell precursors. Thus, these frequencies should be taken only as a relative measure of the diversity within the repertoire.

z

E

TABLE IV FREQUENCY OF ANTIGEN-BINDING CELLS IN LYM~HOID TISSUE Antigens" DNP-Hy DNP-HGG DNP-Lys-Tn DNP-GPA DNP-fiber DNP-MGG DNP-gelatin DNP-MGG HSA BSA HY HY HY HY HY HY TGAL TGAL

Method" 1251 1251 1251 1251

Fiber-binding =51

Affinity column FACS 1251

1251 =51 1251

125.1 1251

F1 FACS 1251 1251

Speciesltissue

ABC/103 cells

F Y 111

Reference

Mouse spleen Mouse spleen Mouse spleen Guinea pig spleen Mouse spleen Mouse spleen Mouse spleen

10 9.9 0.9 0.3 10-20 1.0 5

Rolley and Marchalonis, 1972 Lawrence et al., 1973 Lawrence et al., 1973 Davie and Paul, 1971 Rutishauser et a1 ., 1972 Klinman et al., 1976 Haas and Layton, 1975

Mouse spleen Mouse spleen Mouse spleen Mouse spleen Human PBL Mouse spleen Rat spleen Mouse spleen Mouse spleen Mouse spleen Mouse lymph node

12 0.7 0.5 0.9 0.3 1.4 0.8 7.9 5.0 2.8 4.8

Julius et al., 1976 Sulitzeanu and Naor, 1969 Naor and Sulitzeanu, 1967 Byrt and Ada, 1969 Dwyer and MacKay, 1972a Humphrey and Keller, 1971 Cooper et al., 1972 DeLuca et al., 197% Julius et al., 1976 Humphrey and Keller, 1971 Hammerling and McDevitt, 1974

Flagellin Flagellin Flagellin Flagellin P-gal P-gal P-gal P-gal

NIP lac lac MGH thyro Lysozyme Myelin HRP GO

1251 12.51

3H 1251

Enzyme Enzyme Enzyme F1 Affinity column Affinity column Fiber-binding 1251 1251

Fiber-binding 1251

F1 F1

Human PBL Mouse spleen Mouse spleen Mouse spleen Mouse spleen Mouse spleen Mouse spleen Mouse spleen Mouse spleen

5.0 0.2 0.02 0.04 0.02 0.02 20 13 1

Mouse spleen

0.03

Henry et d . , 1972

Mouse spleen Mouse spleen Human PBL Mouse spleen Guinea pig spleen Mouse spleen Mouse spleen

0.3 0.09 0.2 1.9 0.7 5.6 9.7

D’Eustachio and Edelman, 1975 Unanue, 1971 Bankhurst et al., 1973 D’Eustachio and Edelman, 1975 Coates and Lennon, 1973 DeLuca et al., 197% DeLuca et al., 197%

Dwyer and MacKay, 1972a Ada, 1970 Diener and Paetkau, 1972 Raffet al., 1973 Rotman and Cox, 1971 Modabber et a1 ., 1970 DeLuca et al., 1974 DeLuca et al., 197% Haas and Layton, 1975

4

3

tp n M

r)

r

2 0

rl

4_. M

Abbreviations: BSA, bovine serum albumin; DNP, dinitrophenyl; FACS, fluorescence-activated cell sorter; F1, fluorescein; P-gal, cb P-galactosidase; GO, glucose oxidase; GPA, guinea pig albumin; HGG, human y-globulin; HRP, horseradish peroxidase; HSA, human M serum albumin; Hy, hemocyanin; lac, azophenyllactosidase; MGG, mouse y-globulin; MGH, murine growth hormone; NIP, 4-hydroxy9 5-iodo-3-nitrophenacetyl;PBL, peripheral blood lymphocytes; TGAL, poly-L-(tyrosine, glutamic acid)-poly-DL-alanine-poly-L-lysine; thyro, thyroglobulin.

298

NOLAN H . SIGAL AND NORMAN R. KLINMAN TABLE V FREQUENCY OF ROSETTE-FORMING CELLSIN LYMPHOIDTISSUE

Antigen"

Speciesitissue

SRBC SRBC SRBC SRBC SRBC NNP HGG LPS CGG

Mouse spleen Mouse spleen Mouse spleen Mouse spleen Mouse spleen Mouse spleen Mouse spleen Mouse spleen Mouse spleen Mouse spleen

SIII

RFC/103 cells 1.1 0.4-0.9 0.7-0.9 0.7 2.7

0.6 0.2 0.1 0.07 0.04

Reference Bach, 1973 Decreuseford et al., 1970 Greaves and Moller, 1970 Greaves and Hogg, 1971 Haskill and Axelrod, 1971 Moller and Sjoberg, 1972 Stout and Johnson, 1972 Sjoberg and Moller, 1970 Bankhurst and Wilson, 1971 Howard et al., 1969

Abbreviations: CGG, chicken y-globulin; HGG, human y-globulin; LPS, lipopolysaccharide; NNP, 4-hydroxy-3,5-dinitrophenacetyl; SIII, pneumococcal polysaccharide SIII; SRBC, sheep erythrocytes.

B. TECHNIQUESTO ESTIMATE PRECURSOR FREQUENCIES 1 . In Vivo Splenic-Focus Technique Initial attempts to obtain an assay for antibody-forming cell precursors centered on the in vivo splenic-focus technique (Kennedy et al., 1965; Playfair et aE., 1965; Trentin et al., 1967).This experimental system involved the injection of a small inoculum of donor spleen cells and antigen into lethally irradiated syngeneic mice and, after several days, cutting the recipient spleen into a number of tiny fragments or slices. The pieces were placed on petri dishes containing a layer of agar with embedded red cells. Since antibody diffusing out of the fragments produced complement-mediated lysis, the number of positive foci, each presumably containing a clone of antibody-forming cells, could be enumerated and correlated with the number of cells injected. These studies utilized spleens from unimmunized mice, and therefore the possibility that either the B cells or the T cells were the limiting cell type in this situation made interpretation of these experiments difficult. A number of studies (Playfair et al., 1965; Kennedy et al., 1965; Miller et al., 1967; Gregory and Lajtha, 1968) demonstrated that the number of hemolytic foci in a recipient spleen was directly proportional to the number of cells injected, suggesting that only one cell was limiting. Experiments from other laboratories (Vann and Campbell, 1970; Luzzati et al., 1970) indicated that in some cases, antibody from more than one B cell could be found in a splenic focus.

THE B-CELL CLONOTYPE REPERTOIRE

299

Vann and Campbell (1970)transferred an allogeneic mixture of donor cells to the recipient and showed that some foci contained PFCs of both H-2 types. Luzatti et al. (1970) demonstrated that a single focus can synthesize antibodies that migrate as more than one discrete electrophoretic band, implying that there was more than one B-cell clone per focus. Thus, the frequencies derived from these limiting-dilution experiments are faced with the ambiguity of whether the frequency of specific T cells or B cells or a combination of both had been enumerated.

2. Limiting-Dilution Experiments Cell-transfer limiting-dilution systems all have in common the transfer of a small number of cells to syngeneic irradiated recipients followed by immunization and assay of the antibody titer or splenic PFCs for an “all-or-none” response. The experiments assume that only one donor cell specific for the antigen is transferred to the recipient, and, if the B cell is limiting, it can expand sufficiently to yield detectable serum antibody levels or a large number of PFCs, all synthesizing the same antibody specificity. If one plots the number of responding units versus the number of cells injected, the curve will follow a Poisson distribution and the frequency of responding cells can be calculated. Brown et al. (1966) were the first to use this technique and enumerated the frequency of mouse spleen cells specific for rat erythrocytes. As with the studies using the splenic-focus technique, these experiments were done before the interaction between B and T cells in antibody formation was appreciated and, therefore, are difficult to interpret. This difficulty was overcome by Cudkowicz et al. (1969, 1970) and Miller and Cudkowicz (1970), who transferred a graded number of bone marrow cells as a source of B cells and a large constant number of thymus cells and measured the all-or-none P F C response in the spleen. Although data from the initial studies (Cudkowicz et al., 1969; Miller and Cudkowicz, 1970) deviated from the theoretical Poisson curve, later studies (Cudkowicz et al., 1970)were used to estimate the frequency of precursor cells specific for SRBCs. By injecting a limiting dilution of thymus cells with a constant number of bone marrow cells, this system has also been used to calculate the frequency of T cells specific for an antigen (Shearer et al., 1969; Shearer and Cudkowicz, 1969). Shearer et al. (1973) have also compared the frequency of T and B cells responsive to a number of synthetic polypeptides in high- and low-responder mouse strains. An alternative approach to ensure that the B cell was the limiting

300

NOLAN H. SIGAL AND NORMAN R. KLINMAN

precursor cell was employed b y Moller and Michael (1971). They transferred graded numbers of spleen cells into irradiated mice along with lipopolysaccharide (LPS), which is a thymus-independent antigen, and found that the fraction of nonresponding mice, as assayed b y anti-LPS PFCs in a semilog plot, fit a straight line. Since LPS is a nonspecific B-cell mitogen, the possibility exists that not all the precursor cells stimulated under these conditions would have been triggered under more physiological circumstances. Bosma and Weiler (1970)estimated the frequency of precursor cells responsive to poly-D-alanine in the spleens of (C,H x C57BL)F1 mice. Following a series of immunizations with poly-D-alanine-BSA and polyalanine-ovalbumin, the allotype of the PFC was determined in mice having positive antibody titers. Although nonimmune cells were transferred and an excess of T cells was not provided, the demonstration that all PFCs were of one allotype (Bosma and Weiler, 1970) and that the antibody produced had homogeneous binding characteristics (Bosma et al., 1972),as well as the conformation of the response to Poisson statistics, confinned that only the B cell was limiting in these experiments. The combination of isoelectric focusing (IEF) with transfer of a limiting number of cells to an irradiated recipient has provided a powerful tool to examine the B-cell repertoire. As discussed in Section IV,C these techniques have been utilized to analyze the number of clonotypes responsive to 4-hydroxy-5-iodo-3-nitrophenacetyl (NIP), DNP, and P-galactosidase. The general principle involved in these studies was the use of limiting-cell transfer to isolate a small number of hapten-specific precursor cells in the recipient mouse so that these clonotypes could be analyzed by IEF. In addition, Kohler (1976) has applied Poisson statistics to estimate the frequency of precursor cells responsive to P-galactosidase in nonimmune and immunized mice. Since a small percentage of mice who respond in an all-or-none fashion will receive more than one specific precursor cell, the analysis of responding mice by I E F has the ability to identify those with two clonotypes and correct the precursor frequencies accordingly. A difficulty arising in the use of limiting-dilution cell transfer studies for the estimation of the B-cell repertoire is the relationship between the experimentally derived frequencies and the “cloning efficiency” of this process. One must determine the percentage of the injected cells that home to the spleen (if an all-or-none response is assayed by splenic PFCs) and are stimulated to produce a clone of antibody-forming cells. If this number can be found, the frequency of precursor cells calculated from the Poisson distribution can be cor-

THE B-CELL CLONOTYPE REPERTOIRE

30 1

rected for the cloning efficiency. Estimates of the efficiency for the techniques discussed thus far range from 4% (Playfair et al., 1965) to 10% (Bosma et al., 1968; Miller et al., 1967).

3. Microcul ture Systems Lefkovits (1972) developed a micro method for the culture of small numbers of lymphoid cells in vitro. The all-or-none response of the microcultures to the stimulating antigen was determined by assaying an aliquot of culture fluid on an agar plate embedded with red cells and observing the complement-mediated erythrocyte lysis. Although initial experiments (Lefkovits, 1972) did not ensure that only the B cells were limiting, Quintans and Lefkovits (1973) estimated the frequency of spleen cells specific for SRBCs b y an analysis of graded numbers of spleen cells from nude mice with an excess of allogeneic T cells. Cell density was kept constant by adding graded numbers of irradiated spleen cells along with the immunocompetent nude spleen cells to be assayed. It can be argued that only the B cell is limiting in these cultures since linear dose-response kinetics can b e demonstrated (Quintans and Lefkovits, 1973) and the antibody product appeared homogeneous by IEF (Luzzati et al., 1973a,b). For SRBCs, precursor frequency estimates range from 1 to 3 per lo5 spleen cells (Quintans and Lefkovits, 1973) to 1 per lo4 cells when pokeweed mitogen is added to the cultures (Quintans and Lefkovits, 1974a,b). l t is unclear whether this variation reflects suboptimal stimulation with antigen alone or whether addition of pokeweed mitogen results in antibody production by precursor cells that would not be stimulated under more physiological conditions. LPS added to the microcultures nonspecifically activates precursor cells to antibody production in the absence of antigen and without significant cell proliferation (Quintans and Lefkovits, 1 9 7 4 ~ )In . Section IV,C,3 we will discuss the use of this microculture system in the analysis of the PC-specific repertoire and the streptococcal polysaccharide repertoire. Recently, other methods have been described for isolation and characterization of a single B cell i n vitro. Clones of cells arising from single precursors could be maintained for weeks in mitogen-stimulated suspension cultures (Anderson et a l . , 1977). B cells taken from spleen, lymph node, bone marrow, or thoracic duct of normal or nude mice were stimulated by LPS, purified protein derivative of Mycobacterium tuberculosis (PPD), dextran sulfate, or lipoprotein. While cells cultured at high density allowed cell-to-cell contact and led to maturation and secretion of antibody, the same cells cultured at low density favored growth and expansion of the clones. This latter technique has

302

NOLAN H . SIGAL AND NORMAN R. KLINMAN

the advantage of propagating clones of cells of known specificity for long periods of time, eventually collecting large quantities of antibody for analysis. Kincaide and Ralph (1977)have shown that clonal proliferation of B Iymphocytes in semisolid gel cultures is dependent on one or more polysaccharide mitogens present in laboratory grade agar. When LPS or dextran sulfate was also added, the cloning of B cells is optimized and is not dependent on the presence of T cells. The investigators have utilized this system to investigate the effect of antibodies to cell surface antigens on B-cell activation. Nossal et al. (1977) have determined the precursor frequency of NIP-specific B cells in adult mice (1/15,000 B cells) by stimulating microcultures with NIP-LPS, staining each culture with NIP-rhodamine after 7 days, and looking for rhodamine-stained cells. 4 . In Vitro Splenic-Focus Technique The in vitro splenic-focus technique developed by Klinman and his co-workers (Klinman, 1969; Klinman and Aschinazi, 1971; Klinman, 1972; reviewed in Klinman and Press, 1975a) has been used to isolate individual B cells in culture and analyze the antibody product of the resulting clones. Limiting numbers of donor spleen cells from neonatal or nonimmune or immune adult mice are injected intravenously into lethally irradiated, syngeneic, carrier-primed recipients. Recipient spleens are removed 16 hours after cell transfer, sliced into 1-mm cubes, and placed in vitro in organ fragment culture. Since 5% of the donor cell inoculum is present in the fragments at the time of in vitro stimulation (Press and Klinman, 1973b), and one spleen yields approximately 50 fragments, a cell transfer of 1 X lo6 B cells would result in 103 donor cells per fragment. The number of donor cells injected is adjusted so that there will be 0 or 1antigen-specific B cell per fragment. Antigenic stimulation of the donor B cells occurs in vitro. Culture fluids are changed every 2-3 days, and 9-14 days after antigenic stimulation, clones are assayed for the production of antibody by a sensitive radioimmunoassay. Transfer of graded numbers of cells produced a linear increase in the number of responding foci, suggesting that only the B cell was limiting in this system. The donor origin of the antibody-forming cells (reviewed in Klinman and Press, 1975a) and the monoclonal nature of the response (see Section 11)have been exhaustively demonstrated. The splenic-focus system has a number of advantages over other experimental techniques described. First, one can determine the frequency of cells responsive to an antigen and simultaneously obtain

THE B-CELL CLONOTYF’E REPERTOIRE

303

sufficient quantities of homogeneous antibody to do subsequent assays for isotype, idiotype, isoelectric point, or fine-specificity analysis. Thus, the frequency of an identifiable clonotype can be directly determined. Second, the presence of excess carrier-primed T cells in vitro appears to maximize B-cell responsiveness, even of neonatal and fetal cells, so that the difficulties encountered by other splenic-focus techniques in analyzing nonimmune cell populations have been eliminated. Third, B-cell stimulation in fragment culture closely simulates in vivo conditions, particularly in terms of the specificity requisites of B-cell triggering. Since B-cell clones detected in the splenic-focus system presumably represent those precursor cells which would have been stimulated in vivo, the ambiguity of whether cells stimulated with nonspecific mitogens in other limiting-dilution systems are relevant precursor cells does not arise. Finally, the splenic-focus technique has an advantage over in vivo cloning systems in isoiating the individual B cell, thus eliminating problems created by antigen selection or “clonal dominance.” As discussed previously, it is critical to determine the cloning efficiency of the in vitro splenic-focus technique in order to: (1)establish that subsets of B cells are not systematically excluded from analysis and (2) provide an absolute measurement of the B-cell repertoire. The evidence that the use of carrier-primed recipients maximized the stimulation of all precursor cells has been reviewed (Klinman and Press, 1975a). In addition, the absolute cloning efficiency for this system has been established using two different experimental approaches (Klinman et al., 1976). The cloning efficiency was experimentally determined by measuring the increment in ABCs in normal and immunized donor spleens, and comparing this increment to that obtained from in vitro splenic foci analyses. Enumeration of DNPspecific primary and secondary precursor cells by the in vitro splenicfocus technique demonstrated an increment in the frequency of such cells, which was consistently 3.9% of the increment in the frequency of ABCs. Thus, the efficiency of the in vitro splenic-focus technique for the enumeration of splenic clonal precursor cells is 3.9%. This cloning efficiency is higher than the 1%figure calculated by Askonas et al. (1972) for in vivo transfers and is probably d u e to the use of carrier-primed recipients to maximize stimulation. This efficiency was corroborated by comparing the frequency of cells that bind PC to the number of precursor cells specific for PC in the spenic-focus technique. The number of precursor cells detected represented 3.7-4.3% of the PC-binding cells in spleen cell preparations from both neonatal

304

NOLAN H. SIGAL AND NORMAN R. KLINMAN

TABLE VI FREQUENCY OF SPLENIC B CELLS SPECIFIC FOR HAPTENIC DETERMINANTS IN NONIMMUNE ADULTMICE ~~

Mouse strains

BALBIc

NJ C3H CBA AKR C57BLI6

DNP

TNP

F1

Ars

20" 18 21 16 19 15

20 17 -

13 -

-

-

1.5 2.5 -

-

-

NIP

PC

Dansyl

14

2.3 0.85 1.3

9.4 -

0.94 0.3

-

11 -

-

-

" Numbers are presented as frequency per lo5 B cells (Ig-bearing cells).

and adult mice. Since 5% of donor cells from neonatal or adult mice lodge in recipient fragments, the cloning efficiency of 4% derived from the two experiments indicated that 80% of precursor cells that lodge in the fragments are stimulated to detectable antibody-forming cell clones. Therefore, it is unlikely that major B-cell subsets are being ignored in this analysis. Using this cloning efficiency, the splenic-focus system has been employed to estimate the frequency of precursor cells specific for DNP, TNP, F1, NIP, PC, 5-dimethylaminonaphthalene-l-sulfonyl (dansyl), and p-azophenylarsonate (Ars) in nonimmune BALB/c mice (Klinman, 1972; Press and Klinman, 1974; Sigal et ul., 1975; Klinman et al., 1976; Sigal, 1977b). For some of these antigens, precursor frequencies have been established in AIJ, CBA, AKR, C57BL/6, and C3H mice and in bone marrow, lymph node, blood, and Peyer's patches (Klinman et al., 1976; Cebra et al., 1977; Gearhart et al., 1977; Pickard and Klinman, 1978). The frequency of splenic B cells for these antigens are listed in Table VI. It should be noted that precursor frequencies in adult germ-free mice are identical to those in conventionally reared donors for DNP, F1, and PC (Press and Klinman, 1974; Sigal et ul., 1975). These precursor frequencies, along with the other limiting-dilution studies discussed in this section, suggest that the B-cell repertoire is extremely diverse. However, they do not approach the heart of the matter, since individual clonotypes were not identified and enumerated and, therefore, the true diversity can only be approximated. The following section will discuss studies that combine clonotype identification and precursor frequency analysis in order to arrive at an estimate of the B-cell repertoire.

THE B-CELL CLONOTYPE REPERTOIRE

305

C. ESTIMATES OF THE ADULTB-CELL REPERTOIRE 1 . IEF Studies Kreth and Williamson (1973) estimated the number of NIP-specific clonotypes that inbred CBAlH mice could make. Female CBMH mice were immunized with a single dose of NIP-bovine y-globulin (BGG), and at various times after immunization, a small number of spleen cells from the donors were injected into lethally irradiated recipients. The recipients were immunized with the homologous antigen and bled 10-12 days after transfer; the sera were analyzed by IEF. The donor-cell inoculum was controlled to yield 0,1, or 2 monoclonal antibodies in the recipient sera. Although they did not use the data to calculate precursor frequencies, Kreth and Williamson (1973)found that the distribution of mice that had 0-3 clones followed Poisson statistics. The investigators did show that the number of NIP-specific precursor cells increased significantly from 7 to 28 days after priming and that the precursor pool remained constant in size for approximately the next 100 days. There also appeared to be preferential expansion of certain clonotypes early after immunization with diversity being greatest (fewest repeats among donors) around 100 days after priming. Their ability to use a limiting-dilution spleen cell-transfer system without the addition of excess T cells presumably stems from the utilization of a primed cell population that contained a large number of carrier-primed T cells as well as secondary B cells. Kreth and Williamson’s most important contribution was the detailed I E F analysis of the clones from four donors. Ten weeks after immunization, all spleen cells from the four CBA/H mice were transferred along with NIP-BGG to approximately 400 irradiated recipients, and the IEF spectra were compared in order to detect monoclonal antibodies that were identical between two of the donors. Of the 337 antibodies analyzed, five pairs were found to be identical (Table VII); in no case did three mice share the same clone. Using a statistical model (Wybrow and Berryman, 1973),the most likely pool size of NIP-specific clonotypes was estimated to be about 8000 with 95% confidence that the anti-NIP repertoire is between 3000 and 30,000 different specificities. Pink and Askonas (1974) employed a similar system to assess the number of clonotypes responsive to cross-reactive nitrophenyl haptens. They primed a group of mice with DNP-BGG and, after transfer of a limiting number of the DNP-primed spleen cells to irradiated recipients, boosted the recipients with TNP-ovalbumin. Therefore, the subset of DNP-specific clonotypes that could be stimulated by TNP at

TABLE VII FREQUENCIES OF DEFINABLE MINOR CLONOTWES

Mouse strains

Estimated number of responsive clonotypes

Maximum average frequency of each clonotype (per 10' B cells)"

NIP

CBNH

8000

0.25

DNP

BALBlc

1600

2

P-ml (MT) PC (non TEPC 15) PR8 HA

BALBlc BALBlc BALBlc

1200 > 57 > 50

0.5 0.5 1

Antigen

?

E Reference Kreth and Williamson, 1973 Pink and Askonas, 1974 Kohler, 1976 Sigal et al., 1977a Cancro et al., 1977h

These have been calculated by the authors of this review according to the formula on p. 308. The numbers represent a maximum estimate of the representation of each clonotype in the primary murine B-cell repertoire assuming equal representation of each clonotype.

2 U

z

F

0

z

?j

' z

THE B-CELL CLONOTYPE REPERTOIRE

307

the secondary B-cell level was expanded in the recipients and analyzed by I E F . Instead of examining a large number of clones from a few donors, these investigators compared 140 antisera from 255 different CBA and C3H mice. Since between 5 and 7 identities were found in this pool, the number of anti-DNP clonotypes that could respond to secondary challenges with TNP was estimated to b e approximately 500. In order to estimate the total number of clonotypes responsive to DNP from the number of cross-reactive specificities, one can utilize the demonstration that overlap stimulation by TNP of DNP-primed cells is approximately 30% (Klinman et al., 1973).This would indicate a minimum estimate of 1600 DNP-specific clonotypes in these strains. Although indirect, and based on assumptions that could make this estimate either too high or too low, the studies of Kreth and Wiliamson (1973) and Pink and Askonas (1974) provide two of the few analyses of the repertoire for antigens, such as DNP or NIP, which characteristically give a heterogeneous response. It is possible that these experiments overestimated the NIP-specific repertoire either by (a) failing to account for identities obscured by IgG subclass differences, or (b) missing clonal identities due to variations in isoelectric spectra caused by postsynthetic modification. However, other difficulties may skew their conclusions in the opposite direction. First, the clones analyzed were selected, being secondary and large, and may not have been truly representative of the entire hapten-specific precursor pool. Second, the analysis assumed that 5 x lo4 different clonotypes could be distinguished by the slab-gel IEF technique (Williamson et d., 1973), but recent studies focusing myeloma proteins known to have multiple amino acid sequence differences reveal that the technique may not have the discriminatory ability once assumed (Pink and Skvaril, 1975). Finally, these investigators assumed that all clonotypes occur with the same repeat frequency; in fact, if only the largest clonotypes were identified as repeats, then the conclusion that there are approximately 10,000 clonotypes would be a gross underestimate. While these I E F analyses dramatically illustrate the enormous diversity of the immune system, they do not provide an estimate of the B-cell repertoire. However, given a frequency of cells specific for these haptens, one can derive an estimate of the entire clonotype pool for the species. Between 1/7000 B cells (Press and Klinman, 1974) and 1/15,000 B cells (Nossal et aZ., 1977) is specific for NIP and 1/5000 B cells is specific for DNP (Klinman, 1972). Since there are approximately 1600-8000 clonotypes responsive to these haptens and assuming that these clonotypes are representative of the repertoire as a whole, then there are (-5000 clonotypeshapten) x (5000-

308

NOLAN H . SIGAL AND NORMAN R. KLINMAN

15,000) = 2.5 to 7.5 x lo7clonotypes. With 2 to 3 x loRB cells in the lymphoid system of a mouse (Makinodan et al., 1962; Warner, 1974) and again assuming that each clonotype is represented by the same number of cells, each clonotype, on the average, is represented by approximately (2-3 x 108)/(2.5-7.5x lo7)= 3-12 B cells per mouse. Kohler (1976) has arrived at a similar estimate of the B-cell repertoire in his analysis of the frequency of precursor cells specific for the enzyme P-galactosidase. A method was developed to detect anti-P-galactosidase antibodies after I E F in thin-layer polyacrylamide gels. By use of the wild type (WT) enzyme to catalyze a staining reaction, all antibodies against P-galactosidase were detected, while a subset of antibodies able to activate a mutant (MT) enzyme was detected by staining with that enzyme (Kohler and Melchers, 1972). Kohler (1976) transferred limiting dilutions of P-galactosidase primed or unprimed spleen cells of BALB/c mice together with antigen into sublethally irradiated syngeneic recipients. It was found that the uncorrected frequency of WT-specific precursor cells was 1/420,000 in the primed mice and 1/930,000 cells in the unprimed mice. The subset of B cells responsive to the M T P-galactosidase constituted one-fourth to onefifth of the WT populations. Using an estimated cloning efficiency of 3% and the assumption that one-third of the spleen cells are B cells, the frequency of precursor cells for antibodies activating the MT enzyme was 1/15,000 in the primed and 1/50,000 in the unprimed spleen. Then, Kohler compared 43 anti-MT clones from one primed mouse with 27 clonotypes from another donor and found that only one clone appeared to be shared by both mice. A number of important conclusions could be drawn from this work: (1) since a single animal can make approximately 27-43 different antiM T P-galactosidase antibodies, the maximum frequency of any one of these clonotypes in an individual unprimed mouse would be (27-43) x (50,000) = one precursor in 1.4 to 2.2 x 106 splenic B cells; (2) several clonotypes were identified a number of times within the precursor pool of an individual mouse, while the majority of I E F patterns did not repeat; this implies that clones can vary in size, in this case ranging from 100 to 800 B cells per mouse; (3)the identification of one MT-specific clonotype that was shared between the two donors suggests that the repertoire of the BALB/c strain consists of approximately 1200 different mutant enzyme-activating antibodies; (4) therefore, the total repertoire for the strain is around 5 x 10' antibody specificities, if the same calculations and assumptions employed in the discussion of the Kreth and Williamson data are made. It should b e noted that the criticisms noted for the experiments of Kreth and Williamson (1973)

THE B-CELL CLONOTYPE REPERTOIRE

309

and Pink and Askonas (1974) apply to these studies as well, and thus the estimate of 1200 MT-specific clonotypes is most likely to be minimum estimate. The implications of these findings, most notably the tremendous diversity of the repertoire, the variation in the size of clones, and the possible discrepancy between the size of the individual’s repertoire and that of the strain, have been recently discussed (Sigal et al., 1977b).

2. Splenic-Focus Technique Experiments designed to investigate and enumerate clonotypes using the splenic-focus technique have, for the most part, centered on the analysis of specificities within the PC-specific repertoire. (I. Background. The immune response to PC is restricted in terms of its clonotype repertoire (Sher and Cohn, 1972; Cosenza and Kohler, 1972a; Gearhartet d., 1975a),and this hapten elicits an easily identifiable idiotype (Cosenza and Kohler, 1972a,b; Sher and Cohn, 1972; Claflin et al., 1974a,b; Lieberman et al., 1974). Twelve naturally occurring plasmacytoma proteins with binding specificity for PC have been found in BALBIc mice (Cohn et nl., 1969; Potter and Lieberman, 1970); eight of these are idiotypically identical (Liebeman et nl., 1974) and have the same binding affinity for PC (Leon and Young, 1971; Sher et nl., 1971).The remaining three proteins have non-crossreacting idiotypic determinants (Potter and Lieberman, 1970)and distinct binding affinities for PC (Leon and Young, 1971). Of the nine myeloma H chains sequenced completely, four are identical (TEPC 15, S107, S63 and Y5236) and one (HOPC 8) differs by only one amino acid (Barstad et al., 1974; Hood et al., 1977). The finding of a large number of identical plasmacytomas implies that this antibody specificity must be present in every individual of the strain and coded for b y a germline variable-region gene. A number of investigators (Cosenza and Kohler, 1972a,b; Sher and Cohn, 1972; Claflin, 1974a,b) have shown that clonal precursor cells and PFCs share TEPC 15 idiotypic determinants, and Lieberman et al. (1974) demonstrated that all unimmunized, conventionally reared BALB/c mice possess immunoglobulin with the TEPC 15 idiotype in their sera; germ-free mice, on the other hand, have no natural levels of TEPC 15.The TEPC 15 idiotype segregates with the BALB/c H-chain allotype locus, being inherited as a simple Mendelian codominant gene, as shown by studies with recombinant inbred mice, congenic strains, and F, hybrid mice (Lieberman et al., 1974). There is a considerable body of evidence in the literature that the in zjiuo immune response of BALB/c mice to P C is monoclonal, com-

310

NOLAN H. SIGAL AND NORMAN R. KLINMAN

posed only of the TEPC 15 idiotype. Anti-TEPC 15 serum could specifically suppress both the in uiuo and in uitro immune response to PC (Cosenza and Kohler, 197213; Lee et al., 1974),implying that all B cells responsive to PC bear receptors with the TEPC 15 idiotype and that, even in the absence of the predominant clone, no other B cell can be stimulated. Claflin et al. (1974a) demonstrated that the receptors on all ABCs after immunization with PC were of the TEPC 15 idiotype. In addition, anti-PC serum antibody displayed binding characteristics for PC and choline analogs identical to the TEPC 15 myeloma protein, but distinct from other PC-binding plasmacytomas. More recent evidence (Gearhartet al., 1975a)points out that 25% of the BALB/c clonal precursor cells generated in the splenic-focus technique are not of the TEPC 15 idiotype. This finding has been corroborated by Kluskens et al. (1975), who concluded that there was a heterogeneous, but restricted, array of anti-PC antibodies, using the criteria of disc electrophoresis, IEF, and peptide mapping. Immunization of an A/J mouse with the BALBlc plasmacytoma TEPC 15 yields an anti-idiotypic serum, which appeared to be primarily directed against determinants outside of the combining site region, since the idiotype-anti-idiotype reaction could be inhibited by free hapten (Claflin and Davie, 1974a). As discussed above, the idiotypic determinants detected by this antiserum are found only in BALB/c or closely related strains sharing the a' allotype (Lieberman et at., 1974), although recent studies (Gearhart et al., 1977; Cancro et al., 1978) have shown that clonal precursor cells from other strains may share serological identity. Claflin and Davie (1974a) detected idiotypic determinants shared among all mouse strains, but not other rodent species, using an anti-HOPC 8 serum raised in rabbits. Since the hapten PC can inhibit the interaction between this antiserum and the myeloma protein, the antiserum is presumably directed against hinding-site determinants. These results correlate well with other studies showing similar binding specificity of anti-PC antibodies raised in 17 different strains of mice (Claflin and Davie, 1974b). Moreover, a PC-binding plasmacytoma from a CB20 mouse, CBPCZ, which possessed the C57BLka allotype but shared identical L chains with TEPC 15 by I E F and sequencing through the first hypervariable region, differed from the TEPC 15 H chain by two amino acids through the first hypervariable region, and possessed the rabbit anti-TEPC 15 idiotypic determinants, but not the framework determinants (Claflin et al., 1975). Along with the recent findings (Claflin, 1976a,b; Claflin and Rudikoff, 1976) that L chains identical to HOPC 8, McPc 603, and MOPC 511 by I E F and amino acid sequencing criteria are produced

THE B-CELL CLONOTYPE REPERTOIRE

311

in all mouse strains upon immunization with PC, this work suggests a striking conservation of the hypervariable regions within the mouse strain and implies that all these specificities are coded for by germline genes. b. Dissection of PC-Specific Repertoire. Sigal et al. (1975) analyzed the precursor cell response to PC by stimulating fragment cultures with the PC determinant linked via a tripeptide spacer to Hy, in order to achieve maximal stimulation of PC-specific B cells. An average frequency of one PC-specific clonal precursor cell per 40,000-50,000 splenic B cells was found for both nonimmune, adult, conventionally reared mice and adult germ-free mice. Although the number of PCspecific B cells in germ-free mice varied little among individuals, conventionally reared mice demonstrated a 30-fold fluctuation (Sigal et al., 1975). Idiotypic analysis of the monofocal antibodies revealed that 75% of the B cells specific for PC shared the TEPC 15 idiotype (Gearhart et al., 1975a). TEPC 15 idiotypic identity could be demonstrated in clones producing anti-PC antibody of more than one H-chain class, and except for variations accounted for by changes in isotype, all TEPC 15 monoclonal antibodies had identical I E F spectra: idiotypically positive TgM antibodies had a pZ of 4.6-4.7; IgA, a pZ of 6.0-6.1 (the same as TEPC 15 plasmacytoma protein); and IgG,, a pZ of 7.27.4 (Gearhart et al., 1975b). Since this analysis suggested that the monoclonal antibodies that reacted with the anti-TEPC 15 serum belonged to only one clonotype, it was possible to determine unambiguously an average repeat frequency for the TEPC 15 clonotype in nonimmune BALB/c mice as 1/60,000 B cells. Thus, if the TEPC 15 clonotype were representative of the B-cell repertoire, then from a total B-cell pool of 2 x lo8 cells, there would exist only 60,000 distinct clonotypes in the repertoire, each represented by as many as 3000 cells. Approximately 25% of the PC-specific monoclonal antibodies do not react with the murine anti-TEPC 15 serum (M anti-T15). It was possible to demonstrate a heterogeneous array of clonotypes specific for PC within this subset by anti-idiotypic cross-reactivity with the allogeneic anti-TI5 serum and by hapten inhibition of monoclonal antibody binding to antigen (Gearhart et al., 1977). Using a rabbit antiidiotypic antibody to TEPC 15 (R anti-T15), Gearhart et al. (1977) identified a group of clonotypes which react on a 1: 1 weight basis but have only weak cross-reactivity for the M anti-T15. This group of antibodies apparently represents several clonotypes, since they display heterogeneity by degree of idiotypic cross-reactivity with the allogeneic anti-T15. A maximum estimate of the frequency of this entire

312

NOLAN H. SIGAL AND NORMAN R. KLINMAN

subset could be calculated to be one in 2 x lo6B cells, or 30-40 times fewer than the predominant TEPC 15 clonotype. In AfHe and C3H mice, the R anti-T15 +, M anti-T15- subset represents approximately 1/500,000 splenic B cells, while the total frequency of precursor cells responsive to PC is only slightly less than that in BALB/c mice (Gearhart et al., 1977). Another method of identifying clonotypes within the PC-specific repertoire has been the use of anti-idiotypic antibodies against other PC-binding myeloma proteins, e.g., McPc 603 and MOPC 167. In an analysis of 141 BALB/c non-TEPC 15 monoclonal antibodies, two foci were found that partially cross-reacted with an allogeneic anti-McPc 603 serum (Gearhart et al., 1977).While this result indicates that there were no PC-specific clones reactive on a 1: 1 weight basis with both anti-McPc 603 and anti-Fab, a new clonotype subset (maybe complex) was operationally defined on the basis of this cross-reaction. The frequency of this subset was approximately 1 to 2 in lo7 B cells. With a rabbit anti-MOPC 167 idiotypic serum, an analysis of 220 BALB/c nonTEPC 15 monoclonal antibodies demonstrated 7 clones to be reactive with this anti-idiotype, again illustrating the enormous diversity of the B-cell repertoire (Accolla et al., 1977). In other murine strains, some monoclonal antibodies have been found that react with anti-MOPC 167 on a 1: 1 weight basis, and others have been identified as crossreactive specificities (N. Sigal, M. Cancro and N. Klinman, unpublished observation). A combination of techniques, including idiotypic cross-reactivity, IEF of IgM antibody and hapten inhibition with PC analogs, has been employed to extend the dissection of the PC-specific repertoire (Sigal et al., 1977a). In an analysis of 79 non-TEPC 15 monoclonal antibodies, it was possible to delineate at least 57 distinct clonotypes within the BALB/c strain. Those clonotypes which did not repeat among individual mice were present at a frequency of one per 2 x 107 B cells, a frequency similar to that calculated for most clonotypes responsive to DNP, NIP, and p-galactosidase (see above). While most individuals displayed a diverse array of anti-PC clonotypes, some specificities appeared to repeat several times within an individual mouse and could be represented by as many as 400 cells. Most interestingly, a number of clonotypes occurred among as many as 4 of the 16 donors analyzed. Each of these specificities was represented within the BALB/c strain at a frequency of one per 4 x lo6B cells and once again suggests that not all ubiquitous clonotypes need to be as dominant as the TEPC 15 clonotype. It is obvious that a dichotomy exists in the estimation of the B-cell

THE B-CELL CLONOTYF'E REPERTOIRE

313

repertoire, since extrapolation of the data for the TEPC 15 clonotype suggests that fewer than lo5 specificities are present in the repertoire. On the other hand, analysis of the frequencies of other clonotypes within the PC-specific response, in addition to the data derived from limiting-dilution studies with DNP, NIP, and P-galactosidase, implies that the repertoire consists of greater than lo7 different antibody specificities. The pauciclonal nature of the immune response to PC, as well as the high incidence of plasmacytomas of the TEPC 15 idiotype (Potter, 1972) may suggest that the TEPC 15 clonotype is not typical of the entire specificity repertoire. Thus, direct analysis of repeat frequencies in the PC responses and other experimental systems indicates that two categories of clonotypes can be defined: one group present in all individuals and represented by as many as 3000 cells, and another group represented by fewer than 10 cells per mouse. This division of clonotypes intensifies the difficulties in exactly delineating the repertoire, since an exact extimate of the total antibody diversity will depend on what proportion of the repertoire is accounted for by large clonotypes, e.g., the TEPC 15clonotype, and what proportion by the 1000-fold lower representation characterized by the non-TEPC 15 and NIP responses. That the larger estimate may more closely approximate the actual number of specificities is indicated b y the observation of approximately 10,000 low-frequency clonotypes responsive to NIP (Kreth and Williamson, 1973) and by the finding that non-TEPC 15 PC-specific clonotypes appear to be representative of at least 57 different clonotypes (Sigal et al., 1977b). Another implication of the analysis of the PC-specific repertoire is the suggestion that clonotypes with similar idiotypic specificities can be found in different murine strains, but that each strain expresses the clonotype with a characteristic frequency. For example, AKR mice possessed a specificity that is idiotypically indistinguishable from the predominant TEPC 15 clonotype in BALB/c mice (Gearhart et al., 1977); the clonotype represented 1/400,000 splenic B cells in AKR mice and 1/60,000 B cells in BALBlc mice. The genetic control of the frequency with which a specificity is expressed in the B-cell pool was studied by examining the expression of the TEPC 15 clonotype in mice of various genetic backgrounds, since this specificity is known to be encoded b y a germline gene (Cancro et d . , 1978). The Bailey C x B recombinant inbred strains and other inbred BALB/c and C57BL/6 recombinant strains provided an experimental model to test this question. Results from these studies (Cancro et al., 1978) indicated that the TEPC 15 clonotype could be expressed in recombinant strains of either the a' or a* allotype at frequencies ranging from

3 14

NOLAN €1. SIGAL AND NORMAN R. KLINMAN

1/500,000to 1/60,000 splenic B cells. Thus, a putatively germline antibody specificity may appear as either a major or minor clonotype. The distinction between major clonotypes as defined serologically by cross-reactive and anti-idiotypic sera and minor sporadically occurring clonotypes may, therefore, be an artificial one, reflecting the genetic background of the mouse more than the actual gene complement present. A final use of the anti-idiotypic sera reactive with PC-binding plasmacytoma proteins has been in the study of clonotypes within the DNP-specific repertoire. Since a homogeneous antibody that shows a given idiotypic cross-reactivity can be operationally defined as a distinct clonotype, Sigal (1977a) screened 680 monoclonal antibodies derived from anti-DNP precursor cells for reactivity with the murine anti-TEPC 15 serum and for their ability to bind PC. Two foci of the 680 clones analyzed bound PC, and one of these antibodies reacted with murine anti-TEPC 15 and anti-Fab on a 1 : 1 weight basis. The discovery of a clonotype reactive with M anti-T15 but not with R antiT15, the converse of the R anti-T15 M anti-T15- clonotype identified in the PC-specific repertoire, points to the novel idiotypic relationships which may be found among homogeneous antibodies binding diverse antigens. In addition, the maximum frequency of these clonotypes in the B-cell pool could be calculated to be one in 4 x lo6 splenic B cells. This estimate is, at best, a crude approximation since neither of these specificities repeated in the population analyzed, and therefore may be much more rare than the calculations imply. c. Analysis of the Viral Hemagglutinin (HA)-Speci.c Repertoire. As described in Section 111, a new system for clonotype analysis has been developed by Gerhard et al. (1975), using the murine-adapted HO N 1 influenza strains. Since many HA variants are available, monofocal antibodies can be assessed for fine specificity on a large panel of closely related determinants. Analysis of monoclonal antibodies derived from secondary B cells revealed over 40 distinguishable cross-reactivity patterns (determinants). Recently, Cancro et al. (1977) have carried out a preliminary analysis of monoclonal antibodies derived from primary B cells. Two relevant findings have emerged: (1)the primary HA specific repertoire appears at least as diverse as the secondary; and (2) the frequency of primary B cells in BALB/c spleen responsive to PR8 HA is approximately 1/200,000(50fold lower than in immune mice). Thus, the average frequency of B cells of any given influenza-reactive clonotype is less than one per 10’ and, most important, an individual mouse has fewer than 20 B cells capable of responding to these physiologically “relevant” individual determinants on a protein.

+,

THE B-CELL CLONOTYPE REPERTOIRE

3 15

3. Microcul ture Systems Cosenza et al. (1975)utilized the Lefkovits (1972) microculture system to determine the frequency of B cells responsive to PC. When a heat-killed vaccine of Pneumococcus strain R36A, which was shown to be a T-independent antigen, was added to the microcultures, the frequency of PC-specific precursor cells was estimated to be, on the average, 1/50,000 B cells. Since 85% of the responding clones expressed the TEPC 15 idiotype, these results are in complete agreement with those of Sigal et al. (1975), presented previously. The clonotype repertoire of B cells responsive to streptococcal group A and A-variant antigens was investigated by analyzing the antibodies produced by microcultures of rabbit peripheral blood lymphocytes (PBL) (Braun et al., 1976b). Since the in vitro response was dependent on an optimal cell density, graded numbers of cells could not be used to establish a Poisson distribution. Therefore, the frequency of precursor cells responsive to the A-variant carbohydrate were calculated from the zero term of the Poisson distribution, which provides a minimum estimate of the frequency, since microcultures containing two precursor cells are not accounted for. Braun et al. (197613) estimated that there was one A-variant-specific precursor cell in 35,000 lymphocytes. This information, combined with the fact that three predominant clonotypes that could be identified in the serum of rabbits immunized with streptococcal A variant carbohydrate (clonotypes 3, 8, and 11) were identified in 34, 13, and 9% of the microcultures, by IEF criteria, permitted a calculation of the frequencies of each of the clonotypes. Precursor cells of clonotype 3 were present at a frequency of 1/100,000 PBL; those of clonotype 8, a frequency of 1/250,000PBL; and clonotype 11, 1/400,000 PBL. It should be pointed out that these numbers represent minimum precursor frequency estimates in highly hyperimmunized, and thus, presumably expanded B-cell clones. However, since the analysis depends only on the identification of well-defined clonotypes b y IEF rather than the demonstration of identity among rarely occurring clones, this study escapes many of the pitfalls of other IEF experiments. In addition, the findings again illustrate that not all precursor cells exist at the same clone size, although this may be due to differential expansion after immunization in this case.

4 . Anti-Idiotypic Sera against Minor Clonotypes Ju et al. (1977)investigated private idiotypic specificities that were generated by hyperimmunization after suppression of the predominant anti-azophenylarsonate (Ars) clonotype in A/J mice. Approxi-

316

NOLAN H. SIGAL AND NORMAN R. KLINMAN

mately 20-60% of the antibody made by an A/ J mouse upon immunization with the hapten Ars is of the predominant idiotype (Kuettner et al., 1972). The appearance of the predominant idiotype could be immunologically suppressed by preinoculation of a neonatal or adult A/ J mouse with a rabbit anti-idiotypic serum (Hart et al., 1972). After hyperimmunization, anti-Ars antibody from the suppressed mice could be isolated, and anti-idiotypic antibodies could be made against the “private” idiotypes. Ju et al. (1977) made anti-idiotypic sera against four different private idiotypes, and the sera of 181 immune and nonimmune mice were tested for the presence of these clonotypes. Two of the private idiotypes were not found in any of the other mice tested, one private idiotype was identified in 3 of the 181 mice, and the fourth was demonstrated in 28% of the mice tested. While such studies d o not provide an absolute estimate of the B-cell clonotype pool, they d o illustrate the enormous diversity of the repertoire. Indeed, the authors suggested that the results are most consistent with the somatic generation of all specificities. Difficulty arises with this interpretation, however, because it should not be assumed that the antibody product of clonotypes present at low frequency within the B-cell population will be detected in the serum of all nonimmune and immune mice tested. The experiments do demonstrate the presence of clonotypes of different clone size and point out a unique and valuable probe for dissecting the clonotype repertoire. V. Defining the Neonatal &Cell Repertoire

The acquisition of the specificity repertoire during development poses problems and questions somewhat distinct from analysis of the adult repertoire. Since the fetus and neonate are dynamic systems, one must define the B-cell repertoire at various points in ontogeny. Since the developing mouse perforce must start with fewer cells and hence possess fewer specificities than the adult, the potential to unambiguously identify clonotypes and record their frequency is greater. In addition, the neonatal B-cell population has obvious biological significance and relevance for the analysis of the mechanism responsible for the generation of antibody diversity. Since generative events in the neonate are more likely to be observable than those in the adult due to the smaller number of cells involved, analysis of repertoire acquisition in developing cell populations may provide clues to the diversification process. In particular, how the acquisition of the repertoire varies from individual to individual and whether antigen plays a role in B-cell diversification are crucial questions for under-

THE B-CELL CLONOTYPE REPERTOIRE

317

standing the generation of diversity. Unfortunately, the majority of studies are more qualitative than quantitative, but d o provide some information concerning the development of the repertoire. Cells capable of synthesizing immunoglobulin (Ig) can first be detected in the liver of the mouse on day 12 of gestation (Cooper et al., 1977; Melchers and Phillips, 1977), but cells bearing surface Ig cannot be detected until day 15of gestation when they appear in the liver and the spleen (Spear et al., 1973; Nossal and Pike, 1973; Cooper et al., 1977; Melchers and Phillips, 1977). It would appear that B cells develop independently in both fetal liver and fetal spleen and that there is no requisite for gastrointestinal influences in mammals (Owen et ul., 1975; Owen, 1977).The appearance of Ig-bearing cells in the liver is transient, disappearing within a week of birth. I n the spleen, the number of Ig-bearing cells continues to increase, and reaches a plateau only in the adult animal (Spear et ul., 1973; Gelfand et ul., 1974; Sidman and Unanue, 1975). A. ANTIGEN-BINDING CELLANALYSIS Initial studies detected antigen-binding cells (ABCs) early in development in the embryonic chicken bursa (Dwyer and Warner, 1971) and the fetal mouse and human (Dwyer and MacKay, 1972b). Lymphoid cells binding '251-Iabeled flagellin could b e seen as early as day 14 in the bursa and on day 17 in the spleen of a chicken embryo (Dwyer and Warner, 1971). Dwyer and MacKay (19721)) observed a high frequency of cells binding 1251-labeledflagellin and Hy in the thymns of a 12-week human fetus, in agreement with Hayward and Soothill (1972), but such cells did not appear in the spleen until 16 weeks of gestation. In CBA mice, specific ABC were first present in the 14-day fetal thymus and 17-day fetal spleen, and the frequency of these cells were approximately equal to adult frequencies (Dwyer and MacKay , 197211). Using a rosette assay, Spear et ul. (1973)estimated the frequency of cells binding to three different antigens, TNP-SRBCs, SRBCs, and rabl>iterythrocytes at various points in ontogeny from day 16 of gestation to adulthood in outbred Swiss-L mice. Cells capable of binding to each of the three antigens were first detected between 15 and 16 days of gestation in the spleen. Their numbers increased rapidly and in parallel until about 2 weeks after birth, when they approached the adult plateau. While the absolute number of rosette-forming cells per spleen increased, the frequency of TNP-SRBC binding cells per lo3 Ig-positive cells remained constant at around 5 per lo3 Ig-positive cells throughout development. The neonatal cells resemble adult an-

318

NOLAN H. SIGAL AND NORMAN R. KLINMAN

tigen-binding cells in that free antigen or anti-Ig causes inhibition of antigen binding and the presence of both high- and low-avidity cells could be detected early in development. These studies were extended by D’Eustachio and Edelman (1975), who looked at the antigen-binding properties of cells in the livers and spleens of mice for 11 different hapten and protein antigens as a function of age. The investigators used the fiber-binding assay developed by Edelman et al. (1971). In agreement with the previous experiments, cells specific for each of the 11 antigens (including TNP, F1, lac, sulfanilic acid, myoglobin, and lysozyme) could be detected in the murine fetus by day 17 of gestation, at which time they appeared in both the liver and the spleen. In all cases, these cells disappeared from the liver within a day of birth, but continued to increase in number in the spleen until adulthood. Measurements of the number of fiber-binding cells in single fetal spleens rather than pools of spleens showed no systematic deviation of individual fetuses from this pattern, suggesting that the pooled frequencies were an accurate description of the immunologic development of individual mice. D’Eustachi0 and Edelman (1975) compared the distribution of relative avidities with which fetal and adult cells bound four of the antigens and found that the developing mice expressed the full range of specificities as assessed by avidity early in ontogeny. Thus, these investigators concluded that, although neonatal mice do not become immunocompetent until approximately 2 weeks of age, the mice appear to have the full range of antibody diversity present by the time of birth or before. The work of Decker et al. (1974) also suggests that lymphoid cells, early in development, can recognize a diverse array of antigenic determinants. They discovered ABCs capable of binding five diverse antigens (@-galactosidase,Hy, horse spleen ferritin, horseradish peroxidase, and ovalbumin) in the lymphoid tissue of mice, chick embryos, and rabbits at approximately the same frequency as in the adult. Particularly surprising was the observation that mouse yolk sacs had large numbers of ABCs b y day 10 of gestation, since other studies (Cooper et al., 1977; Melchers and Phillips, 1977) have shown that cells capable of synthesizing Ig first appear on the day 12 of gestation and cells bearing surface Ig cannot be detected until day 15 in the mouse. However, Decker et al. (1974) showed that binding of antigen to the cells could be inhibited by anti-Ig. Moreover, a proportion of the cells were able to bind two antigens simultaneously, suggesting that at least a subset of developing cells is multipotential. In marked contrast to the data discussed thus far, Lydyard et al. (1976) observed the sequential acquisition of ABCs specific for SRBC,

THE B-CELL CLONOTYF’E REPERTOIRE

319

Hy, and TGAL in the chick embryo bursa. Cells bearing surface Ig were detected first in the bursa on day 12 of gestation. Hy-binding cells and TGAL-binding cells were found by day 16, and SRBC rosette-forming cells were not seen until day 18. This pattern of development was not altered in germ-free embryos or by the exposure of embryos to exogenous antigens. The most compelling evidence for the sequential expression of the repertoire in this study came from an analysis of individual bursal follicles, where specific ABCs were observed to develop in multiple follicles as small foci of ABCs among the much larger total population of Ig-positive cells within the follicle. While the late expression of SRBC-specific cells may have been due to the differential sensitivity of the rosetting assay versus the binding of 1251-labeledantigen, the appearance of ABC foci in multiple bursal follicles is not subject to this criticism. These studies are in agreement with the discovery of the sequential acquisition of flagellin and SRBC-binding cells in CBA mice repopulated with a single hematopoietic stem cell (Yung et al., 1973).Using a double-transfer system consisting of transfer of 12-day fetal liver cells into an irradiated syngeneic recipient followed by transfer of a single hematopoietic focus from the primary recipient into another irradiated recipient, these investigators demonstrated that flagellin-binding celIs appeared 18-20 days after transfer, but SRBC-binding cells were not present until days 20-23. The differential sensitivity of the binding assays used again could complicate interpretation of these experiments. The recent findings of D’Eustachio et al. (1976) may also be explained by the sequential expression of clonotypes in the developing repertoire. Since earlier work from that laboratory (Spear et al., 1973; D’Eustachio and Edelman, 1975) indicated that for individual SwissL fetal mice, the frequency of ABCs specific for a number of antigens varied significantly more than was explained by sampling fluctuation. D’Eustachio et al. (1976) examined the frequency of cells binding TNP and SRBC in two inbred strains, CBA and BALBk. They found that the frequency of ABC varied little from individual to individual within an inbred strain, but the number of specific ABC increased at a characteristic rate for each strain, suggesting that the development of these cells may be subject to strong genetic controls. B. In Vivo STUDIESOF CLONOTWEDEVELOPMENT Since antigen-binding cell analyses are relatively nonspecific (see Section 11) and do not have the capacity to delineate developing specificities at the clonotype level, a number of experiments have at-

320

NOLAN H. SIGAL AND NORMAN R. KLINMAN

tempted to examine the capacity of the neonatal B cell to make specific antibody and compare the neonatal repertoire with the adult. Such experiments are complicated, however, by the fact that antibody responses are the products of complex interactions between T cells, B cells, and macrophages. A number of studies have shown that both the macrophage (Argyris, 1968; Landahl, 1976) and T cell (Chiscon and Golub, 1970; Arrenbrecht, 1973; Spear and Edelman, 1974) are functionally immature at birth, and, in fact, there is a high level of suppressor T cell activity in the neonatal period (Mosier and Johnson, 1975). Therefore, it is not surprising that numerous investigators demonstrated that neonates were unresponsive to a wide variety of antigens (reviewed in Sterzl and Silverstein, 1967). The requirement for mature T cells and macrophages should also be kept in mind with respect to in vivo experiments where adult T cells are not provided (Silverstein et al., 1963; Playfair, 1968; Montgomery and Williamson, 1972; Yung et al., 1973),since the differential responsiveness seen by these investigators may be due to a T-cell defect. Nevertheless, since the experiments are of historical interest and foreshadow similar findings in more controlled systems, they will be discussed briefly in the following paragraphs. The fetal lamb was an ideal model to investigate the development of immune responsiveness, since maternal IgG does not cross the placenta and the incidence of twinning is high, providing a genetic and aged-matched control. Silverstein et al. (1963)utilized this system by immunizing fetuses 65-90 days old (total gestation, 150 days) with a complex of antigens: bacteriophage 4x174, ferritin, ovalbumin, diphtheria toxoid, Salmonella, and BCG. They found that the 65-70-dayold fetus was able to respond vigorously to 4x174 and less well to ferritin, and in 90-day-old fetuses, ovalbumin elicited an antibody response in approximately 50% of the animals. Thus, despite an ignorance of T-cell cooperation and the variable sensitivity of the antibody assays, the phenomenon of sequential acquisition of the repertoire was established. In the mouse, there was a sequential acquisition of responsiveness to polymerized flagellin and SRBC after transfer of a single hematopoietic stem cell to an irradiated recipient (Yung et al., 1973).In addition, Rowlands et al. (1974)demonstrated a hierarchy of responsiveness to a variety of antigens in the opposum, whose fetus is externally suckled and available for experimentation early in gestation. The question of the ordered appearance of antigen-reactivity has been examined in the chicken (Ivanyi, 1975).It might have been predicted that if the hierarchy of responsiveness were not simply due to a

THE B-CELL CLONOTYF’E REPERTOIRE

32 1

T-cell deficiency, bursectomy at various points in development would induce unresponsiveness to some antigens, but not others. Chickens were surgically bursectomized 1, 4, or 7 days after hatching and immunized at 7 weeks of age with four antigens. Ivanyi (1975)found that the response to Bordetella pertussis was eliminated by bursectomy at any age, the anti-influenza response was unaffected b y neonatal bursectomy, and the anti-SRBC response was diminished in chickens bursectomized at 1 day of age, but normal in day 4-bursectomized animals. The results of the earlier studies were, therefore, confirmed by this work. Another approach to resolving the limitations of Silverstein’s original experiments was to supply the neonatal B cells with adult T cells and macrophages in an adoptive cell-transfer system (Sherwin and Rowlands, 1975). Previous experiments (Sherwin and Rowlands, 1974) had established that, in agreement with Silverstein et al. (1963), the response of fetal liver cells transferred to an irradiated adult recipient followed a distinct hierarchy for 8 unrelated antigens. The sequence of antigen reactivity was invariant among three inbred mouse strains, BALB/c, C3H, and AKR, except that BALB/c and C,H differed in their acquisition of responsiveness to myoglobin (Sherwin and Rowlands, 1975). When these investigators transferred fetal liver cells to carrier-primed irradiated recipients, which presumably could provide optimal T-cell help, and immunized with DNP and F1 on the homologous carrier, they found that the antibody response to DNP was restored soon after reconstitution, but F1-specific precursor cells did not respond until 21 days after transfer. Although the work discussed thus far appears to establish that the sequential acquisition of immune responsiveness is an intrinsic property of developing B cells rather than due to secondary effects, it provides only qualitative information about the neonatal repertoire. Thus, the “Silverstein phenomenon” suggests that clonotypes reactive to certain antigens develop in ontogeny before other specificities, but the experiments do not identify clonotypes within the repertoire and consequently do not establish a coherent picture of the neonatal repertoire or its acquisition. The first study which attempted to delineate clonotypes within the neonatal repertoire utilized IEF to examine the anti-DNP response of neonatal rabbits (Montgomery and Williamson, 1972). Rabbits were immunized at birth with DNP-BGG, boosted with the same antigen at approximately 15 days of age, and the IgG from the immune serum was analyzed b y slab gel IEF several days later. In contrast to the heterogeneous anti-DNP antibody from adult rabbits, the IgG from a neonatal rabbit was highly restricted, and

322

NOLAN H. SIGAL AND NORMAN R. KLINMAN

within a litter the same spectrotype pattern tended to occur in a number of individuals. The results suggest that the neonatal anti-DNP repertoire is less diverse than the adult repertoire and may be under genetic control, since similar clonotypes tended to repeat in related individuals. Difficulty arises in interpretation of these experiments, however, since a source of immunocompetent T cells was not present and since it was unclear whether the IgM antibody was as restricted as the IgG. Goidl and Siskind (1974) used hapten inhibition of PFC formation to assess the heterogeneity of the anti-DNP response of neonatal and fetal mice. Lymphoid tissue (liver, spleen, bone marrow) from fetal or neonatal mice was transferred to adult, irradiated recipients along with syngeneic adult thymocytes as a source of T cells, and the recipient was immunized with DNP-BGG. It was found that B cells from 17-day fetal liver or neonatal liver are highly restricted and produce only low avidity PFCs as compared with adult spleen or bone marrow, which make both high- and low-avidity PFCs. By 2 weeks of age, the B cells achieved an adult character with respect to heterogeneity of affinity, the spleen maturing in affinity earlier than the bone marrow. Subsequent investigations have shown that the heterogeneity of PFC from fetal liver can be increased by administration of LPS to the recipient at the time of cell transfer (Goidl et al., 1976). For the past several years, DuPasquier and his co-workers have been investigating the antibody response of larval and adult amphibians (Haimovich and DuPasquier, 1973; reviewed in DuPasquier, 1973; DuPasquier and Wabl, 1976). Young larvae possessing a small number of lymphocytes ( 1 to 5 x lo6) were shown to synthesize specific antibodies against a wide range of antigens. Haimovich and DuPasquier (1973) employed hapten-inhibition techniques to demonstrate that the anti-DNP antibody of the IgM class had affinities similar to that of goat anti-DNP IgM and that the affinity of the tadpole anti-DNP increased after immunization. I E F studies of the immune fluid from the tadpoles revealed a restricted spectrotype pattern in most individuals, but there were few identical clonotypes observed in an outbred population. A recent advance in the study of this experimental system has been the use of a Xenopus hybrid which gives rise to large clones of fully identical, isogenetic animals (Kobe1 and DuPasquier, 1975). The anti-DNP response in isogenetic tadpoles is extremely restricted both in terms of affinity and clonotype repertoire (DuPasquier and Wabl, 1976). I E F patterns obtained from isogenetic animals showed that all the animals of one genotype share a small number of anti-DNP spectrotypes, whereas this was not the case in

THE B-CELL CLONOTYPE REPERTOIRE

323

outbred individuals. Because so little is known about immune responsiveness in amphibians, it is premature to claim that the clonotype identity in these tadpoles reflect the inheritance of a small set of germline genes, but these fascinating results mirror the findings of Montgomery and Williamson (1972) and Goidl and Siskind (1974) as well as the results in the murine system discussed below.

C. In Vitro STUDIESOF CLONOTYPE DEVELOPMENT The splenic focus technique has been very helpful in defining the neonatal repertoire and its acquisition during ontogeny. Initial reports demonstrated that fetal and neonatal B cells from spleen, liver, or bone marrow were immunologically competent when the ancillary mechanisms for stimulation were provided b y the milieu of the adult carrier-primed recipient (Press and Klinman, 1973a,b). The splenicfocus technique has demonstrated that neonatal precursor cells are unipotential and give rise to clonal progeny producing homogeneous monofocal antibodies (Press and Klinman, 1973a,b; Klinman and Press, 1975b,c) and therefore could be used to enumerate the frequency of neonatal clonal precursor cells responsive to a variety of antigens (Press and Klinman, 1973b, 1974; Sigal et al., 1976, 1977c; Sigal, 1977). The picture that emerges from these studies appears to reflect a sequential acquisition of the repertoire at the clonotype level. Because of the importance of this work both in terms of defining the repertoire and for theories concerning the generation of antibody diversity, the experimental findings for each antigen system analyzed will b e discussed.

1 . DNP, T N P , Dunsyl, und F1-Specijc Neonatal Repertoires When the frequencies of neonatal precursor cells specific for DNP, TNP, dansyl, and F1 were analyzed, it was observed that a significant disparity existed in the rates of development of F1-specific B cells when compared to dansyl, DNP- or TNP-specific B cells (Press and Klinman, 1974; Sigal, 1977h). The neonatal frequencies for DNP, TNP, and dansyl, when corrected for cloning efficiency and number of B cells present in the neonatal spleen, were approximately the same as the adult frequency (1/40OO for DNP and TNP; 1/10,000 for dansyl), while the neonatal frequency for F1 was significantly lower than the adult (1/25,000 in the neonate vs 1/8000 in the adult). This discrepancy may explain the hierarchy of immune responsiveness documented b y so many investigators, including the observation of the appearance of anti-DNP reactivity prior to F1 responsiveness in the mouse (Sherwin and Rowlands, 1975).

324

NOLAN H. SIGAL AND NORMAN R. KLINMAN

In order to further delineate the neonatal specificity repertoire, the monoclonal antibodies derived from DNP- and TNP-specific neonatal precursor cells were analyzed by sucrose gradient I E F (Klinman and Press, 1975b). The results indicated that, early in development, when only lo5 to lo6 B cells are present in the BALB/c mouse spleen, the neonatal specificity repertoire is relatively restricted. Preliminary evidence on the I E F patterns of anti-dansyl monoclonal antibodies has confirmed this notion (Owen et al., 1978). During the first 4 days of neonatal life, almost all of the DNP-specific clones produced IgM antibody of pl 5.05, 5.25, or 5.55. These could be distinguished from clonotypes responding to TNP, which were also predominantly of three distinct PIS: 5.00, 5.15, or 5.40. The DNP-specific clonotypes could also be shown to be distinct from the TNP-specific clonotypes by the additivity of the monofocal response when stimulation was carried out with a mixture of DNP- and TNP-containing antigens (Klinman and Press, 197513). Since the neonatal spleen cell population, as a whole, repeatedly expressed all three TNP and all three DNP clonotypes, these predominant clonotypes fulfill the classical definition of germline antibody expressions. Although an individual neonate expressed in a random fashion a preponderance of only one of the three DNP and one of the three TNP specificities, the predominant clonotypes, on the average, represented one-third of the total DNP- or TNPspecific neonatal B cells. Therefore, the frequency of each clonotype in the B-cell population is one per 12,000 B cells, implying that approximately lo4 clonotypes are present in the repertoire early in development. Owing to the relative insensitivity of the sucrose gradient IEF, this is only a minimum estimate of the number of DNP- and TNP-specific clonotypes, and the neonatal repertoire may be larger than the postulated 10,000 clonotypes. By 2-3 days of life, each predominant clonotype can be represented by 50-200 cells. Since individual neonates possess many precursor cells of the same clonotype, this would indicate that there was a conservative clonal expansion of a given specificity in the individual. If the precursor doubling time is 24 hours (Spear et aZ., 1973; Gelfand et al., 1974; Nossal and Pike, 1973), then the original clonotype precursor cell must have arisen by day 15 of fetal life. Thus, at least lo4 different clonotypes may be expressed early in development and can expand in the absence of antigenic contact (in u t e r o ) to reach a clone size of up to 200-400 cells. Klinman and Press (1975b) found that the predominant clonotypes constituted the major expression of the DNP and TNP specificity repertoire until day 6 after birth. At this time, the predominant clonotypes

THE B-CELL CLONOTYPE REPERTOIRE

325

became a minority of the population and were replaced by a more heterogeneous array of specificities. These “sporadic” clonotypes could be viewed as arising later in development (i.e., after day 18 of gestation) than the early predominant ones. The expansion of the anti-DNP and anti-TNP repertoires from the three neonatal clonotypes to the 1500-8000 clonotypes found in the adult (see Section IV) presumably occurs during the next several weeks, and it is obvious then that the vast majority of clonotypes in the repertoire arise after the first week of life.

2 . PC-Specijic Neonntal Repertoire It was of considerable interest to identify within the neonatal repertoire PC-specific clonal precursor cells which bear the same idiotype as the TEPC 15 plasmacytoma protein, since the TEPC 15 clonotype is present in high frequency in all individuals of the BALB/c strain, and therefore is presumably encoded by germline variable region genes (see Section IV). The time of its appearance in the repertoire may have implications for the generation of antibody diversity, since the acquisition of a well-defined clonotype can be mapped. Precursor cell analyses indicated that PC-specific B cells, and the TEPC 15 clonotype in particular, did not appear in the BALB/c spleen until approximately 1 week of age (Sigal et al., 1976). This was a highly predictable process in that between 6 and 8 days of age essentially every BALB/c neonate acquired PC-specific precursor cells. In addition, by the criterion of susceptibility to tolerance induction, PC-specific B cells arising during the second week of life are, indeed, immature cells that have been recently generated from the stem cell pool (Metcalf et al., 1977). The results demonstrated that at 7-10 days after birth, when over 90% of the DNP-specific splenic B cells are resistant to tolerance induction, the majority of PC-specific B cells were tolerizable. The data illustrated the use of tolerance susceptibility as a characteristic of developing clones and confirmed the late acquisition of the PCspecific repertoire. In a more recent report, the observation that the TEPC 15 clonotype does not appear in the spleen until 1 week after birth was extended (Sigal et ul., 1977~). Antigenic influences did not appear to play a role in the development of TEPC 15 precursor cells since the clonotype followed similar kinetics of appearance in germfree neonates. Analysis of (C57BL/6? x BALB/c)F, neonates raised by C57BL/6 mothers, whose serum had no natural levels of TEPC 15 (as do BALB/c mothers), revealed that these neonates acquire the TEPC 15 clonotype at the same time in ontogeny as neonates raised by BALB/c

326

NOLAN H. SIGAL AND NORMAN R. KLINMAN

mothers, suggesting that maternal influences and genetic background play a minor role in development of the repertoire. Sigal et al. (1977~) also found that the lack of PC-specific precursor cells shortly after birth was reflected in a dearth of PC-specific antigen binding cells as well. Finally, no PC-specific B cells were observed in 19-day fetal liver or in bone marrow until 7 days of life, coincident with their appearance in the spleen. The relatively late appearance of the TEPC 15 clonotype has profound implications for the acquisition of the B-cell specificity repertoire. Although this specificity is present in high frequency in every adult BALB/c mouse and serves as the prototype of a germline antibody specificity, it displays none of the characteristics of the six DNP and TNP predominant clonotypes. In addition to its late acquisition, the appearance of the TEPC 15clonotype is a highly predictable process in that by day 8 of life, at a time when the TNP and DNP-specific repertoires are already extremely heterogeneous, essentially every individual BALBtc neonate has acquired the TEPC 15 clonotype. Thus, within a period of a few days, the TEPC 15 clonotype “spontaneously,” but predictably, expands from an average of one cell on day 4 after birth to 30 cells by day eight. These kinetics mirror events in the acquisition of the DNP and TNP predominant clonotypes but occur approximately 7 days later in ontogeny. Thus, the generation of not only the TEPC 15 clonotype but perhaps all specificities may be the result of highly ordered, rigorously predetermined events.

3. Ars-SpeciJk Neonatal Repertoire Sigal(1977b) has recently analyzed the frequency of precursor cells responsive to p-azophenylarsonate (Ars) in BALB/c neonatal mice. The frequency of Ars-specific B cells at birth was identical to the DNP- and TNP- specific frequencies (1/4000 B cells), but the frequency of splenic B cells responsive to Ars fell during the first week of life to a level of approximately one per 200,000 splenic B cells at 7 days after birth, which is only slightly higher than the average frequency of PC-specific B cells at this age. Examination of individual neonates revealed that this was a highly reproducible and predictable process such that there was essentially no overlap between the frequencies seen in 1-day-old and 3-day-old mice. When the precursor cell frequencies were replotted as the total number of hapten-specific cells in the spleen at various ages, the number of Ars-specific B cells was observed to remain essentially constant from 1to 4 days of age, at the same time that the number of cells responsive to most other antigens (and splenic B cells in general) were increasing linearly. Thus,

THE B-CELL CLONOTYPE REPERTOIRE

327

the dramatic decrease in precursor frequency could be accounted for by the fact that while the Ars-specific precursor pool remained constant, the neonatal spleen was increasing in size as more and more clonotypes reactive to other antigens were being generated. The constancy in the size of the Ars-specific precursor pool serves as an additional argument in favor of the programmed acquisition of the B-cell repertoire. If one postulates that the expansion of the DNPspecific repertoire from the three predominant clonotypes to the heterogeneous array found in the adult is due to antigen-driven mutational events, then the same process should take place to expand the repertoire of anti-Ars specificities. Instead, the increase in the Arsspecific repertoire is predictably delayed for approximately 5 days in every BALB/c neonate analyzed. VI. Conclusion

In spite of some persistent inconsistencies in data concerning elements of the clonal selection hypothesis, general agreement with its basic veracity has engendered a great deal of productive investigation of the specificity repertoire. In summary, the available information indicates that the murine B cell repertoire probably consists of greater than lo7 clonotypes and that individual mice express a large proportion of these. While the degree of repertoire homology from individual to individual remains in the realm of speculation, those clonotypes which are represented by multiple B cells within the primary repertoire of an individual (major or dominant clonotypes) are invariably expressed by most or all members of an inbred strain. Preliminary evidence indicates that even relatively minor clonotypes can be found repeatedly within a strain. However, this does not yet prove that all clonotypes are shared, a situation which would suggest that random somatic events do not play a significant role in repertoire diversification. The neonatal B cell repertoire is significantly more restricted and, preliminarily, appears repetitive within a strain. However, because neonatal responses are more difficult to obtain, less information is available pertaining to the neonatal repertoire. Since questions of diversity can be viewed in the context of (a) what is present in the germline (ontogenic expression) and (b) what genetic and regulatory mechanisms control the expansion to the adult repertoire, it may be assumed that much effort will be focused on this area. Already many interesting findings have arisen which indicate that clonotype expression is controlled at many levels and that many assumed polymor-

328

NOLAN H. SIGAL AND NORMAN R. KLINMAN

phisms in variable-region structural genes between strains may more closely reflect disparities in control of clonotype expression (Cancro et al., 1977a,b).Given the foundation of repertoire definition reviewed here, the availability of more and more sophisticated cloning techniques, including cellular hybridization and the possibility for microsequencing analysis of selected clonotypes, the future should yield important clarifications of repertoire expression.

REFERENCES Aasted, B., and Kindt, T. J. (1976a).Eur. J . Immunol. 6,721. Aasted, B., and Kindt, T.J. (1976b).Eur. J . Immunol. 6,727. Abney, E. R., Keeler, K. D., Parkhouse, M. E., and Wilcox, N. A. (1976).Eur. J . Immunol. 6,443. Accolla, R., Gearhart, P. J., Sigal, N. H., Cancro, M. P., and Klinman, N. R. (1977).Eur. J . Immunol. 7,876. Ada, G. L. (1970).Transplant. Reo. 5, 105. Ada, G. L., and Byrt, P. (1969).Nature (London)222, 1291. Adler, F. L., Fishman, M., and Dray, S. (1966). J . Immunol. 97,554. Anderson, J,, Coutinho, A., Melchers, F., and Watanabe, T. (1977).Cold Spring Harbor Symp. Quant. Biol. 41,227. Argyris, B. F. (19ss).J.Erp. Med. 128,459. Arrenbrecht, S. (1973).Eur. J . Immunol. 3,506. Askonas, B. A., and Williamson, A. R. (1972a).Nature (London)238,339. Askonas, B. A., and Williamson, A. R. (197213). Eur. J . Immunol. 2,487. Askonas, B. A., Williamson, A. R., and Wright, B. (1970).Proc. Natl. Acad. Sci. U.S. A. 67, 1398. Askonas, B. A., Cunningham, A. J., Kreth, H., Roelants, G., and Williamson, A. R. (1972). Eur. J . Immunol. 2,494. Attardi, G., Cohn, M., Horibata, K., and Lennox, E. S. (1964).J . Immunol. 92,346. Avrameus, S . , and Guilbert, B. (1971).Eur. J . Immunol. 1,394. Awdeh, Z.L., Williamson, A. R., and Askonas, B. A. (1968).Nature (London)219,66. Awdeh, Z.L., Williamson, A. R., and Askonas, B. A. (1970).Biochem. J . 116,241. Bach, J. F. (1973).Contemp. Top. Immunobiol. 2, 189. Bach, J. F., Reyes, F., Dardenne, M., Foumier, C., and Muller, J. Y. (1971).In “Cell Interactions and Receptor Antibodies in Immune Responses” (0.Makela, A. Cross, and T. U. Kosunen, eds.), p. 111. Academic Press, New York. Bankhurst, A. D., and Wilson, J. D. (1971).Nature (London),New Biol. 234, 154. Bankhurst, A. D., Tomgiani, G., and Allison, A. C. (1973).Lancet 1,226 Barstad, P., Rudikoff, S , Cohn, M., Konisberg, M., and Hood, L. (1974).Science 183, 962* Basten, A., and Howard, J. G. (1973).Contemp. Top. Immunobiol. 2,265. Basten, A,, Miller, J. F. A. P., Warner, N. L., and Pye, J. (1971).Nature (London),New Biol. 231, 104. Bell, C., and Dray, S. (1971). J . Immunol. 107,83. Berek, C., and Eichmann, K. (1977).lmmunogenetics 4,417. Biozzi, G., Stiffel, C., Mouton, D., Liacopoulas, B. M., Decreuseford, C., and Bouthiller, Y. (1966).Ann. Inst. Pasteur, Paris 110,7. Birshtein, B. K., Preud’homme, J.-L., and ScharfT, M. D. (1974).In “The Immune Sys-

T H E B-CELL CLONOTYPE REPERTOIRE

329

tem: Genes, Receptors, Signals” (E. E. Sercarz, A. R. Williamson, and C. F. FOX, eds.), p. 339. Academic Press, New York. Blomberg, B., Geckeler, W., and Weigert, M. (1972). Science 177, 178. Bona, C. A., Trebiciavsky, I., Antevnis, A., Meuclin, C., and Robineaux, R. (1972).Eur. J. Zmmunol. 2,434. Bosma, M. J., and Weiler, E. (1970).J.Zmmunol. 104, 203. Bosma, M. J., Perkins, E. H., and Makinodan, T. (1968).J. Zmmunol. 101,963. Bosma, M. J., Davis, G., and Bosma, G. (1972).J.Zmmunol. 109, 506. Bosma, M. J., DeWitt, C., Hausman, S . J., Marks, R., Potter, M., and Taylor, B. (1977). Znimunogenetics 4,418. Braun, D. G., and Kelus, A. S . (1973).J.E x p . Med. 138, 1248. Braun, D. G., Huser, D. G., and Reisen, W. F. (1976a).Eur. J. Zmmunol. 6, 570. Braun, D. G., Quintans, J., Luzzati, A., Lefiovits, I., and Read, S. (1976b).J.E x p . Med. 143,360. Bridges, S. H., and Little, R. J. (1971).Biochemistry 10,2525. Brient, B., and Nisonoff, A. (1970).J.E x p . Med. 132, 951. Brown, R. A., Makinodan, T., and Albright, J. F. (1966).Nature (London)210, 1382. Burnet, F. M. (1959). “The Clonal Selection Theory of Acquired Immunity.” Vanderbuilt Univ. Press, Nashville, Tennessee. Byrt, P., and Ada, G. L. (1969). Zmmunology 17,503. Cancro, M. P., Sigal, N. H., and Klinman, N. R. (1978a).J. Erp. Med. 147, 1. Cancro, M. P., Gerhard, W., and Klinman, N. R. (1978b).J.E x p . Med. 147, 776. Capra, J. D., and Kehoe, J. M. (1975).Ado. Zmmunol. 20, 1. Capra, J. D., Tung, A., and Nisonoff, A. (1975).J.Zmmunol. 115,414. Capra, J. D., Berek, C., and Eichman, K. (1976).J.zrnmunol. 117, 7-10. Capra, J. D., Klapper, D. G., Tung, A., and Nisonoff, A. (1977). Cold Spring Harbor S y m p . Quant. Biol. 41,847. Carson, D., and Weigert, M. (1973). Proc. Natl. Acad. Sci. U.S.A. 70,235. Cehra, J. J. (1977).In “Antibodies in Human Diagnosis and Therapy” (E. Haber and R. Krause, eds.), p. 79. Raven, New York. Cehra, J. J., Gearhart, P. J., Kamat, R., Robertson, S., and Tseng, J. (1977). Cold Spring Harbor Symp. Quant. Biol. 41, 201. Cecka, J. M., Stratton, J. A,, Miller, A,, and Sercarz, E. E. (1976).Eur.J.Immunol. 6,639. Chiscon, M., and Golub, E. (1970).J. Zmmunol. 108, 1379. Choi, T. K., Sleight, D., and Nisonoff, A. (1974).J.E x p . Med. 139,761. Civan, C. I., Levine, H. B., Williamson, A. R., and Schlossman, S. F. (1976).J.Zmmunol. 116, 1400. Claflin, J. L. (1976a).Eur. J. Zmmunol. 6,666. Claflin, J . L. (1976b). Eur. J. Zmmunol. 6, 669. Claflin, J. L. (1977). Cold Spring Harbor Symp. Quant. B i d . 41,725. Claflin, J. L., and Davie, J. M. (1974a)J.E x p . Med. 140,673. Claflin, J. L., and Davie, J. M. (1974b).J.Zmmurwl. 113, 1678. Claflin, J . L., and Davie, J. M. (1975).J.Erp. Med. 141, 1073. Claflin, J. L., and Rudikoff, S. (1976).J.Exp. Med. 144, 1294. Claflin, J . L., Lieberman, R., and Davie, J. M. (1974a).J.E x p . Med. 139,58. Claflin, J . L., Lieberman, R., and Davie, J. M. (1974b).J.Zmmunol. 112, 1747. Claflin, J . L., Rudikoff, S., Potter, M., and Davie, J. M. (1975).J.Erp. Med. 141,608. Coates, A. S., and Lennon, V. A. (1973).Zmmunology 24,425. Cohn, M. (1971).Ann. N . Y. Acad. Sci. 190,529. Cohn, M., Notani, G., and Rice, S. (1969) lmmunochemistry 6, 111.

330

NOLAN H. SIGAL AND NORMAN R. KLINMAN

Cohn, M., Blomberg, B., Geckeler, W., Raschke, W., Riblet, R., and Weigert, M. (1974). I n “The Immune System: Genes, Receptors, Signals” (E. E. Sercarz, A. R. Williamson, and C. F. Fox, eds.), p. 89.Academic Press, New York. Coombs, R. R. A,, Gurner, B. W., Janeway, C. A., Jr., Wilson, A. B., Gel], P. G., and Kelus, A. S. (1970).Immunology 18,417. Coons, A. H. (1958). J. Cell. Physiol. 52, Suppl. 1, 55. Cooper, M. D., Kearney, J., Lydyard, P., Grossi, C., and Lawton, A. (1977).Cold Spring Harbor Symp. Quant. Biol. 41, 139. Cooper, M. G., Ada, G. L., and Langman, R. E. (1972).Cell. Immunol. 4,289. Cosenza, H., and Kohler, H. (1972a).Proc. Natl. Acad. Sci., U.S.A.69,2701. Cosenza, H.,and Kohler, H. (1972b).Science 176, 1027. Cosenza, H., Quintans, J., and Lefkovits, I. (1975).Eur. J . Zmmunol. 5,343. Cosenza, H., Augustin, A. A., and Julius, M. H. (1977).Cold Spring Harbor Symp. Quant. Biol. 41,709. Couderc, J., Bleux, C., Birrien, J. L., and Liacopoulos, P. (1973).J. Zmmunol. 111,1155. Couderc, J., Bleux, C., and Liacopoulos, P. (1975).Immunology 29,665. Coutinho, A,, Gronowicz, E., Bullock, W., and Moller, G. (1974).J. E x p . Med. 139,74. Cramer, M., and Braun, D. G. (1974).J. E x p . Med. 130, 1513. Cramer, M.,Schwartz, M., Moses, E., and Sela, M. (1976).Eur. J . Immunol. 6,618. Cudkowicz, G., Shearer, G. M., and Priore, R. L. (1969).J. E x p . Med. 130,481. Cudkowicz, G . , Shearer, G. M., and Ito, T. (1970).J. E x p . Med. 132,623. Cunningham, A. J. (1976a).Ann. Zmmunol. (Paris) 127c, 531. Cunningham, A. J. (1976b).In “The Generation of Antibody Diversity: A New Look” (A. J. Cunningham, ed.), p. 89.Academic Press, New York. Cunningham, A. J,, and Fordham, S. (1974).Nature (London)250,669. Cunningham, A. J., and Pilarski, L. M. (1974a).Eur. J. Immunol. 4,319. Cunningham, A. J., and Pilarski, L. M. (1974b).Eur. J . Zmmunol. 4, 757. Cunningham, A. J., and Pilarski, L. M. (1974c).in“The Immune System: Genes, Receptors, Signals” (E. E. Sercarz, A. R. Williamson, and C. F. Fox, eds.), p. 367.Academic Press, New York. Davie, J. M., and Paul, W. E. (1971).J. E x p . Med. 134,495. Davie, J. M., and Paul, W. E. (1972).J. E x p . Med. 135,660. Davie, J. M.,and Paul, W. E. (1974).Contemp. Top. Immunobiol. 3, 171. Davie, J. M., Rosenthal, A. S., and Paul, W. E. (1971).J. E x p . Med. 134,517. Decker, J., Clarke, J., Bradley, L., Miller, A., and Sercarz, E. (1974).J. Zmmunol. 113,

1823.

Decreuseford, C., Mouton, D., Binet, J. L., Pavlovsky, S., Stiffel, C., Bouthilier, Y., and Bozzi, G. (1970).Ann. Znst. Pasteur, Paris 119,76. DeLuca, D., Decker, J., Miller, A., and Sercarz, E. (1974).Cell. Immunol. 10,1. DeLuca, D., Miller, A., and Sercarz, E. (1975a).Cell. Imrnunol. 18,255. DeLuca, D., Miller, A., and Sercarz, E. (1975b).Cell. Zmmunol. 18,274. DeLuca, D., Miller, A., and Sercarz, E. (1975~). Cell. Zmmunol. 18,286. D’Eustachio, P., and Edelman, G. M. (1975).J . Exp. Med. 142, 1078. D’Eustachio, P., Cohen, J., and Edelman, G. M. (1976).J. E x p . Med. 144,259. Diener, E., and Paetkau, V. H. (1972).Proc. Natl. Acad. Sci. U.S.A. 69,2364 Diener, E.,Kraft, N., and Armstrong, W. D. (1973).Cell Immunol. 6,80. Diener, E., Lee, K. L., Langman, R. E., Kraft, N., Paetkau, V. H., and Pernis, B. (1974). In “Control of Proliferation in Animal Cells” (B. Clarkson and R. Baserga, eds.), p. 411.Cold Spring Harbor Lab. Press, Cold Spring Harbor, New York. DiPauli, R. (1976).Eur. J . Immunol. 6,385.

T H E B-CELL CLONOTYPE REPERTOIRE

331

DuPasquier, L. (1973). Curr. Top. Microbiol lmmunol. 61,37. DuPasquier, L., and Wabl, M. R. (1976).In “The Generation of Antibody Diversity: A New Look” (A. J. Cunningham, ed.), p. 151. Academic Press, New York. Dwyer, J. M., and MacKay, J. R. (1972a). Clin. E x p . Immunol. 10,581. Dwyer, J. M. and MacKay, J. R. (197211).Immunology 23, 871. Dwyer, J. M., and Warner, N. L. (1971).Nature (London),New Biol. 229,210. Edelman, G. M., Rutishauser, U., and Millette, C. F. (1971). Proc. Natl. Acad. Sci. U.S.A. 68,2153. Eichmann, K. (1972).E u r . J . lmmunol. 2,301. Eichmann, K. (1973).1. E x p . Med. 137,603. Eichmann, K. (1975).Immunogenetics 2,491. Eichmann, K. (1977).i n “Regulation of the Immune System: Genes and the Cells in which they Function” (E. E. Sercarz and L. A. Herzenbert, eds.). Academic Press, New York (in press). Eichmann, K., and Berek, C. (1973). Eur. J . Immunol. 3,599. Eichmann, K., and Kindt, T. J. (1971).J.E x p . Med. 134,532. Elson, C. L., Singh, J., and Taylor, R. B. (1973).Scund. J. lmmunol. 2, 143. Fazekas d e St. Groth, S., and Webster, R. G. (1966).J.E x p . Med. 124,331. Ferrarini, M., Kent, S. P., Munio, A., Kelius, A. S., Cathy, D., and Coombs, R. A. (1973). Eur. J . lminunol. 3,213. Fraser, B. A., Johnstone, A. P., and Kindt, T. J. (1977). Cold Spring Horbor Symp. Quant. B i d . 41, 689. Fu, S. M., Winchester, R. M., and Kunkel, H. G . (1974).J.E x p . Med. 139,451. Gearhart, P. J., Sigal, N. H., and Klinman, N. R. (1975a).j.E x p . Med. 141,56. Gearhart, P. J., Sigal, N. H., and Klinman, N. R. (197%).Proc. Natl. Acad. Sci. U.S.A. 72, 1707. Gearhart, P. J., Sigal, N. H., and Klinman, N. R. (1977).J . E x p . Med. 145,876. Gelfand, M. C., Elfenbein, F. M., and Paul, W. E. (1974).J. E x p . Med. 139, 1125. Gerhard, W. (1976).J . E x p . Med. 144,985. Gerhard, W. (1978).In “Topics in Infectious Diseases,” Vol. 3. Springer-Verlag (in press). Gerhard, W., Braciale, T. J., and Klinman, N. R. (1975).’Eur.J . Immunol. 5,720. Gershon, H., Bruminger, S., Sela, M., and Feldman, M. (1968).J . E x p . Med. 128,223. Gibson, D. (1976).J . E x p . Med. 144,298. Goding, J. W., Warr, G. W., and Warner, N. L. (1976).Proc. Natl. Acad. Sci. U.S.A. 73, 1305. Goidl, E., and Siskind, G. W. (1974).J.E x p . Med. 140, 1285. Goidl, E., Klass, J., and Siskind, G. W. (1976).J.E x p . Med. 143, 1503. Gottlieb, P. D., and Durda, P. J. (1977).Cold Spring Harbor Symp. Quant. Biol.41,805. Greaves, M. F., and Hogg, N. M. (1971).Prog. lmmunol. 1, 111. Greaves, M., and Moller, E. (1970). Cell. lmmunol. 1,372. Green, I., Vassalli, P., Nussenzweig, V., and Benacerraf, B. (1967).J.E x p . Med. 125,511. Gregory, C . J., and Lajtha, L. G. (1968).Nature (London)218, 1979. Grey, H. M., Mannik, M., and Kunkel, H. G. (1965).1.E x p . Med. 121,561. Haas, W. (1975).J. E x p . Med. 141, 1015. Haas, W., and Layton, J. E. (1975).J.E x p . Med. 141, 1004. Haber, E. (1964).Proc. Natl. Acad. Sci. U.S.A. 52, 1099. Haber, E., Margolies, M. N., and Cannon, L. E. (1977). Cold Spring Harbor Symp. Quant. B i d . 41,647. Haglund, H. (1971).Methods Biochem. Anal. 19, 1. Haimovich, J., and DuPasquier, L. (1973).Proc. Natl. Acad. Sci. I1.S.A. 70, 1898.

332

NOLAN H. SIGAL AND NORMAN R. KLINMAN

Hammerling, G., and McDevitt, H. 0. (1974).J. Zmmunol. 112, 1726. Hart, D. A., Wang, A. L., Pawlak, L. L., and Nisonoff, A. (1972).J.E x p . Med. 135,1293. Haskill, J. S . , and Axelrod, M. A. (1971).Nature (London),New B i d . 231,219. Hayward, A. R., and Soothill, J. F. (1972).In “Ontogeny of Acquired Immunity,” p. 261. Assoc. Sci. Pub., Amsterdam. Henry, C., Kimura, J., and Wofsy, L. (1972). Proc. Natl. Acad. Sci. U S A . 69,34. Herzenberg, L. A. (1973). In “Genetic Control of Immune Responsiveness” (H. 0. McDevitt and M. Landy, eds.), p. 171. Academic Press, New York. Hill, S. W., and Sercarz, E. E. (1975). Eur. J. Zmmunol. 5,317. Hiramoto, R. N., and Hamlin, M. (1965).J.Zmmunol. 95,214. Hood, L., Barstad, P., Loh, E., and Nottenburg, C. N. (1974).In “The Immune System; Genes, Receptors, Signals” (E. E. Sercarz, A. R. Williamson, and C. F. Fox, eds.), p. 119. Academic Press, New York. Hood, L., Barstad, P., Eaton, B., Farnsworth, V., Loh, E., Hubert, J., Silver, J., and Weipert, M. (1977).Cold Spring Harbor Symp. Quant. Biol. 41,817. Howard, J. G., Elson, J., Christie, G. H., and Kinsky, R. G. (1969). Clin. E x p . Immunol. 4,41. Humphrey, J. H., and Keller, H. V. (1971).In “Developmental Aspects of Antibody Formation and Structure” (J. Sterzl and I. hiha, eds.), 2nd ed., Vol. 2, p. 485. Academic Press, New York. Humphrey, J. H., Roelants, G., and Willcox, N. (1971).In “Cell Interactions and Receptors Antibodies in Immune Responses” (0. Makela, A. Cross, and T. Kensuen, eds.), p. 123. Academic Press, New York. Imanishi, T., and Makela, 0. (1973). Eur. J. Zmmunol. 3,323. Imanishi, T., and Makela, 0. (1974).J.E x p . Med. 140, 1498. Imanishi, T., and Makela, 0. (1975).J.E x p . Med. 141,840. Ivanyi, J. (1975). Immunology 28, 1007. Jaton, J.-C., Klinman, N. R., Givol, D., and Sela, M. (1968).Biochemistry 7,4185. Jerne, N. K. (1955). Proc. Natl. Acad. Sci. U.S.A.41,849. Ju, S.-T., Gray, A,, and Nisonoff, A. (1977).J. E x p . Med. 144, 1294. Julius, M. H., and Henenberg, L. A. (1974).J.E x p . Med. 140,904. Julius, M. H., Masuda, T., and Herzenberg, L. A. (1972).Proc. Natl. Acad. Sci. U.S.A. 69, 1934. Julius, M. H., Janeway, L., and Henenberg, L. A. (1976). Eur. J . Zmmunol. 6,288. Karush, F. (1962).J.Pediatr. 60, 103. Kennedy, J. C., Till, J., Siminovitch, L., and McCulloch, E. A. (1965). Proc. SOC. E x p . Biol. Med. 120,868. Kenney, J. J. (1977).J. Supramol. Struct. 6, Suppl. 1,224 (abstr.). Kincaide, P. W., and Ralph, P. (1977). Cold Spring Harbor Symp. Quant. B i d . 41,245. Kindt, T. J. (1975).Ado. Immunol. 21,35. Kindt, T. J., Seide, R. K., Bokisch, V. A., and Krause, R. M. (1973).J.Erp. Med. 138,522. Klein, E., Esekland, T., Inoue, M., Strom, R., and Johansson, B. (1970).Exp. Cell Res. 62, 133. Klinman, N. R. (1969). Immunochemistry 6, 757. Klinman, N. R. (1971a).J.Immunol. 106, 1330. Klinman, N. R. (1971b).J.Zmmunol. 107,934. Klinman, N. R. (1971c).J.Zmmunol. 106, 1345. Klinman, N. R. (1972).J.Exp. Med. 136, 241. Klinman, N. R., and Aschinazi, G. (1971).J.Immunol. 106, 1338. a i n m a n , N. R., and Press, J. L. (1975a). Transplant. Reu. 24,41. Klinman, N. R., and Press, J . L. (1975h).J. E x p . Med. 141, 1133.

T H E B-CELL CLONOTYPE REPERTOIRE

333

Klinman, N. R., and Press, J. L. ( 1 9 7 5 ~ )Fed. . Proc., Fed. Am. Soc. E x p . Biol. 34,47. Klinman, N. R., Press, J. L., and Segal, G. (1973).J. E x p . Med. 138, 1276. Klinman, N. R., Pickard, A,, Sigal, N. H., Gearhart, P. J., Metcalf, E. S., and Pierce, S. K. (1976).Ann. Zmmunol. (Paris) 127c, 488. Klinman, N. R., Sigal, N. H., Metcalf, E. S., Pierce, S. K., and Gearhart, P. J. (1977). Cold Spring Harhor Symp. Quant. Biol. 41, 165. Kluskens, L., Lee, W., and Kohler, H. (1975).Eur. J. Zmmunol. 5,489. Kobel, H . R., and DuPasquier, L. (1975).Zmmunogenetics 2,87. Kohler, G . (1976).Eur. J. Zmmunol. 6, 340. Kohler, G., and Melchers, F. (1972). Eur. J . Immunol. 2,453. Kreth, H. W., and Williamson, A. R. (1973).Eur. J . Zmmunol. 3, 141. Kuettner, M. G., Wang, A., and Nisonoff, A. (1972).J . E x p . Med. 135,579. Kunkel, H. G . , Mannik, M., and Williams, R. C. (1963). Science 140, 1218. Landahl, C. (1976).Eur. J. Zmmunol. 6, 130. Lawrence, D. A., Spiegelbery, H. L., and Weigle, W. 0. (1973).J . E x p . Med. 137,470. Lee, W., Cosenza, H., and Kohler, H. (1974).Nature (London)247,55. Lefkovitz, I. (1972).Eur. J. Zmmunol. 2, 360. Lefkovitz, I. (1975).Contemp. Top. Microbiol. Zmmunol. 65, 21. Leon, M., and Young, N. (1971). Biochemistry 10, 1424. Liacopoulos, P., Couderc, J., and Bleux, C. (1976).Ann. Zmmunol. (Paris) 127c, 519. Lieberman, R., and Humphrey, W. (1971).Proc. Natl. Acad. Sci. U . S. A . 68,2510. Lieberman, R., and Potter, M. (1976). Zmmunogenetics 4,414. Liebeman, R., Potter, M., Mushinski, E., Humphrey, W., and Rudikoff, S . ‘(1974).J. E x p . Med. 139,983. Lieberman, R., Potter, M., and Humphrey, J. (1976). Zmmunogenetics 4,417. Loor, F., Forni, L., and Pernis, B. (1972). Eur. J . Zmmunol. 2,203. Luzzati, A., Lefkovitz, I., and Pernis, B. (1973a).Eur. J. Zmmunol. 3,632. Luzzati, A., Lefkovitz, I., and Pernis, B. (19731~). Eur. J . Zmmunol. 3,636. Luzzati, A. L., Tosi, R. M., and Carbonara, A. 0. (1970).]. E x p . Med. 132, 199. Lydyard, P. M., Grossi, C., and Cooper, M. (1976).J.E x p . Med. 144,79. McCarthy, M. M., and Dutton, R. W. (1975).J . Zmmunol. 115,1327. McCarthy, M . M., Dutton, R. W., and Taylor, B. A. (1976). Zmmunogenetics 4,418. Mchlichael, A. J., Phillips, J. A,, Williamson, A. R., Imanishi, T., and Makela, 0. (1975). Zmmunogenetics 2, 161. Makela, 0. (1964a). Immunology 1, 19. Makela, 0. (1964b).Ann. Med. E x p . B i d . Fenn. 42, 152. Makela, O . , and Cross, A. (1970). Prog. Allerg!{ 14, 145. Makela, O., and Karjalainen, K. (1977).Transplant. Rev. 34, 119. Makela, O., and Nossal, G. J. V. (1961).J. Zmnmnol. 87,447. Makela, O., Julin, M., and Becker, M. (1976).J.E x p . Med. 143,316. Makela, O., Karjalainen, K., Julin, M., and Catton, A. (1977a). Cold Spring Harhor Symp. Qunnt. Biol. 41, 735. Makela, O., Julin, M., and Karjalainen, K. (1977b). Zmmunogenetics 4,417. Makinodan, J., Kastenbaum, M., and Peterson, W. J. (1962).J.Zmmunol. 88,31. Marchalonis, J . J. (1974).In “The Immune System: Genes, Receptors, Signals” (E. E. Sercarz, A. R. Williamson, and C. F. Fox, eds.), p. 141. Academic Press, New York. Marchalonis, J. J., and Cone, R. (1973).Transplant. Reu. 14, 3. Marchalonis, J. J., Cone, R., and Atwell, J. (1972).J. E x p . Med. 135, 956. Melchers, F. and Phillips, R. A. (1977).Cold Spring Harbor Symp. Quant. B i d . 41 (in press). Metcalf, E. S., and Klinman, N. R. (1976).J. E x p . Med. 143, 1327.

334

NOLAN H. SIGAL AND NORMAN R. KLINMAN

Metcalf, E. S., Sigal, N. H., and Klinman, N. R. (1977).J . E x p . Med. 145, 1382. Miller, A., DeLuca, D., Decker, J., Ezzell, R., and Sercarz, E. E. (1971).Am. J . Pathol. 65,451. Miller, H. C., and Cudkowicz, G. (1970).J.E x p . Med. 132, 1122. Miller, J. F. A. P., Mitchell, G . F., and Weiss, N. S. (1967). Nature (London)214,992. Milstein, C., Cotton, R. G. H., and Secher, D. S. (1974). Ann. Zmmunol. (Paris) 125c, 287. Milstein, C., Adetugbo, K., Kerr, D., Kohler, G., and Secher, D. S. (1977). Cold Spring Harbor Symp. Quant. Biol. 41,793. Mitchison, N. A. (1967). Cold Spring Harbor Symp. Quant. Biol. 32,431. Modabber, F., Morikawa, S., and Coons, A. H. (1970). Science 170, 1102. Moller, E., and Sjoberg, 0. (1972). Transplant. Rev. 8,26. Moller, E., Bullock, W. W., and Makela, 0. (1973).Eur. J . Zmmunol. 3, 172. Moller, C. (1961).J.E x p . Med. 114,415. Moller, G . (1975). Transplant. Reu. 23, 126. Moller, C., and Michael, C. (1971). CeEl. Zmmunol. 2,309. Montgomery, P. C., and Wiliamson, A. R. (1972).j. Zmmunol. 109, 1036. Montgomery, P. C., Rockey, J. H., and Williamson, A. R. (1972). Proc. Nut. Acad. S c i . U.S.A. 69,228. Montgomery, P. C., Kahn, R. L., and Skandera, C. H. (1975a).J.Zmmunol. 115,904. Montgomery, P. C., Skandera, C. A., and Kahn, R. L. (1975b).Nature (London)256,138. Morrison, S. L., Baumal, R., Birshtein, B. K., Kuehl, W. M., Preud’homme, J.-L., Frank, L., Jasek, T., and Scharff, M. D. (1974).In “Cellular Selection and Regulation in the Immune Response” (G. M. Edelman, ed.), p. 253. Raven, New York. Mosier, D., and Johnson, B. (1975).J.E x p . Med. 141,216. Mudd, S. (1932).J . Zmmunol. 23,423. Naor, D., and Sulitzeanu, D. (1967).Nature (London)214,687. Neauport-Sautes, C., Lilly, F., Silvestre, D., and Kourilsky, F. M. (1973).J . E x p . Med. 137, 511. Nisonoff, A. (1977). Reported at the Homogeneous Antibody Workshop, Bethesda, Maryland. Nisonoff, A., and Bangasser, S. A. (1975).Transplant. Reu. 27, 100. Nisonoff, A., Hopper, J. E., and Spring, S. B. (1975).The Antibody Molecule, pp. 444496. Academic Press, New York. Nisonoff, A., Ju, S.-T., and Owen, F. L. (1977). Transplant. Reu. 34,89. Nossal, G. J. V. (1960). B r . J . E x p . Pathol. 41, 89. Nossal, G. J. V., and Lederberg, J. (1958).Nature (London) 181, 1419. Nossal, G. J. V., and Pike, B. (1973). Immunology 25,33. Nossal, C. J. V., Stocker, J. W., Pike, B., and Coding, J. W. (1977). Cold Spring Harbor Symp. Quant. Biol. 41,237. Nota, N. R., Liacopoulos-Briot, M., Stiffel, C., and Biozzi, G. (1964).C. R . Hebd. Seances Acad. S c i . 259, 1277. Osoba, D. (1970).J.E x p . Med. 132,368. Oudin, J. (1966). Proc. R. SOC. London, Ser. B 166,207. Owen, J., Sigal, N. H., and Klinman, N. R. (1978).In preparation. Owen, J. J. T. (1977). Cold Spring Harbor Symp. Quant. Biol. 41, 129. Owen, J. J. T., Raff, M., and Copper, M. (1975).Eur. J . Immunol. 5,468. Paraskevas, F., Lee, S.-T., and Israels, L. G. (1971).J.Immunol. 106, 160. Paul, W. E., Yoshida, T., and Benacerraf, B. (1970).J.Zmmunol. 105,314. Pauling, L. (194O).J.Am. Chem. S O C . 62, 2643.

T H E B-CELL CLONOTYPE REPERTOIRE

335

Pawlak, L. L., Mushinski, E., Nisonoff, A., and Potter, M. (1973).J.Exp. Med. 137,22. Phillips, B., and Roitt, J. M. (1973).Nature (London)New Biol. 241,254. Pickard, A., and Klinman, N. R. (1978). In preparation. Pilarski, L. M., and Cunningham, A. J . (1974).E u r . J . Zmmunol. 4, 762. Pilarski, L. M., and Cunningham, A. J. (1975).E u r . J . Zmmunol. 5, 10. Pink, J. R. L., and Askonas, B. (1974). Eur. J . Immunol. 4, 426. Pink, J. R. L., and Skvaril, F. (1975).FEBS Lett. 58,207. Pisetsky, D., Fathman, C. G., and Sachs, D. H. (1977).Zmmunogenetics 4,415. Playfair, J. H. L. (1968). Zmmunology 15,35. Playfair, J. H . L., Papermaster, B., and Cole, J . J. (1965). Science 149, 998. Potter, M. (1972).Physiol. Reo. 52,631. Potter, M., and Liebennan, R. (1970).J. E x p . Med. 132, 737. Press, J. L., and Klinman, N. R. (1973a).Zmmunochemistry 10, 621. Press, J. L., and Klinman, N. R. (1973b).J.Zmmunol. 111,829. Press, J. L., and Klinman, N. R. (1974). E u r . J. Zmmunol. 4, 155. Preud’homme, J. L., Neauport-Sautes, C., Piaf, S., Silvestre, D., and Kourilsky, F. M. (1972).E u r . J . Zmmunol. 2,297. Quintans, J., and Lefkovits, I. (1973). E u r . J. Zmmunol. 3,392. Quintans, J., and Lefkovits, I. (1974a).E u r . J . Zmmunol. 4, 614. Quintans, J., and Lefkovits, I. (197413).E u r . J . Zmmunol. 4, 617. Quintans, J., and Lefkovits, I. ( 1 9 7 4 ~J) . Zmmunol. 113, 1973. Raff, M . C. (1977). Cold Spring Harbor Symp. Quant. Biol. 41, 159. RafF, M . C., Steinberg, M., and Taylor, R. B. (1970). Nuture (London) 225,553. RafT, M. C., Feldmann, M., and DePetris, S. (1973).J.Exp. Med. 137, 1024. Richards, F. F., Amzel, L. M., Konigsberg, W. H., Manjula, B. N., Poljak, R. J., Rosenstein, R. W., Saul, F., and Varga, J. M. (1974).In “The Immune System: Genes, Receptors, Signals” (E. E. Sercarz, A. R. Williamson, and C. F. Fox, eds.), p. 53. Academic Press, New York. Richards, F. F., Konigsberg, W. H., and Rosenstein, R. W. (1975).Science 187, 130. Roelants, G. (1972). Contemp. Top. Microbiol. Zmmunol. 59, 135. Roelants, G., and Askonas, B. A. (1971).E u r . J . Zmmunol. 1, 151. Roelants, G., Farni, L., and Pernis, B. (1973).J . E x p . Med. 137, 1060. Rolley, R. T., and Marchalonis, J. (1972). Transplantation 14, 118. Rotman, B., and Cox, D. R. (1971).Proc. Natl. Acad. Sci. U.S.A. 68,2377. Rowe, D., Hug, K., Forni, L., and Pernis, B. (1973).]. E x p . Med. 138, 965. Rowlands, D. T., Jr., Blakeslee, D., and Angala, E. (1974).J.Zmmunol. 112,2148. Rutishauser, U., Millette, C., and Edelman, G. M. (1972).Proc. Natl. Acad. Sci. U.S.A. 69, 1596. Rutishauser, U., D’Eustachio, P., and Edelman, G. M. (1973). Proc. Natl. Acad. Sci. U.S.A. 70,3894. Sakato, N., and Eisen, H. (1975).J.E x p . Med. 141, 1411. Schlossman, S. F., and Hudson, J . (1973).J.Zmrnunol. 110, 313. Schreiner, G. F., and Unanue, E. R. (1976).Adu. Zmmunol. 24,38. Scott, D. W. (1976).J.E x p . Med. 144,69. Secher, D. S., Cotton, R. G., Cowan, N. J., and Milstein, C. (1974). In “The Immune System: Genes, Receptors, Signals” (E. E. Sercarz, A. R. Williamson, and C. F. Fox, eds.), p. 353.Academic Press, New York. Shearer, G. M., and Cudkowicz, G. (1969).J . E x p . Med. 129,935. Shearer, G. M., Cudkowitz, G., Connell, M. S. J., and Priore, R. L. (1968).J.E x p . Med. 128,437.

336

NOLAN H. SIGAL AND NORMAN R. KLINMAN

Shearer, G. M., Cudkowicz, G., and Priore, R. L. (1969).J.E x p . Med. 129, 185. Shearer, G. M., Moses, E., and Sela, M. (1973).In “Genetic Control of Immune Responsiveness” (H. 0. McDevitt and M. Landy, eds.), p. 35. Academic Press, New York. Sher, A., and Cohn, M. (1972).Eur. J . Zmmunol. 2,319. Sher, A., Lord, E., and Cohn, M. (1971).J.Zmmunol. 107, 1226. Sherwin, W. K., and Rowlands, D. T., Jr. (1974).]. Zmmunol. 113, 1353. Sherwin, W. K., and Rowlands, D. T., Jr. (1975).J:Zmmunol. 115, 1549. Sidman, C. L., and Unanue, E. (1975).J.Zmmunol. 114, 1730. Sigal, N. H. (1977a).J . E x p . Med. 146,282. Sigal, N. H. (1977b).J.Zmmunol. 119, 1129. Sigal, N. H., Gearhart, P. J., and Klinman, N. R. (1975).J.Zmmunol. 68, 1354. Sigal, N. H., Gearhart, P. J., Press, J. L., and Klinman, N. R. (1976). Nature (London) 259, 51. Sigal, N. H., Owen, J., and Klinman, N. R. (1977a). In preparation. Sigal, N. H., Cancro, M. P., and Klinman, N. R. (1977b). In “Regulation of the Immupe System: Genes and the Cells in which They Function” (E. E. Sercarz and L. A. Herzenberg, eds.), p. 217. Academic Press, New York. Sigal, N. H., Pickard, A. R., Metcalf, E. S., Gearhart, P. J., and Klinman, N. R. (1977~). J . E x p . Med. 146,933. Silverstein, A. M., Uhr, J. W., Kraner, K. L., and Lukes, R. J. (1963).J.E x p . Med. 117, 799. Sirishinha, S., and Eisen, H. (1971). Proc. Natl. Acad. Sci. U.S.A. 68,3130. Sjoberg, J., and Moller, E. (1970).Nature (London)228,780. Smith, G., Hood, L., and Fitch, W. (1971).Annu. Reu. Biochem. 40,969. Spear, P. A., and Edelman, G. M. (1974).J.E x p . &fed.139,249. Spear, P. A., Wang, A., Rutishauser, U., and Edelman, G. M. (1973).J.Exp. Med. 138, 557. Sterzl, J., and Silverstein, A. M. (1967). Ado. Zmmunol. 6,337. Stout, R. D., and Johnson, A. G. (1972).J . E x p . Med. 135,45. Sulitzeanu, D., and Naor, D. (1969). Znt. Arch. Allergy Appl. Zmmunol. 35,564. Taylor, R. B., Duffus, W. P., Raff, M. C., and DePetris, S. (1971).Nature (London)New Biol. 233,225. Trenker, E., and Riblet, R. (1975).J.E x p . Med. 142, 1121. Trentin, J., Wolf, N., Chang, V., Fahlberg, W., Weiss, D., and Bonhag, R. (1967).J.Zmmunol. 98, 1326. Tung, A. S., and Nisonoff, A. (1975).J . Exp. Med. 141, 112. Unanue, E. R. (1971).J.Zmmunol. 107, 1168. Unanue, E. R., Perkins, W., and Karnovsky, M. (1972)./. E x p . Med. 136,885. Vann, D. C., and Cambell, P. A. (1970).J.Zmmunol. 105, 1584. Vitetta, E. S., and Uhr, J. W. (l972).J.E x p . Med. 136,676. Vitetta, E. S., and Uhr, J. W. (1973). Transplant. Reo. 14,50. Vitetta, E., Forman, J., and Kettman, J. (1976).]. E x p . Med. 143, 1055. Vrana, M., Rudikoff, S., and Potter, M. (1977). Zmmunogenetics 4,424. Walters, C. S., and Wigzell, H. (1970).J . E x p . Med. 132, 1233. Warner, N. L. (1974).Ado. Zmmunol. 19,67. Whitney, P. L., and Tanford, L. (1965). Proc. Natl. Acad. Sci. U.S.A. 53,524. Wigzell, H., and Anderson, B. (1969).J . E x p . Med. 129,23. Wigzell, H., and Makela, 0. (1970).J. E x p . Med. 132, 110. Williams, R. C., Jr., Kunkel, H. G . , and Capra, J. D. (1968). Science 161,379. Williamson, A. R. (1971). Eur. J . Zmmunol. 1,390.

THE B-CELL CLONOTYPE REPERTOIRE

337

Williamson, A. R. (1976).Annu. Reu. Biochem. 3, 146. Williamson, A. R., and Askonas, B. A. (1972).Nature (London) 238,337. Williamson, A. R., Salaman, M., and Kreth, H. (1973).Ann. N . Y. Acad. Sci. 209,210. Wofsy, L., Kimura, J., and Truffa-Bachi, P. (1971).J.Immunol. 107, 725. Wybrow, G . M., and Berryman, I. L. (1973). Eur. /. Immunol. 3, 146. Yung, L., Wyn-Evans, T., and Diener, E. (1973). Eur. J . Immunol. 3,224. Zaalberg, 0. B. (1964).Nature (London)202, 1231.

This Page Intentionally Left Blank

Subiect Index A

Anaphylatoxin(s) characterization chemical properties, 21-38 isolation, 14-16 physical properties, 17-21 formation and control, 6-14 functional behavior cellular effects, 30-34 chemotaxis, 34-38 effects on smooth muscle and othei tissues, 38-42 systemic effects, 42-44 historical, 2-6 Anaphylatoxin(s) C3a and C5a, role in inflammation and acute shock 44-47 B

B-cell adult, defining the repertoire, 294-295 antigen-binding cell analysis, 295298 estimates of repertoire, 305-316 techniques to estimate precursor frequencies, 298-304 neonatal, defining the repertoire, 316317 antigen-binding cell analysis, 317319 in oitro studies of clonotype development, 323-327 in oioo studies of clonotype development, 319-323 C

Chemotaxis, anaphylatoxins and, 34-38 Clonal selection hypothesis, 256-257 contradictions to, 271-276 immunoglobulin nature of cell receptor. 257-258 integrity of B-cell clone, 269-271

precommitment of precursor cell, 263269 specificity of cell’s receptor, 258-263 Clonotype identification methods, 276-277 fine-specificity analysis, 280-284 idiotype, 285-291 isoelectric focusing, 277-280 other probes, 294 sequence, 291-294 H

H - 2 mutations biochemistry of, 96-99 contributions to immunology, 137-139 effect on immune functions allograft reaction and immunological tolerance, 108-112 CML to alloantigens, 112-118 CML to associative antigens, 118121 genetic control of immune response, 136-137 graft-versus-host disease, 127 MLR and GVHR, 121-126, production of allogeneis supernatant, 126-127 serology of H - 2 mutants, 127-136 Of

description, 95-96 distribution among regions, 91 genetic localization, 91-95 origin intragenic crossing-over, 99-101 point mutations, 105-106 regulator gene mutation, 101-105 rates, 84-87 interpretation of, 87-90 H - 2 syllabus chromosome 17,60 H-2 map, 60-62 H - 2 L locus, 62-63 Histocompatibility mutation rates, 79-84 339

340

SUBJECT INDEX

mutation study methods complementation test, 69-73 detection, 64-66 determination of mutation type, 6869 isolation, 66 linkage test, 66-67 naming the mutation, 67-68 progeny test, 66 I

Inflammation, anaphylatoxins and, 44-47 1

Lymphocytes analysis of B-cell receptor idiotypes, 211-217 analysis of T-cell receptor idiotypes, 2 17-234 anti-idiotypic reagents to idiotypic subspecificities, 199205 preparation against cell-surface receptors, 209-21 1 preparation of isogeneic reagents, 205-209 functional role of receptor idiotypes, 234-236 consequences of idiotype recognition, 24 1-246 idiotype recognition in immune system, 236-240 M

&Microglobulin, 148- 149 distribution of, 150-151 function of, 150 genetics of, 149-150 structure of, 151-154

Mutation(s), polymorphism and, 106-108 Mutation genetics, basic terms, 63-64 Mutation rates, general considerations, 73-79 5

17th chromosome I region, 167 functional relationships between IR genes and Ia antigens, 169-170 genetics, 168-169 structure, 171-177 tissue distribution of Ia antigens, 170- 171 K and D regions, 157-158 distribution, 159 function, 158 genetics, 158 structure, 160-167 region between H-2D and TLa, 181-

182 H-26,27,28 and 29, 182 Qa-1, 182 Qa-2, 182 S region, 177-178 distribution, 179 function, 178-179 genetics, 178 relation between murine Ss protein and human C-4, 180 structure of S c protein, 179-181 TLa antigenic modulations, 183-184 biochemistry of antigens, 184-185 definition and genetics, 182-183 tissue distribution, 183 T/t region genetics, 154-156 structure, 156-157 Shock, acute, anaphylatoxins and, 44-47 Smooth muscle, anaphylatoxins and, 3842

CONTENTS OF PREVIOUS VOLUMES

Volume 1

Antibody Production by Transferred Cells

CHARLESG . COCHRANE AND FRANK J. DIXON

Transplantation Immunity a n d To1era nce

M . HASEK, A. T. HRABA

LENGEROVA,

AND

Phagocytosis

DERRICKROWLEY

Immunological Tolerance o f Nonliving Antigens

Antigen-Antibody Reactions i n Helminth Infections

RICHARD T. SMITH

E. J. L. SOULSBY

Functions o f the Complement System

ABRAHAMG. OSLER

Embryological Development o f Antigens REED A. FLICKINCER

In Vitro Studies of the Antibody Response

ABRAM B. STAVITSKY

AUTHOR INDEX-SUBTECT INDEX

Duration of Immunity i n Virus Diseases

J. H . HALE Volume 3

Fate a n d Biological Action of Antigen-Antibody Complexes

In Vitro Studies o f the Mechanism of Anaphylaxis

WILLIAM 0. WEIGLE

K . FRANK AUSTEN A N D H . HUMPHREY

Delayed Hypersensitivity to Simple Protein Antigens

P. G. H. CELLA N D B. BENACERKAF The Antigenic Structure o f Tumors

JOHN

The Role o f Humoral Antibody in the Homograft Reaction

CHANDLER A. STETSON

P. A. CORER

Immune Adherence

AUTHOR INDEX-SUBJECT INDEX

D. S. NELSON Reaginic Antibodies

Volume 2

D. R. STANWORTH

Immunologic Specificity a n d Molecular Structure

N a t u r e of Retained Antigen a n d i t s Role i n immune Mechanisms

FREDKARUSII

DANH . CAMPBELL AND JUSTINE S. GAHVEY

Heterogeneity o f y-Globulins IOlIN

L.

FAHEY

Blood Groups i n Animals Other Than M a n

W. €I. STONE A N D M . R. I R W I N

The Immunological Significance of the Thymus

Heterophile Antigens a n d Their Significance i n the Host-Parasite Relationship

J . F. A. I? M I L L E R ,A. H . E. bfAHSHAI,L, A N D R. G . WHITE

C . R. JENKIN

Cellular Genetics of Immune Responses

C;. J . V. NOSSAL

AUTHOR INDEX-SUBJECT INDEX 34 1

342

C O N T E N T S O F PREVIOUS VOLUMES

Volume 4

Volume 6

Ontogeny a n d Phylogeny of Adaptive Immunity ROBERTA. GOOD A N D

Experimental Glomerulonephritis: Immunological Events a n d Pathogenetic Mechanisms

EMIL R. UNANUE AND FRANK J. DIXON

BEN W. PAPERMASTER Cellular Reactions i n Infection

EMANUEL SUTERA N D HANSRUEDYRAMSEIER Ultrastructure of Immunologic Processes

Chemical Suppression of Adaptive Immunity ANN E . GABRIELSON AND ROBERT A. GOOD

JOSEPH D . FELDMAN Cell Wall Antigens of Gram-Positive Bacteria

MACLYNMCCARTYAND STEPHENI. MORSE

WERNERBRAWN In Vifro Studies of Immunological Responses of lymphoid Cells

Structure a n d Biological Activity of Immunoglobulins SYDNEYCOHENAND

RODNEYR. PORTER Autoantibodies a n d Disease H. G . KUNKEL AND E. M.

Nucleic Acids as Antigens OTTO I. PLESClA A N D

RICHARDW. DUTTON Developmental Aspects of Immunity

JAROSLAVSTERZLAND ARTHURM. SILVERSTEIN

TAN

Effect o f Bacteria a n d Bacterial Products on Antibody Response

J. MUNOZ AUTHORINDEX-SUBJECTINDEX

Anti-anti bodies

PHILIP G . H . GELLA N D ANDREWS. KELUS Conglutinin a n d lmmunoconglutinins

P. J. LACHMANN AUTHORINDEX-SUBJECTINDEX

Volume 5 N a t u r a l Antibodies and the Immune Response

STEPHEN V. BOYDEN Immunological Studies with Synthetic Polypeptides

MICHAELSELA Experimental Allergic Encephalomyelitis a n d Autoimmune Disease

PHILIP Y. PATERSON The Immunology of Insulin

C. G . POPE Tissue-Specific Antigens

D. C. DUMONDE AUTHOR INDEX-SUBJECTINDEX

Volume 7 Structure a n d Biological Properties of Immunoglobulins

SYDNEYCOHENA N D CESAR M ILSTEIN Genetics o f Immunoglobulins i n the Mouse

MICHAELPOTTER A N D ROSE LIEBERMAN Mimetic Relationships between Group A Streptococci a n d Mammalian Tissues

JOHN B. ZABRISKIE Lymphocytes a n d Transplantation Immunity DARCY B. WILSON AND

R. E. BILLINCHAM

C O N T E N T S OF PREVIOUS VOLUMES Human Tissue Transplantation

Phy Io g eny of ImmunogIobul ins

JOHN P. MERRILL

HOWARDM. GREY

AUTHOR INDEX-SUBJECTINDEX

343

Slow Reacting Substance o f Anaphylaxis

ROBERTP. ORANGEAND K. FRANK AUSTEN Volume 8 Chemistry a n d Reaction Mechanisms of Complement

HANSJ . MULLER-EBERHARD Regulatory Effect of Antibody on the Immune Response JONATHAN W. UHR AND GORAN MOLLER The Mechanism of immunological Paralysis

D . W. DRESSER AND N. A. MITCHISON In Vitro Studies of Human Reaginic Allergy

Some Relationships among Hemostasis, Fibrinolytic Phenomena, Immunity, and the inflammatory Response

OSCARD. RATNOFF Antigens o f Virus-Induced Tumors

KARL HABEL Genetic a n d Antigenetic Aspects of Human Histocompatibility Systems

D. BERNARDAMOS AUTHORINDEX-SUBJECTINDEX

ABRAHAM G. OSLER,

Volume 1 1

LAWRENCEM. LICHTENSTEIN, A N D DAVIDA. LEVY

Electron Microscopy of the immunoglobulins

AUTHOR INDEX-SUBJECTINDEX

N. MICHAELGREEN Genetic Control o f Specific Immune Responses HUGH0. MCDEVITT AND

Volume 9 Secretory Immunoglobulins

BARUJBENACERRAF

THOMAS n. TOMASI, JR., A N D JOHNBIENENSTOCK

The Lesions in Cell Membranes Caused b y Complement

Immunologic Tissue Injury M e d i a t e d b y Neutrophilic Leukocytes

JOHN H . HUMPHREYA N D ROBERTR. DOURMASHKIN

CHARLES G. COCHRANE The Structure a n d Function of Monocytes a n d Macrophages

ZANVILA. COHN

PETER PERLMANN A N D HOLM

GORAN

Transfer Factor

The Immunology and Pathology o f NZB Mice

J . B. HOWIEA N D

Cytotoxic Effects of Lymphoid Cells In Vitro

n. J. HELYER

AUTHOR INDEX-SUBJECTINDEX

H. S. LAWRENCE immunological Aspects o f Malaria Infection

IVOR N. BROWN AUTHOR INDEX-SUBJECTINDEX

Volume 10 C e l l Selection by Antigen i n the immune Response

GREGORYW. SISKINDA N D BARUJBENACERPAF

Volume 12 The Search for Antibodies with Molecular Uniformity

RICHARDM. KRAUSE

344

CONTENTS O F PREVIOUS VOLUMES

Structure a n d Function of yM Macroglobulins

HENRYMETZCER Transplantation Antigens R. A. REISFELT AND

B. D. KAHAN

The Role of Bone Marrow i n the Immune Response

NABIH I. ABDOU AND MAXWELLRICHTER Cell Interaction in Antibody Synthesis

D. W. TALMAGE, J. h D O V I C H , A N D H. HEMMINCSEN The Role o f Lysosomes i n Immune Responses

GERALDWEISSMANN A N D PETERDUKOR Molecular Size a n d Conformation of Immunoglobulins KEITH J. DORFUNGTON AND

Volume 14 lmmunobialogy o f Mammalian Reproduction ALAN E. BEER AND

R. E. B~LLINCHAM Thyroid Antigens and Autoimmunity

SIDNEY SHULMAN Immunological Aspects o f Burkitt’s Lymphoma

GEORGEKLEIN Genetic Aspects of the Complement System CHESTER A. ALPER AND

FREDS. ROSEN The Immune System: A M o d e l for Differentiation in Higher Organisms L. HOODAND J. PRAHL

AUTHOR I N D E X ~ U B J E INDEX CT

CHARLES TANFORD AUTHOR INDEX-SUBJECT INDEX

Volume 15 Volume 13 Structure a n d Function o f Human Immunoglobulin E HANSBENNICH A N D GUNNAR 0. JOHANSSON

s.

Individual Antigenic Specificity of Immunoglobulins JOHN E. HOPPERAND

ALFRED NISONOFF In Vitra Approaches to the Mechanism o f Cell-Mediated Immune Reactions

The Regulatory Influence o f Activated T Cells on B C e l l Responses t o Antigen DAVID ti. KATZ AND

BARUJ BENACERPAF The Regulatory Role o f Macrophages i n Antigenic Stimulation

E . R. UNANUE Immunological Enhancement: A Study o f Blocking Antibodies JOSEPH D. FELDMAN

BARRY R. BLOOM Immunological Phenomena in leprosy a n d Related Diseases J. L. TURK AND

A. D. M. BRYCESON Nature a n d Classification of Immediate-Type Allergic Reactions

ELMERL. BECKER AUTHOR INDEX-SUBJECT INDEX

Genetics a n d Immunology o f Sex-Linked Antigens

DAVIDL. GASSERAND WILLYSK. SILVERS Current Concepts of Amyloid

EDWARDc. FRANKLIN AND DOROTHEA ZUCKER-FRANKLIN

AUTHOR INDEX-SUBJECT INDEX

CONTENTS OF PREVIOUS VOLUMES Volume 16 Human Immunoglobulins: Classes, Subclasses, Genetic Variants, a n d ldioty pes

J. I3. NATVIGA N D H . G . KUNKEL Immunological Unresponsiveness

WILLIAM0. WEIGLE Participation o f Lymphocytes i n Viral Infections

E. FREDERICK WHEELOCK AND STEPHEN T. TOY Immune Complex Diseases in Experimental Animals a n d Man C. COCHRANE AND D . KOFFLER

c.

The lmmunopathology of Joint Inflammation i n Rheumatoid Arthritis

345

Cell-Mediated Cytotoxicity, A l l o g r a f t Rejection, a n d Tumor Immunity

JEAN-CHARLES CEROTTINI AND K. THEODORE BRUNNER Antigenic Competition: A Review o f Nonspecific Antigen-Induced Suppression

HUGH F. PROS A N D DAVIDEIDINGER Effect o f Antigen Binding on the Properties o f Anti body

HENRYMETZGER Lymphocyte-Medi a t e d Cytotox icity a n d Blocking Serum Activity t o Tumor Antigens

KARL ERIK HELLSTROMAND INGEGERDHELLSTXOM AUTHOR INDEX-SUBJECTINDEX

NATHANJ. ZVAIFLER AUTHOR INDEX-SUBJECTINDEX Volume 19 Volume 17 Antilymphocyte Serum

EUGENE M. LANCE,P. B. MEDAWAR, A N D ROBERT N. TAUB In Vitro Studies of Immunologically Induced Secretion o f Mediators from Cells a n d Related Phenomena ELMER L. BECKER A N D

PETER M. HENSON Antibody Response to Viral Antigens

KEITH M. COWAN Antibodies t o Small Molecules: Biological a n d Clinical Applications VINCENT P. BUTLER, JR., AND

Molecular Biology o f Cellular Membranes with Applications to Immunology

S. J. SINGER Membrane Immunoglobulins a n d Antigen Receptors on B a n d T lymphocytes

NOEL L. WARNER Receptors for Immune Complexes on Lymphocytes

VICTOR NUSSENZWEIG Biological Activities of Immunoglobulins of Different Classes a n d Subclasses

HANS L. SPIEGELBERG SUBJECT INDEX

SAM M. BEISER AUTHORINDEX-SUBJECTINDEX

Volume 20

Volume 18

Hypervariable Regions, Idiotypy, and Antibody-Combining Site J. DONALDCAPRA AND

Genetic Determinants of Immunological Responsiveness DAVIDL. GASSER AND

WILLYS K. SILVERS

J. MICHAELKEHOE Structure a n d Function of the J Chain

MARIAN ELLIOTT KOSHLAND

346

CONTENTS O F PREVIOUS VOLUMES

Amino Acid Substitution a n d the Antigenicity of Globular Proteins

MORRIS REICHLIN The H-2 Major Histocompatibility Complex a n d the I Immune Response Region: Genetic Variation, Function, and Organization

DONALD c. SHREFFLER AND CHELLA S. DAVID

Delayed Hypersensitivty i n the Mouse

ALFRED J. CROWLE SUBJECTINDEX

Cellular Aspects of Immunoglobulin A

MICHAEL E. LAMM Secretory Anti-Influenza Immunity YA.

s. SHVARTSMAN AND

M . P. ZYKOV

SUBJECTINDEX Volume

23

Cellular Events i n the IgE Antibody Response

KIMISH~GEISHIZAKA Chemical ond Biological Properties of Some Atopic Allergens

Volume 21 X-Ray Diffraction Studies of Immunoglobulins

ROBERTO J. POLJAK Rabbit Immunoglobulin Allotypes: Structure, Immunology, a n d Genetics

THOMAS J. KINDT Cyclical Production of Antibody as a Regulatory Mechanism i n the Immune Response

WILLIAM 0. WIECLE Thymus-Independent B-Cell Induction a n d Paralysis

ANTONIO COUTlNHO AND GORAN M ~ L L E R

SUBJECTINDEX

Volume 22 The Role of Antibodies i n the Rejection a n d Enhancement of Organ Allografts

CHARLES B. CARPENTER, ANTHONYJ. F.D’APICE, AND ABUL K. ABBAS Bioaynthesis o f Complement

HARVEY R. COLTEN Graft-versus-Host Reactions: A Review

STEPHENC. GREBEAND J. WAYNE STREILEIN

T. P. KING Human Mixed-lymphocyte Culture Reaction: Genetics, Specificity, and Biological Implications

B o DUPONT, JOHN A. HANSEN, EDMOND J. YUNIS

AND

lmmunochemical Properties o f Glycolipids a n d Phospholipids

DONALDM. MARCUS AND GERALD A. SCHWARTINC SUBJECT INDEX

Volume 24 The Alternative Pathway of Complement Activation

0. GBTZE AND H. J. MULLER-EBERHARD Membrane a n d Cytoplasmic Changes i n B lymphocytes Induced by LigondkSurface lmmunag lobuli n Interaction

GEORGEF. SCHREINER AND EMILR. UNANUE lymphocyte Receptors for Immunoglobulin

HOWARDB. DICKLER Ionizing Radiation a n d the Immune Response

ROBERT E. ANDERSON AND NOEL L. WARNER

SUBJECTINDEX

347

CONTENTS OF PREVIOUS VOLUMES Volume 25

Current Status of Rat lmmunogenetics

DAVIDL. GASSER Immunologically Privileged Sites

F. BARKERAND R. E. BILLINCHAM CLYDE

M a j o r Histocompatibility Complex Restricted Cel I-Med iated Immunity GENE M. SHEARERAND ANNE-MARIE SCHMITT-VERHULST

n B

c a D 9 E O

F a 6 2 H 3 1 4

I S

Antigen-Binding Myeloma Proteins of Mice

MICHAELPOTTER Human lymphocyte Subpopulations

L. CHESS

AND

SUBJECTINDEX

s. F. SCHLOSSMAN

This Page Intentionally Left Blank

Advances in Immunology Volume 45

Read more

Advances in Immunology Volume 34

Read more

Advances in Immunology Volume 1

Read more

Advances in Immunology Volume 10

Read more

Advances in Immunology Volume 30

Read more

Advances in Immunology Volume 82

Read more

Advances in Immunology Volume 31

Read more

Advances in Immunology Volume 52

Read more

Advances in Immunology Volume 48

Read more

Advances in Immunology Volume 56

Read more

Advances in Immunology Volume 12

Read more

Advances in Immunology Volume 21

Read more

Advances in Immunology Volume 41

Read more

Advances in Immunology Volume 71

Read more

Advances in Immunology, Volume 97

Read more

Advances in Immunology Volume 84

Read more

Advances in Immunology Volume 43

Read more

Advances in Immunology Volume 73

Read more

Advances in Immunology, Volume 61

Read more

Advances in Immunology Volume 14

Read more

Advances in Immunology Volume 46

Read more

Advances in Immunology, Volume 106

Read more

Advances in Immunology Volume 19

Read more

Advances in Immunology Volume 99

Read more

Advances in Immunology Volume 81

Read more

Advances in Immunology Volume 4

Read more

Advances in Immunology Volume 80

Read more

Advances in Immunology Volume 64

Read more

Advances in Immunology Volume 62

Read more

Advances in Immunology Volume 28

Read more

Recommend Documents

Advances in Immunology Volume 45

ADVANCES IN Immunology EDITED BY FRANK J. DIXON Scripps Clinic and Research Foundation La Jolla, California ASSOCIATE E...

Advances in Immunology Volume 34

ADVANCES IN Immunology V O L U M E 34 CONTRIBUTORS TO THIS VOLUME MARILYN L. BALTZ ROBERT M. FRIEDMAN A. GONWA THOM...

Advances in Immunology Volume 1

...

Advances in Immunology Volume 10

CONTRIBUTORS TO THIS VOLUME D. BERNARD AMOS K. FRANK AUSTEN BARUJ BENACERRAF HOWARDM. GREY KARL HABEL ROBERTP. ORAN...

Advances in Immunology Volume 30

ADVANCES IN Immunology V O L U M E 30 CONTRIBUTORS TO THIS VOLUME M. J. BUCHMEIER HENRYN. CLAMAN PAULJ. CONLON F. ...

Advances in Immunology Volume 82

/RJLQ + 5HJLVWHU %URXJKWWR\RXE\ 8QLYHUVLW\RI0HOERXUQH,QIRUPDWLRQ'LYQ %URZVH 6HDUFK 0\6HWWLQJ...

Advances in Immunology Volume 31

CONTRIBUTORS T O THIS VOLUME HOWARD FLEIT PAULW. KINCADE ROBERT W. KOZAK IRAS. MELLMAN CHARLES E. MOODY, JR. FRANCIS W....

Advances in Immunology Volume 52

ADVANCES IN Immunology VOLUME 52 This Page Intentionally Left Blank ADVANCES IN Immunology EDITED BY FRANK J. DI...

Advances in Immunology Volume 48

ADVANCES IN Immunology EDITED BY FRANK J. DIXON Scripps Clinic and Research Foundation La Jolla, California ASSOCIATE ...

Advances in Immunology Volume 56

ADVANCES IN Immunology EDITED BY FRANK J. DIXON The Scripps Research Institute La Jolla, California ASSOCIATE EDITORS ...