Advances in Immunology Volume 23

ADVANCES IN

Immunology

VOLUME 23

CONTRIBUTORS TO THIS VOLUME Bo DUPONT A. HANSEN KIMISHIGEISHIZAKA T. P. KING DONALD M. MARCUS GERALD A. SCHWARTING EDMOND J. YUNIS JOHN

ADVANCES IN

Immunology E D I T E D BY

HENRY G. KUNKEL

FRANK J. DIXON

The Rockefeller University

Scrippr Clinic and Research Foundation

N e w York, N e w York

La Jolla, California

VOLUME 23

1976

ACADEMIC PRESS New York

Sun Francisco

A Subsidiary o f Harcourf Broce Jovanovich, Publishers

London

COPYRIGHT 0 1976, 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 published by ACADEMIC PRESS, INC. (LONDON) LTD. 24/28 Oval Road. London N W l

LIBRARY OF CONGRESS CATALOG

CARD

NUMBER:6 1 - 17057

ISBN 0-12-022423-2 PRINTED IN THE UNITED STATES OF AMERICA

CONTENTS LIST OF CONTRIBUTORS . PREFACE

.

.

.

.

. .

. .

.

. .

.

. .

. .

. .

. .

. .

.

.

vii

.

.

ix

.

.

.

.

.

. .

. .

. .

. . . .

1 3 12 20

Cellular Events in the IgE Antibody Response

KIMISHIGE ISHIZAKA

I. Introduction

.

.

.

.

.

.

11. Immunoglobulin E Antibody Formation in Viuo and in Vitro 111. Immunological Factors Essential for IgE Antibody Responses . . IV. Cellular Basis of IgE Antibody Responses .

V. Regulation of IgE Antibody Responses VI. Discussion and Summary . . . . . . . . . References

. . .

.

. . .

. . .

.

.

.

. . . . .

.

. . .

.

45

67 70

Chemical and Biological Properties of Some Atopic Allergens

T. P. KING I. 11. 111. IV. V. VI.

Introduction . . . . . . . . . Allergen Assay . . . . . . . . Chemical and Biological Properties of Some Allergens General Observations on Allergens . . . . Uses of Purified Allergens . . . . . . . . . . . . Concluding Remarks . References . . . . . . . . .

. . . . . .

. . . . . . . . .

. .

. .

. . .

. . . . . .

77 78 80 92 96 100

101

Human Mixed-lymphocyte Culture Reaction: Genetics, Specificity, and Biological Implications

Bo DUPONT,JOHNA. HANSEN,AND EDMONDJ. YUNIS I. Introduction: Major Histocompatibility System in Man . . 11. Serology of Human Leukocyte Alloantigens (HLA-A,B,C) . 111. Cell-Mediated Allogeneic Reactions in Vitro . . . . IV. Measurement of Antigenic Differences in Mixed-Lymphocyte C u h r e Reaction . , . . . . . . . . V. Single-Locus Concept for Mixed-Lymphocyte Culture Stimulation (HLA-D Locus) . . . . . . . . . . V

. .

. .

. .

108 110 119

.

.

124

.

.

130

vi

CONTENTS

.

VI Mixed-Lymphocyte Culture ( HLA-D ) Specificities Defined by HLA-D-Homozygous Typing Cells . . . . . . . VII . Genetic Control of Immune Response Related to Histocompatibility VIII . Mixed-Lymphocyte Culture As a Histocompatibility Test for Clinical Transplantation . . . . . . . . . IX Genetic Mapping of the HLA Complex on Chromosome C-6 . X . Conclusions . . . . . . . . . . . . References . . . . . . . . . . . .

.

. .

135 169

.

177 183 185 187

.

203 204 229 233 233

. . .

lmmunochemical Properties of Glycolipids a n d Phospholipids

DONALDM . MARCUSAND GERALDA. SCHWARTING

I. I1. 111. IV

.

Introduction . . Glycolipids . . Phospholipids . . Concluding Remarks References . .

SUBJECTINDEX.

.

.

. .

. .

. .

. .

. .

. .

. .

. .

. .

. .

. .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

241

CONTENTS OF PREVIOUS VOLUMES.

.

.

.

.

.

.

.

.

.

243

LIST

OF CONTRl BUTORS

Nurnlxm in parentheses indicate the pages on which the authors’ contributions begin.

Bo DUPONT,Tissue Typing Laboratory, Sloan-Kettering Institute for Cancer Research, New York, New York (107)

A. HANSEN, Tissue Typing Laboratory, Sloan-Kettering Institute for Cancer Research, New York, New York (107)

JOHN

KIMISHICE ISHIZAKA, Department of Medicine, The Johns Hopkins University School of Medicine at the Good Samaritan Hospital, Baltimore, Maryland (1) T. P. KING, The Rockefeller University, New York, New York (77)

DONALD M. MARCUS,Departments of Medicine, Microbiology and Zmmunology, Albert Einstein College of Medicine, Bronx, New York (203) GERALD A. SCHWARTING, Department of Microbiology and Immunology, Albert Einstein College of Medicine, Bronx, New York (203) EDMOND J. YUNIS, Department of Pathology and Laboratory Medicine, University of Minnesota Hospitals, Minneapolis, Minnesota ( 107)

This Page Intentionally Left Blank

PREFACE

The familiar and somewhat tiresome debate over the relative merits of fundamental versus applied research has if anything intensified in the last few years. This has occurred largely as a result of the greatly increased competition for funds that exists today. It has been fostered to a considerable degree by the “somewhat snobbish attitude of many academics to applied research.” The distinction is purely arbitrary; scientific knowledge is a continuum in which every component part can and does feed back on every other. Nowhere is this more clearly apparent than in the field of immunology, as exemplified by the articles in Volume 23. The first paper is by Dr. Kimishige Ishizaka, the individual primarily responsible for the basic work on IgE antibodies and their role in reaginic hypersensitivity. The initial definitive work was carried out in the human system, and the extension to the cellular regulation of IgE antibodies, the main topic of the review, was continued in various experimental animals. The important role of both helper and suppressor T cells in this regulation is quite apparent. I t is still uncertain whether the same cells are involved as those defined for the major immunoglobulin classes. Promising approaches to therapy derived from the animal-model work are discussed. The work of Dr. T. P. King, author of the second article, has centered on the chemistry of the allergens, a subject which has advanced markedly in the last few years, largely through his efforts. Ragweed pollen allergens have received the most attention, and antigen E, the dominant antigen involved in hypersensitivity, has been isolated and characterized in considerable detail. It consists of two non-identical polypeptide chains with molecular weights of approximately 26,000 and 13,000. Additional ragweed allergens have been isolated, but their significance relative to antigen E remains to be defined. Many other types of allergens have been isolated as well. Of special interest is the current active work on the chemical modification of these isolated proteins for possible therapeutic immunization. The third article is written by Drs. Dupont, Hansen, and Yunis, and deals primarily with the new and exciting developments in MLC typing in human histocompatibility studies. These workers have played a major role in placing this system on a firm scientific basis. The use of homozygous cells from specific individuals has made it possible to delineate a t least six different distinct MLC antigens, and there are clearly more. Some of these can also be recognized by B-cell-specific alloantisera and clearly relate to the Ia antigens of the murine system. I t is of special iX

X

PREFACE

interest that certain disease associations, as well as the genes involved in certain of the complement components, appear more closely linked to the MLC genes than to the other components of the HLA system. The last paper covers the somewhat neglected area of the immunology of lipids and glycolipids. The authors, Drs. Marcus and Schwarting, have had wide experience in this field and their contributions have played a major role in current recognition of the significance of these antigens. Suddenly, with the great expansion of interest in cell membranes, the glycolipids have assumed a particular importance and their study by immunological procedures as specific moieties of the cell membrane is receiving great emphasis. Much remains to be learned about the many different types of lipid antigens and their cross reactions, but this review provides the many interested investigators with an up-to-date treatment of the subject. HENRYG. KUNKEL FRANKJ. DIXON

ADVANCES IN

Immunology

VOLUME 23

This Page Intentionally Left Blank

Cellular Events in the IgE Antibody Response’ KlMlSHlGE ISHIZAKA Department of Medicine, The Johns Hopkinr University School of Medicine ot the Good Samaritan Hospital, Baltimore, Marylond

.

.

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

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

I. Introduction . . . . . . . 11. Immunoglobulin E Antibody Fomiation in Vioo and in Vitro . A. Kinetics of IgE Antibody Responses in Various Animal Species B. Helminth Infection and IgE Responses . . . C. Distribution of IgE-Forming Cells . . . D. Immunoglobulin E Antibody Response in Vitm . 111. Immunological Factors Essential for IgE Antibody Responses . A. Genetic Control of IgE Responses . . . B. Adjuvant for IgE Antibody Response . . . . C. Nature and Dose of Antigen . . . . . N.Cellular Basis of IgE Antibody Responses . A. Requirement for T and B Lymphocytes . . . . B. Type B Lymphocytes in IgE Antibody Response . C. Generation of a Helper Function for IgE Antibody Response D. Mechanisms of T Cell-B Cell Collaboration . . . V. Regulation of IgE Antibody Responses . A. Suppression by Humoral Antibodies . . . . . . , B. Unresponsiveness in IgE-B Cells C. Regulation by T Cells D. Experimental Model for Immunotherapy . . . . . VI. Discussion and Summary References . . .

.

.

.

. . . . . .

. .

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

. . . . . . . .

.

. . . .

.

1 3 3 6 9 11 12 12 15 17 20 20

23 28 36 45

45 48 50 62 67 70

I. Introduction

Since the discovery of IgE in the serum of hay fever patients (45), much progress has been made in the field of reaginic hypersensitivity. It is now established that reaginic hypersensitivity reactions in atopic diseases are mediated by IgE antibody [reviewed by Ishizaka and Ishizaka (%)I. Meanwhile, homocytotropic antibodies, which are similar to human IgE antibodies, were detected in experimental animals. Mota (109) and Binaghi et al. (11) first described production of rat “reaginic” antibodies after immunization with antigen plus Bordetella pertussis vaccine. Subsequently, antibodies that were capable of sensitizing homologous skin ‘Supported by research grants AI-11202 from the U.S. Public Health Service, GB-41443 from National Science Foundation, and a grant from John A. Hartford Foundation. This is publication No. 223 from the O’Neill Laboratories at the Good Samaritan Hospital.

1

2

KIMISHIGE ISHIZAKA

were found in rabbit (46, 91, 189), dog (133), mouse (110, 112, la), monkey (47), guinea pig (88), pig ( 7 ) , and cattle (39). The physicochemical properties of the reaginic antibodies in experimental animals are similar to those of human IgE, and their molecular sizes are distinct from those of immunoglobulins of the other isotypes. It was also found in each species that the antigenic structure of the immunoglobulin class to which the reaginic antibody belongs was different from IgG, IgA, and IgM. More recently, Bazin et al. ( 9 ) reported that the inbred Lou/WST rat strain presented a high incidence of spontaneous ileocecal immunocytoma, which secreted monoclonal immunoglobulins, and that nearly one-third of them represented a unique isotype to which reaginic antibody belonged. From the biological viewpoint, human IgE and reaginic antibodies in experimental animals share common characteristics. Once skin sites of homologous species are passively sensitized with the antibody, sensitization persists for 2 to 3 weeks. This property and the molecular size of reaginic antibodies are distinct from those of another type of skinsensitizing antibodies that belong to a subclass of IgG. A crucial role of IgE antibody in atopic diseases suggested that prevention or suppression IgE antibody formation is beneficial for atopic individuals. Identification of IgE antibodies in experimental animals provided an important tool for studying this problem. Fortunately, the scope of our knowledge on the mechanisms of antibody response has considerably broadened in the past decade [reviewed by Katz and Benacerraf (64)]. It is firmly established that collaboration of two distinct types of lymphocytes, i.e., bone marrow-derived precursors of antibody-forming cells ( B cells) and thymus-derived lymphocytes ( T

4

8 1970

1 2 4

8 1971

1 2 4 1972

FIG.1. Titers of IgE and IgC antibody in the serum of ragweed-sensitive patient. Both IgE ( A )and IgC ( 0 )antibody titers are expressed by units. The IgE antibody unit corresponds to the minimal concentration of the antibody required to give a positive Prausnitz-Kiistner reaction. [From Ishizaka and Ishizaka ( 4 4 ) .I

CELLULAR EVENTS I N THE IGE ANTIBODY RESPONSE

3

cells), is essential for the induction of antibody responses to most protein antigens. This principle obtained with IgM and IgG antibody responses has been proved to be the case also in IgE antibody formation. From the immunological viewpoint, however, it became clear that IgE antibody responses in experimental systems have certain characteristic features that are not easily demonstrated in IgG antibody response. The purpose of the present review is to analyze the cellular events involved in the IgE antibody response in different experimental systems in comparison with the IgG antibody response. It is hoped that elucidation of the mechanisms for induction and suppression of IgE antibody response will provide a clue to future therapy for atopic diseases. II. Immunoglobulin E Antibody Formation in Vivo a n d in Vitro

A. KINETICS OF IGE ANTIBODYRESPONSESIN VARIOUSANIMAL SPECIES Many years ago, Sherman et al. (142) followed reaginic antibody titers in hay fever patients by recording Prausnitz-Kiistner reactions and showed that antibody titers persisted in the sera of ragweed-sensitive individuals. The results were recently confirmed by quantitative measurement of IgE antibody by a radioimmunoassay (RAST technique), which was developed by Wide et al. (185). Application of this method to measure serum IgE antibody in untreated ragweed-sensitive patients revealed that the antibody level persisted and that most patients showed secondary IgE antiragweed antibody responses after the ragweed season (Fig. 1 ) (44). Because the catabolic rate of IgE is very fast, with an average halflife of 2 to 3 days (178), persistence of IgE antibody titers in the sera of atopic patients indicates that IgE antibody is being formed continuously. Such a pattern of antibody formation, however, is not characteristic only for IgE. Titration of IgG antiragweed antibody in the sera of the same untreated patients by double antibody radioimmunoassay showed that IgG antibody formation also persisted, and the antibody titer definitely increased after the ragweed season. As shown in Fig. 1, the time course of IgG antibody produced to ragweed antigen E paralleled that of IgE antibody. Several investigators injected allergen into non-atopic individuals in the course of their studies of hyposensitization treatment. Some normal individuals who received parenteral injections of alum-precipitated allergen developed IgE antibody against the allergen. The IgE antibody in the sera of these individuals disappeared within 2 to 3 months; however, many of them showed secondary IgE antibody responses after the

4

KIMISHIGE ISHIZAKA

pollen season (99). De Weck (20) has shown a similar pattern of IgE antibody response in patients with hypersensitivity to penicillin. In many patients, the IgE antibody was detected when they had clinical symptoms but disappeared within several weeks after the administration of penicillin. Obviously, these patients will show secondary IgE antibody responses after reexposure to the drug. Although the physicochemical properties and biological function are similar for IgE antibodies from various animal species, the kinetics of IgE antibody responses are different depending on the species and strains of the animals. Immunization of rats with usual protein antigens, such as ovalbumin (OA) (109) or human IgG (11) together with pertussis vaccine or aluminum hydroxide gel (alum) as adjuvants resulted in the formation of IgE antibody, but the antibody response was transient in nature. Maximum IgE antibody titer was reached at 10 to 14 days after the immunization and rapidly declined thereafter. A booster injection of the same antigen 4 to 5 weeks after the primary immunization did not elicit secondary IgE antibody response. Even when a secondary response was observed, maximum IgE antibody titer after a booster injection was lower than the maximum titer after primary immunization. As will be described later, the dose of antigen and nature of adjuvant employed for the primary immunization appear to be important factors in obtaining a secondary antibody response, By using a purified antigen from Ascaris suum extract (Asc-1), Strejan et al. (151) have shown a definite secondary IgE antibody response after a booster injection. More recently, Jarrett et al. (60) immunized Hooded Lister strain rats with 1 to 10 pg OA or keyhole limpet hemocyanin (KLH) together with 10'O pertussis vaccine and then gave a booster injection of homologous antigen without adjuvant at 30 days after primary immunization. This immunization schedule gave a definite secondary IgE antibody response. Other strains of rats, e.g., Sprague-Dawley, Wister, and Lewis, however, failed to show secondary IgE antibody response after a booster injection of antigen without adjuvant. A unique system for obtaining an IgE antibody response in the rat was described by Tada et al. (155). Their immunization schedule was based on previous observations of Strejan and Campbell ( 148), who found that two closely spaced injections of A. mum extract ( Asc) were effective in obtaining a high titer of reaginic antibody in the rat. Tada et al., injected 1 mg of dinitrophenyl derivatives of A. suum extract ( DNP-Asc) together with 10'O Bordetella pertussis vaccine into footpads of Wistar rats, followed by an intramuscular injection of 0.5 mg of DNP-Asc on day 5. In most animals, IgE antibody to homologous antigen was detected after the second injection. The IgE antibody titer reached a

CELLULAR EVENTS I N THE ICE ANTIBODY RESPONSE

5

maximum at 3 days after the second injection and declined thereafter. I n their experiment, neither the first injection of antigen with pertussis vaccine nor the injection of antigen alone induced IgE antibody response. An average maximum antibody titer, which was determined by the homologous passive cutaneous anaphylaxis (PCA) reaction, was on the order of 1:80, and the antibody became undetectable at about 4 weeks. I t was also found that a booster injection of the same antigen 4 weeks after the immunization failed to give a secondary IgE antibody response. Their immunization regimen is unique in that a large dose of antigen was used to obtain an IgE antibody response and that a single injection of antigen with pertussis vaccine failed to elicit the antibody response. As will be discussed later, usually a small dose of antigen is favorable for the IgE antibody response, and a single injection of an adequate dose of antigen with either pertussis vaccine or alum gives primary IgE antibody response. Subsequently, Tada et al. ( 162) succeeded in sustaining the IgE antibody response by irradiation of rats with sublethal doses (200400 R ) of X-ray, 1 day before or 1 day after the initial injection of DNP-Asc with pertussis vaccine. In the irradiated rats, IgG antibody was undetectable, but serum IgE antibody titer was higher than that obtained in nonirradiated animals, and the titer was maintained more than 3 weeks. This immunization schedule was frequently used by Tada and his associates when they wished to analyze the mechanisms involved in the IgE antibody response. Unfortunately, irradiation abolished rather than sustained IgE antibody responses in some other strains such as Sprague-Dawley and Lewis (see Section V,C,1). Rabbit IgE antibody was first described by Zvaifler and Becker (189) who had immunized animals with a relatively high dose of DNP-bovine y-globulin ( BGG ) included in complete Freunds adjuvant ( CFA) . The antibody did not persist for long, and these animals failed to show secondary IgE antibody responses after a booster immunization. Subsequently, it became clear that immunization with a relatively small dose of antigen precipitated with alum (132) gave a primary IgE antibody response and that the animals immunized by this procedure frequently gave secondary IgE antibody responses after a booster immunization. In our experience, more than one-half of rabbits immunized with DNP-Asc showed secondary antihapten IgE antibody responses in which maximal antibody titers were higher than primary responses (48 ). Mota and Peixoto (112) detected reaginic antibody in the mouse, after they had immunized outbred mice with a relatively high dose (50-100 p g ) of antigen included in CFA, alum, or with pertussis vaccine. The IgE antibody response was transient in nature, and antibody became

6

KIMISHIGE ISHIZAKA

undetectable in the sera within 3 weeks after the immunization. Similar results were obtained by Revoltella and Ovary (131) in several inbred strains of mice using DNP-KLH as antigen. Thus, the kinetics of reaginic antibody formation in the mouse were believed to be different from that observed in hay fever patients. Such a difference was overcome in a model developed by Levine and Vaz (go), who immunized several inbred strains of mice with 0.1-1.0 pg of protein antigens absorbed to alum. Repeated immunization at 4-week intervals resulted in a secondary response with a high titer of reaginic antibodies. Subsequently, Vaz et al. (173) succeeded in obtaining a persistent reaginic antibody response by injecting alum-absorbed OA (0.1 pg) into SW-55 strain mice. The reaginic antibody titer persisted for several months without booster injections. So far, the IgE antibody response in this system is the best model for reaginic antibody formation in humans. A persistent IgE antibody response has now been achieved with several different combinations of antigens and inbred strains of mice. For example, a minute dose of OA (0.05-0.2 pg) adsorbed to alum produced a persistent antibody response in DBA/1 and (C57B1/6 x DBA/2)F, mice (176). Immunization of these strains with 1-2 pg DNP-KLH absorbed to 1-2 mg of alum gave a persistent anti-DNP antibody response (120). An injection of alumabsorbed ragweed antigen E into the A/ J strain gave a similar pattern of IgE antibody response ( 5 2 ) .

B. HELMINTH INFECTION AND ICE RESPONSFS It has been known for a long time that an intracutaneous injection of an extract of Ascaris lumbricoides into normal individuals frequently elicits a positive erythema wheal reaction, suggesting that IgE antibody is formed following Ascaris infection. Johansson et al. as well as others reported that total IgE levels in sera increased in most individuals infected with any one of a variety of helminths including A. lumbricoides (61), Capillaria phillipinensis ( 183), and Ancylostoma ( 6). Infected individuals’ other serum immunoglobulins, such as IgG, IgM, IgA, and IgD, were usually in the normal range or were elevated only slightly emphasizing the strong relationship of helminth infections with IgE. The IgE antibody formation following helminth infection was established in experimental animals such as the rat, mouse, and rabbit (108, 115, 118, 137, 190). Nematodes, cestodes, trematodes, as well as arthropods all share this immunogenic characteristic ( 117). A typical example was shown in the rat by Ogilvie (115), who demonstrated IgE antibody formation after infection with Nippostrongylus brasiliensis larvae. The IgE antibody against worm extract became detectable 3 4 weeks after the infection, and antibody persisted for a longer period of time than that

CELLULAR EVENTS I N THE IGE ANTIBODY RESPONSE

7

obtained by an artifical immunization with protein antigen included in an appropriate adjuvant. Furthermore, the animals showed a definite secondary IgE antibody response upon reinfection (116, 188). Recently, Jarrett and Bazin (58) determined total IgE levels in rats infected with N . brasiliensis. Their results showed that total IgE levels in the sera of normal Hooded Lister rats were less than 0.35 pg/ml, but these levels increased to 250-500 pg/ml at 12 days after infection. Recently, we studied the relationship between total IgE and IgE antibody against worm antigen, following the infection of Sprague-Dawley rats with N . brasiliensis larvae ( 5 8 ) . The results showed that total IgE level began to increase about 10 days after the infection and reached a maximum on the fourteenth day. On the other hand, IgE antibody against worm antigen became detectable at 3 to 4 weeks after the infection, when total IgE level had already begun to decline (Fig. 2 ) . It is apparent that the kinetics TOTAL Ig E Nl/ml

100

PCA

-

I

I I I

10 -

-80

II

p--I

I

-20

I

I

I

I

I

I

I II

I

3

5

7

9

WEEKS

t

INF

t

INF

FIG.2. Total IgE (0) and IgE antibody in the serum of a rat infected with Nippostrongylus brasiliensis. The IgE antibody titer ( A ) was determined by PCA reactions using an extract of worm as antigen. [From Ishizaka et al. ( 5 6 ) . ]

8

KIMISHIGE ISHIZAKA

of the IgE antibody formation did not parallel the total IgE synthesis. Lack of correlation between total IgE and IgE antibody was confirmed in Hooded Lister rats (57). Another interesting finding in parasite infection is that infection of rats with N . brasiliensis or Fasciola hepatica causes nonspecific potentiation of unrelated IgE antibody responses to antigens such as OA and KLH (12, 59, 126, 127). Orr and Blair (126) first described this phenomenon following the infection of OA-primed animals with N . brasiliensis. Bloch et al. (12) found that augmentation of the antibody response after parasite infection was directed only to IgE antibodies: Neither the IgGl nor IgG2 antibody response was altered following the infection. There are some requirements for obtaining the potentiation. First of all, rats have to be primed in such a way as to produce IgE antibody prior to the infection. Second, there should be an appropriate interval between the priming immunization with antigen and infection. In Sprague-Dawley rats, which were employed by Orr and Blair (127), an interval of 1week to 10 days was optimal for the potentiation. Neither the infection prior to the immunization nor late infection after the primary IgE antibody response gave potentiation. This interval, however, did not appear to be critical when Hooded Lister rats were used in the experiments. Jarrett and Bazin (58) immunized these rats with OA together with pertussis vaccine and infected them 20 days after the priming for successful potentiation. The difference among the strains may be related to the fact that the primary IgE antibody response to OA in Hooded Lister rats was more persistent than that observed in the other strains. It is also known that the Hooded Lister strain show a secondary IgE antibody response to OA without adjuvant, whereas Sprague-Dawley rats fail to respond to a booster injection. In both strains, potentiation of the IgE antibody response was observed at 12 to 14 days after the infection when the total IgE increase was maximum. These results suggest that potentiation is due to nonspecific stimulus on B cells that have been programmed for IgE antibody production by previous immunization. This idea is supported by the finding of Jarrett et al. (59), who demonstrated that IgE antibodies against both OA and KLH were potentiated following parasitic infection if the rat had been primed with both antigens. The potentiation of an IgE antibody response after N . brasiliensis infection was observed in the mouse as well (82). In this species, however, infection with parasites 5 to 14 days prior to primary immunization was mogt effective for potentiation, whereas the infection after the immunization was ineffective. The reasons for these differences between rats and mice are unknown at the present time.

9

CELLULAR EVENTS I N THE IGE ANTTBODY RESPONSE

C. DISTRIBUTION OF IGE-FORMING CELLS Th IgE-forming cells were first detected in primate lymphoid tissues by using a fluorescent antibody technique ( 154). In nonatopic individuals, recurrently infected tonsils and adenoids removed by surgery possessed a large number of plasma cells that stained with anti-IgE. Some germinal centers in these tissues also stained. Bronchial and peritoneal lymph nodes contained IgE-forming plasma cells as well as germinal centers. By contrast, IgE-forming cells were scarce in spleen and subcutaneous lymph nodes. The IgE-forming cells were detected in respiratory and gastrointestinal mucosa. In nasal mucosa, some of the plasma cells under epithelial cells stained with anti-IgE. Immunoglobulin E-forming cells were found in the bronchial mucosa especially around the mucous serous glands. In the stomach, small intestine, colon, and rectum, IgE-forming cells were observed in the lamina propria, especially around the crypts of Lieberkuhn. Lymphoid cells in bone marrow, lung tissues, and peripheral blood from nonatopic individuals did not stain with anti-IgE. The distribution of plasma cells and germinal centers that stained with anti-IgE is summarized in Table I, which also shows the distribution of IgE-forming cells in monkey tissues. It would appear that the IgE-forming cells predominate in the respiratory and gastrointestinal mucosa and in the regional lymph nodes. TABLE I DISTRIBUTION OF IGE-FORMING CELLSI N LYMPHOID TISSUES Monkeya

Humana

Lymphoid tissues Tonsil Adenoid Bronchial and peritoneal Subcutaneous lymph node Spleen Respiratory mucosa Gastrointestinal mucosa Lung Blood Bone marrow

Plasma cells

Germinal center

Plasma cclls

+ + - +++ + - + + +-+++ + - + + ++ ++ (+) 5 +-+ +-++ +-+ + + - -+ + -

Parentheses indicate negative in some cases; nd, not determined. Plus in Peyer’s patches.

Germinal center

++ (+)

-

+

10

KIMISHIGE ISHIZAKA

Morphologically, most of the cells stained by anti-IgE had the characteristics of plasma cell with an eccentric nucleus, abundant cytoplasm, and a clearly defined Golgi apparatus. The others appeared to be lymphoid cells that are normally present in the germinal centers. The IgEforming cells were different from plasma cells forming other immunoglobulins such as IgA or IgG, and germinal centers stained by anti-IgE were distinct from those stained by either anti-IgA or anti-IgG. The distribution of IgE-forming cells in primate lymphoid tissues suggests that spleen and peripheral lymph nodes are not the source of IgE. This idea was supported by the finding of Mota (111) that splenectomy of rats prior to immunization with OA did not affect the IgE antibody level in their sera. Recently, we studied IgE-forming cells in spleens and mesenteric lymph nodes of rats infected with Nippostrongylus brasiliensis. Smears of spleen and mesenteric lymph node cells were treated with rabbit anti-rat IgE antibody and then with fluoresceinated antirabbit IgG (56). Both mesenteric lymph nodes and spleen contained IgEforming plasma cells, however, the number of IgE-forming cells in the spleen was significantly less than that observed in the mesenteric lymph nodes. After these rats were reinfected with the parasite to induce secondary IgE antibody responses, suspensions of their spleen cells, mesenteric lymph nodes, and parathymic lymph node cells were injected intracutaneously into Sprague-Dawley rats for passive sensitization. On the basis of the number of mononuclear cells, parathymic lymph nodes had the highest sensitizing activity, and mesenteric lymph node cells gave a lower PCA titer, whereas spleen cells failed to give the PCA reaction. In the rabbit, however, IgE-antibody forming cells were detected in the spleen. As described, an anti-DNP IgE antibody response was obtained by intraperitoneal injections of a minute dose of DNP-Asc included in alum. In order to see the distribution of IgE antibody-forming cells in these animals, cell suspensions were prepared from their spleens, mesenteric lymph nodes, popliteal lymph nodes and thymuses, and serial dilutions of the cell suspensions were injected intracutaneously into outbred normal rabbits. Challenge of the recipients with DNP-human serum albumin (HSA) at 48 hours after sensitization showed that spleen cells gave the highest PCA titers and mesenteric lymph node cells gave twoto four-fold lower titers, Neither thymus cells nor popliteal lymph node cells were capable of sensitizing homologous skin for a positive PCA reaction. Because the same cells killed by freezing and thawing failed to sensitize rabbit skin, it appears that IgE-forming cells released the antibody in the skin tissues. Kind and Macedo-Sobrinho ( 7 1 ) performed similar experiments using mouse lymphoid tissues. They injected cell suspensions of spleen, bone marrow, and lymph nodes from immunized

CELLULAR EVENTS I N THE ICE ANTIBODY RESPONSE

11

animals into rat skin to exclude possible proliferation of immunocompetent cells in the recipients. Their data showed that spleen cells gave the highest PCA titer when the same number of viable cells were used for passive sensitization. Taken collectively, it appears that distributions of IgE antibodyforming cells are different depending on the method and route of immunization. Thus, preferential distribution of IgE-forming cells in the respiratory and gastrointestinal tracts in the primate might result from frequent exposure of the lymphoid tissues to small doses of environmental antigens, which may be favorable for IgE synthesis. D. IMMUNOGLOBULIN E ANTIBODY RFSPONSEin Vitro In view of recent progress in tissue culture techniques and successful antibody formation in vitro (95, l05), attempts were made to form IgE antibody in cell culture. So far, the primary IgE antibody response has not been obtained in vitro; however, a secondary antihapten IgE antibody response was observed using mesenteric lymph node cells of rabbits that had been properly immunized for IgE antibody response (48).An experimental design for the culture system is shown in Fig. 3. Rabbits were immunized with 10 pg DNP-Asc included in 10 mg alum and boosted with the same dose of alum-absorbed antigen 4 weeks after the primary immunization. Animals giving a PCA titer of 1:160 or more against DNPHSA were sacrificed 2 weeks after the booster injection to obtain mesenteric lymph nodes. The cells were suspended in minimum essential medium, enriched with 20%fetal calf serum and 2 mM L-glutamine, and incubated for 24 hours at 37°C with homologous antigen. After being washed to remove free antigen, 1 ml of the cell suspension containing 1-2 x lo7 nucleated cells was cultured for 6 days by the method of Marbrook (95). By this procedure, primed mesenteric lymph node cells formed anti-DNP IgE antibody together with the IgG and IgM antibodies specific for the DNP group, whereas unstimulated cultures of the same cells failed to produce the antibody. Kinetic studies of antibody formation showed that both IgE and IgG antibodies became detectable after 3 to 4 days in culture, and their concentration in the culture fluid increased exponentially. It was also found that an optimal concentration of antigen for maximum antibody formation was comparable for the three immunoglobulin classes. Because the rabbits used in the experiments were outbred, the limitation of this system was that lymphocytes from two different animals could not be mixed in the culture. Nevertheless, this system was useful for studying cell requirements for a secondary IgE antibody response, and for analyzing the mechanisms involved in T cell-R cell collaboration.

12

KIMISHIGE ISHIZAKA

Ill.

Immunological Factors Essential for IgE Antibody Responses

A. GENETICCONTROL OF IGE RESPONSES Immunization of inbred strain of mice with several protein antigens and a hapten-protein conjugate by Levine and Vaz (90) revealed that the reaginic antibody response is controlled by genes at a single autosoma1 locus, closely linked to the H-2 system on the 1X linkage group, which controls immune responses per se. They observed marked differences among the strains in the production of both IgGl and IgE antibodies, if minute doses (0.1 pg) of the immunogens were used for immunization. When large doses of the immunogens (100 pg) were used, such a strain difference became less apparent. As the result of systematic experiments in many inbred strains with three different antigens, i.e., benzyl penicilloyl-BGG ( BPO-RGG ), OA, and chicken ovomucoid in low doses, they found that the responsiveness of a given strain was antigen-specific. One strain was a good responder to one antigen but a poor responder to another antigen. When good and poor responders to a given antigen were grouped, responsiveness correlated with H-2 geno-

CELLULAR EVENTS IN THE IGE ANTIBODY RESPONSE

13

types. For example, strains A/He, CBA/J, C E / J and C3H, whose H-2 genotypes is either a or k, were good responders to BGG, BPO-BGG, and ovomucoid, but poor responders to OA. On the other hand, DBAI1 and C57BL/6 having type q or b were good responders to OA but poor responders to ovomucoid. Subsequent breeding experiments and immunization of strain pairs congenic at the H-2 locus established the genetic control. The experiments have shown that the F1 generation between A/He (H-2") and C57BL/6J ( H-2b) were good responders to ovomucoid. The backcross (F1 x poor responder) yielded approximately 50% responders, indicating that the immune response is controlled by a single gene. The results of the experiments by Levine and Vaz (90) as well as Vaz et al. (175, 177) are in agreement with previous studies by McDevitt and Chinitz (101) who established that the H-2-linked Ir-I gene controls the immune response to synthetic branched polypeptide (TG ) -AL. Comparisons between the two series of experiments indicated that responsiveness to ovomucoid correlated with that to (TG)-AL, whereas responsiveness to OA correlated with that to ( HG ) -AL. The experiments of Levine and Vaz (90) clearly showed that strain difference was related to carrier protein when mice were immunized with hapten-protein conjugates. As will be discussed later, both T and B cells are required for an IgE antibody response. Correlation of the immune response with carrier protein suggested that the genetic control observed in their experiments is on T cells. The mechanisms by which allele products of the 17-1 locus influence immune responsiveness are unknown; however, evidence has accumulated that they are expressed on T lymphocytes ( 100). Although the genes linked to the H-2 locus control a major portion of the responsiveness of the strains, there are some variations that cannot be due to the presence of a particular combination of H-2 alloantigen specificities. Thus, among H-2" mice, strain AKR/J responded better to both ovomucoid and OA than the other H-2rcstrains. Similarly, (C57BL/6 X DBA/B)F, (BDF1) mice gave better responses to OA than both parental strains and genetically related strains such as B10.D2. Vaz et d. (174) suggested that a portion of the heightened responsiveness of BDFl mice to OA might be due to genes present in the DBA/2 background. An entirely different kind of genetic factor found in mice uniquely controls the immune response of IgE antibody (87). When a large number of mouse strains were immunized with different antigens in various doses, all strains produced IgG antibody to one or another antigen. The IgE antibody responses were high in some strains but low or absent in others. For example, the SJL strain showed a poor IgE antibody response

14

KIMISHIGE ISHIZAKA

to six different antigens, but a reasonably high titer of IgG antibody to some of these antigens. In this situation, there was neither dose effect nor antigen specificity and the effect was on IgE antibody alone. Breeding experiments showed that the genetic control is by more than one locus and not linked to H-2. Thus, it appears that two different genetic controls operate in determining whether or not a given mouse strain will produce IgE antibody to a given antigen. Genetic control of the IgE antibody response was found in humans as well. In the typical case of pollen allergy, the individual became sensitized to extremely low doses of allergens within the pollen. Even if all allergens were extracted completely from pollen in the mucosa of the respiratory tract, a dose of a major allergen in ragweed pollen (antigen E ) or in grass pollen (Group I antigen) would be less than 1 pglpollen season for an individual in Baltimore (96). It is quite conceivable that the antibody response to such a minute dose of antigen is induced only in high responders. Indeed, most individuals who are not allergic to these pollen antigens lack both IgE and IgG antibodies, but almost all allergic individuals have antibodies of both immunoglobulin classes. Levine et al. (89) performed family studies in order to demonstrate Ir genes for ragweed antigen. Their results showed that members of a family having the HL-A1, HL-A8 (18) haplotype had intense immediate skin reactivity to antigen E, whereas none of the subjects having the other HLA haplotypes had immediate skin reactivity. Data for seven families, in which ragweed hay fever occurred in more than one member showed that 22 out of 26 of the members having the hay fever-associated haplotypes had ragweed hay fever and skin reactivity to antigen E. By contrast, non of the 11 members who had the other haplotypes had clinical hay fever. In the seven families, however, the hay fever-associated haplotype was different from one family to another. It is not known whether the antibody response to antigen E is controlled by only one or several different Zr genes. Nevertheless, the authors have shown that sensitivity to the other allergens has no relationship to the ragweed hay fever-associated haplotype, and suggested that the HLA-linked control of immune responsiveness to antigen E has antigenic specificity. As expected from their studies, population studies that were carried out by Marsh et al. (97) did not show a clear-cut relationship between a single HLA haplotype and the sensitivity to ragweed antigen. Genetic studies of Marsh et al. (97), however, showed a relationship between HLA haplotype and the sensitivity to a minor ragweed allergen, Ra-5, which has a relatively simple structure comprised of only 43 amino acids. Their approach was to examine skin sensitivity of ragweed-sensitive individuals to both antigen E and Ra-5, and to classify the patients into Ra-5-sensitive and Ra-5-insensitivegroups.

CELLULAR EVENTS IN THE I G E ANTIBODY RESPONSE

15

All patients were sentitive to antigen E, indicating that the patients received enough allergen to produce IgE antibody against antigen E. Statistical analysis of the two groups with respect to HLA types indicated that the control of IgE responsiveness to Ra-5 is determined by a single Zr gene that is intimately associated with genes controlling the expression of HL-A7 Creg antigens. Marsh et al. (98) studied the possibility that IgE synthesis may be genetically controlled. Statistically, allergic persons have high IgE levels compared with nonallergic individuals ( 3 2 ) , but some nonallergic persons have atypically high and some allergic individuals have atypically low IgE levels. The authors set up a cutoff point between high and low IgE at the level at which the combined percentages of such atypical subjects were minimized. This cutoff point was 95 & 5 international units ( I U ) / ml which was in excellent agreement with a value of 91 f 5 IU/ml calculated by replotting data of Gleich et al. (32). Thus, they analyzed three types of mating in 28 families: ( A ) low x low giving at least one high IgE, ( B ) low x high giving at least one high IgE, and ( C ) high x high. If one assumes recessive inheritance of high IgE, types A, B, and C would be RrX Rr, Rr X r r , and rr x rr, respectively. As predicted, all siblings of type C families had high IgE levels, and the incidences of high IgE level children in types A and B families were close to, but slightly higher than 1 standard deviation ( S D ) above the theoretically predicted values. From these studies they speculated that the inheritance of a high serum IgE level could be a simple Mendelian recessive trait and that there is no linkage between HLA haplotype and IgE level. An interesting observation in their studies is an implication that hay fever patients with high IgE levels have allergies to multiple allergens, whereas those with low IgE levels are usually sensitive to few allergens. Thus, in most allergic families, a gene regulating serum IgE level appears to mask the role played by hypothetical Zr genes linked to an HLA haplotype in controlling the expression of a specific IgE antibody response to different allergens. On might speculate that a genetic factor capable of controlling total IgE synthesis in humans may correspond to a factor in mice that uniquely controls the immune response of IgE antibody. It appears that IgE antibody responsiveness to low doses of specific allergens may be controlled by two distinct genetic factors in mice and humans.

B. ADJUVANT FOR IGE ANTIBODYRESPONSE The nature and dose of adjuvant are critical factors in the IgE antibody response to protein antigens. Injections of various doses of soluble antigen without adjuvant into rodents failed to give an IgE antibody response. In many different animal species, such as rabbits, guinea pigs, mice, and

16

KIMISHIGE ISHIZAKA

rats, alum is a better adjuvant than CFA for the IgE antibody response. Revoltella and Ovary (132) as well as others (147) have reported that the percentage of rabbits that produce reaginic antibodies is greater with alum than with CFA. It was also found that immunization of rabbits with protein antigens included in CFA failed to establish a memory for the IgE antibody response. Yet, repeated immunization with an adequate dose of alum-absorbed antigen elicited secondary IgE antibody responses in many of the animals immunized. In the rat, it is difficult to obtain a secondary IgE antibody response whichever adjuvant is employed; however, a primary IgE antibody response obtained with alum is higher than that obtained with CFA as an adjuvant. In certain inbred strains of mice, both alum and CFA are effective for inciting an IgE antibody response. It is generally observed, however, that the minimum dose of an appropriate antigen required for giving an IgE antibody response is less if alum is employed as an adjuvant. Using DNP conjugates of three different carriers, i.e., KLH, OA, and BGG, Hamaoka et al. ( 3 8 ) confirmed that alum-absorbed antigen favored IgE rather than IgG responses, whereas the reverse was true for antigen included in CFA. Another important adjuvant used for IgE antibody response is Bordetella pertussis vaccine. This adjuvant is effective in the rat (109) and mouse (112) but essentially useless in the rabbit. It is known that pertussis vaccine induces severe inflammation in the lung and lymphocytosis in peripheral blood (15). It is also known that the sensitivity of the rat and mouse to histamine increases after the administration of pertussis vaccine (113). Attempts were, therefore, made to isolate active substance with an adjuvant effect from the vaccine. Clausen et al. (16) obtained a saline extract from the B . pertussis organisms at an alkaline pH and showed that the extract exerted an adjuvant effect for IgE antibody production in the mouse. In their experiment, the active component was not distinguishable from the histamine-sensitizing factor. They have also shown that endotoxin from B. pertussis was not effective in stimulating IgE antibody response. Tada et al. (157) studied the effect of “lymphocytosis promoting factor,” which was obtained from a culture filtrate of B. pertussis, and showed that less than 1 pg of this component had a definite adjuvant effect on IgE antibody production in the rat. Because this substance induced lymphocytosis in the peripheral blood and depleted small lymphocytes in thymus-dependent areas of lymphoid tissues, the authors speculated that treatment with lymphocytosis-promoting factor caused a depletion of a certain subpopulation of T cells from the lymphoid tissue that regulate the IgE antibody response. Recent studies by Lehrer et al. (85) suggested that both histamine-sensitizing activity

CELLULAR EVENTS IN THE ICE ANTIBODY RESPONSE

17

and lymphocytosis-promoting activity were associated with the same molecule. Adjuvant other than alum and pertussis vaccine gave inconsistent results for IgE antibody responses. Clausen et al. (15) have reported that endotoxin ( lipolysaccharide, LPS ) from Salmonella minesota was not as effective as pertussis vaccine in stimulating an IgE antibody response in the mouse. Newberger et al. (114), however, showed that administration of LPS with DNP-Asc into irradiated recipient mice that received DNP-Asc-primed spleen cells enhanced antibody responses in both IgE and IgG classes. It appeared that CFA, LPS, and Poly A:U enhanced antibody responses of both IgE and IgG classes in the mouse and did not preferentially affect the IgE antibody response. The requirement of adjuvant for IgE antibody response in the mouse is limited to the primary response. If the mouse was primed with an appropriate antigen absorbed to alum for IgE antibody response, a booster injection of homologous antigen without adjuvant gave a secondary IgE antibody response (120). Within 2 weeks after the booster injection, the IgE antibody level declined to the same level as that observed before the booster injection. As described before, secondary IgE antibody responses were not observed in most rat strains. However, in the Hooded Lister strain, a secondary IgE antibody response was easily obtained by injecting antigen without adjuvant (60). The antibody titer increased within 4 days after the booster injection and rapidly declined by the seventh day. Recently, we primed the same strain with DNP-OA either with pertussis vaccine or with alum for primary antihapten antibody responses and gave a booster injection of homologous antigen without adjuvant. The magnitude of the primary IgE antibody response was higher with alum than with pertussis vaccine. By contrast, the secondary antihapten IgE antibody response was readily observed when pertussis vaccine was employed for the primary immunization. When alum was employed for priming, essentially no secondary IgE antibody response was observed. I t would appear that pertussis vaccine is a better adjuvant than alum to establish immunological memory for an IgE antibody response in the rat.

C. NATUREAND DOSEOF ANTIGEN In all animal experiments, IgE antibody was formed against T-dependent antigens. Attempts to produce IgE antibody in mice by the immunization with T-independent antigens were unsuccessful. TO date, immunization of mice with Salmonella bacilli ( 143), pneumococcus polysaccharide SIII, or the DNP-derivative of an acidic copolymer of D-tyro-

18

KIMISHIGE ISHIZAKA

sine, glutamine, and lysine ( DNP-D-TGluL) failed to induce IgE antibody responses ( 121 ). Although some atopic individuals were sensitive to some T-independent antigens such as dextran (63), there was no evidence that the T-independent antigens were actually immunogens. It is quite possible that such T-independent antigens were in the form of complexes with a carrier protein in the natural state and that the complex was immunogenic with respect to the IgE antibody response. This possibility is conceivable in view of the work by Paul et al. (128), who demonstrated an anti-SIII antibody response in the rabbit to immunization with SIIIBGG conjugates. In this system, BGG served as a carrier to which helper cells were directed. Because hay fever patients have IgE antibodies against a variety of allergens, and helminth infection induces IgE antibody responses in many animal species, pollen allergens as well as Ascaris extract were frequently used to induce IgE antibody responses in experimental animals. It has been shown that both ragweed antigen E and extract of Ascaris suum produced IgE antibody responses in rabbits (48, 76, 150), rats ( 139, 149), and mice ( 5 2 ) . Strejan et al. (151) compared the immunogenicity of Asc with other protein antigens, such as OA, KLH, and bovine serum albumin (BSA), for the IgE antibody response in the rat, and found that Asc was superior to the other antigens. Purified Ascaris antigen ( Asc-1) (42) was a strong immunogen for an IgE antibody response in the rat, whereas BSA and BGG were unable to produce reaginic antibody. Under the same immunization regimen, OA and KLH were less immunogenic than Ascaris antigen but better than BSA for the formation of IgE antibody. It should be noted that all of these protein antigens were capable of producing large amounts of IgG antibody. Potent antigens for IgE antibody formation are useful as carrier proteins for the formation of antihapten IgE antibody. Strejan and Marsh ( 152) compared various DNP-coupled protein antigens for their ability to induce anti-DNP IgE antibody in the rat, and found that DNP coupled to Ascaris extract was a potent immunogen and superior to DNP conjugates with KLH or BGG. Similar results were obtained in the rabbit as well. Both DNP-Asc and DNP coupled to ragweed Fraction D (Rag), in which antigen E is the major immunogen, are superior to DNP-BGG or DNP-KLH for the formation of antihapten IgE antibody (76). As both Asc and ragweed antigen ( Fraction D and purified antigen E ) themselves are excellent immunogens for IgE antibody response in the rabbit, one may generalize that allergens serve as excellent carriers for producing antihapten IgE antibody responses. The peculiar properties of allergens stimulating IgE antibody formation suggest that they have physicochemical properties in common; however,

CELLULAR EVENTS I N THE IGE ANTIBODY RESPONSE

19

no common structure for allergen has been established. On the other hand, an extensive study on IgE antibody formation in inbred strains of mice suggests that the common immunological property of potent allergens is high immunogenicity. Levine and Vaz (90) showed that almost all inbred strains of mice gave a transient IgE antibody response to protein antigens or hapten-protein conjugates when they were immunized with 50 to 100 pg antigen included in alum, but a booster injection of the same dose of antigen failed to give a secondary IgE antibody response. If the same strains were immunized with a low dose (0.1-1 p g ) of antigen in responder mice to OA, such as Sw-55 or DBA/1, the IgE antibody responses, and IgE antibody titers increased upon booster injections. In high responder mice to OA, such as Sw-55 or DBA/l, the IgE antibody response to a minute dose of the antigen was comparable to or higher than that obtained by immunization with allergens such as Asc or ragweed antigen. In these animals, DNP-OA and DNP-KLH induced a higher antihapten IgE antibody response than did DNP-Asc. Thus, IgE antibody responses of high responder strains of mice to certain antigens or haptencarrier conjugates were similar to those of individual outbred animals to potent allergens. Another important factor in the IgE antibody response is the dose of antigen used for immunization. Strejan et a2. (151) immunized rats with different doses of alum-absorbed purified Asc-1 every 3 to 4 weeks and showed that the secondary IgE antibody response was obtained by a l-pg dose rather than by a 10-pg dose. Similarly, Jarrett et al. (60) reported that 1-10 pg of OA with pertussis vaccine was optimal for establishing memory for the IgE antibody response. Rats immunized with 100 pg to 1 mg OA gave the antibody response, but secondary IgE antibody responses of these animals were lower than those observed in rats primed with 0.1-1 pg OA. The dose dependence of IgE antibody formation was also shown in the rabbit by Revoltella and Ovary (132). They found that a low dose of antigen in alum favored reagin production. As already described, Levine and Vaz (90) showed that in high-responder strain mice a low dose of antigen gave higher and longer-lasting IgE antibody responses than did a high dose. It is known that a high dose of antigen is required for producing antibody responses in poor responder mice. One might speculate that the failure of poor responder mice to give persistent IgE antibody responses may be owing to the fact that optimal conditions for the IgE antibody response, i.e., a low dose of antigen with a relatively high dose of adjuvant cannot induce an immune response in a low responder. The dose of antigen is particularly important for the induction of an antihapten antibody response ( 51). When DBA/ 1mice were immunized

20

KIMISHICE ISHIZAKA

with DNP-OA included in alum, a minimum immunogenic dose (0.05 pg) of the antigen gave a persistent antihapten IgE antibody response. If the antigen dose was increased from 0.2 to 1 p g , the anti-DNP antibody titer reached a maximum soon afterward and then declined between 2 to 3 weeks. An interesting observation was that the anticarrier (OA) IgE antibody response was persistent even with 0.2-1 p g of DNP-OA (51). The reasons for the discrepancy between the time course of an antihapten antibody response and an anticarrier antibody response are unknown. Because carrier-specific helper cells must be common to both anti-DNP and anti-OA antibody responses, the decline of antihapten antibody titers may not simply be ascribed to some possible changes in helper cell population in the course of the response. IV. Cellular Basis of IgE Antibody Responses

A. REQUIREMENTFOR T

AND

B LYMPHOCYTES

It has been established that the collaboration of two distinct types of lymphocytes, i.e., bone marrow or bursa-derived ( B ) lymphocytes and thymus-derived ( T ) lymphocytes, is essential for induction of an antibody response by mice to certain antigens, such as sheep erythrocytes and proteins (14, 104). An analogous cooperation of two lymphocyte cell lines was demonstrated in the mouse (106) as well as in the guinea pig and rabbit (69) in the formation of antihapten antibodies. In the haptenspecific antibody response, T cells are usually primed with the determinants present on the carrier, whereas B cells are precursors of antibodyforming cells and have the same specificity as the antibody formed by their progeny. Naturally, a question arose as to whether T cells are required for the IgE antibody response. As will be described below, several attempts were made to establish the requirement of T and B cells for the IgE antibody response. Okumura and Tada (122) reported that rats thymectomized within 24 hours of birth failed to produce IgG and IgE antibodies upon subsequent immunization with DNP-Asc. Supplementation of the neonatally thymectomized rats with normal thymocytes restored the ability to produce IgE antibody to the antigen. Similarly, Michael and Bernstein (103) reported that congenitally athymic (nulnu) mice were unable to produce IgE antibody against OA, but with supplementation by thymocytes from nu/ mice the nude mice were able to produce the antibody. In the antihapten IgG and IgM antibody responses, Mitchison (106) and Katz et al. (69) have shown that priming of mice or guinea pigs with free carrier enhanced the primary antihapten antibody response to

+

CELLULAR EVENTS I N THE ICE ANTIBODY RESPONSE

21

hapten-homologous carrier conjugate. This phenomenon, i.e., carrier effect, is due to priming of carrier-specific T cells. The same principle was reproduced in the IgE antibody response. For example, immunization of rats with Asc extract and pertussis vaccine on day 0, followed by an intramuscular iiijectioii of DNP-Asc on day 5, resulted in the formation of IgE antibody to DNP-Asc (156). Similarly, priming of rabbits with ASC extract included in alum 4 weeks prior to the immunization with DNP-Asc induced both IgE and IgG anti-DNP antibody responses (75). Under the experimental condition employed, a single injection of DNPAsc in alum was insufficient to induce an antihapten IgE antibody response. Although IgE antibody responses in the rabbit and rat were inconsistent, an enhancing effect of carrier priming was demonstrated in DBA/1 mice that were primed with a subimmunogenic dose (0.02-0.05 p g ) of OA included in CFA (51).The animals did not have a detectable amount of either IgE or IgG anti-OA antibody; but priming definitely enhanced the anti-DNP antibody response of both IgE and IgG classes to a subsequent immunization 2 weeks later with DNP-OA. Thus, a carrier effect on the IgE antibody response has been demonstrated in three animal species. However, the dose of carrier, the adjuvant vehicle employed for priming, and/ or the intervals between carrier-priming and immunization with hapten-carrier conjugate were critical for the demonstration of a carrier effect in the IgE antibody response. Priming with an immunogenic dose of OA, which enhanced the antihapten IgG antibody response in the mouse, suppressed the IgE antibody response to DNP-OA. Evidence was obtained for the participation of carrier-specific helper cells in the secondary IgE antibody response by rabbit mesenteric lymph node cells in uitro (82). In this experiment, rabbits were primed with DNP-Asc in alum and some of the animals received supplemental immunization of partially purified ragweed pollen extract ( Rag) [Fraction D by King et al. (72)]. Four weeks after the priming immunization, all animals received a booster injection of DNP-Asc in alum and were sacrificed 2 weeks later. Mesenteric lymph node cells were stimulated by either DNP-Asc or DNP-Rag for the antibody response. As shown in Table 11, the lymph node cells of rabbits that did not receive a supplemental immunization formed anti-DNP IgE antibody upon stimulation with the homologous antigen ( DNP-Asc) but failed to form the antibody upon stimulation with DNP-Rag. On the other hand, lymph node cells from the animals that received a supplemental immunization of alumabsorbed Rag formed both IgE and IgG anti-DNP antibodies upon stimulation with either DNP-Asc or DNP-Rag. Comparisolls between the two groups indicated that the DNP-specific B cells raised by the

22

KIMISHIGE ISHIZAKA

TABLE I1 HELPER FUNCTION OF CARRIER-SPECIFIC CELLS A G A I N S T PRIMARY A N D SECONDARY CARRIERS FOR IGE A N D IGE ANTIBODY RESPONSES Anti-DNPb Supplemental immunization0 None Ragweed Ag in alum Ragweed Ag in CFA

Antigen in vitro None IINP-ASC DNP-Rag None IINP-Asc DNP-Rag None DNP-ASC DNP-Rag

IgG bglml) 0.29 39.0 0.35 0.32 18.5 23.5 0.40

IgE (PCA)

40.0

<2.5 40.0 <2.5 <2.5 20.0 60.0 <2.5 40.0

30.0

<2.5

Supplemental immunization was given to DNP-Asc-primed rabbits. The IgG antibody was measured by radioimmunoassay.

immunization with DNP-Asc collaborated with Rag-specific helper cells to form anti-DNP antibodies of both IgG and IgE isotypes. Similar experiments were carried out in the mouse to show that haptenspecific B cells and carrier-specific helper cells are both required for the maximum antihapten IgE antibody response (119). Strain DBA/1 mice were primed with DNP-Asc included in alum. After anti-DNP IgE antibody titers declined to a low level, a number of DNP-Asc-primed mice received a supplemental immunization with a subimmunogenic dose of OA included in CFA. Two weeks later, these mice as well as those that had received either DNP-Asc or OA alone were challenged with DNPOA included in alum. Anti-DNP antibody responses in these animals and unprimed control mice are shown in Table 111. The results indicate that DNP-specific B memory cells and OA-specific helper cells collaborated to induce antihapten IgE antibody responses to DNP-OA. More definitive evidence for T cell-B cell collaboration in IgE antibody response was obtained by using adoptive transfers into the mouse. Hamaoka et al. (37) immunized two groups of donors with DNP-KLH or Asc extract included in appropriate adjuvant and transferred their spleen cells into irradiated syngeneic mice. Adoptively transferred, DNPKLH-primed spleen cells produced high levels of anti-DNP antibodies of both IgE and IgG classes in response to challenge with DNP-KLH but developed meager responses to DNP-Asc. When spleen cells from Asc-

CELLULAR EVENTS I N THE IGE ANTIBODY RESPONSE

23

TABLE I11 COLLABORATION BETWEIGN CARRIER-SPECIFIC HELPERCELLS A N D HAPTEN-SPECIFIC MEMORY CELLSFOR THE ANTIHAPTEN ICE ANTIBODY RESPONSE Priming immunization.

Supplemental immunizationb

Secondary challengec

Anti-DNP IgE antibodyd

DNP-ASC DNP-ASC None DNP-ASC None

OA None OA OA None

DNP-OA DNP-OA DNP-OA None DNP-OA

640 40 40

5 <5

Immunization with 1 pg DNP-Asc absorbed to alum. A supplement of 0.02 pg OA in CFA was given 40 days after priming immunization. Secondary stimulus with 0.02 pg DNP-OA in alum was given 14 days after supplemental immunization. The PCA titer at 7 days after secondary challenge. a

primed animals were concomitantly transferred with DNP-KLH-primed cells, secondary IgE and IgG antihapten antibody responses were consistently obtained upon challenge with DNP-Asc. Furthermore, the capacity of Asc-primed spleen cells to cooperate with the DNP-KLHprimed lymphocytes in the adoptive secondary response was abolished by treatment of the spleen cells with anti-8 antiserum plus complement. Thus, the adoptive transfer experiments provided direct evidence that carrier-specific helper cells for IgE antibody response are &bearing T lymphocytes.

B. TYPEB LYMPHOCYTES IN IGE ANTIBODYRESPONSE The precursors of antibody-forming cells bear immunoglobulin on their surfaces. Considering the differentiation events for the development of direct precursors of IgG antibody-forming cells from their progenitor cells, one may speculate that heavy chain determinants of immunoglobulin on B memory cells and/or direct precursors of antibody-forming cells represent the immunoglobulin class of the antibody that will be produced by their progeny. Pierce et al. (129) studied the suppressive effect of antiimmunoglobulin heavy chain on the anamnestic plaque-forming cell response of primed mouse spleen cells and showed that anti-a-chain and anti-p-chain, but not anti-y-chain, suppressed the IgA antibody response, whereas either anti-y,- or anti-y,, but not anti-a-chain, suppressed both IgGl and IgG2 antibody responses. These findings suggested the possibility that B memory cells for IgE antibodies, which may be called IgE-B

24

KIMISHIGE ISHIZAKA

cells, are distinct from those for IgG/IgM antibodies with respect to their surface immunoglobulin. The question was answered by fractionating antigen-primed rabbit mesenteric lymph node cells with anti-immunoglobulin immunosorbent (74). Thus, mesenteric lymph node cells were obtained from rabbits immunized with DNP-Asc, and the cell suspension was passed through either anti-y-chain- or anti-p-chain-coated Sepharose columns to remove 7-chain- (or p-chain) -bearing lymphocytes. Effluent cells from these columns as well as unfractionated cell suspensions were stimulated with homologous antigen and cultured. The results of the experiments showed that effluent cells from anti-y- or anti-p-chain columns gave poor IgG and IgM antibody responses compared with the unfractionated control, whereas all of the three cell preparations gave comparable IgE antibody responses (Table IV). It appears that memory B cells for IgG and IgM antibody formation bear one or both 7- and p-chain determinants, but those for IgE formation do not. Since B cells bear light-chain determinants together with heavy-chain determinants, the same DNP-Asc primed lymph node cells were passed through an immunosorbent coated with anti-Fab antibody, which removed more than 90% of immunoglobulinbearing ( B ) cells, and the effluent cells were stimulated with DNP-Asc. AS shown in Table IV, the anti-Fab column removed not only the ability to form IgG and IgM antibodies but also the cells essential for the IgE antibody response. Although Wigzell ( 187) has reported that carrierspecific helper T cells were not removed by an anti-immunoglobulin column, Sell and Sheppard (140) claimed that rabbit peripheral lymphocytes responding to T-cell mitogens bore immunoglobulin determinants. This suggested the possibility that poor antibody responses by the effluent TABLE IV EFFECT OF DEPLETION OF IMMUNOGLOBIJLIN-BI<;ARING CELLSO N ANTIHAPTEN ANTIBODY FORMATION Anti-DNP antibody DNP-AX (pg/ml) 0 0.1 0.1 0.1 0.1 a

Immunosorbent -

Anti-y Anti-g Anti-Fab

The PCA titer.

IgM IgG (pg/ml) (pglml) 1.15 6.20 2.50 1.9 0 1.0 5

0.4 20.0 4.4 6.0 1.1

IgE” <2.6 40 40 40 10

CELLULAR EVENTS IN THE IGE ANTIBODY RESPONSE

25

cells of all immunoglobulin classes from the anti-Fab column might be due to removal of carrier-specific helper cells. However, in a separate experiment, an anti-Fab column failed to remove carrier-specific cells that were responsible for a DNA synthetic response to Asc (19). Incubation of DNP-Asc-primed cells with either Asc- or DNP-rabbit serum albumin (RSA) induced an increase of DNA synthesis. Treatment of the cells with anti-Fab immunosorbent removed the cells responding to DNP-RSA, but the DNA synthetic response of the effluent cells to Asc was comparable to that of the original cell preparation. The results suggest that poor antihapten antibody responses of the effluent cells from the anti-Fab column are due to the removal of B cells rather than helper cells. Taken collectively, it appears that B memory cells for the IgE antibody response bear light-chain determinants but neither 7-chain nor p-chain determinant. Although we were unable to perform an experiment to remove selectively B memory cells for IgE by using an anti-€-chain immunosorbent, it would appear that B memory cells for IgE are distinct from B memory cells for IgG/IgM antibodies. So far, lymphocytes bearing 6 determinants have not been detected either in the mouse or in the rabbit, Recently, however, IgE-bearing lymphocytes were demonstrated in rats infected with Nippostrongylus brasiliensis by using a fluorescent antibody technique and radioautography (56). Mesenteric lymph node cells or spleen cells from normal animals did not contain a significant number of IgE-bearing lymphocytes. Such cells appeared in the mesenteric lymph nodes on the eighth day after infection and gradually increased. On the fourteenth day of infection, approximately 4-6% of small lymphocytes in mesenteric lymph nodes and 3 4 %of lymphocytes in the spleen bore €-chain determinants. The appearance of IgE-B cells in the lymph nodes preceded the appearance of IgE-forming plasma cells and correlated with enhanced IgE synthesis following Nippostrongylus infection. Evidence was obtained that IgE on the small lymphocytes was synthesized by these cells. Treatment of these cells with pronase by the method described by Johns et uZ. (62) removed the surface immunoglobulin, but €-chain determinants reappeared on these cells after 24 hours in culture. More recent experiments indicated that a significant number of IgE-forming cells appeared after 5 days of culturing the lymph node cells in the presence of Nippostrongcjlus antigen. Since IgE-forming cells in the lymph nodes disappeared after the culture of the same cell suspension in the absence of antigen, it appears that the antigen enhanced the differentiation of IgE-B cells into IgE-forming cells. These findings support the hypothesis that precursors of IgE-forming cells bear €-chain determinants on their cell surfaces. The present knowledge on the maturation of B lymphocytes suggests

26

KIMISHIGE ISHIZAKA

that IgE-B cells originate in stem cells in the bone marrow. The virgin B cells probably bear p-chain and/ or &chain determinants on their surfaces and some of these cells differentiate to B cells bearing y-chain or a-chain determinants ( 134, 181). Two alternative views have been considered in inducing differential immunoglobulin class expression in lymphoid cells. The first view is an extension of the clonal selection concept that receptor equals antibody and assumes that in the immunocompetent animal separate specific precursor cells exist for each immunoglobulin class prior to antigen presentation (93, 94). Cooper et al. (17) proposed a model of plasma cell differentiation that incorporates a sequential activation of p-, y-, and a-chain genes with cell differentiation principally in the fetal and neonatal period. The second view differs from the first view in that a cell ultimately producing IgG antibody is not necessarily derived from a precursor cell bearing surface IgG, but rather from a cell bearing IgM (180). According to this concept, antigen is one of the essential components in inducing expression of heavy-chain classes other than IgM. Upon stimulation with antigen, virgin B cells bearing p-chain determinants would either directly differentiate into IgM antibody-forming cells or convert into B cells bearing either 7-chain or both p- and y-chain determinants. In any event, the ontogenic sequence from IgM to IgG to IgA is supported by the fact that treatment of newborn mice with anti-pchain suppressed the synthesis of not only IgM but also of IgG and IgA (sea). Quite recently, however, Dwyer et al. (22a) reported that the anti-p-chain treatment suppressed both IgM and IgG antibody responses but failed to affect IgE antibody response, and suggested that an ontogenic development for IgE-bearing B cells may be independent of the IgM to IgG to IgA sequence. Evidence was presented suggesting that T cells are involved in the generation of IgG-B cells. Cheers and Miller (13) showed that mice primed to horse erythrocytes (HRBC) produced a greatly enhanced 3,5dinitro-4-hydroxyphenylacetic ( NNP ) -specific, indirect ( IgG ) , plaqueforming cell response when given NNP-HRBC but did not show any difference in the hapten-specific, direct ( IgM ), plaque-forming cell response in comparison to non-carrier-primed mice. Furthermore, when spleen cells from HRBC-primed mice were transferred to irradiated recipients, there was an enhanced IgG antibody response to NNP-HRBC in the recipients, The results suggested that T cells not only influence the amount but also exert some selective pressure on the emergence of IgG-B cells. Because the IgE antibody response is T-cell-dependent, one can expect that T cells may be involved in the generation of IgE-B cells. However, definite evidence for this hypothesis is lacking. With respect to the generation of TgE-B memory cells, it was found in

CELLULAR EVENTS I N THE IGE ANTIBODY RESPONSE

n

adoptive transfer experiments that immunization of DRA/ 1 mice with 0.05 pg of DNP-OA plus a high dose of alum was more favorable than immunization with 10 pg of the same antigen without adjuvant. Comparisons of spleen cells from the two groups as the source of haptenprimed B cells showed that recipients of the spleen cells from the alumimmunized donors gave a higher antihapten IgE antibody response ( 120). If the animals were immunized with alum-absorbed 0.05- or 10-pg OA without adjuvant and their spleen cells were used as a source of helper cells in adoptive transfer experiments, spleen cells from the two groups were comparable with respect to their helper function for the IgE antibody response. The results were interpreted that alum had a significant effect on the generation of IgE-B cells. If carrier-primed T cells were involved in the generation of these IgE-B cells, however, it is possible that alum might have enhanced proliferation of a certain population of T cells and created favorable conditions for the generation of IgE-B cells. On the other hand, recent experiments on the rat infected with N . brasiliensis suggested that T cells may not be required for the generation of B cells bearing € determinants. It was found that neonatal thymectomy, which abolished the capacity of the animals to form antihapten antibodies against DNP-OA, did not affect the generation of IgE-B cells following infection. At 7 to 10 days after infection of the thymectomized animals, their mesenteric lymph nodes contained many B cells bearing c determinants. The proportion of IgE-B cells in total lymphocytes was higher in the thymectomized animals than in nonthymectomized rats that were infected with N . brasiliensis. Such results suggest that selective proliferation of IgE-B cells occurred without T cells. However, the possibility still remains that a small number of T cells might have disseminated into peripheral lymphoid organs before thymectomy or that some substance released at the site of infection might have stimulated prethymocytes to generate T cells. Differentiation of IgE-B cells to IgE-forming cells is highly dependent on T cells. However, with respect to the T-cell dependency, IgE-B cells may be heterogeneous. An interesting observation was made by Okudaira and Ishizaka (121) on DNP-KLH-primed spleen cells in DBA/1 mice. In an adoptive transfer experiment, depletion of T cells from DNP-KLHprimed cells by anti-8 antiserum and complement abolished the antihapten antibody response of the spleen cells to homologous antigen. If the same T-depleted spleen cells were cultured with DNP-KLH for 24 hours and then transferred into irradiated animals, anti-DNP IgE antibody was formed in the recipients, although an antihapten IgG antibody response was not obtained. It was also found that DNP-KLH-primed spleen cells were triggered for an IgE antibody response if they were

28

KIMISHIGE ISHIZAKA

cultured for 24 to 48 hours with a DNP-heterologous carrier conjugate, such as DNP-BGG or T-independent DNP-D-TGluL. Transfer of cells cultured with the antigen into irradiated recipients resulted in the formation of anti-DNP IgE antibody. Furthermore, DNP-KLH-primed cells cultiired for 24 hours in the absence of antigen and then treated with DNP-BGG at 0°C also gave an adoptive IgE antibody response. The results indicated that the T-cell dependency of IgE-B cells diminished during the culture period. The mechanisms by which IgE-B cells lose their T-cell dependency cannot be explained at present. One possibility is that cultivation of the primed cells caused removal of an unknown regulatory device in B cells or of other regulatory cells that prevented the differentiation of B cells. Alternatively, some of the primed B cells were at a stage when the help of T cells was not required, and such cells might have proliferated in culture. Although the explanation for this phenomenon is incomplete, it was suggested that T cells’ help may be required only at a critical stage of maturation of IgE-B cells. It should be noted that the results described above did not occur when DBA/1 mice were immunized with DNP-OA instead of DNP-KLH. The difference may be related to the fact that anti-DNP IgE antibody response was persistent when the animals were immunized with DNP-KLH, but the antibody response to DNP-OA declined within 3 to 4 weeks. One might speculate about the possibility that the presence of T-independent B cells may be related to a persistent antibody response.

C. GENERATION OF

A

HELPERFUNCTION FOR ICE ANTIBODYRESPONSE

It is clear from previous work that induction of an IgE antibody response is highly dependent on dose of antigen and adjuvant vehicle employed. In rabbits and rats, use of a certain adjuvant, such as alum and/or pertussis vaccine, with antigen was required for both IgE and IgG antibody responses. Immunization with the same antigen included in CFA gave a high IgG antibody response but a transient or no IgE antibody response. In order to study the role of T cells in distributing antibody production among different isotypes, attempts were made to reproduce this dissociation between IgG and IgE antibody responses by using different adjuvants for carrier priming. It was found that priming of rabbits with Asc in alum prior to immunization with DNP-Asc enhanced both IgG and IgE antihapten antibody responses, whereas priming with the same carrier included in CFA enhanced only the IgG antibody response (75). Tada et al. ( 160) showed that immunization of rats by repeated injections of a high dose (200 p g ) of Asc included in CFA suppressed the IgE antibody response to subsequent immunization with

CELLULAR EVENTS I N T H E IGE ANTIBODY RESPONSE

29

DNP-Asc plus pertussis vaccine. These results suggested strongly that the effect of adjuvant is on the priming of T cells and that carrier-prinied T cells actually control the distribution of antibodies among different immunoglobulin classes. A quite interesting observation was made by Okumura and Tada (123) who demonstrated that the transfer of thymocytes or spleen cells from animals hyperimmunized with Asc included in CFA depressed on-going IgE antibody formation to DNP-Asc in the recipients. In this experiment, recipients were irradiated with a sublethal dose of X-rays immediately before immunization with DNP-Asc plus pertussis vaccine to sustain IgE antibody responses. Donors were hyperimmunized with Asc included in CFA, and 10y-lO’ thymocytes or spleen cells were transferred into the recipients which were producing high titers of IgE antibody. As shown in Fig. 4, the production of IgE antibody was drastically reduced within 2 days after the transfer. From these findings, the authors suggested that carrier-primed T cells not only collaborate with hapten-specific B cells for the IgE antibody response but also regulate or terminate the antihapten IgE antibody response. They postulated that the presence of too large a population of carrier-specific helper T cells may be unfavorable for the formation of IgE antibodies ( 158). One may question whether thymocytes enhance or suppress the IgG antibody response. Unfortunately, the findings were limited to the IgE 512

1

Cell transfer

1

2 56 128 64

4

2 I

0

5

10

15

20

25

30

DAYS

FIG.4. Inhibitory effect of thymus ( T ) and spleen (Sp) cells obtained from rats immunized with unconjugated Ascaris antigen ( As ) but not from those immunized with hapten ( DNP) coupled to heterologous carrier ( DNP-immunized). [From Okuniura and Tada (123, 158).]

30

KIMISHIGE ISHIZAKA

antibody response, because these X-irradiated animals failed to form IgG antibody against DNP-Asc. Nevertheless, the hypothesis was supported by subsequent observations that the transfer of thymocytes from donors preimmunized with Asc in CFA restored the ability of tolerant recipients to produce IgE antibody against DNP-Asc (160). Thus, a high dose of Asc was injected into neonatal rats within 24 hours of birth to make them tolerant to subsequent immunization with DNP-Asc. These animals did not form IgE antibody upon immunization with DNPAsc plus pertussis vaccine. However, supplementation of the animals with thymocytes of Asc-immunized animals prior to the immunization restored the capacity to form IgE antibodies to DNP-Asc. The thymocyte preparation transferred into the tolerant animals was obtained by hyperimmunization of donors with Asc included in CFA, the same preparation that suppressed on-going IgE antibody responses in the irradiated animals. Thus, the same thymocyte preparation could either enhance or suppress the IgE antibody response under different circumstances. The hypothesis proposed by Tada et al. (158) may explain the fact that priming of mice with a subimmunogenic dose of carrier enhanced the antihapten IgM antibody response to hapten-homologous carrier conjugate, but an immunogenic dose of carrier did not show this carrier effect (26). Indeed, an optimal dose of carrier priming was essential for the enhancement of an antihapten IgE antibody response. When DBA/ 1 strain mice were primed with various doses of OA included in CFA and then immunized with DNP-OA, priming with a subimmunogenic dose of carrier enhanced both IgE and IgG antihapten antibody responses, whereas priming with an excess amount of carrier was inhibitory for antihapten antibody responses of both immunoglobulin classes (Table V ) . An intermediate dose for carrier priming, such as 1.0 pg, enhanced the IgG antibody response but suppressed the IgE antibody response (51). Similar observations were made by Hamaoka et al. ( 3 8 ) in A/ J strain mice using KLH as antigen. Following the hypothesis i f Tada et al. (158), this discrepancy between IgE and IgG antibody responses may be explained if the optimal number of helper T cells for the IgE antibody response is less than that for the IgG antibody response, and if too many helper cells suppress the IgE antibody response. However, one should not overlook the fact that antihapten antibody is not the only antibody formed against a haptencarrier conjugate. When an immunogenic dose (O.OS10 p g ) of OA was used for carrier priming, injection of DNP-OA induced not only a primary response to the DNP group but also a secondary antibody response to OA, and carrier priming increased anticarrier antibody responses of both IgG and IgE classes. Thus, an apparent suppression of antihapten

31

CELLULAR EVENTS IN THE IGE ANTIBODY RESPONSE

TABLE V EFFECTOF CARRIERPRIMING O N ANTI-DNP A N D ANTI-OA ANTIBODYRESPONSES AGAINST DNP-OAnsb Priming with carrier. (pg) 0.02d

0.05 1 .o

10.0

An ti-D N P

Anti-OA

IgE

IgG

IgE

IgG

Enhanced Accelerated Suppressed Suppressed

Enhanced Enhanced Accelerated Suppressed

Accelerated Accelerated Secondary Secondary

No effect No effect Secondary Secondary

From Ishizaka and Okudaira (51). Enhancement: antibody was detected in the serum earlier and maximum titcr was higher than in control mice that were not primed with carrier. Acceleration: antibody appeared and reached maximum earlier than control. Maximum antibody titer in primed mice was not significantly higher than that in control. Secondary: anti-OA antibody was detectable before immunization with DNP-OA. c Intramuscular injection of OA included in CFA 2 weeks beforc immunization with 0.0.5 pg DNP-OA in alum. Dose of DNP-OA for challenge was 0.02 pg in this group. a

b

antibody response following high-dose carrier-priming may well be due to the presence of carrier-specific B memory cells that compete for common helper T cells with unprimed hapten-specific B cells. It is also possible that anti-carrier antibody might have prevented an antihapten response. In order to eliminate a possible role of carrier-specific B cells in an apparent suppression of the antihapten antibody response, Hamaoka et al. (38) analyzed helper cells in carrier-primed animals first exposed to a sublethal dose of X-rays. The experimental design was based on the fact that a high degree of T helper function was maintained after sublethal X-irradiation, but B memory cells were susceptible to irradiation (36). Strain A/ J mice were primed with varying doses of KLH included in CFA. Three weeks later, these animals were irradiated and then received DNP-Asc-primed spleen cells as the source of hapten-specific B cells. Immediately after the transfer, the mice were challenged with DNP-KLH, and antihapten IgG and IgE antibody responses were observed. An increase of KLH dose from 0.01 to 10 pg resulted in a parallel increase of helper functions for both IgE and IgG antibody responses. Further increase of KLH to 100 pg for carrier priming diminished helper functions for both IgE and IgG antibody responses. The results confirmed the point that maximal helper function was obtained by an optimal dose of carrier priming but failed to support the hypothesis that an opti-

32

KIMISHIGE ISHIZAKA

ma1 number of helper T cells for an IgE antibody response is less than that for an IgG antibody response. The helper function for IgG and IgE antibody responses was explored in the in vitro antibody response of rabbit lymph node cells as well (76). Rabbits were immunized with alum-absorbed DNP-Asc and given a supplemental immunization of varying doses of BGG included in CFA to raise BGG-specific helper cells. Aliquots of the mesenteric lymph node cells from these animals were stimulated with either DNP-Asc or DNPBGG and then cultured for the formation of antihapten antibodies (Table VI). As expected, a supplemental immunization with BGG enTABLE V I ANTI-DNP ANTIBODY FORMATION I N LYMPHNODE CELLS PRIMED WITH DNP-Asc A N D BGG" Supplemental immunizationb BGG CFA BGG CFA BGG CFA BGG CFA BGG CFA BGG Alum None

Anti-DNP

Helper effcctc

Antigen in vitro

IgM Gcg/ml)

IgG (/Jcrg/ml)

IgE PCA)

-

0.75 4.35 4.2 3.6 8.2 8.4 1.8 6.8 4.8 1.25 6.10 4.0 1.07 7.0 0.6 0.75 2.60 1.60 0.15 9 .3 0.6

0.7 40.0 50.0 0.04 86.0 80.0 0.04 39 29.2 0.2 35.0 17.0 0.27 42.5 11.0 0.7 34.0 13.5 0.1 46.6 2.6

<2.5 40 <2.5 <2.5 80 <2.5 <2.5 100 <2.5 <2.5 30 <2.5 <2.5 40 <2.5 <2.5 40 10 <2.5 320 <2..5

DNP-AX DNP-BGG DNP-ASC DNP-BGG DNP-ASC D N P-B G G -

DNP-ASC I) N P-B G G -

DNP-Asc DNP-BGG -

DNP-ASC DNP-BGG DNP-ASC DNP-BGG

IgG (%)

IgE (%)

125

0

93

0

75

0

49

0

26

0

40

2-5

5.6

0

From Kishimoto and Ishizaka (76). 1 or 10 pg of BGG included in CFA or alum. For each immnnoglobulin class, antibody formed by the stimulation with DNPBGG was compared with that formed by the stimulation with IINP-Asc: a

* Either

helpcr effect (%)

=

AhnNp-mx - Abo A ~ D N Y --AAbo ~~

x

100

CELLULAR EVENTS IN THE ICE ANTIBODY RESPONSE

33

hanced both IgM and IgG antihapten antibody responses to DNP-BGG. The results in Table VI clearly show that DNP-Asc induced both IgE and IgG antibody responses, whereas DNP-BGG elicited the formation of IgG antibody but not IgE antibody. With respect to the IgG antibody response, DNP-BGG was more effective than DNP-Asc in some cases but less effective in others. Because the IgG-13 cells stimulated by DNPAsc and DNP-BGG are common to both, the ratio of anti-DNP IgG antibody formed by the two antigens would represent the helper function of BGG-specific cells relative to the Asc-specific cells. In the experiments shown in the Table VI, the size of the BGG-specific helper cell population, as judged by helper function for the IgG antibody formation, was 25125%of the Asc-specific helper cells. On the other hand, formation of anti-DNP IgE antibody upon stimulation with DNP-Asc indicated that all of the primed cell suspensions contained a sufficient number of IgE-B cells for the antibody response. Therefore, failure of the IgE antibody response to DNP-BGG indicated that BGG-specific helper cells, which collaborated with IgG-B cells, failed to collaborate with IgE-B cells. If alum was used with BGG for supplemental immunization, however, the mesenteric lymph node cells of the DNP-Asc-primed animals responded to both DNP-Asc and DNP-BGG for IgE antibody synthesis (Table VI ) . Comparison of the results obtained with alum immunization and CFA immunization indicates that the ability of BGG-specific helper cells to collaborate with DNP-specific IgE-B cells had no relationship to their helper function for the IgG antibody response. The results were reproduced using Rag as a secondary carrier. As already shown in Table 111, supplemental immunization of DNP-Ascprimed animals with aluni-absorbed Rag provided Rag-specific helper cells for both IgE and IgG antihapten antibody responses to DNP-Rag. If the animals were supplemented with the same ragweed antigen included in CFA, their lymph nodes did not form IgE antibody upon stimulation with DNP-Rag, although the same cell suspension responded to DNP-Asc to form the antibody. Again, no difference was observed between alum immunization and CFA immunization with respect to the generation of helper cells for an IgG antibody response. In order to confirm the sizes of the Rag-specific cell populations in the mesenteric lymph nodes of both the alum group and CFA group, lymph node cells from both groups were stimulated with Rag, and thymidine-3H incorporation was determined. The experiment was based on the fact that a T-cell population is responsible for the DNA synthetic response to carrier in rabbit lymph node cells (19). The results showed that the sizes of the Rag-specific T-cell population were comparable irrespective of whether

34

KIMISHIGE ISHIZAKA

alum or CFA was employed for the carrier (supplemental) immunization and excluded the possibility that the failure of CFA immunization to generate helper function for IgE antibody response was due to too large a population of helper cells. One might consider that CFA immunization could induce proliferation of carrier-specific B cells that might have interfered with the antihapten antibody response (21). In this system, however, the size of the Rag-specific IgG-B cell population was comparable in the alum group and CFA group. Taken collectively, these data suggest strongly that carrier-specific helper cells collaborating with IgE-B cells are different from those collaborating with IgG-B cells. As will be described later, this hypothesis is supported by subsequent experiments showing that distinct enhancing soluble factors are involved in IgE and IgG antibody responses. Because the mouse is the best animal species for the formation of IgE antibodies, Hamaoka et al. (37) used mice to perform similar experiments with an adoptive transfer technique. However, none of the experiments carried out by these investigators or by us (119, 120) provided evidence for the presence of distinct T-cell subpopulations for different immunoglobulin classes, Investigators of both groups have shown that a helper function for the IgE antibody response was obtained by immunization of mice with an appropriate carrier included in CFA. When donors of helper cells were primed with OA either in alum or in CFA, the transfer of their spleen cells together with DNP-KLH-primed spleen cells into irradiated animals followed by challenge with DNP-OA induced both IgE and IgG antihapten antibody responses in the recipients. Hamaoka et al. (38) analyzed the effect of immunogenic strength and dose of antigen as well as the nature of the adjuvant vehicle for immunization on the magnitude of helper function generated for IgG and IgE classes. A series of experiments compared helper functions generated by three carrier proteins, i.e., KLH, OA, and BGG, differing in their inherent immunogenicity in A/ J mice. Carrier-primed spleen cells were obtained from mice immunized with one of the three proteins varying in log increment doses from 0.01 to 100 pg and administered either in alum or CFA. The cells were then transferred into irradiated mice together with DNP-specific B cells. Evaluation of these carrier-specific cells for relative IgG vs IgE helper activity revealed that with equally effective antigens, such as KLH and OA, the T-helper cell function for both IgE and IgG secondary antibody responses could be generated with combinations of either alum or CFA. Although the authors confirmed that alum was a better adjuvant than CFA for the IgE antibody response relative to the IgG antibody response, dissociation of the two responses was less

CELLULAR EVENTS IN THE IGE ANTIBODY RESPONSE

35

ObViOUS in analyses of the ability of helper cell populations generated by carrier immunization. Subsequently, Katz et al. (68) analyzed ( i ) allogeneic effects on IgE and IgG antibody responses and ( i i ) adoptive secondary IgE and IgG responses produced by administration of hapten and carrier determinant on separate molecules (see Section IV,D,l). As a result of these studies, they proposed that differences observed in the degree of helper function for one or the other antibody class, depending on the conditions of T-cell priming, may relate to differences in sensitivities of IgE- vs IgG-B cells to the regulatory influences of the same T cells rather than to the existence of different T cells for different immunoglobulin classes. Experiments by Okudaira and Ishizaka (120) as well as by Hamaoka et al. (38) showed that adjuvant was not required for priming T helper cells for the IgE antibody response if an appropriate dose of a potent immunogen was injected into high-responder mice. Using an adoptive transfer technique, they showed that immunization of donors with OA in saline, which induces neither IgG nor IgE antibody responses, resulted in a high level of helper function for both IgE and IgG classes in the adoptive response of DNP-KLH-primed cells to DNP-OA. More recently, Ishizaka and Adachi (43) succeeded in priming helper cells for both IgE and IgG antibody responses by culturing normal splenic T cells of high-responder ( BDFl ) mice on OA-bearing peritoneal macrophages for 5 days. Transfer of the cultured lymphocytes together with DNPKLH-primed spleen cells into irradiated syngeneic recipients followed by challenge with DNP-OA resulted in both IgE and IgG antihapten antibody responses. Helper cells were not generated if normal splenic lymphocytes were cultured with OA in the absence of macrophages. The helper function of the cultured lymphocytes was abolished by treatment with anti4 antiserum and complement and was specific for OA. These findings indicated that OA-bearing macrophages primed normal T cells in uitro and developed antigen-specific helper T cells for both IgE and IgG antibody responses, Thus, extensive studies on splenic lymphocytes of high-responder mice failed to provide evidence for the existence of distinct T cells for different immunoglobulin classes. The question remains how can one reconcile the conflicting findings obtained in the rabbit and mouse. From our present knowledge, there is no way to justify one or the other hypothesis. An important fact to be considered is that the IgE antibody response is readily induced in certain strains of high-responder mice that were employed in almost all studies described, but the antibody response in the rabbit was obtained only when the animals were immunized with an adequate dose of antigen

36

KIMISHIGE ISHIZAKA

together with suitable adjuvant. In other words, an immunization regimen in the rabbit was selected so that the IgE antibody response was readily obtained. Such difficulty in including an IgE antibody response in the rabbit may denote some critical factor for IgE antibody response in this species that is minimally important or even bypassed in the mouse strains in which an IgE antibody response is induced as readily as an IgG antibody response. However, under some extreme conditions, such as helminth infection, there is a possibility that T cells may be stimulated in a special way to facilitate an IgE antibody response even in the mouse. Kojima and Ovary (81) studied helper cells elicited by infection with Nippostrongylus brasiliensis and indicated that T cells for IgE-B cells might be different from those for IgG-B cell. They were able to demonstrate by an adoptive transfer experiment that, upon challenge with DNP-worm extract (Nb antigen), spleen and lymph node cells of infected mice specifically augmented the synthesis of anti-DNP IgE antibody in the irradiated host that had received DNP-KLH-primed cells. Since the potentiating effect of parasite infection is mostly, if not exclusively, directed to IgE antibody formation, one can hypothesize that the parasites infection might stimulate a certain population of T cells that augments; specifically or nonspecifically the differentiation of IgE-B cells ( 82) Further analysis of the IgE response following parasite infection might: provide a clue for determining whether or not helper T cells for IgE andl IgG-B cells belong to distinct subpopulations of T cells. I

D. MECHANISMS OF T CELL-BCELLCOLLABORATION 1. Enhancing Soluble Factors Choosing from a number of hypotheses to explain the mechanisms of T cell-B cell collaboration for the hiimoral antibody response, Dutton et al. (22) suggested that specifically activated T cells may release a nonspecific soluble factor that enhances the differentiation of B cells bound with antigen through their receptors. In the past 5 years, evidence has accumulated to support this hypothesis (2, 35, 136, 138, 182). Gorczynski et al. (35) as well as Rubin and Coons (136) showed that culture fluids of spleen cells primed with antigen enhanced the direct plaque-forming cell response of normal spleen cells to sheep erythrocytes (SRBC), Because we have established a culture system in which anamnestic IgE and IgG antibody responses can be observed in vitro, experiments were carried out to study whether the helper effect of carrierspecific cells is mediated by a soluble factor. As suggested by Dutton et al. (22), a critical experiment to demonstrate enhancing soluble factor

CELLULAR EVENTS I N THE I G E ANTIBODY RESPONSE

37

is to follow the immune response upon stimulation of T and B cells by two independent antigen molecules. Thus, rabbits were immunized with alum-absorbed DNP-Asc, and their mesenteric lymph node cells were stimulated with a mixture of DNP-coupled to an unrelated carrier, i.e., DNP-KLH, and a homologous carrier ( Asc). As shown in Fig. 5, stimulation of DNP-Asc-primed cells with DNP-KLH alone resulted in the formation of a small amount of antihapten IgG antibody but essentially no IgE antibody. The addition of free carrier to DNP-KLH enhanced the IgG antibody response by three- to six-fold and induced a high titer of IgE antibody. The IgG and IgE antibodies formed by a mixture of DNP-KLH and homologous carrier amounted to 40-70% of those formed by stimulation with DNP-Asc. As expected, incubation of the same primed cell suspension with homologous carrier alone failed to induce an anti-DNP antibody response. A critical condition for enhancement of the antihapten antibody response was the concentration of the homologous carrier added to the system. Although an optimal concentration of DNP-Asc for maximal antibody response was 0.01-0.1 pg/ml, enhancement of the antihapten antibody response to DNP-KLH was obtained by the addition of 10 to 100 pg/ml of Asc. This concentration corresponded to the optimal concentration of Asc for a maximum DNA synthetic response, indicating that stimulation of carrier-specific cells is involved in the enhancement. The results suggest that a soluble factor was released from carrier-specific cells and enhanced the response of DNP-specific B cells to the DNP-heterologous carrier conjugate. Similar observations were made by Hamaoka et al. (36) in the adopANTIGEN FOR STIMULATION

ANTI- DNP Ig E ANTIBODY IPCA)

<5

160

40

640

DNP-ASC

DNP-KLH

DNP-KLH Asc

t

4

8

16

32

64

ANTI- DNP Iq G ANTIBODY(pq/ml)

FIG.5 . Enhancement of anti-DNP antibody responses of DNP-Asc-primed lymph node cells to DNP-heterologous carrier (DNP-KLH) by free carrier (Asc). Both IgE antibody titer (blank bars) and IgG antibody titer (hatched bars) in culture fluid are shown.

38

KIMISHIGE ISHIZAKA

tive secondary response of the mouse. They transferred DNP-Asc-primed spleen cells and Asc-primed spleen cells into irradiated syngeneic mice and then challenged them either with DNP-KLH alone or with a mixture of DNP-KLH and Asc. Both IgE and IgG antibody responses of the recipients to DNP-KLH were enhanced by Asc. Katz et al. (68) extended this observation through the demonstration that maximum enhancement of the IgE antibody response was obtained if Asc was injected 1-2 days prior to the injection of DNP-KLH. If DNP-KLH and Asc were injected 1 4 days apart, then enhancement of the IgE antibody response did not parallel that of IgG antibody. An interval of 2 days between administration of carrier and hapten determinant was the limit for elicitation of an IgG antibody response, whereas IgE-B cells were triggered even after 3 days interval. Their explanation for this difference was that IgE-B cells can respond to a lower threshold of T-cell activity than is required for IgG-B cells. Kojima and Ovary (82) carried out similar adoptive transfer experiments using spleen and mesenteric lymph node cells from mice infected with N . brasiliensis. Irradiated recipients of DNP-KLH-primed spleen cells and lymphocytes from the infected mice were challenged with either DNP-OA or a mixture of DNP-OA and Nb antigen. Challenge with DNP-OA elicited a very low titer of anti-DNP IgE antibody. However, concomitant administration of Nb antigen with DNP-OA significantly erlhanced the IgE but not the IgG antibody response. All of the experiments both in tiitro and in tiitio strongly suggested that the helper function of T cells for the IgE antibody response was mediated by an enhancing soluble factor. Definitive evidence lor the hypofhesis was dbtaineh ‘by using cdture supernatant of carrier-primed lymphocytes ( 77). In the experiment shown in Fig. 6, lymph node cells of a Rag-primed rabbit were incubated with Rag for 24 hours and cell-free supernatant (CFS) was obtained. Independently, lymph node cells from DNP-Asc-primed rabbit were incubated with DNP-D-TG~uL,which is a T-independent antigen (25), to stimulate DNP-specific B cells. After washing to remove the DNP conjugate, aliquots of these cells were resuspended in either fresh medium or a CFS preparation obtained from Rag-primed cells, and both cell suspensions were cultured. It is apparent that anti-DNP antibody responses of both IgG and IgE classes were significantly enhanced by CFS, indicating that the culture supernatant contained enhancing soluble factor( s ) . It should be noted that the factor(s) itself does not have the carrier specificity, because the differentiation of DNP-specific cells raised by the immunization with DNP-Asc was enhanced by CFS of Rag-primed cells. In this respect, the factor( s ) detected in this experiment is a “nonspecific factor( s ) .” The release of the soluble factor( s )

CELLULAR EVENTS IN THE IGE ANTIBODY RESPONSE

Rag-p r ime d Lymph node c e l l s

DNP-Asc primed cells

+ Rag

I

-1

24 hr Cell free sup. \ \ \ \ \ \ \

1 ;

Suspend in ‘Ace11 f r e e sup.

1.

6 days culture

culture f l u i d IgE IgG 10.oug 10

Suspend in fresh medium

1

6 days culture

culture f l u i d IgG IgE <2.5 1 . 4 ~

FIG. 6. Experimental design to demonstrate nonspecific enhancing soluble factor for both IgE and IgC antibody responses. Donors of Rag-primed cells were immunized with alum-absorbed Rag, and donors of DNP-primed cells were immunized with alum-absorbed DNP-Asc.

was not inhibited by cytosine arabinoside, which is a strong mitotic inhibitor, but was inhibited by pactamycin. The results indicated that cell proliferation was not required, but de novo synthesis of protein was essential for the formation of soluble factor( s ) (79). The experimental system was applied to studying whether the enhancing factor for an IgE antibody response is different from the factor for an IgG antibody response. As already described, the helper function of carrier-specific cells for an IgE antibody response varies depending on the adjuvant vehicle employed for supplemental immunization (see Section IV,C). If the helper function of T cells is actually mediated by enhancing soluble factor, one can expect that the enhancing activity of CFS for an IgE antibody response may depend on the adjuvant used for carrier priming. In order to study this possibility, we compared the enhancing activity of different CFS preparations using the same DNPprimed B cells that were raised by immunization of a rabbit with DNPAsc (78). Two donors of carrier-primed cells were immunized with Rag included in CFA (CFA group) and the other two were immunized with the same antigen absorbed to alum (alum group). Lymph node cells from each of the animals were incubated with Rag to recover CFS.

40

KIMISHIGE ISHIZAKA

Mesenteric lymph node cells from DNP-Asc-primed animals were stimulated with DNP-KLH, and aliquots of the cell suspension were cultured in one of the four CFS preparations or in fresh medium. The results showed that all four preparations of CFS enhanced the IgG antibody response. With respect to the IgE antibody response, however, only the CFS preparations derived from alum group enhanced the antibody response. There was no relationship between the enhancing activity for IgE and IgG antibody responses. Because the IgE-B cells and IgG-B cells were common to all of the cultures, the difference in the IgE antibody response must be due to enhancing factors present in the CFS preparations rather than the sensitivity of B cells. Because the results suggested that enhancing soluble factors for IgE and IgG antibody responses may be different, CFS preparations were fractioned by density gradient ultracentrifugation, gel filtration, and block electrophoresis, and then the fractions were tested for the presence of enhancing activities ( 79). Electrophoretically, fractions corresponding to a- and p-globulin regions had enhancing activities for both IgE and IgG classes, After the CFS preparation was fractionated by gel filtration through a Sephadex G-200 column or by density gradient ultracentrifugation, the IgE antibody response was enhanced by the 7 S fraction, whereas the IgG antibody response was enhanced by a fraction containing molecules with molecular weight of 20,000 to 40,000. The presence of enhancing activity in the two different fractions correlated with the activity of the original CFS preparations. When donors of carrier-specific cells were immunized with carrier included in CFA, then CFS preparations obtained from the lymph node cells enhanced IgG antibody responses but not IgE antibody responses. Fractionation of such a CFS preparation by gel filtration showed that the 20,00040,000 mol wt fraction enhanced the IgG antibody response but the 7 s fraction failed to enhance the IgE antibody response. The results indicated that IgEenhancing factor is distinct from IgG-enhancing factor, and that the failure of some CFS preparations to enhance an IgE antibody response is due to lack of IgE-enhancing factor. Because CFS preparations contained soluble antigen ( carrier) that was used for stimulation, the possibility was considered that the enhancing factors might be antigen-T-cell receptor complexes. However, the following experiment showed that antigen does not appear to be an integral part of the factor. It was found that incubation of carrier-primed lymphocytes with carrier-coated Sepharose released the enhancing factors for both IgE and IgG antibody synthesis (79). Absorption of the CFS either with antigen-coated immunosorbent or with anticarrier antibodycoated immunosorbent failed to remove the enhancing activities. The en-

CELLULAR EVENTS IN THE IGE ANTIBODY RESPONSE

41

hancing factors were susceptible to trypsin but not to DNAse. It was also found that neither the IgE- nor IgG-soluble factor contained immunoglobulin determinants. When CFS was absorbed with anti-imniunoglobulin ( anti-y, anti-p, and anti-Fab ) immunosorbent, essentially all activity was recovered in the effluent. Enhancing soluble factors were not absorbed by IgG-coated immunosorbent, indicating that the factors do not have affinity for immunoglobulin. This, nonspecific enhancing factors obtained from rabbit lymph node cells are similar to those for IgG and IgM antibody responses in the mouse. Indeed, the molecular size of the IgG-enhancing factor was cornparable to that of allogeneic effect factor ( 2 ) obtained from mouse spleen cells. The molecular size of IgE-enhancing factor is comparable to that of T-cell-replacing factor described by Watson ( 182). It appears that enhancing factors are heterogeneous with respect to their molecular sizes and that IgE enhancing factor is not unique in this respect. An antigen-specific T-cell factor that replaced T cells in IgM antibody responses was described by Feldmann (27, 28) and by Taussig et al. ( 170). According to Feldmann, his soluble factor contained both lightchain and p-chain determinants and might represent antigen receptors on T cells ( IgT), but soluble factors described by Taussig et al. did not contain immunoglobulin determinants ( 107). With respect to the IgE antibody response, Taniguchi and Tada (167) described an antigenspecific enhancing factor in the extract of thymocytes and spleen cells of rats hyperimmunized with DNP-Asc or Asc. As described in a previous section, the same investigators showed that thymocytes or spleen cells of rats hyperimmunized with Asc in CFA restored the ability of neonatally tolerized rats to form IgE antibody against DNP-Asc (160). Thus, they disrupted the thymocytes by freezing and thawing and injected the CFS derived from these cells into neonatally thymectomized rats along with immunizing them with DNP-Asc and pertussis vaccine. The results showed that the extract restored the ability of the thymectomized rats to forin IgE antibody against DNP-Asc. As the thymus extract from (either normal rats or those hyperimmunized with DNP-BSA failed to reconstitute the thymectomized animals for the IgE antibody response to DNP-Asc, the authors concluded that the T-cell-replacing component in the extract had carrier specificity. Fractionation of the thymus extract showed that the inductive component had a molecular size comparable to IgG and contained both light-chain and p-chain determinants but was not absorbed by antithymocytes. All of these properties were similar to those of IgT reported by Feldmann ( 27). h4ore recently, however, Taniguchi and Tada found that monomeric IgM produced by reduction and alkylation of serum IgM antibody against Asc was capable of recon-

42

KIMISHIGE ISHIZAKA

stituting the IgE antibody response of neonatally thymectomized rats to DNP-Asc ( 153). Because the T-cell-replacing component extracted from the thymus had the same molecular size and antigenic characteristics as monomeric IgM and lacked the antigenic marker of thymocytes, the authors wondered if the active materials in the thymus extract might be monomeric IgM antibody. 2. Zmmunogenic Specificity of T Cells It is generally accepted in the antihapten antibody response that B cells have receptors specific for haptenic determinant, and T helper cells are specific for the carrier. Theoretical background obtained in the hapten-carrier systems may be applied in natural antigen systems as well. Evidence is now available that the specificity of T-cell receptor for some antigens appears to be different from the specificity of B-cell receptor. Using glucagon, Senyk et al. (141) demonstrated that antibody against whole glucagon molecules had a specificity primarily for the amino-terminal heptadecapeptide and showed little or no binding with the carboxyl-terminal dodecapeptide. However, when antigen-primed lymph node cells were stimulated with the fragment, the C-terminal fragment induced blast transformation and synthesis of DNA, whereas the amino-terminal peptide failed to do so. Alkan et al. (1) also demonstrated that a l-tyrosine-azobenzene arsonate-poly-yD-glutamic acid conjugate reacted with B and T cells through different portions of the molecules. In view of these findings, we have studied the possibility that immunogenic determinants for helper T cells in ragweed antigen E molecules may be different from the major antigenic determinants for B cells. The experiments were based on the fact that the secondary anti-DNP antibody response of DNP-Rag-primed rabbit lymph node cells was induced by the stimulation of B and T cells with a mixture of two separate molecules; i.e., DNP-KLH and antigen E, indicating that helper T cells in this system are specific for antigen E. With respect to antigen E molecules, King et al. (73) have shown that antigen is irreversibly denatured in 8 M urea because of its dissociation into two polypeptide chains, (Y and p. The denatured antigen preparations (UD antigen) as well as the and p-polypeptide chains failed to combine with rabbit or human IgG antibody against the native antigen or to induce erythema wheal reactions in ragweed-sensitive individuals. However, these modified antigens enhanced anti-DNP antibody responses of DNP-Rag-primed rabbit lymph node cells to DNP-KLH (49). As shown in Fig. 7, stimulation of the primed cells with DNP-KLH alone resulted in the formation of a small amount of IgG antibody but essentially no IgE antibody response. (Y-

CELLULAR EVENTS IN THE ICE ANTIBODY RESPONSE

43

IyGpq/rnl

40

5 20 10

o '

w

~

W

o

n

n

nY

Y n

~

o

FIG.7 . Enhancement of anti-DNP antibody responses of DNP-Rag primed rabbit lymph node cells to DNP-KLH ( D K ) by either native antigen E or modified antigen. The UD and RC represent urea-denatured and reduced-carboxylated antigen E ( A g E ) . The 01 and p are the dissociated polypeptide chains resulting from UD of AgE. Both IgC and IgE antibodies in the culture fluids are shown [From Ishizaka et al. (49).]

If the primed cells were treated with DNP-KLH for 24 hours and then cultured in the presence of one of these modified antigens, both IgG and IgE anti-DNP antibodies were formed in uitro. The ability of the modified antigens to stimulate carrier-specific helper cells was comparable to that of native antigen, indicating that the modified antigens were capable of stimulating antigen E-specific helper T cells for the release of enhancing soluble factor. From the theoretical viewpoint, the results excluded the possibility that carrier-specific B cells were involved in the release of enhancing factor( s ) . Since the rabbits immunized with DNPRag did not have antibody reacting with modified antigen E, B cells specific for the modified antigen would not be present in the lymph node cells. Evidence was obtained also that these modified antigens stimulated human T cells specific for antigen E. All of the UD antigen, a- and p-chain, induced DNA synthetic responses of the lymphocytes from ragweed-sensitive patients but failed to enhance thymidine incorporation of lymphocytes from normal individuals. The stimulation indices obtained with the same concentration of modified antigens and native antigen E were comparalde. Because the patients had no antibody against the modified antigen, it appears that the DNA synthetic response was due to the stimulation of T cells rather than B cells.

44

KIMISHIGE ISHIZAKA

The UD antigen and a-chain are poor immunogens, but hyperimmunization of A/J strain mice with these antigens included in CFA resulted in the formation of antibodies. The antiserum was specific for a-chain but did not react with the native antigen. On the other hand, immunization of the mouse with native antigen E resulted in the formation of both IgE and IgG antibodies but the antiserum did not react with a chain. The results showed that anti-a-chain and anti-antigen E had different specificities and suggested that B cells specific for a chain are different from those specific for native antigen (52). In spite of the different specificities at the level of B cells, both a-chain and U D antigen could prime T cells specific for the native antigen. If one primes mice with the modified antigen prior to immunization with native antigen E, the antibody response to the native antigen is enhanced. The best evidence for T-cell priming by the modified antigen was obtained by using an adoptive transfer technique (166). As shown in Fig. 8, spleen cells from a-chain- or UD-antigen-primed mice collaborated with DNP-primed B cells for the adoptive secondary anti-DNP antibody response to DNP-Rag in irradiated recipients, and the helper function of a-chain- or UD-antigen-primed spleen cells for the adoptive antibody response was abolished by treatment with anti-8 antiserum and complement. The difference between immunogenic determinants for T cells and antigenic determinants for B cells is not unique for antigen E. Essentially Helper C e l l s Priming Treatment -~ i n vitro

r-chain

Anti-B+C

Ig E ANTI-DNP ANTIBODY (PCA TITER 0 1 (10 10 20 40 80

P 05 10 2 0 40 IgG ANTI-DNP ANTIBODY (ABA pg Pro1 /ml & A ‘ 1

FIG. 8. Priming of antigen E-specific T cells by immunization with a-chain. Donors of helper cells were primed with 0.5 p g a-chain. 10’ spleen cells from the donors were mixed with lo’ DNP-KLH-primed spleen cells and transferred into irradiated recipients. A portion of a-chain-primed spleen cells were treated with anti-0 antiserum and complement before transfer. The control group received normal spleen cells and DNP-KLH-primed cells. All recipients were challenged with DNP-Rag. Both IgE and IgC anti-DNP antibody titers on the tenth day are shown in the figure. ABA = antigen-binding activity. [From Takatsu et d. (ISS).]

CELLULAR EVENTS I N THE ICE ANTIBODY RESPONSE

4s

the same results were obtained by denaturation of OA in 8 M urea. Ureadenatured OA did not combine with the antibodies specific for native OA but reacted with OA-specific T cells. Immunization of BDFl mice with UD-OA induced helper cells specific for OA ( 163). Thus, in antigen molecules, such as antigen E and OA, the B-cell and T-cell determinants seem to play the same role as haptenic and carrier determinants in hapten-carrier conjugates. V. Regulation of IgE Antibody Responses

As described in Section 11, the IgE antibody response is induced under limited immunization regimens and the kinetics of the antibody response vary depending on the immunization. Dissociation between IgE and IgG antibody responses raised possibilities for the suppression of IgE antibody response. Several different approaches described below suggest that IgE antibody response is more susceptible to regulatory mechanisms than IgG antibody response.

A. SUPPRESSION BY HUMORAL ANTIBODIES It is well known that antibody synthesis is regulated by humoral antibody against the same antigen (172). An injection of antibody prior to or simultaneously with immunization suppressed IgG and IgM antibody responses in various animal species. As expected, the same principle was reproduced in the IgE antibody response. Strannegard and Belin (146) have shown that an injection of rabbit antiserum against KLH 1 day before or 1 day after primary immunization with the antigen resulted in complete suppression of the IgE antibody response and partial suppression of the formation of hemagglutinating antibody. Similar results were obtained in mice that were immunized with a low dose (0.2-1.0 p g ) of OA included in alum (50). Passive administration of a large dose of mouse IgG antibody against OA suppressed primary antibody responses of both IgE and IgGl classes. However, suppression of the antibody response in the mouse was transient, since both IgE and IgGl antibodies appeared at a later date, At 5 weeks after immunization, IgE antibody levels in the suppressed animals became comparable to those in control animals that did not receive passive antibody. It was also found that an increased interval between immunization and passive antibody administration resulted in progressively decreased inhibition of the IgE antibody response. If the antibody was injected 3-5 days after immunization, animals behaved as if the immunization were merely delayed for a few days to a couple of weeks. These findings were in agreement with previous observations on the IgG antibody response

46

KIMISHIGE ISHIZAKA

in rabbits ( 179) that received passive antibody immediately before injection of antigen included in CFA. It appeared that the passive antibody initially prevented immune responses, but, as the antibody levels fell with antigen still present, an essentially normal immune response followed. In the anti-OA antibody response in the mouse, administration of passive antibody did not affect the on-going IgE antibody response nor suppress the secondary IgE antibody response ( 5 0 ) . Strannegard and Belin ( 146) also showed that passive antibody administered at 8 days after immunization failed to affect the IgE antibody response in the rabbit. These results are similar to previous observations on the IgG antibody response by Wigzell (186) that suppression was limited if passive antibody was administered a long time after immunization and those by Uhr and Bauman (171) that the secondary response was more difficult to suppress with antibody than the primary response. As suggested by Siskind and Benacerraf (144), such results may be explained by the presence of a large population of specific cells of high affinity in primed as compared to virgin animals. In principle, the suppressive effect of passive antibody on IgE antibody responses in high-responder mice is essentially the same as that observed in the IgG antibody response, suggesting that the mechanism for suppression is probably the neutralization of antigen at the peripheral level ( 172). Tada and Okumura (155) studied the effect of passive antibody on the IgE antibody response in rats immunized with DNP-Asc. In this system, the IgE antibody response was suppressed even if IgG antibody was administered at 5 days after the primary immunization. If the antibody was administered on day 8, when the IgE antibody titer reached a maximum, on-going IgE antibody formation terminated. The average PCA titer in the sera dropped to one-sixteenth within 2 days. If we consider the half-life of IgE antibody, its production should have terminated within 12 hours after injection of antibody. There are several unique features in the termination of rat IgE antibody formation by passive antibody. When the antibody was administered simultaneously with antigen for immunization, ( i ) the amount of antibody required for suppression was much less than that for neutralization of antigenic determinants in the immunizing antigen, (ii) suppression was restricted to the IgE antibody response and the formation of hemagglutinating antibody was not suppressed, and ( iii ) the antiserum against DNP-Asc, but neither the antibody against Asc nor anti-DNP antibody suppressed the IgE antibody response. These results suggest that mechanisms involved in the suppression of IgE antibody formation in the rat are entirely different from those in the mouse. Rowley and Fitch ( 135) proposed that passively administered antibody may directly

CELLULAR EVENTS I N THE IGE ANTIBODY RESPONSE

47

inactivate immunocompetent cells. Feldmann and Diener ( 29) reported that antigen-antibody complexes paralyzed the responsiveness of immunocompetent cells to antigen. As the dose of DNP-Asc used for immunization of rats was quite high (1.5 mg) in Tada’s experiment (155), passive antibody might have formed antigen-antibody complexes that inactivated immunocompetent cells. This possibility is supported by the findings of Strannegard (145), who immunized rabbits with mixtures of hemocyanin and antihernocyanin at different ratios. These animals did not form IgE antibody and failed to respond to subsequent immunization with free antigen to produce IgE antibody, In the mouse, however, such a phenomenon has not been observed. Priming of DBA/1 mice with OA-anti-OA complexes formed in antibody excess did not induce IgE antibody responses but caused enhancement of IgE antibody responses to a booster injection of free antigen. The failure of the anti-Asc antiserum and anti-DNP antibody to suppress IgE antibody responses in rats is difficult to explain unless the majority of IgE antibodies against DNP-Asc is directed toward newly developed determinants in the hapten-protein conjugate. This is conceivable in DNP-Asc-primed rats because subsequent experiments of Tada et al. (160) showed that PCA titers of rat anti-DNP-Asc serum differed, depending on the antigen used for challenge. If DNP-BSA was employed, the PCA titer was one-quarter of that obtained with DNP-Asc. An important difference between the IgE antibody response in DBA/1 mice and that in the rat is the pattern of antibody formation. The IgE antibody formation in the rat immunized with DNP-Asc was transient and no secondary response was obtained, whereas DBA/1 mice immunized with a low dose of OA showed a persistent IgE antibody response and a booster injection elicited the secondary IgE antibody response, Such a difference suggests the possibility that the transient pattern and lack of memory in the rat system may be related to the termination of preexisting IgE antibody synthesis by passive antibody. The findings obtained in the mouse do not exclude the possibility that, in hay fever patients, IgG antibody may regulate the IgE antibody response by neutralizing antigen, but a preferential effect of the antibody on the IgE antibody response would not be expected. It is known that the IgE antibody response in atopic individuals persists for a long time and includes a secondary response (see Fig. 1).Since IgG antibody concentrations in the sera of atopic patients are not high, even after hyposensitization treatment, the existence of a suppressive effect by the IgG (blocking) antibody on IgE antibody synthesis is doubtful. The time course of IgE and IgG antibody titers during immunotherapy showed that IgG antibody titers increased thirty- to forty-fold within 2

48

KIMISHIGE ISHIZAKA

to 3 months after initiation of the treatment (44).However, the increase of IgG antibody titer was not accompanied by depression of IgE antibody formation or suppression of a secondary IgE antibody response. AS will be described in a later section, IgG antibody is probably not responsible for the suppression of secondary IgE antibody responses after long-term treatment.

B. UNRESPONSIVENESS IN IGE-l3 CELLS Collaboration between T and B cells for the formation of IgE antibody immediately raised the possibility that specific immunological tolerance in B cells would lead to suppression of IgE antibody response. Based on the induction of hapten-specific tolerance for IgM and IgG antibody responses in inbred guinea pigs (65) and mice (66) by the treatment of such animals with a nonimmunogenic DNP-conjugate of a copolymer of D-glutaminic acid and D-lysine ( D-GL),Katz et al. (67) tried to establish conditions for the induction of tolerance in specific B cells of the IgE class. They found that two injections of 500 pg DNP-D-GL given to A/J mice on successive days before the primary immunization with DNPAsc caused a striking suppression of DNP-specific IgE and IgG antibody responses. Furthermore, they demonstrated that DNP-D-GL induced DNP-specific tolerance in the DNP-primed B-cell population. In this experiment, mice were primed with DNP-Asc in alum for the form at’1017 of antihapten IgE antibody, and some of the immunized mice were treated with two successive daily injections of DNP-D-GL. Spleen cells obtained from treated and untreated mice were transferred into irradiated recipients that were then secondarily challenged with DNP-Asc plus alum. The results clearly showed that recipients of cells from untreated donors developed high levels of anti-DNP antibodies of both IgG and IgE classes, whereas the recipients of spleen cells from treated mice manifested marked suppression of antibody production in IgE and IgG classes. Since D-GL does not have the capacity to stimulate T cells, the authors concluded that the tolerance was induced in DNP-specific B cells. Their experiment also showed that the tolerant state induced in the donor cell population did not change after administering to these donor mice an immunogenic dose of DNP-Asc in alum. Recipients of spleen cells from such donors, which had received booster injections of DNP-Asc after a tolerogenic dose of DNP-D-GL, failed to produce an anti-DNP antibody response after injections of DNP-Asc. The tolerant state was not removed by the treatment of the cells with trypsin (66), indicating that cell reactivity to DNP determinants was irreversibly inhibited. Induction of tolerance with DNP-D-GL in the IgG antibody response

CELLULAR EVENTS I N THE IGE ANTIBODY RESPONSE

49

of guinea pigs resulted in a preferential depression of high-affinity antibody-forming cells (18). Experiments by Katz et al. (67) indicated that IgE-B cells having high-affinity receptors for the hapten were more easily tolerized by DNP-D-GL. In a similar approach, Lee and Sehon (83) used DNP coupled to isologous serum protein in mice. The experiments were based on the fact that presentation of hapten on isologous serum protein induced partial or complete hapteti-specific tolerance ( 33 ) . They injected various doses of DNP-coupled mousc y-globulin (DNP-MGG) 2 4 hours prior to primary immunization with alum-absorbed DNP-OA and observed that 100 pg or more of DNP-MGG suppressed anti-DNP antibody responses but not anti-OA antibody responses. If one immunized mice with DNP-OA and then administered 2.5 mg of DNP-MGG into the mice that were then forming high titers of anti-DNP IgE antibody, the IgE antibody titer declined. Different from B-cell tolerance induced by DNP-D-GL, DNP-MGG treatment failed to abolish the secondary antiDNP IgE antibody response to DNP-OA unless the tolerogen was given at the time of secondary immunization. Their results showed that an injection of DNP-MGG prior to immunization with DNP-OA was less effective than simultaneous injection for the induction of tolerance. These findings suggested the possibility that the apparent unresponsiveness might result from thc neutralization of antihapten IgE antibody by DNP-MGG. However, Lee and Sehon’s subsequent experiment ( 84) using adoptive transfer excluded this possibility. In this experiment, donors of spleen cells were tolerized by the intravenous injection of DNP-MGG followed by immunization with DNP-OA. Spleen cells from these animals were transferred into irradiated recipients that were then challenged with DNP-OA. Even when DNP-MGG was injected into donors 58 days before the cell transfer, anti-DNP IgE antibody responses of the recipients were significantly suppressed. It was also found that such an unresponsive state of hapten-specific B cells was maintained even after serial cell transfer into a second host. The results indicated that the initial exposure of immunocompetent cells with the hapten conjugated to isologous globulin resulted in long-lasting immunosuppression. HOWever, the tolerized cells affected neither antihapten nor anticarrier IgE antibody responses of cells from immune or normal mice. The admixture of tolerized cells with primed cells did not affect the expression of the IgE antibody response of the latter on adoptive transfer. As will be described later, transfer of suppressor T cells depressed the antibody response of normal or primed cells. Thus, hapten-specific unresponsiveness induced by treatment with DNP-MGG appears to involve elimination or

50

KIMISHIGE ISHIZAKA

inactivation of hapten-specific IgE-B cells or blockade of the receptors of these cells. It was speculated that immunological tolerance may be induced more easily in IgE-B cells than in IgG-B cells and that IgE-B cells have a higher affinity for antigen than B cells of other immunoglobulin classes (153). Based on such a hypothesis, it was anticipated that failure of the IgE antibody response after immunization with a large dose of antigen may be due to B-cell tolerance. Indeed, Maia et al. (92) showed that IgE antibody responses to an optimal dose (0.1 pg) of alum-absorbed OA were suppressed in BDFl mice by three intravenous injections of 100 pg of OA shortly after immunization, although IgGl antibody responses in these animals were not suppressed. Such selective suppression of the IgE antibody response was interpreted by Tada (153) as the preferential induction of immunological tolerance in IgE-B cells. As will be described later, however, it was shown that the three successive injections of OA into DNP-OA-primed animals suppressed the anti-DNP IgE antibody response and that the same procedure induced suppressor T cells specific for OA. To date, there is no evidence that immunological tolerance is more easily induced in IgE-B cells than in IgG-B cells.

C. REGULATIONBY T CELLS I. Enhancement of 1gE Antibody Formation by Immunosuppressive Treatment It is generally accepted that neither adult thymectomy nor splenectomy affect immune responsiveness. Mota ( 111) reported that splenectomy of rats did not affect the IgE antibody response to OA and suggested that the spleen was not the major lymphoid organ for IgE antibody formation. Subsequently, Okumura and Tada (122) found that adult thymectomy, splenectomy, or both significantly enhanced IgE antibody responses of rats immunized with DNP-Asc plus pertussis vaccine. In the thymectomized or splenectomized animals, maximum IgE antibody titers were higher and antibody responses persisted longer than those in sham-operated controls, which gave transient IgE antibody responses. In their experiment, adult thymectomy and splenectomy failed to affect the formation of agglutinating antibody, indicating that the effect was exclusively on the IgE antibody response. Subsequently, however, Taniguchi and Tada ( 168) reported that adult thymectomy enhanced IgG antihapten antibody responses of rabbits to DNP-BGG. Since the average association constant of anti-DNP antibody produced by the thymectomized animals was higher than that of controls, the authors suggested that relative diminution of T cells results in a stimulation of antibody-forming cells with a high affinity.

CELLULAR EVENTS IN THE ICE ANTIBODY RESPONSE

51

The same investigators (162) showed that irradiation of Wistar rats with a sublethal dose of X-rays enhanced IgE antibody responses. Doses of 400 R X-irradiation given 1 day before or 1 day after the primary immunization with DNP-Asc plus pertussis vaccine enhanced the production of IgE antibody. In the irradiated animals, the IgE antibody titers persisted more than 30 days. The animals formed IgM antibody but not IgG antibody against the immunizing antigen. The enhancement of IgE antibody response was observed even with 100-200 R X-irradiation. They also showed that the irradiated animals were capable of producing IgE antibody upon immunization with DNP-Asc without adjuvant. Because their observation was quite interesting and useful in obtaining a persistent IgE antibody response in the rat, the experiments were repeated in two separate laboratories. However, the results were not reproduccd in either the Sprague-Dawley or Lewis (inbred) strains. In these strains, 200 R or more of X-irradiation 1 day before or 1 day after immunization completely abolished the capacity of the animals to form IgE antibodies to antigen included in alum or injected with pertussis vaccine. It appears that the effect of irradiation is different depending on the strain of rat tested. A favorable effect of partial depletion of T cells on IgE vs IgG antibody response was observed in the mouse as well. Michael (102) reported an interesting dose effect of T cells on the IgE antibody response in athymic nude mice. These mice were unable to produce IgE antibody, but transfer of 5 x 105 thymocytes from nu/+ mice reconstituted them for the IgE antibody response, which persisted for more than 4 weeks. If 5 x 10' thymocytes were transferred, the recipients formed IgE antibody, but the antibody response terminated earlier. On the other hand, IgM/ IgG antibody responses against SRBC occurred only in mice repopulated with higher numbers of thymocytes. Thus, Michael suggested that the supraoptimal number of thymocytes, which was adequate for IgG/ IgM antibody responses, seemed to suppress IgE antibody responses. Recently Okudaira and co-workers (120) obtained similar results in an adoptive secondary response in the mouse. They showed that immunization of DBA/1 mice with DNP-KLH plus a low dose of alum failed to induce an antihapten IgE antibody response, but the transfer of their spleen cells into irradiated animals followed by injection of DNP-KLH gave an IgE antibody response. If the same cells were transferred into nonirradiated animals and recipients were challenged by DNP-KLH, no IgE antibody response was obtained. Since irradiated and nonirradiated recipients gave comparable IgG antibody responses, it appears that irradiated mice have favorable conditions for IgE antibody responses. In

52

KIMISHIGE ISHIZAKA

order to analyze the effect of X-irradiation on recipients, these authors reconstituted irradiated mice with normal spleen cells before the transfer of DNP-KLH-primed spleen cells. Adoptive secondary responses showed that transfer of lox normal spleen cells considerably reduced IgE antibody responses of the recipients and that the cells’ suppressive activity was partially removed by treatment with anti4 antiserum and complement. It was also found that normal splenic T cells but not B cells suppressed IgE antibody responses of DNP-KLH-primed cells. Quite recently, Watanabe et al. ( M a ) showed definitive evidence for the presence of nonspecific suppressor T cells in SJL strain mice. As described by Kojima and Ovary (81), both Balb/c and A/ J mice primed with DNP-KLH and infected with N . bradiemis induced IgE antihapten antibody response upon challenge with DNP-Nb. They found that the IgE antibody response was low and transient in SJL strain but irradiation of the mice one day after DNP-Nb challenge sustained the IgE antibody response. In this system, the transfer of 5 x lo7 normal spleen cells or thymocytes from normal SJL mice terminated the antibody response. Another paradoxical effect of immunosuppressive treatment on IgE antibody formation was reported relative to antithymocyte serum ( ATS ) and antilymphocyte serum ( ALS ). The immunosuppressive activity of ALS or ATS was shown to be the result of direct injury to certain lymphoid cells especially T lymphocytes (86, 169). Their immunosuppressive effect has been definitely demonstrated in cell-mediated immunity; however, these antisera showed an enhancing effect on humoral antibody responses in some antigen systems. Baker et al. (4) demonstrated that treatment of mice with ALS simultaneously with an immunization with pneumococcal polysaccharide SIII potentiated the antibody response and suggested that ALS may inhibit a population of lymphocytes that regulate, rather than amplify, the antibody response. Subsequently, Baker ( 5 ) showed that ALS treatment had no effect on the response of nude mice to SIII and indicated that suppressor cells represent a subpopulation of T cells. Since an enhanced antibody response of thymus-bearing littermates after ALS treatment was much greater than the response of nude mice, Baker proposed that helper function and suppressive activity were mediated by separate subpopulations of T cells. Based on the findings by Baker and his associates, Okumura and Tada (125) studied the effect of ATS on the IgE antibody response in the rat. They found that an injection of ATS before or simultaneously with primary immunization with DNP-Asc inhibited the antibody response of a11 IgG, IgM, and IgE classes, but the same dose of ATS given 5-8 days after the immunization enhanced the IgE antibody response and accel-

CELLULAR EVENTS IN THE IGE ANTIBODY RESPONSE

53

erated the IgG antibody response. The quantity of ATS was also important for the regulation of antibody responses. If a small dose of ATS was given to the animals simultaneously with primary immunization, IgE antibody appeared with some delay, but the maximal PCA titer was comparable to that obtained in control animals. Furthermore, moderately high IgE antibody titers persisted for a longer period of time in the treated animals than in controls. In the animals treated with a small dose of ATS, not only IgE but also IgM antibody titers persisted for a long period of time, unlike the normal sequential switch from an IgM to an IgG antibody response. A similar enhancing effect of ALS on the IgE antibody response was described by White and Holm (184) in the rat immunized with KLH. Kind and his associates (70) showed that an injection of horse antimouse lymphocyte serum without adjuvant initiated an IgE antibody response to horse serum proteins. Taken collectively, the findings indicate that treatment with an adequate dose of ALS at a certain time seems to enhance IgE antibody responses and changes the distribution of antibodies in different immunoglobulin classes. 2. Suppression of ZgE Antibody Response by Antigen-Specific T Cells As a result of the series of experiments described above, Okumura and Tada (123) anticipated that T cells, which are necessary for antigen recognition, may negatively regulate IgE antibody formation. In order to prove this hypothesis, they studied whether thymocytes and spleen cells from hyperimmunized rats terminated the on-going IgE antibody response. Since they succeeded in obtaining enhanced and sustained production of IgE antibody in the rat bv X-irradiation and simultaneous immunization with DNP-Asc together with pertussis vaccine, such animals were used as recipients, Donors of thymocytes and spleen cells were immunized by repeated injections of DNP-Asc or Asc in CFA, and loR to lo9 thymocytes or spleen cells from these animals were transferred into recipients that were producing high titers of IgE antibody. As already illustrated in Fig. 4, the amount of IgE antibody against DNP-ASC was ‘drastically reduced within 2 days after the transfer of thymocytes or spleen cells from the donors hyperimmunized with DNP-Asc or Asc, but neither cells from the donors immunized with DNP-BSA nor normal thymocytes showed this inhibitory effect. Therefore, it appears that the suppressive effect of regulatory T cells is carrier-specific. The results also showed that the suppressive activity of spleen cells was completely abrogated by in vitro treatment of the cells with ATS and complement ( 125). This finding as well as the fact that thymocytes themselves showed suppressive activity indicated that the regulator cells were thymus-de-

54

KIMISHICE ISHIZAKA

rived. Tada et al. (158) proposed a tentative hypothesis that helper T cells first assist in IgE antibody formation, but after further proliferation or differentiation, the large number of helper T cells negatively regulate the on-going IgE antibody response against hapten on the same carrier. They pointed out, however, that the results were equally well explained if regulator T cells are different from helper T cells. Recently, evidence has accumulated indicating that T cells may regulate IgM and IgG antibody responses (30, 31). With an intravenous injection of a tolerogenic dose of fowl 7-globulin, Basten et al. ( 8 ) induced suppressor T cells. Similar results were reported by Benjamin (10) using monomeric human IgG as a tolerogen. The transfer of spleen cells from tolerant animals into normal mice abrogated the response of the recipients to subsequent immunization with immunogenic ( aggregated ) human IgG. Takemori and Tada (161) reported that spleen cells from mice that received two intravenous injections of KLH suppressed antihapten antibody responses of nonirradiated recipients to DNP-KLH. In view of such evidence in the mouse, Tada (153) is now considering that his group’s previous observations in the rat, i.e., suppression of the ongoing IgE antibody formation by the transfer of thymocytes and spleen cells, were probably caused by suppressor T cells. More recently, however, Eardley and Gershon (24) suggested that the transfer of immune T cells might induce suppressor T cells in normal recipients. Since, in Okumura’s experiment, the donors of spleen cells and thymocytes had been hyperimmunized with carrier included in CFA, it is possible that the depression of IgE antibody formation was due to induction of suppressor cells in vivo. Evidence was recently obtained in the mouse that antigen-specific suppressor T cells are actually involved in the regulation of IgE antibody response. A series of experiments was initiated based on an interesting observation by Maia et al. (92) that successive injections of a high dose of OA without adjuvant into OA-primed mice selectively depressed the IgE antibody response. They immunized BDFl mice with a minute dose of alum-absorbed OA and then gave three intravenous injections of 100 pg of OA at 3, 5, and 7 days after the immunization. The treatment suppressed IgE antibody responses but rather enhanced IgG antibody responses (Fig. 9 ) . Because we obtained evidence in a separate experiment that UD-OA stimulated OA-specific T cells but not OA-specific B cells (see Section IV,D,2), OA-primed mice were treated with UD-OA instead of OA. The results showed that three successive injections of UD-OA suppressed both IgE and IgG antibody responses to the priming antigen (Fig. 9 ) . It was also found that intravenous injections of OA or UD-OA suppressed IgE and IgG antihapten antibody responses to

CELLULAR EVENTS IN THE ICE ANTIBODY RESPONSE

0

7

1,ALUM I"

10

14

55

21

DAYS AFTER IMMUNIZATION

FIG. 9. Effect of antigen injections on IgE and IgC antibody responses. All animals were primed with 0.2-pg OA included in alum. Two groups of animals received either 100 pg OA or the same dose of urea-denatured OA (UD-OA) at 3, 5, 7 days after priming. Anti-OA IgC antibody titer was shown by antigenbinding activity ( A B A ) . [From Takatsu and Ishizaka ( 163).]

DNP-OA but failed to suppress the antibody response to DNP-KLH, indicating that the suppressive effect obtained by the treatment was carrier-specific. Thus, we studied the effect of the treatment on the development of helper T cells and B memory cells in the spleen by using an adoptive transfer technique ( 163). The results showed that UD-OA treatment suppressed the development of helper T cells and of B memory cells in the spleen. Since UD-OA reacted with T cells but not with B cells specific for native antigen, one may speculate that the primary effect of the treatment is suppression of the development of T helper cells. Subsequent experiments showed that UD-OA treatment induced the generation of suppressor T cells (165). It was found that intravenous

56

KIMISHIGE ISHIZAKA

injections of UD-OA into OA-primed animals suppressed the antibody response not only to the priming antigen but also to subsequent immunization with DNP-OA. In order to investigate whether the suppression of antihapten antibody responses in the treated animals was an active process or resulted from depletion of OA-specific immunocompetent cells, the treated animals were supplemented with OA-primed spleen cells and then immunized with DNP-OA. Indeed, the transfer of OA-primed spleen cells into normal animals enhanced antihapten antibody responses of IgE and IgG classes to DNP-OA; however, supplementation of UD-OA-treated animals with the OA-primed cells did not restore the capacity to respond to DNP-OA (Fig. 10). The results indicate that suppression of the antibody response is an active process. Furthermore, the transfer of splenic T cells from UD-OA-treated animals into normal nonirradiated mice diminished antibody responses of the recipients to DNP-OA. In the experiments shown in Fig. 11, splenic T cells from either UDOA-treated mice or those from OA-primed mice were transferred into normal mice and recipients were immunized with alum-absorbed DNP-OA. It should be noted that antihapten and anticarrier antibody responses were both suppressed by the transfer of T cells from UD-OAtreated mice. Another important finding was that not only IgE but also IgG antibody responses were significantly suppressed by the transfers. Control animals that received splenic T cells from either OA-primed mice or normal mice gave high IgE and IgG antibody responses. If the

RECIPIENT

CELLS TRANSFERRED(5

Ig E ANTI - DNP 20

80

320

Ig G 1280 M I - D N P ANTI*

7-

NORMAL

NORMAL

NORMAL

OA-PRIMED

OA-PRIMED UD- TREATED OA - PRIMED UD-TREATED

17

44

32

89

NORMAL


23

OA-PRIMED

( 0 12

28

<5

20

80

320

1280

ANTI - OA

FIG. 10. Failure of OA-primed spleen cells to reconstitute suppressed animals. Either nonnal or OA-primed UD-OA-treated animals were supplemented with 2 x 10' normal or OA-primed spleen cells and then immunized with 0.2 fig DNP-OA included in album. Both anti-DNP (blank bars) and anti-OA (hatched bars) IgE antibody titers at 14 days after immunization are shown. [From Takatsu and Ishizaka ( 165).]

CELLULAR EVENTS IN THE IGE ANTIBODY RESPONSE

57

-

I-

>

i 2

a

z a I

320-

8020-

I-

z a w

s

5' 5 ,

'b------* 1

1

I

I

I

1

1280t-

> a

8

320-

I-

z Q

a

80-

? -

5

a W IT

Y

20-

5<5

1

DAYS AFTER IMMUNIZATION

DNP-OA in ALUM

FIG. 11. Suppressive effect of splenic T-cell fraction of UD-OA-treated mice. T cells ( 2 x 10') from nornial ( a ) , OA-primed (O), or OA-primed UD-OAtreated mice ( A ) were transferred into each nornial mouse and the recipients were immunized with alum-absorbed DNP-OA. Anti-DNP ( top ) and anti-OA (bottom) IgE antibody responses are shown. [From Takatsu and Ishizaka ( 165).]

same T-cell fractions were examined for helper activity by adoptive transfers into irradiated mice, the helper function of T cells from UD-OAtreated mice was much less than that of the T cells from OA-primed mice. An inverse relationship between suppressive effect and helper function indicated that spleens of UD-OA-treated mice contained suppressor T cells that were distinct from helper T cells. It was also found that the suppressive activity of the T cells was specific for OA. The transfer of T cells from UD-OA treated mice into normal mice failed to suppress anti-DNP antibod), responses of the recipients to DNP-KLH. Subsequent experiments showed that denaturation of OA in urea or loss of the major antigenic determinant of the antigen molecule was not essential for the generation of suppressor T cells. In mice primed with alum-absorbed OA and then injected with 100 ,.~g of OA at 3, 5, and 7 days after priming, the splenic T cells exerted a suppressive effect on the primary antibody response of nornmal mice to DNP-OA. Comparisons

58

KIMISHIGE ISHIZAKA

between splenic T cells from OA-treated animals and those from UDOA-treated animals indicated that UD-OA treatment was more effective than OA-treatment with respect to the generation of suppressor T cells. Because we succeeded in generating antigen-specific suppressor T cells in high-responder mice that readily showed a persistent IgE antibody response to OA, the effect of suppressor T cells on the on-going IgE antibody response was studied. Recipients were immunized with a minute dose of OA included in alum. When the IgE antibody titer reached a maximum, splenic T cells of UD-OA-treated animals were transferred, and IgE antibody titers in the recipients were followed. The results clearly showed that suppressor T cells depressed on-going IgE antibody formation, but splenic T cells from OA-primed animals failed to do so (Table VII). Thus, it appears that suppressor T cells obtained by UD-OA treatment depressed both the primary IgE antibody response and on-going IgE antibody formation. The findings described above were confirmed by more recent experiments in which suppressor cells were generated in uitro (43). As described in Section IV, helper T cells for both IgG and IgE antibody responses were obtained by the culture of normal splenic T cells on antigen ( OA) -bearing macrophages. If either splenic lymphocytes or T cells are cultured with soluble antigen (OA) in the absence of macrophages, the cultured' lymphocytes do not show the helper function but contain suppressor T cells. Transfer of the cultured cells into normal animals prior to immunization with alum-absorbed DNP-OA depressed the primary antihapten and anticarrier antibody responses of both IgE and IgG classes. On the other hand, the lymphocytes cultured on OA-bearing macrophages and having helper function failed to suppress the primary antiTABLE 1 7 1 1 EFFECTO F SUPIW-SSOR CELLSO N T H E ON-GOINGANTIBODYRESPONSES~ -

Anti-OA antibody Donors of T cells

I@* on day: Priming

Treatment

14

21

28

35

None OA OA

Saline Saline UI)-OA

1280 1280 1280

1810 1280 450

1810 903 450

1810 90.5 320

IgG (pglml) on day: 14 21 28 3R 1.9 1.7 1.6

7.2 6.4 3.6

14.2 13.6 11.6

19.0 17.0 17.0

a Three groups of mice were primed with alum-absorbed OA to induce a persistent IgE antibody response. On the fourteenth day, 2 f 107 splenic T cells from various sources were transferred into each mouse. * Determined by PCA reaction.

CELLULAR EVENTS I N THE ICE ANTIBODY RESPONSE

59

body response. The function of suppressor cells obtained in the cultures was antigen-specific, and these cells were depleted by the treatment with anti4 antiserum and complement. The results confirmed that macrophages are required for priming helper T cells (23) and that suppressor T cells are distinct from helper T cells. Throughout the experiments described above, it was noticed that suppressor T cells depressed the IgE antibody response more effectively than the IgG antibody response. As shown in Fig. 11, anti-DNP and anti-OA IgE antibodies in the recipients of suppressor cells constituted one-sixteenth to one-eighth of the antibodies in control mice that received normal T cells. The IgG antibody levels in the recipients of suppressor cells were about 408 of those in control animals. The higher susceptibility of the IgE antibody response than of the IgG antibody response to suppressor T cells may be related to previous observations by Katz et ul. (68) who suggested that IgE-B cells are more sensitive than IgG-B cells to the complex regulatory influences of the T-cell populations. An alternative explanation for this high susceptibility of the IgE antibody response is that the size of IgE-B cell population is much smaller than that of IgG-B cells. I t is frequently observed in adoptive transfer experiments that an increase of hapten-specific B cells counterbalances a decrease or carrier-specific helper cells. Thus, helper function diminished by the presence of suppressor cells is easily demonstrated when the size of the B cell population is limited. In any event, the fact that the IgE antibody response is more susceptible than the IgG antibody response to suppressor T cells may be related to previous observations that the IgE antibody response is more affected by the dose of antigen and the adjuvant vehicle employed. One might also speculate that the transient nature of IgE antibody responses following immunization with a high dose antigen may be due, at least in part, to the generation of suppressor T cells late in the course of primary immunization. As shown in Fig. 9, priming of mice with alum-absorbed OA followed by successive injections of OA, which actually generate suppressor T cells, resulted in a transient IgE antibody response. It is not known how suppressor T cells regulate the IgE antibody response. Based on the previous finding that on-going IgE antibody formation in the rat was depressed by the transfer of thymocytes and spleen cells from hyperimmunized animals, Tada et al. (159) mechanically disrupted these cells and tested the cell-free extract for its ability to terminate the IgE antibody response. Thymocytes and spleen cells were obtained from rats that had been hyperimmunized with DNP-Asc or Asc included in CFA, and the extract from these cells was injected into animals that had been 400 R X-irradiated and immunized with DNP-Asc

60

KIMISHIGE ISHIZAKA

plus pertussis vaccine. The results clearly showed that on-going IgE antibody formation was terminated by the extract from 2 x 10" thymocytes. The antibody titers dropped to one-eighth of preinjection titers within 6 days. Thymocyte extracts obtained from rats hyperimmunized with DNP-BSA or with CFA alone failed to depress the antibody formation. The results indicated that thymocytes or spleen cell extracts of carrierimmunized animals contained a subcellular component that inhibited IgE antibody formation against a hapten on the carrier. The suppressive component in the extract was absorbed with Asc-coated immunosorbent, indicating that the component contained receptors for the carrier. In spite of the specific binding with antigen, the component was not absorbed with an immunosorbent coated with anti-Fab, anti-p, or polyvalent anti-rat immunoglobulin but was absorbed with antiserum against rat thymocytes. The active component was not affected by DNase or RNase but was destroyed by digestion with trypsin (124). The component had a molecular size of 35,000 to 60,000, as judged by gel filtration through a Sephadex G-100 column, and electrophoretic mobility corresponding to a- and ,&globulins. These findings clearly showed that the suppressive component in thymocytes is not immunoglobulin and suggested that the component may be a unique molecule containing the antigen receptor on T cells. It is reasonable to speculate that such a component in T cells is involved in the regulation of the IgE antibody response; however, evidence is not available as to whether the component is naturally released from the thymocytes and/ or spleen cells. So far, this suppressive component has not been detected in culture fluids of thymocytes or spleen cells that were incubated either in the presence or absence of carrier. Nevertheless, Tada et al. (153) speculatcd that the component would combine with antigen through the receptor and that the complex would interact with B cells to suppress the B-cell response. An immediate decrease in IgE antibody after injection of the extract suggested that the suppressive component acts on B cells that are already at a certain stage of differentiation to terminate their antibody synthesis. Recently, Herzenberg et al. (40) obtained evidence suggesting that suppressor T cells involved in allotype suppression (41) regulate helper T cells rather than B cells. Since allotype suppression as well as depression of IgE antibody synthesis in the mouse by the transfer of suppressor T cells are slow processes (cf. Table VII), the suppression could be explained by either the regulation of helper T cells or inhibition of B-cell differentiation. Because the termination of rat IgE antibody formation by a thymocyte-derived suppressive component is so rapid, probably neither regulation of helper cells nor prevention of B-cell differentiation explains this phenomenon ( cf. 153).

CELLULAR EVENTS I N THE IGE ANTIBODY RESPONSE

61

A nonspecific suppressive factor ( s ) was described by Kishimoto and Ishizaka (78) using an in uitro system in which IgE and IgG antibodies were formed by DNP-primed rabbit lvmph node cells. They obtained CFS from the culture of mesenteric fyniph node cells primed with a homologous carrier. As described, the CFS preparation exerted enhancement of antibody formation if it was added to DNP-primed cells that had been treated with DNP-heterologous conjugate (cf. Fig. 6 ) . If the same CFS was added to the cells simultaneously with DNP-homologous conjugate or with DNP-KLH, some of the CFS preparations suppressed the antibody response. In order to study whether a single substance is responsible for the suppression and enhancement, both activities were titrated in four CFS preparations using the same DNP-Asc-primed lymph node cells as the source of B cells. The CFS was prepared from Rag-prinied cells that had been raised by different immunization schedules. The suppressive effect of CFS was determined by suspending the DNP-Asc-primed cells in each CFS preparation and then stimulating with homologous antigen. The enhancing effect of the same CFS preparations was determined by the stimulation of the same primed cells with DNP-KLH, followed by culture in CFS. The results of such experiments showed that the suppressive activity of CFS had no correlation with the enhancing activity and that the observed suppression was not due to a supraoptimal dose of enhancing factor, Thus, it appears that suppressive factor is different from enhancing factor. The multiplicity of T-cell factors involved in the humoral antibody response raised the question whether the enhancing factor for one immunoglobulin class may be a suppressive factor for another immunoglobulin. However, comparisons among several preparations of CFS with respect to the enhancing and suppressive activities for IgE and IgG antibody responses indicated that enhancing factor for IgG has no suppressive effect on the IgE antibody response. It is not known how nonspecific suppressor factor affects the differentiation of B cells. The findings described above suggest that the sensitivity of B cells to the enhancing and suppressive factors differs depending on the stage of their differentiation. It appears that the suppressive factor displays its function when B cells are stimulated by antigen, or before the exposure to antigen, whereas antigen-stimulated B cells are more sensitive to the enhancing factor. Indeed, more recent experiments by Kishimoto et al. (80) suggest strongly that receptors for nonspecific enhancing factor were generated after rabbit B cells were activated by binding of antigen through immunoglobulin receptors. Nevertheless, the presence of both enhancing and suppressive factors supports the dual regulatory role of T cells in the IgE antibody response. It is likely that

62

KIMISHIGE ISHIZAKA

the balance of these two factors is important for the magnitude of IgE antibody response in the rabbit. One might speculate that the joint action of multiple T-cell factors having different sites of action regulates the magnitude and class distribution of antibody responses. D. EXPERIMENTAL MODELFOR IMMUNOTHERAPY Hyposensitization treatment or immunotherapy has been used in the past 40 years to treat hay fever patients. The clinical effect of this treatment was believed to be due to the formation of blocking ( IgG) antibody that may compete with reaginic IgE antibody for binding with allergen. Extensive studies on the mechanisms of reaginic hypersensitivity reactions revealed that IgG does not have affinity for mast cells and basophile granulocytes, to which IgE binds (54, 55), and, therefore, that blocking antibody cannot compete with IgE antibody at the cellular level. Furthermore, IgG is not the major immunoglobulin in respiratory secretions. Even though the IgG antibody level in the serum increased as a result of the treatment, the amount of IgG antibody secreted in the respiratory tract would be limited. These considerations suggested that the increase of blocking antibody may not be responsible for the clinical effect of immunotherapy for pollinosis (44 ) . Once quantitative measurements of IgE and IgG antibodies became available, the effect of immunotherapy on the antibody level was studied by using radioimmunoassay. In ragweed-sensitive patients who received immunotherapy, IgG antibody titers against purified ragweed antigen (antigen E ) increased thirty- to forty-fold within 1 to 2 months. During this period, the IgE antibody titer usually increased significantly, and then gradually declined in some cases (Fig. 12). After 2 years treatment, the IgE antibody level became significantly lower than the pretreatment level in more than half of the treated patients. The most significant immunological effect of the treatment was suppression of secondary IgE antibody response. After long-term treatment, most of the patients failed to show the secondary IgE antibody response after the ragweed season, whereas almost all untreated patients showed the response (cf. Fig. 1). Thus, the major immunological effects of long-term immunotherapy are ( a ) increase of IgG antibody, ( b ) gradual decline of IgE antibody level, and ( c ) suppression of secondary IgE antibody response. However, analysis of the results in several patients, indicated that an increase of IgG antibody level was not responsible for the suppression of a secondary IgE antibody response. Thus, it appears that suppression of the IgE antibody response is probably due to some change in the memory cell population (44). Since T cells have a regulatory role in the antibody response, we

CELLULAR EVENTS I N THE IGE ANTIBODY RESPONSE

I l.Oo0-

200

I

-."'\

- 1,000

*-4I

I

63

1

I

I'

I

I

I

-

I

200

speculated that the effect of immunotherapy on the IgE antibody response might be due to some change in the T-cell population. We tested the hypothesis in mice that were immunized with antigen E. In view of the evidence that UD antigen E and &-chain contain immunogenic determinants for T cells but lack the major antigenic determinants for B cells specific for the native antigen (49), these modified antigens were utilized to stimulate T cells ( see Section IV,D,2). Thus, three groups of A/J mice were immunized with antigen E included in alum for a persistent IgE antibody response, and two groups received 10 pg U D antigen or &-chain once every week ( 5 2 ) . The injections were initiated at 2 weeks after priming when the IgE antibody titer against antigen E reached a maximum, As shown in Fig. 13, the IgE antibody titer began to decline after four injections of either a-chain or U D antigen. After 7 weeks treatment, all of the animals including untreated controls were boosted with 10 pg antigen E. Secondary IgE antibody responses to the native antigen were suppressed in the two groups that were treated with modified antigen, However, the treatment did not modify the on-going IgG antibody response. The time course of IgE

64

KIMISHIGE ISHIZAKA

IY

W

r, Ia V

a

I

2

4

2

4

1

I

1

1 12

6

8

10

6

8

1 0 1 2

I00

10 W

z

W

9 -i z a

10 WEEKS

FIG. 13. Effect of weekly injections of a-chain or UD antigen on the on-going antibody response to antigen E ( AgE ). All animals were immunized with 1 pg antigen E plus 1 mg alum and boosted with 10 pg antigen E in saline at week 9. Either 10 pg UD antigen or a-chain was injected once a week beginning from week 2. Control mice received saline. [From Ishizaka et al. ( 5 2 ).I

antibody production in the treated mice was similar to that of IgE antibody response by ragweed-sensitive hay fever patients who were treated with ragweed antigen (cf. Fig. 12). The major difference between murine and human reactions was an increase of IgG antibody in the treated patients. If antigen E-primed animals were treated by weekly injections of 10 pg of native antigen E, instead of a modified antigen, IgG antibody titers increased 1000-fold, but the treated animals showed IgE antibody titer comparable to or slightly higher than that of untreated mice (166). In order to learn the mechanisms for the suppression of IgE antibody formation by the treatment with UD antigen E, we evaluated the effect of treatment on the helper function of splenic T cells by adoptive transfers into A/ J mice ( 166). Splenic T cells from treated or untreated mice were mixed with DNP-KLH-primed spleen cells and transferred into irradiated mice followed by challenge with DNP-Rag. The results clearly showed that the helper function of antigen E-primed spleen cells diminished with weekly injections of either native antigen E or modified antigen (Table VIII). A decrease of helper function after the treatment

CELLULAR EVENTS IN THE IGE ANTIBODY RESPONSE

65

Anti-IINP

Treatment for donors. Untreated Antigen 15 UI) antigen a-Chain Control

IglV at d a y : 7 10 120 10

30 30 <.i

Ig(i (pg/ml) at d a y : 10 14

xo

6 1 2 3

10


160 30 60

X

9 9 9

12 5 3.1

8 7 8 1 2 0

From Takatsu cl a / . (166). function was evaluated b y adoptive transfer experiment: 2 x 106 DNP-KLII primed spleen cells were transferred into all recipients. Ilonors were primed with 1 pg antigen E before weekly injections of respective antigen. PCA titer.

* IIelper

was accompanied by a diminished DNA synthetic responsiveness of splenic T cells to antigen E, as determined by thymidine-3H incorporation. The effect of antigen E treatment on B cells was studied in adoptive transfer experiments. Spleen cells from treated and untreated animals were treated with anti4 antiserum and complement to remove T cells. The B-cell-rich fraction from each group was mixed with a constant number of spleen cells from p-chain-primed animals as a source of antigenspecific T cells and transferred into irradiated animals. The IgE and IgG antibody responses to antigen E showed that the spleens of UD antigentreated mice had less IgE-B cells and IgG-B cells than those of untreated mice. By contrast, treatment with antigen E increased the amount of IgG-B cells and IgE-B cells in the spleens. Such changes in the T cell and B cell populations partially explain how repeated antigen injections affect on-going antibody responses. Thus, decreases of both helper function and IgE-B cells by the treatment with modified antigen appear to be responsible for depressed IgE antibody responses. An increase of the IgE-B cell population after antigen E treatment might overcome a diminished helper function of T cells to maintain the IgE antibody level. Similarly, a great increase of IgG-B cells, in spite of decreased helper function, might explain an increase of IgG antibody level in the antigen E-treated animals.

66

KIMISHIGE ISHIZAKA

A problem yet to be answered is the discrepancy between the IgG antibody response in modified antigen-treated mice and the number of IgG-B cells in their spleens. In the animals treated with a modified antigen, IgG antibody titers were slightly higher than in untreated animals, but B-cell function in the spleen cells of the treated animals was less than that in untreated mice. One might consider that the spleen is only a portion of the lymphoid tissues involved in the IgG antibody response and that changes in the spleen do not totally reflect all lymphoid cell compartments. The results obtained in the antigen E system were reproduced in BDFl mice that were immunized with alum-absorbed OA for a persistent IgE antibody response ( 164). In this experiment, three injections of 100-pg UD-OA were given to the OA-primed mice between 2 and 3 weeks after the immunization, when the anti-OA IgE antibody titer reached a maximum. With this treatment, the IgE antibody titer diminished, but the IgG antibody response was not significantly affected. Independently, Bach and Brashler ( 3 ) reported that treatment of OAprimed DBA/1 mice with a high dose of acetylated OA prevented the secondary IgE antibody response to OA. Since the major antigenic determinants in OA molecules degraded by acetylation, the principle of acetylated OA treatment would be the same as treatment with UD antigen. The helper function of spleen cells from the UD-OA-treated animals and untreated controls was examined during the adoptive transfer experiment. In the UD-OA-treated animals, the helper function of spleen cells was significantly lower after treatment than of cells obtained before treatment. As the helper function of splenic T cells from the untreated animals did not change during this period, it is apparent that the function of T cells diminished after the treatment. Evaluation of OA-specific IgE-B cells in the spleens of untreated mice at 2 and 6 weeks after the priming immunization showed that the activity of the B-cell popul at'ion increased during this period. In the UD-OA-treated animals, however, the activity of IgE-B cells did not change after treatment. Thus, it appears that the treatment suppressed the development of, but did not diminish, the IgE-B cell population. Since UD-OA does not contain the major antigenic determinants, prevention of B-cell development in the treated animals was probably due to diminished T-cell activity. The experiments described above indicate that the primary effect of repeated injections of either native or modified antigen is depression of helper function that accounts for suppression of the secondary antibody response. Because immunotherapy for hay fever patients consists of weekly or biweekly injections of increasing doses of allergen, treatment

CELLULAR EVENTS IN THE IGE ANTIBODY RESPONSE

67

causes depressed helper function. A gradual decrease of IgE antibody titer after long-term immunotherapy may be expected if depressed helper function overcomes an increase of IgE-B cells. It is not established whether depression of the IgE antibody response has clinical value. If this is the case, modified antigen, such as UD antigen, may be more effective than native antigen for treatment, because injections of the modified antigen depress the helper function of T cells without increasing the number of IgE-B cells. Another advantage is that the modified antigen does not induce allergic reactions in ragweed-sensitive patients. One may expect that a high dose of modified antigen can be used for treatment without allergic side effects. A basic question to be asked is why helper function was depressed by repeated injections of allergen or modified antigen. A diminished DNA synthetic response of T cells after treatment excluded the possibility that the diminished helper function was due to supraoptimal number of helper T cells. As previously described, injections of OA or UD-OA into OA-primed BDFl mice induced the generation of OA-specific suppressor T cells, and the transfer of these cells into OA-primed mice depressed ongoing IgE antibody formation (see Table VII). These findings suggest strongly that depressed helper function by the treatment is probably due to generation of suppressor T cells. VI. Discussion and Summary

It is clear that the IgE antibody response is highly dependent on antigen-specific T cells, and collaboration between T and B cells is required for the formation of IgE antibody. The B memory cells involved in the IgE antibody formation are different from the memory cells for IgG/IgM antibodies. With respect to the maturation from stem cells to IgE-B cells, questions remain as to whether ( a ) antigen is required for the maturation and whether ( b ) a subset of T cells is involved. Further studies on the helminth infection that uniquely enhances IgE synthesis and proliferation of IgE-B cells may provide important information on this process. One of the mator questions yet to be answered is whether or not the antigen-specific T-helper cells for IgE-B cells are distinct from those for IgG-B cells. In the in uitro IgE antibody response by rabbit lymph node cells, helper function for IgE-B cells was obtained when the animals were immunized with an appropriate dose of alum-absorbed carrier. Such a function was not obtained by immunization with the same carrier included in CFA. Analysis of the IgG and IgE antibody responses in this system suggested strongly that helper cells for IgE antibody response

68

KIMISHIGE ISHIZAKA

are different from the cells for IgG antibody response. In the highresponder mouse strains, however, helper functions for both IgE and IgG antibody responses were obtained by immunization with a highly immunogenic antigen without adjuvant. Furthermore, culture of normal splenic T cells on antigen-bearing macrophages induced generation of helper cells that collaborated with both IgE and IgG antibody responses. In this process, neither antigen-specific B cells nor adherent cells other than macrophages were involved. Even in the high-responder mice, the magnitude of IgE antibody response vs IgG antibody response was different, depending on the nature and dose of antigen and adjuvant vehicles employed. Such a quantitative difference may be explained if IgE-B cells are more sensitive than IgG-B cells to both enhancing and suppressive control mechanisms of T cells. Since the IgE antibody response in the rat is greatly influenced by the adjuvant vehicles employed, it would be interesting to know whether the results obtained in the rabbit system may be reproduced in the rat. It has been shown both in uitro and in uiuo that the helper function of T cells for IgE antibody formation is mediated by soluble factors. Since soluble factors are released by stimulation with a modified antigen which has immunogenic determinants for T cells but lacks antigenic determinants for B cells, it appears that B cells are not involved in the release of soluble factors. In the in uitro system, evidence was obtained that enhancing soluble factor for IgE antibody formation is different from the factor for the IgG antibody response. The release of an enhancing factor for IgE antibody formation correlates with a helper function of the cells for IgE antibody response. When the carrier-specific cells had a helper function for the IgG antibody response but failed to collaborate with IgE-B cells, stimulation of these cells with carrier released an IgG-enhancing factor but not the IgE-enhancing factor. These findings indicate that the enhancing factor derived from T cells actually mediates B-cell differentiation and strongly support the hypothesis that T-cell populations regulate not only the amount of the antibodies but also the distribution of antibodies in different immunoglobulin classes. The IgE antibody response may be regulated by several factors. As expected, tolerance induction in B cells completely suppresses the formation of antibody associated with all immunoglobulin classes including IgE. Humoral antibodies regulate the antibody response. In high-responder mouse strains, IgG antibody failed to depress on-going IgE antibody formation or to suppress the secondary IgE antibody response. In the rat system, however, administration of passive antibody suppressed on-going IgE antibody formation without affecting the IgG antibody

CELLULAR EVENTS I N THE IGE ANTIBODY RESPONSE

69

response. The discrepancy between the two animal spccics may be related to the fact that memory .for the IgE antibody response was lacking in the rat system employed. It is quite clear that the IgE antil)od\ response is regulated by suppressor T cells. In the mouse system, repeated injections of antigen without adjuvant induced generation of antigen-specific suppressor T cells that suppressed primary antibody responses of both IgE and IgG classes. Since splenic T-cell preparations having the suppressiLe effect had a very low helper function, as determined by adoptive transfer experiment, and yet antigen-primcd T cells having a high helper function failed to suppress the primary antibody response in nonirradiated mice, therefore suppressor T cells appear to be distinct from helper T cells. It was also found that the transfer of suppressor T cells into antigen-primed mice suppressed the on-going IgE antibody response. It was shown in the rat that the transfer of thymocytes or spleen cells from hyperirnmunized animals terminated the on-going IgE antibody response. Extracts of thymocytes and spleen cells showed the same effect on the on-going IgE antibody response, and a subcellular component responsible for the suppression was identified. A qiiestion still remains as to whether the suppression is unique for IgE antibody formation and whether the T-cell preparations have helper function for the IgG antibody system. The antigen-specific T cells involved in this process may well be suppressor T cells. For this conclusion, however, evidence is required that hyperimmunization of rats with antigen in CFA is a favorable way to generate suppressor T cells. It is still possible that too large a population of hclper T cells might have suppressed the IgE antibody response. If the helper T cells for IgG-B cells are different from the helper T cells for IgE-B cells, the former cell population might have been responsible for the suppression of IgE antibody formation. Further studies in the rat system may clarify another regulatory mechanism for the IgE antibody response. Studies on high-rcsponder inbred mice provided an experimental model of immunotherapy for hay fever. Antigen-primed mice were treated I)y weekly injections of either native antigen or modified antigen, which react with T cells but do not possess the major antigenic determinants in the native antigen molecules. Analysis of the antigen-specific T cells and B cells before and after treatment revealed that a decline of helper function of the T-cell population was responsible for the depression of IgE antibody formation and suppression of the secondary IgE antibody response. In view of the fact that the same treatment induced generation of suppressor T cells, decline of helper function appears to be

70

KIMISHIGE ISHIZAKA

caused by suppressor T cells. One might expect that generation of suppressor T cells may be a possible approach toward immunotherapy for hay fever patients.

ACKNOWLEDGMENTS The author is grateful to his colleagues Drs. T. Ishizaka, T. Kishimoto, H. Okudaira, and K. Takatsu, who carried out our experiments cited in this review. He also acknowledges the collaboration of Dr. T. P. King, Rockefeller University.

REFERENCES Alkan, S. S., Nitecki, D. E., and Goodman, J. S. (1971). J. Immunol. 107, 353. Armerding, D., and Katz, D. H. (1974). J. E x p . Med. 140, 19. Bach, M. K., and Brashler, J. R. (1975). J. Immunol. 114, 1799. Baker, P. J., Barth. R. F., Stashak, P. W., and Amsbaugh, D. F. (1970). J. Immunol. 104, 1311. 5. Baker, P. J., Reed, N. P., Stashak, P. W., Amsbaugh, D. F., and Prescott, B. (1973). J. Exp. Med. 137, 1431. 6. Ball, P. A. J., and Barlett, A. (1969). Trans. R . SOC. Trap. Med. H y g . 63, 362. 7. Barrett, M.E.T. (1969). Immunology 22, 601. 8. Basten, A., Miller, J. F. A. P., Sprent, J., and Cheers (1974). J. Exp. Med. 140, 199. 9. Bazin, H., Beckers, A,, and Querinjean, P. ( 1974). Eur. J. lmmunol. 4, 44. 10. Benjamin, D. C. (1975). J. Exp. Med. 141, 635. 11. Binaghi, R. A., and Benacerraf, B. ( 1964).J. Immunol. 92, 920. 12. Bloch, K. J., Ohman, J. L., Jr., Waltin, J., and Cygan, R. W. (1973). J. Immunol. 110, 197. 13. Cheers, C., and Miller, J. F. A. P. (1972). J . E x p . Med. 136, 1661. 14. Claman, H. N., and Chaperon, E. A. (1969). Transplant. Reu. 1,92. 15. Clausen, C. R., Munoz, J., and Bergman, R. K. (1968). J. Bacterial. 96, 1484. 16. Clausen, C. R., Munoz, J., and Bergman, R. K. (1969). J. Immunol. 103,78. 17. Cooper, M. D., Lawton, A. R., and Kincade, P. W. (1972). Clin. Exp. lmmunol. 11, 143. 18. Davie, J. M., Paul, W. E., Katz, D. H., and Benacerraf, B. (1972). J. Exp. Med. 136, 426. 19. Delespesse, G., Ishizaka, K., and Kishimoto, T. (1975). J . Zmmunol. 114, 1065. 20. De Weck, A., personal communication. 21. Dutton, R. W. (1973). In “Immunologic Intervention” (J. W. Uhr and M. Landy, eds.), p. 67. Academic Press, New York. 22. Dutton, R. W., Falkoff, R. Hirst, J. A,, Hoffman, M., Kappler, J. W., Kettman, J. R., Lesley, J. F., and Vann, D. ( 1971). Prog. Zmmunol. 1, 355. 22a. Dwyer, J. W., Rosenbaum, J. T., and Lewis, S. (1976). J. Exp. Med. 143, 781. 23. Erb, P., and Feldman, M. (1975). CeZl. lmmunol. 19, 356. 24. Eardley, D. D., and Gershon, R. K. (1975). J. Exp. Med. 142, 524. 25. Falkenberg, F., Mozes, E., and Sela, M., unpublished data. 26. Falkoff, R., and Kettman, J. (1972). J. Immunol. 108, 54. 27. Feldmann, M. (1972). J. Exp. Med. 136, 737. 28. Feldmann, M., and Basten, A., (1972). J. Exp. Med. 136,722. 29. Feldmann, M., and Diener, E. (1970). J. Exp. Med. 131,247. 1. 2. 3. 4.

CELLULAR EVENTS IN THE IGE ANTIBODY RESPONSE

71

30. Gershon, R. K., and Kondo, K. (1970). Immunology 18,723. 31. Gershon, R. K. (1974). Contemp. Top. Mol. Immunol. 3, 1. 32. Gleich, G. J., Averbeck, A. K., and Swedlund, H. A. ( 1971). J. Lab. Clin. Med. 77, 690. 33. Golan, D. T., and Borel, Y. (1971). J. Exp. Med. 134, 1046. 34. Goldstein, G., Scheid, G. M., Hammering, U., Boyse, E. A., Schlesinger, D. H., and Nicoll ( 1975). Proc. Natl. Acud. Sci. U.S.A. 72, 11. 35. Gorczynski, R. M., Miller, R. G., and Phillips, R. A. (1972). J. Immunol. 108, 547. 36. Hamaoka, T., Katz, D. H., and Benacerraf, B. (1972). Proc. Natl. Acad. Sci. U.S.A. 69, 3453. 37. Hamaoka, T., Katz, D. H., and Benacerraf, B. (1973). J. Exp. Med. 138,538. 38. Harnaoka, T., Newburger, P. E., Katz, D. H., and Benacerraf, B. (1974). J. Immunol. 113, 958. 39. Hammer, D. K., Kirkhofen, B., and Schmid, T. (1971). Eur. J. Immunol. 1, 249. 40. Herzenberg, L. A., Okumura, K., and Metzler, C. M. (1975). Transplant. Rev. 27, 57. 41. Herzenberg, L. A., Chan, E. L., Ravitch, M. M., Riblet, R. J., and Herzenberg, L. A. (1973). J. Erp. Med. 137, 1311. 42. Hussain, R., Strejan, G., and Campbell, D. H. (1972). J. Immuml. 109, 638. 43. Ishizaka, K., and Adachi, T. (1976). J. Immunol. 117. 44. Ishizaka, K., and Ishizaka, T. (1973). In “Asthma: Physiology, Imrnunopharmacology, and Treatment” ( K . F. Austen and L. M. Lichtenstein, eds.), p. 55. Academic Press, New York. 45. Ishizaka, K., Ishizaka, T., and Hornbrook, M. (1966). J. Immunol. 97, 840. 46. Ishizaka, K., Ishizaka, T., and Hornbrook, M. M. (1970). Immiinochemistry 7, 515. 47. Ishizaka, K., Ishizaka, T., and Tada, T. ( 1969). J. Immunol. 103, 445. 48. Ishizaka, K., and Kishinioto, T. (1972). J. Zmmunol. 109, 65. 49. Ishizaka, K., Kishimoto, T., Delespesse, G., and King, T. P. (1974). J. Immunol. 113, 70. 50. Ishizaka, K., and Okudaira, H. (1972). J. Immrtnol. 109, 84. 51. Ishizaka, K., and Okudaira, H. (1973). J. Immunol. 110, 1067. 52. Ishizaka, K., Okudaira, H., and King, T. P. (1975). J. Immiinol. 114, 110. 53. Ishizaka, T., and Ishizaka, K. (1975). Prog. Allergy 19, 60. 54. Ishizaka, T., Ishizaka, K., and Tomioka, H. ( 1972). J. Immunol. 108, 513. 55. Ishizaka, T., Konig, W., Kurata, S., Mauser, L., and Ishizaka, K. (1975). I. Zmmrrtiol. 115, 1078. 56. Ishizaka, T., Urban, J., and Ishizaka, K. ( 1976).Cell. Immunol. 22, 248. 57. Jarrett, E. E. E., Haig, D. M., and Bazin, H. (1976). Clin. Exp. Immtcnol. (in press). 58. Jarrett, E. E. E., and Bazin, H. (1974). Nature (London) 251, 613. 59. Jarrett, E. E. E., and Stewart, D. C. (1972). Zmmtcnology 23, 749. 60. Jarrett, E. E. E., and Stewart, D. C. (1974). Immunology 27, 365. 61. Johansson, S. G. O., Mellbin, T., and Vahlquist, B. (1966). Lancet 1, 1118. 62. Johns, P. P., Cebra, J. J., and Herzenberg, L. A. (1974). J. Exp. Med. 139, 581. 63. Kabat, E. A,, Turino, G. M., Barrow, A. B., and Maurer, P. H. (1057). J. Clin. Invest. 36, 1160. 64. Katz, D. H., and Benacerraf, B. (1972). Adv. Immunol. 15, 1. 65. Katz, D. H., Davie, J. M., Paul, W. E., and Benacerraf, B. (1971). J. Exp. Med. 134, 201. 66. Katz, D. H., Hamaoka, T., and Benacerraf, B. (1972). J. Erp. Med. 136, 1404.

72

KIMISHIGE ISHIZAKA

67. Katz, I). H., Hamaoka, T., and Benacerraf, B. (1973). Proc. Natl. Acad. Sci. U.S.A. 70, 2776. 68. Katz, D. H., Hamaoka, T., Newburger, P. E., and Benacerraf, B. ( 1974). J. Immunol. 113, 974. 69. Katz, L). II., Paul, W. E., Goidl, E. A., and Benacerraf, B. (1970). J. EX^. Aled. 132, 261. 70. Kind, L. S., and Macedo-Sobrinho, B. (1973). Int. Arch. Allergy Appl. Immunol. 45, 780. 71. Kind, L. S., and Macedo-Sobrinho, B. (1973). 1. Immrrnol. 111, 638. 72. King, T. P., and Norman, P. S. (1962). Biochemistry 1, 709. 73. King, T. P., Norman, P. S., and Tao, N. (1974). lmmunochemistry 11, 83. 74. Kishimoto, T., and Ishizaka, K. (1972). J. Immunol. 109, 1163. 75. Kishimoto, T., and Ishizaka, K. (1973). J. Immunol. 111, 1. 76. Kishimoto, T., and Ishizaka, K. (1973). J. Immunol. 111,720. 77. Kishimoto, T., and Ishizaka, K. (1973). J. Immunol. 111, 1194. 78. Kishimoto, T., and Ishizaka, K. ( 1974). J. Immunol. 112, 1685. 79. Kishimoto, T., and Ishizaka, K. ( 1975). J. Immunol. 114, 1177. 80. Kishimoto, T., Miyake, T., Nishizawa, Y., Watanabe, T., and Yamaniura, Y. (1975). J. Immunol. 115, 1179. 81. Kojima, S., and Ovary, Z. (1975). Cell. Immunol. 15, 274. 82. Kojima, S., and Ovary, Z. (1975). Cell. Immunol. 17, 383. 82a. Lawton, A. R., Asofsky, R., Hylton, M. B., and Cooper, M. D. (1972). J. E r p . Med. 135, 277. 83. Lee, W. Y., and Sehon, A. H. (1975). J. Immunol. 114,829. 84. Lee, W. Y., and Sehon, A. H. ( 1975). J. Immunol. 114,837. 85. Lehrer, S. B., Tan, E. M., and Vaughan, J. H. (1974). J. Immunol. 113, 18. 86. Levey, R. H., and Medawar, P. B. (1966). Proc. R. SOC.London, Ser. B 174, 155. 87. Levine, B. B. ( 1971), In “Biochemistry of the Acute Allergic Reactions” ( K. F. Austen and E. L. Becker, eds.), p. 1. Blackwell, Oxford. 88. Levine, B. B., Chang. H., and Vaz, N. M. (1971). J. Imrnunol. 106,26. 89. Levine, B. B., Stember, R. H., and Fotino, M. (1972). Science 178, 1201. 90. Levine, B. B., and Vaz, N. M. (1970). Int. Arch. Allergy Appl. Immunol. 39, 156. 91. Lindquist, K. ( 1968). Immunochemistry 5, 525. 92. Maia, L. C. S., Vaz, N. M., and Vaz, E. M. (1974). Int. Arch. Allergy Appl. Immunol. 46, 339. 93. Makela, 0. (1970). Transplant. Reu. 5, 3. 94. MWkelW, O., and Cross, A. M. (1970). Prog. Allergy 14, 145. 95. Marbrook, J. (1967). Lancet 2, 1279. 96. Marsh, D. G. ( 1976). In “Childhood Hay Fever” ( M . G. Rhyne, ed.). Thomas, Springfield, Illinois ( in press ). 97. Marsh, D. G., Bias, W. B., Hsu, S. H., and Goodfriend, L. ( 1973). In “Mechanisms in Allergy” ( L . Goodfriend, A. H. Sehon, and R. P. Orange, eds.), p. 113. Dekker, New York. 98. Marsh, D. G., Bias, V. B., and Ishizaka, K. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 3588. 99. Marsh, D. G., Lichtenstein, L. M., and Norman, P. S. (1972). Immunology 22, 1013. 100. McDevitt, H. O., and Benaceraff, B. (1969). Adu. Immunol. 11, 31.

CELLULAR EVENTS IN THE IGE ANTIBODY IIESPONSE

73

hlcDevitt, H. O., and Chinitz, A. (1969). Science 163, 1207. Michael, J. G., and Bernstein, I. L. (1074). J. Allergy C h i . Zrnnitrnol. 53, 67. Michael, J. G., and Bernstein, I. L. (1973). J. Immcrnol. 111, 1600. Miller, J. F. J. A., and Mitchell, G . F. (1969). Transplant. Reu. 1, 3. Mischell, R. I., and Dntton, R. W. ( 1967). J Erp. Med. 126, 423. Mitchison, M. A., Rajcwsky, K., and Taylor, R.B. (1971). In “Developmental Aspects of Antibody Formation and Structure” ( J . Sterzl and I. Riha, eds.), Vol. 2, p. 547. Acadeniic Press, New York. 107. Monro, A. J., Taussig, M.J., Campbell, R., Williams, H., and Lawson, Y. (1974). J. Exp. Med. 140, 1579. 108. Mota, E. H., Sadun, E. €I., Bradshaw, R. M., and Gore, R. W. (1969). Immunology 16, 71. 109. hlotn, I. (1964). Zmmicnology 7, 681. 110. hlota, I. (1967). Zmmrrnology 12, 343. 111. hlota, I. (1972). Immunology 11, 137. 112. Mota, I., and Peixoto, J. M. (1966). Life Sci. 5, 1723. 113. Munoz, J., Schuchardt, L. F., and Verwey, W. F. (1958). J. Zmmunol. 80, 77. 114. Newberger, P. E., Hamaoka, T., and Katz, D. H. ( 1974). J. Zmrnnnol. 113, 824. 115. Ogilvie, B. M. (1961). Nature (London) 204, 91. 116. Ogilvie, B. M. (1967). Zmmtrnology 12, 113. 117. Ogilvie, B. M., and Jones, V. E. (1973). Prog. Allergy 17, 93. 118. Ogilvie, B. M., Smithers, S. R., and Terry, R. J. (1966). Natirre (London) 209, 1221. 119. Okudaira, H., and Ishizaka, K. (1973).J. Zinmrmol. 111, 1420. 120. Okudaira, H., and Ishizaka, K. ( 1974 ). J. Zmmtinol. 113, 563. 121. Okudaira, H., and Ishizaka, K. (1975).1. Zmmzcnol. 114, 615. 122. Okuniura, K., and Tada, T. (1971). J. Zmmunol. 10G, 1019. 123. Okumura, K., and Tada, T. ( 1971). J. Zmmtcnol. 107, 1682. 124. Okumura, K., and Tada, T. (1974). J. Zmnwnol. b12, 783. 125. Okumura, K., Tada, T., and Ochiai, T. ( 1974). Zmrnrcnology 26, 257. 126. Orr, T. S. C., and Blair, A. M. J. N. (1969). Life Sci. 8, 1073. 127. Orr, T. S. C., Riley, P. A., and Doe, J. E. ( 1971). Zmmrcnology 20, 185. 128. Paul, W. E., Katz, D. H., and Benacerraf, B. (1971). J. Zmmnnol. 107, 685. 129. Pierce, C. W., Solliday, S. M., and Asofsky, R. (1972). J. Exp. Med. 135, 698. 130. Prouvost-Danon, A,, Binaghi, R., Rochas, S., and Boussac-Aron, Y. ( 19721. Immunology 23, 481. 131. Revoltella, R., and Ovary, Z. (1969). Z~~imunology 17, 45. 132. Revoltella, R., and Ovary, Z. ( 1969) , Znt. Arch. Allergy Appl. Zrnmunol. 36, 282. 133. Rockey, J. €I., and Schwartznian, R. M. (1967). J. Zmmunol. 98, 1143. 134. Rowe, D. S., Hug, K., Faulk, W. P., McCormick, J. N., and Gerber, H. (1973). Nature (London) 242, 155. 135. Rowley, D. A,, and Fitch, F. W. ( 1964). J. Erp. Med. 120, 987. 136. Rubin, A. S., and Coons, A. H. (1972).J. Zmmtrnol. 108, 1597. 137. Sadun, E. H., Mota, I., and Gore, R. W. (1968). J. Parasitol. 54, 814. 138. Schimple, A,, and Wacker, E. (1973).J. Exp. Med. 137,547. 139. Sehon, A. H. ( 1972). In “The Biological Role of the Immunoglobulin E System” ( K . Ishizaka and D. H. Dayton eds.), 11. 297. Nat. Inst. Child Health Hum. Dev. Bethesda, Md. 140. Sell, S., and Sheppard, H. (1973). Science 182, 586. 141. Senyk, G., Williams, E. B., Nitecki, D. E., and Goodman, J. W. (1971). J. Erp. Med. 133, 1294.

101. 102. 103. 104. 105. 106.

74

KIMISHIGE ISHIZAKA

142. Sherman, W. B., Stull, A., and Cooke, R. A. (1940). J . Allergy 11,225. 143. Shinohara, N., and Tada, T. (1974). Int. Arch. Allergy Appl. Immunol. 47, 762. 144. Siskind, G. W., and Benacerraf, B. ( 1969). Ado. Immunol. 10, 1. 145. Strannegard, O. ( 1971). Clin. Exp. Immunol. 8, 963. 146. Strannegard, O., and Belin, L. (1970). Immunology 18, 773. 147. Strannegard, O., and Chan, P. C. Y. (1969). J. Allergy 43, 224. 148. Strejan, G. H., and Campbell, D. H. (1967). J. Immunol. 98, 893. 149. Strejan, G. H., and Campbell, D. H. (1968). J. Immunol. 101, 628. 150. Strejan, G. H., and Campbell, D. H. (1970). J . Immunol. 105, 1264. 151. Strejan, G. H., Hussain, R., and Bradburg, S. (1973). In “Mechanisms in Allergy’’ (L. Goodfriend, A. H. Sehon, and R. P. Orange, eds.), p. 33. Dekker, New York. 152. Strejan, G. H., and Marsh, D. G. ( 1971). J. Immunol. 107,306. 153. Tada, T. (1975). Prog. Allergy 19, 122. 154. Tada, T., and Ishizaka, K. (1970). J . Immunol. 104, 377. 155. Tada, T., and Okumura, K. (1971). J . Immunol. 106, 1002. 156. Tada, T., Okumura, K. (1971). J . Immunol. 107, 1137. 157. Tada, T., Okumura, K., Ochiai, T., and Iwasa, S. (1972). Int. Arch. Allergy Appl. Immunol. 43, 207. 158. Tada, T., Okumura, K., and Taniguchi, M. ( 1973). In “Mechanisms in Allergy” (L. Goodfriend, A. H. Sehon, and R. P. Orange, eds.), p. 43. Dekker, New York. 159. Tada, T. Okumura, K., and Taniguchi, M. (1973). J . Immunol. 111, 952. 160. Tada, T., Okumura, K., and Taniguchi, M. ( 1972). J . Immunol. 108, 1535. 161. Tada, T., and Takemori, T. (1974).J. Exp. Med. 140,329. 162. Tada, T., Taniguchi, M., and Okumura, K. (1971). 1. Immunol. 106, 1012. 163. Takatsu, K., and Ishizaka, K. (1975). Cell. Immunol. 20, 276. 164. Takatsu, K., and Ishizaka, K. (1975). Fed. Proc., Fed. Am. SOC.Exp. Biol. 34, 1000. 165. Takatsu, K., and Ishizaka, K. (1976). J . Immunol. 116, 1257. 166. Takatsu, K., Ishizaka, K., and King, T, P. ( 1975). J. Immunol. 115, 1469. 167. Taniguchi, M., and Tada, T. (1974). 1. Immunol. 113, 1757. 168. Taniguchi, M., and Tada, T. (1974). J . E x p . Med. 139, 108. 169. Taub, R. N. (1970). Prog. Allergy 14,208. 170. Taussig, M. J., Mozes, E., and Isaac, R. (1974). J . Exp. Med. 140, 301. 171. Uhr, J. W., and Bauman, J. (1961). J . Exp. Med. 113,935. 172. Uhr, J. W., and Moller, G. (1972). Adu. Immunol. 15, 95. 173. Vaz, E. M., Vaz, N. M., and Levine, B. B. (1971). Immunology 21, 11. 174. Vaz, N. M., deSouza, C. M., and Maia, S. L. C. ( 1974). Int. Arch. Allergy Appl. Immunol. 46, 275. 175. Vaz, N. M., Phillips-Quagliata, J. M., Levine, B. B., and Vaz, E. M. (1971). J . Exp. Med. 134, 1335. 176. Vaz, N. M., deSouza, C. M., and Marx, L. C. S . (1974). Int. Arch. Allergy Appl. Immunol. 46, 275. 177. Vaz, N. M., Vaz, E. M., and Levine, B. B. (1970). J. Immunol. 104j 1572. 178. Waldmann, T. A. ( 1969). N. Engl. J. Msd. 281, 1170. 179. Walker, J. G., and Siskind, G. W. (1968). Immunology 14,21. 180. Warner, N. L. (1972). Contemp. Top. Immunobiol. 1, 87. 181. Warner, N. L. (1974). Ado. Immunol. 19, 67. 181a. Watanabe, N., Kojinia, S., and Ovary, Z. (1976). J. Exp. Med. 143, 833. 182. Watson, J. ( 1973). J . Immunol. 111, 1301.

CELLULAR EVENTS I N T H E IGE ANTIBODY RESPONSE

75

183. Whalen, C. E., Rosenberg, E. B., Strickland, G . T., Gutnian, R. A., Cross, J. H., and Waken, R. H. (1969). Lancet 1, 13. 184. White, G . J., and Holm, M. S. ( 1973). J. Imrnunol. 110, 327. 185. Wide, L., Bennich, H., and Johansson, S. G. 0. (1968). Lancet 2, 1105. 186. Wigzell, H. (1966). J. Exp. M e d . 124, 953. 187. Wigzell, H. (1971). Prog. Ztnmnnol. 1, 1105. 188. Wilson, J. M., and Bloch, K. J. (1968). J. Inmntiol. 100, 628. 189. Zvaifler, N. J., and Becker, E. L. ( 1966). J. E x p . Med. 123, 935. 190. Zvaifler, N. J., Sadun, E. H., Becker, E. L., and Schoenbechler, M. J. (1967). Erp. Parasitol. 20, 278.

This Page Intentionally Left Blank

Chemical and Biological Properties of Some Atopic Allergens T.

P. KING

The Rockefeller University, N e w York, N e w York

I. Introduction . . . . . . . . . 11. Allergen Assay . . . . . . . . . 111. Chemical and Biological Properties of Some Allergens . A. Ragweed Pollen Allergens . . . . . . B. Grass Pollen Allergens . . . . . . . C. Tree Pollen Allergens . . . . . . . D. Mammalian Dander Allergens . . . . . E. Mite and House Dust Allergens . . . . F. Food Allergens . . . . . . . . G. Honeybee Venom Allergens . . . . . IV. General Observations on Allergens . . . . . V. Uses of Purified Allergens . . . . . . VI. Concluding Remarks . . . . . . . . References . . . . . . . . . .

.

. . . .

.

. . . . . . . .

.

. . . .

.

. . . . . . . .

.

. . . .

.

. . . . . . . .

.

. . . .

.

. . . . . . . .

77 78 80 80

85 86 87 88 88 91

92 96 100 101

I. Introduction

All of us are constantly exposed to a variety of antigens froin various sources in our environment by inhalation, ingestion, injection, or contact. However, only certain genetically predisposed persons, about 10% of our population, become hypersensitized ( allergic) to these antigens under the natural conditions of exposure ( Cooke, 1947). Antigens that can induce immediate and/or delayed types of hypersensitivity in man are termed allergens. This article is concerned with the allergens causing the immediate type of hjyersensitivity in man. One form of immediate allergy in man is anaphylaxis which includes the manifestations of hay fever, asthma, and hives. This form of allergy is also known as atopy. There are many cnvironniental sources of atopic allergens, such as pollens of grasses, trees, and weeds, animal danders, fungi, insects, foods, certain drugs and chemicals and, recently, laundry detergents containing additives of bacterial enzymes ( Gutman and Miyamoto, 1972; Norman, 1971). It is generally accepted that the antibodies involved in atopic allergy belong primarily, if not solely, to the class of immunoglobulins of IgE. Prior to the isolation and characterization of IgE (Ishizaka, 1973), this class of antibody was known as reagin, skin-sensitizing or homocytotropic antibody and these terms are still used. The IgE from man or 77

78

T. P. KING

animals has the special property of binding firmly to mast cells and basophiles of the same or of closely related species. On combination of cell-bound IgE and the specific allergen, certain cellular events take place and chemical mediators, such as histamine, are released into the medium ( Austen, 1974; Lichtenstein, 1974). These mediators, in turn, cause contraction of smooth muscles and permeability changes of blood vessel walls to give the allergic symptoms. In this way the final biological expression of the combination of IgE and specific allergen is greatly amplified. For example, intradermal injection of 10-l6 mole of the ragweed pollen allergen, antigen E, is sufficient to give a wheal and erythema reaction in most ragweed-sensitive individuals (King et al., 1964). Many studies have been made to purify and characterize chemically the atopic allergenic components from various sources, and these studies have been reviewed by several authors (Berrens, 1971; Marsh, 1975; Stanworth, 1973). Most purified allergens have been shown to be proteins, usually in the molecular weight range of 10,000 to 60,000. The chemical properties of the purified protein allergens do not indicate any obvious differences from those of other protein antigens. Polysaccharides, such as dextran (Kabat et al., 1957) and pneumococcal polysaccharide (Leskowitz and Lowell, 1961) can be allergens, but they are much less common. Many simple but reactive chemical compounds that form covalent linkages with proteins, such as 2,4-dinitrofluorobenzene, can also serve as allergenic determinants. The purpose of this article is twofold: one, to examine the chemical and biological properties of some purified allergens and, the other, to indicate some practical uses of the purified allergens in our understanding and control of atopic allergy. II. Allergen Assay

The mean serum concentration of total IgE for allergic individuals is about 1.5 pglml whereas that for normal individuals is about one-fifth of this value (Bennich and Johansson, 1971). The serum level of allergenspecific IgE is clearly less than that of total IgE. For example, in 13 of 15 untreated ragweed-sensitive persons, the IgE specific for the ragweed allergen antigen E represented 3-22%of the total seivm IgE; of the other 2 persons, in one it was less than 0.3%,and in the remaining one, 41% (Zeiss et aZ., 1973). Therefore, the reaction of allergen with its specific IgE can best be detected by biological assays or by radiochemical procedures. Yman (1975) has presented a general review of the procedures commonly used for allergen assay. The two most frequently used biological assays for measuring the

PROPERTIES OF SOME ATOPIC ALLERGENS

79

allergenic activities of samples are ( a ) the direct skin test on allergic persons and ( b ) the histamine release assay using leukocytes of allergic donors (Osler et al., 1968). Leukocytes from normal donors may also be used for the histamine release assay after passive sensitization with allergic sera. Histamine release assays are more complicated to perform than the skin tests are, but the results are more quantitative. The higher the concentration of skin mast cell- or leukocyte-bound specific IgE, the lower is the concentration of allergen required for a positive skin reaction or for histarnine release. The amount of cell-bound IgE of any given specificity is proportional to its relative concentration in the serum. Thus, the allergenic activity measured by these two biological methods is a reflection of the serum concentration of allergen-specific IgE. This relationship has been demonstrated experimentally with the ragweed allergen antigen E system (Norman et al., 1973; Zeiss et al., 1973). Allergic persons show varying responses to the same allergen as well as to different allergens. Therefore, in order to obtain a reliable estimate of the relative allergenic activities of different preparations, it is necessary to carry out the tests in a relatively large group of persons, preferably greater than 10. For a mixture of allergens or crude allergen extracts, the activity measured by either of the two biological tests will depend on the relative concentration of the allergens and their specific IgEs. Usually the activity measured will reflect primarily that of the allergen component present in highest concentration. Haptens are inactive in these two bioassays, as only multivalent allergens on combination with IgE can trigger the cellular reactions to give positive tests (Levine, 1965, 1966; De Weck, 1974, Ishizaka, 1973). The radioallergosorbent test (RAST) developed by Wide et al. (1967) can also be used for allergen assays. The allergens are chemically coupled to a solid support, such as paper discs (Ceska et al., 1972a). The solidphase allergens serve as immunosorbent for reaginic sera, and the amount of allergen-specific IgE absorbed is then determined by its reaction with radiolabeled IgE. A useful modification is to run inhibition of RAST; that is, allergens in solution compete with solid-phase allergens for the binding of specific IgE (Foucard et al., 1972; Yman et al., 1973; Gleich et al., 1974). This modification is particularly useful for allergen purification studies as only one solid-phase allergen has to be prepared and is used to test all fractions. The method can be used to test the purity as well as the identity of different allergens. For a multiallergen system, the maximal inhibition by each allergen will be less than that attained by the mixture of allergens. The extent of its maximal inhibition will depend on the abundance of its specific IgE in the reaginic serum sample used. The RAST test is simple in principle, but there are still unanswered

80

T. 1’. KING

questions regarding its use to measure the concentration of specific IgE. For example, when a mixture of allergens of different sizes and shapes is coupled to the solid-phase support, do the immobilized allergens have the same affinities for their specific IgEs as the allergens have in solutions? If not, RAST is selective in its measurement for allergens whose affinities for IgE are less affected than those with sharply reduced affinities. Nonetheless, RAST has obvious advantages over the biological tests as a routine laboratory method for following allergen purification. A useful in vitro method to determine the number of allergens in a complex mixture is by crossed rndioimmunoelectrophoresis ( Weeke and Lplwenstein, 1975). The mixture is first analyzed by crossed immunoelectrophoresis, i.e., the mixture is first separated on agarose gel electrophoresis, followed by rocket immunoelectrophoresis in the second dimension into a gel containing precipitating antibodies against the mixture. Duplicate gels are run. One is stained with dye to reveal the different antigen peaks; the other is exposed to sera from persons who are allergic to the mixture, followed with radiolabeled anti-IgE. Radioautography will then show which antigen peaks are active as allergens by their binding of IgE. A survey of serum samples from different individuals will establish the extent of variation in the individual response to different allergens.

Ill. Chemical and Biological Properties of Some Allergens

A. RAGWEEDPOLLENALLERGENS 1. Characterization The chief causative agent for late-summer hay fever in the eastern United States and Canada is ragweed pollen. It is for this reason that the allergens of ragweed pollen have received more studies by different laboratories than any other pollen allergens, The studies prior to 1960 were reviewed by Richter and Sehon (1960). The two most abundant species of ragweed plants are Ambrosia trifida and Ambrosia elatior, commonly known as giant or tall and short or dwarf ragweed, respectively. The pollen proteins of these two ragweed species were shown to be antigenically related by immunodiffusion against rabbit antisera ( Wodehouse, 1954; Goldfarb et al., 1958) and to be allergenically related by RAST inhibition tests using human reaginic sera (Yunginger and Gleich, 1972). Nearly all reported studies on ragweed pollen allergen were carried out with short ragweed pollen. Immunodiffusion with rabbit antisera had shown that aqueous extracts of short ragweed pollen contain at least ten different antigens

PROPERTIES OF SOME ATOPIC ALLERGENS

81

( Wodehouse, 1954; Cebra, 1957), but these antigens are not equally active as allergens in sensitive individuals ( King and Norman, 1962). Five of these antigens of short ragweed pollen have been purified. The two major allergens, designated a s antigens E and K, are acidic proteins with molecular weights of about 38,000 daltons and PIS in the range of 4.8 to 6.0 (King et al., 1964, 1967a; King, 1972). These two antigens represent 6 and 3%of pollen proteins, respectively, Both antigens E and K can be obtained in multiple electrophoretic forms which are separable on ion exchange chromatography. The purity of the isolated antigens E and K was demonstrated on isoelectric focusing (King, 1972) and on disc electrophoresis (King et al., 1974). Both antigens were found to contain less than 0.6%of carbohydrate, and this amount is probably due to a contaminant. In each case, at least 99%of the weight of the sample can be accounted for as common amino acid residues, so it is not likely that these proteins contain other prosthetic groups. Antigens E and K were found to share some antigenic determinants using rabbit and human antisera (King et al., 1967a). Studies by other workers (Callaghan and Goldfarb, 1962; Robbins et al., 1966) also led to the isolation of antigen E-rich fractions as the major allergen. The other three allergens of ragweed pollen are designated as antigens Ra3 ( Underdown and Goodfriend, 1969), Ra5 ( Goodfriend and Lapkoff, 1974; Roebber et al., 1975), and BPA-R or Ra4 (Griffiths and Brunet, 1971; Griffiths, 1972; Roebber, 1975). They are basic proteins with molecular weights of 11,000, 4970, and 23,000 daltons, respectively, having PIS in the range of 8.0 to 9.6. They represent, respectively, 0.4, 0.1 and 0 . a of pollen proteins. Antigens Ra3 and Ra4 may be glycoproteins, as these preparations contain small amounts of carbohydrate. Antigen Ra4 was found to share common antigenic determinants with antigen E, but antigens Ra3 and Ra5 are antigenically distinct from antigen E (Roebber, 1975). The complete amino acid sequence of antigen Ra5 has been determined (Mole et al., 1975), and partial sequences in the amino terminal region of antigens Ra3 and Ra4 ragweed have been reported ( Roebber, 1975). By direct skin tests on ragcveed-sensitive individuals (King et al., 1964) and by histamine release assay with leukocytes from sensitive donors (Lichtenstein et al., 1966), it was established that antigen E was the most active allergen of all the fractions obtained in the course of the fractionation of ragweed pollen extract. Direct skin tests on 184 sensitive individuals showed that the median minimal amount of antigen E required to elicit n positive skin reaction was about lo-'? gm (correspondPNU),' whereas similar tests on 114 individuals with ing to 1.7 x PNU stands for protein nitrogen unit (10 conimonly used to standardize allergen extracts.

gn/nil), a measure that is

82

T. P. KING

the unfractionated ragweed pollen proteins ( i.e., whole extract) showed PNU). In about the median amount required to be gm (1.3 x 5% of the individuals tested, antigen E was less active than the whole extract. Antigen K is about half as active as antigen E (King et al., 1967a). Its activity in sensitive individuals is closely related to that of antigen E, probably because these antigens share common determinants. Comparative studies on the allergenic activities of antigens Ra3 (Lichtenstein et al., 1973), Ra4 (Santilli et al., 1975) and Ra5 (Marsh et al., 1973) with antigen E have appeared in several reports. The results show that these antigens are highly allergenic in about 20 to 30%of the sensitive individuals as contrasted to the high activity of antigen E in the majority of individuals tested (95%).The results also show that antigens Ra3 and Ra4 are distinct allergens, as people show varying responses to these two antigens and to antigen E. A more detailed description of the results on antigens Ra3 and Ra5 is given below. The allergenic activity of antigen Ra3 relative to that of antigen E was studied by the histamine release assay from leukocytes ( Lichtenstein et al., 1973). Of the 22 individuals studied, 3 were 2- to 10-fold more sensitive to Ra3 than to E, 13 were 2- to 10,000-fold less sensitive to Ra3 than to E, and 6 showed no response whatsoever at the highest concentration of Ra3 (10 pg/ml) tested. The concentration of antigens Ra3 and E required for 50%histamine release in the individuals studied was in the range of to pg/ml. Similar results were found by end-point skin test titrations on 19 individuals. Comparative skin tests of antigens Ra5 and E were made (Marsh et al., 1973). Of the 105 persons tested, 18 (17%)were 1- to 10-fold more reactive to Ra5 than to E, 10 (10%) were 100- to 1000-fold less reactive to Ra5 than to E, and 77 (73%)were more than 100,000-fold less reactive to Ra5 than to E. Full details on the studies of Ra4 have not yet appeared in print ( Santilli et al., 1975). In addition to the five purified allergens, ragweed pollen probably contains other allergens yet to be identified. This is indicated by the distribution of allergenic activity among the various fractions of different antigens from ragweed pollen ( King and Norman, 1962). Recently, Marsh (1975) reported that a short 16-minute aqueous extraction of ragweed pollen released less than 1%of the total extractable antigen E but nearly all the Ra5. Yet this extract, when expressed in terms of its antigen E content, was found to have a higher allergenic activity than antigen E did, as measured by leukocyte histamine release and intradermal skin tests on 38 patients. His results would suggest the presence of other ragweed allergen ( s ) having higher allergenic activity than antigen E. A similar report was made by Goldfarb (1968).

PROPERTIES OF SOME ATOPIC ALLERGENS

83

Characterization of these rapidly extracted components of ragweed pollen is of interest, as these reports are difficult to reconcile with the following observations that implicate antigen E as the major allergen of ragweed pollen. ( a ) The allergenic activity of ragweed pollen extract, with its antigen E removed by absorption with specific rabbit antisera, was decreased 10-fold, or more, when skin tested on 11 of the 13 patients, and the decrease was less than 10-fold in the remaining 2 patients (King et al., 1964, 1967a). ( b ) In ragweed-sensitive patients, antigen-E-specific IgE accounted for an average of 45% of all IgE antibodies specific for ragweed allergens. This is shown by the data of Gleich and Jacob (1975) and Zeiss et al. (1973). Gleich and Jacob showed that the IgE antibodies specific for ragweed allergens accounted for an average of 29% of total serum IgE (range 0-553 for 10 patients), while Zeiss et al. showed that the IgE antibody specific for antigen E accounted for an average of 13% of total serum IgE (range 0.3-41% for 15 patients). 2. Studies on Antigen E Antigen E consists of two nonidentical polypeptide chains ( Griffiths, 1972, 1973; King et al., 1974), LY and p, with molecular weights of about 26,000 and 13,000 daltons, which are held together in the native molecule by noncovalent forces.' On dissociation under denaturing conditions, such as low pH, urea or detergent, and subsequent removal of denaturants, the chains do not recombine to form the native molecule but they form aggregates. The and p chains of antigen E are separable by chromatography on agarose columns employing a pH 4.7 acetate buffer containing 6 M guanidine hydrochloride. The two chains differ in their amino acid compositions and their electrophoretic mobilities. The electric charge difference of the various isoelectric forms of antigen E was found to be localized in the p chain, since the N chain from different forms of antigen E showed identical electrophoretic mobilities. The denatured antigen E and its separated LY and p chains have conformations different from that of the native antigen, as indicated by their altered physical properties. They also show altered antigenic properties. Thc LY and p chains and the urea-denatured antigen E are all about 0.001 to 0.0001 as active as native antigen when skin tested on sensitive individuals. The chains and the denatured antigen neither reacted with sheep antiantigen E sera nor inhibited the reaction of antigen E with aiitisera to a significant degree. When compared to the native antigen, the denatured antigen is a poor immunogen in rabbits and mice and the (Y

* Originally, we reported their molecular weight as being about 22,000 and 16,000 (King et d., 1974), but recent unpublished experiments have given slightly different values, as indicated above.

84

T. 1’. KING

weak antisera produced cross-reacted poorly with the native antigen (King et al., 1974; Ishizaka et al., 1975). Native antigen E is stable toward proteolytic digestion at neutral pH by trypsin, chymotrypsin, and papain, but it is readily digested by the bacterial proteinase nagarse at neutral pH (King et al., 1967b). The nagarse-digested antigen E showed less than 0,001 of the allergenic activity of intact antigen E. Reduction of the three disulfide bonds of antigen E followed by carboxymethylation of the liberated thiol groups also abolished its antigenic and allergenic activities. The reduced and alkylated derivative had less than 0.0001 of the activity of the native molecule as shown by direct skin tests on patients. Also, this derivative was found to be a very poor inhibitor, if at all, of the reaction between antigen E and its specific human or rabbit IgG antibody. Other chemical modifications of antigen E were carried out, namely acetylation, succinylation, and butyrimidination of its amino groups and coupling of its carboxyl groups with glycinamide or with taurine (King et al., 1967b, 1974). Sixteen out of the total 18 amino groups of antigen E are readily accessible for modification, and 18 out of the total 34 carboxyl groups are readily accessible. The derivatives with the modified accessible amino or carboxyl groups showed small changes in their allergenic, antigenic, and physical properties. However, the extensively modified derivatives had lost their biological activities and formed soluble aggregates indicative of conformational changes of the molecule. Induction of antibody formation to a variety of immunogens is known to require the collaboration of thymus-derived ( T ) and bone marrowderived ( B ) lymphocytes (Katz and Benacerraf, 1972). The helper T cells in some way regulate the proliferation and differentiation of B cells into antibody-secreting plasma cells. The antigenic and the immunogenic properties described above for the urea-denatured antigen E and its isolated chains and p indicate that the B-cell determinants of antigen E are strongly conformation-dependent, as the modified derivatives neither reacted with nor elicited the formation of antigen E-specific antibodies. The modified derivatives have B-cell determinants different from those of native antigen as the antisera raised in animals did not cross-react with the native antigen. However, these derivatives were found to possess helper T-cell determinants common with those of the native antigen E, as shown by several different experiments (Ishizaka et al., 1974, 1975). For example, priming of mice with the denatured antigen enhanced both IgG and IgE responses to the native antigen E. Thus the helper T-cell determinants of antigen E, in contrast to its B-cell determinants, are conformation-independent. This property is not unique to antigen E. An extensively modified bacterial a-amylase showed (Y

PROPERTIES OF SOME ATOPIC ALLERGENS

83

immunological properties identical to those of denatured antigen E, namely nonimmunogenic for the native enzyme but having the priming effect for the native enzyme (Nakashima et al., 1974). It is known that T-cell determinants for protein antigens in cell-mediated immunity are conformation-independent ( Gel1 and Benacerraf, 1961; Goodman, 1975).

B. GRASSPOLLENALLERGENS Grass pollens are another common cause of hay fever. There are several thousand species of grasses but only twenty or so are sufficiently abundant to be important in allergy ( Wodehouse, 1971). Like ragweed pollen, grass pollens are antigenically complex, both timothy ( Phleum pratense ) and orchard ( Dactylis glomerata) grass pollens containing at least fifteen antigens ( Augustin and Hayward, 1962). Using the more sensitive technique of crossed immunoelectrophoresis, the presence of twenty-seven antigens was detected in timothy pollen (Lgwenstein et al., 1974), and seven of these antigens were allergenic in man, as shown by crossed radioimmunoelectrophoresis ( Weeke and Lgwenstein, 1975). Four of the seven timothy antigens bound IgE from an serum samples tested (30 patients), whereas the others bound IgE from less than half of the sera analyzed. The allergens of timothy pollen have been studied by several laboratories ( Augustin and Hayward, 1962; Malley and Harris, 1967; Lgwenstein et al., 1974). Malley and associates isolated two timothy pollen allergens designated as antigens A and B. These two antigens are acidic proteins with molecular weights of about 30,000 and 16,000 (Malley and Harris, 1967; Marsh, 1974). They are in part antigenically related on immunodiffusion. Both antigens are highly skinreactive in timothy-allergic persons at a concentration of gm/ml. Four different pollen allergens from ryegrass ( Loliurn perenne ) have been isolated (Johnson and Marsh, 1965, 1966a; Marsh et al., 1970a; Marsh, 1975). These allergens are designated as Groups I, 11, 111, and IV. With the exception of Group IV antigen, each one is isolated in multiple electrophoretic forms. Groups I and I1 are acidic proteins with molecular weights of about 27,000 and 11,000 daltons, respectively, whereas Groups I11 and IV are basic proteins with molecular weights of about 11,000 and 50,000 daltons, respectively. The four ryegrass allergens are antigenically distinct as shown by immunodiffusion studies with rabbit antirye sera (Marsh, 1974) and by leukocyte histamine release assays with sensitive patients ( Lichtenstein et al., 1969). Group I antigen was found to be the major allergen as established by skin and leukocyte sensitivities of over 250 grass-sensitive individuals; over 95% tested were highly reactive to Group I. The other three antigens, Groups 11, 111, and IV, showed variable responses in

86

T. P. KING

individuals sensitive to Group I antigen ( Marsh, 1975). Group I1 antigen was as active as, or more so than Group I antigen in 45 (60%)of the 75 allergic individuals tested; Group I11 antigen was so in 43 (70%) of 61 individuals tested; Group IV antigen was so in 10 (20%)of 50 individuals tested. Chemical modifications of the major ryegrass pollen allergen Group I have been reported (Johnson and Marsh, 1966b). Removal of the carbohydrate present in Group I by enzymatic digestion with cellulase and galactosidase produced no change in its allergenicity, ruling out the possibility of the carbohydrate moiety being an important antigenic determinant. In contrast to the ragweed Ellergen antigen E, rye Groups I and I1 allergens are readily digested by trypsin and chymotrypsin. Proteolytic digestion of Group I allergen completely removed its antigenic and allergenic activities, nor did the digest show haptenlike inhibitory activity. Chemical modifications of the amino groups and the disulfide bridges of Group I allergen also led to loss of allergenic activity. Comparative immunological studies of rye pollen antigens with those from six other common grass pollens were made (Marsh and Haddad, 1968; Marsh et al., 1970a). Orchard (Dactylis glomerata), meadow fescue (Festuca elatior), meadow velvet (Holcus lanatus), and sweet vernal ( Anthoxanthum odoratum) pollens were found to contain groups of antigens that cross-react strongly, but to differing degrees, with rye Groups I, 11, and I11 antigens on immunodiffusion with the appropriate rabbit antisera. Timothy ( Phleum pratense ) pollen probably contains an antigen that is immunologically related to rye Group I antigen. The antigens of Bermuda (Cynodon dactylon) pollen are not related to the rye antigens. Skin tests in patients showed that the relative allergenic activities of Group I, 11, and I11 antigens of orchard, meadow fescue, meadow velvet, and sweet vernal follow closely the reactivity pattern found for rye Group I, 11, and I11 antigens.

C. TREE POLLENALLERGENS Hay fever due to tree pollens is believed to be less of a clinical problem than hay fever due to weed or grass pollens because of the shorter tree pollinating season. There are few reports on the characterization of tree pollen allergens. Pollen extracts from five different species of birch were examined by polyacrylamide gel isoelectrofocusing followed by RAST test with reaginic sera; the fractions located in the PI region of 4.5 to 5.5 were most active in the binding of IgE (Ceska et al., 1972b). The major allergen component of birch pollen (Betula verrucosa) has been partially purified (Belin, 1972). It is an acidic protein of molecular weight of about 20,000. Its allergenic activity depends on the intactness

PROPERTIES OF SOME ATOPIC ALLERGENS

87

of the protein structure, as glutaraldehyde treatment of birch allergen (Belin, 1972) or proteolytic digestion (Ceska and Ponterius, 1973) largely destroyed its activity. Inimunodifl'usion with rabbit antibirch sera showed cross-reaction of the major birch allergen with the alder (Alnus glutinosa) allergen. The allergen( s ) of alder pollen have been partly purified ( Herbeston et al., 1958).

D. M A h f M A L I A N DANDER ALLERCENS Animal danders are frequent elicitors of asthma and rhinitis. The incidence of positive skin reactions to cat and dog dander has been reported to be about 25%in an allergic population (Fontana et al., 1963). Studies have shown that all animal danders, cat dander (Varga, 1972; Ohman et aZ., 1973), cow dander (Ceska and Hultkn, 1972), dog dander (Varga, 1972; Glovsky et al., 1975), and horse dander (Squire, 1950; Ceska, 1972; Aaronson and Wide, 1974 ) are multiallergen systems. In each case the major dander allergen is found in the PI region of 4.3 to 5.5. The major allergens from horse dander (Stanworth, 1957a,b; Ponterius et d., 1973) and cat dander (Ohman et d., 1973, 1974; Brandt et al., 1973) have been purified and partially characterized chemically. The major allergen from both animal sources has a molecular weight of about 34,000 daltons. Persons who are sensitive to animal danders are also often sensitive to animal serum proteins. The presence of serum protein in extracts of horse dander (Stanworth, 1973; Ponterius et ul., 1973), cat dander (Ohman et al., 1973; Brandt et a/., 1973), and dog dander (Yman et aZ., 1973; Glovsky et nl., 1975) is well documented. Surveys of cat-sensitive patients by radioallergosorbent tests showed that about 66%of the serum samples were only positive with dander extracts, 27%were positive with both dander and serum proteins, and 7% were only positive with serum proteins (Brandt et nl., 1973). A similar finding was reported by Ohman et al. (1973); 5 of 12 subjects with skin reactivity to cat pelt extract also reacted strongly to cat serum. Squire (1950) showed that asthma patients sensitive to horse dander were also sensitive to horse serum proteins, and that among the serum proteins albumin was more allergenic than the globulin fraction. His skin prick test data on 2 asthmatic persons showed that albumin was about 0.2 as active as the whole dander extract in 1 patient and about 0.001 in the other. Sensitivity of horse dander allergic persons to horse serum albumin has also been reported by others (Fink et nl., 1966; Ponterius et al., 1973). Serum albumins of other mammalian species, dog and cat (Yman et al., 1973; Ohman et al., 1973, 1975), rat ( Frankland, 1974), and mice (Levy, 1975) have also been shown to be allergenic in man.

88

T. P. KING

E. MITEAND HOUSEDUSTALLERGENS Allergy to house dust is an important problem in many countries. House dust is a mixture of materials such as various kinds of fibers, molds, animal and human danders, and insect emanations. Studies by various workers have suggested that mites are an important source of allergen in house dust; these early studies have been reviewed by Gutman and Miyamoto (1972), Miyamoto ( 1974), and Ricci et al. ( 1974). Convincing evidence on the association of mite and house dust allergies in Japanese and Italian subjects have been presented by Miyamoto and Ricci. However, another study among American subjects (Kawai et al., 1972) failed to show such an association. More than thirty-six species of mites have been identified in Japanese house dust. The most common species of mite found in the Japanese and Italian house dust samples is Dermatophagoides pteronyssinus. Another species of mite found in house dust is Dermatophagoides farinae. At least one and probably two antigens common to D. pteronyssinus and D . farinae are present (Dasgupta and Cunliffe, 1970). The allergens from these two species of mites have been investigated by Miyamoto and Ricci and their colleagues. The allergenic activity was distributed in fractions in the molecular weight range of 10,000 to 60,000 daltons. The most active ones are acidic proteins with molecular weights of 20,000 to 30,000 daltons. The allergenic activity of mite extract was not altered on treatment with trypsin, chymotrypsin, papain, and pepsin.

F. FOOD ALLERGENS

1. Codfish Allergens Fish allergens occupy a special place in the history of allergy. The classic passive transfer experiment of reaginic serum was carried out by Prausnitz and Kustner (1921) with fish allergens. Like most food allergies, fish allergy is more common with children. In a study of 61 Norwegian chrildren with asthma and/or urticaria due to fish allergy (Aas, 1966), cod (Gadus callarias L. ) and haddock were the most common offenders of the eleven different species of fish tested. There are two main groups of fish white muscle proteins: ( 1 ) sarcoplasmic proteins and ( 2 ) myofibrillar proteins including actomyosin, myosin, and actin (Hamoir, 1955). All these proteins from codfish were allergenic on skin tests in sensitive persons, and the sarcoplasmic protein fraction was the most active ( Aas, 1967). A highly active allergen was isolated from the sarcoplasmic protein fraction ( Aas and Jebsen, 1967;

PROPERTIES OF SOME ATOPIC ALLERGENS

a9

Elsayed and Aas, 1971a). This purified codfish allergen is designated as allergen M. It gave positive skin reactions in 6 sensitive individuals at concentrations of about 10-9 gmlml. Codfish allergen M is a protein of molecular weight of 12,328 daltons with PI of 4.75. It contains 113 amino acid residues and 1 glucose residue. Its amino acid sequence has been established (Elsayed et al., 1974; Elsayed and Bennich, 1975). The single glucose residue is bound to the single cysteine residue at position 18 of the molecule. The molecule has no half-cystine residues and, therefore, no disulfide bridges. The antigenic and the allergenic activities of codfish allergen M were lost on extensive proteolytic digestions ( Aas and Elsayed, 1969). Selective tryptic cleavage of allergen M at the single arginine residue of position 75 was made following reversible blocking of the susceptible lysine residues (Elsayed et al., 1972). The two fragments, containing 75 and 58 amino acid residues, were both allergenically about one-tenth as aotive as the intact allergen M is. Treatment of allergen M with urea or guanidine hydrochloride did not alter its allergenic activity ( Elsayed and Aas, 1971b). This was taken to indicate that the allergenic determinants were dependent only on the linear sequence of the protein ( Elsayed and Bennich, 1975). Many proteins are reversibly denatured. This may also be the case with allergen M; therefore their results do not prove that the determinants of allergen M are independent of conformation of the molecule. Allergen M is a minor protein component of cod muscle relative to actomyosin, which accounts for about 60% Yet allergen M is a more potent allergen than actomyosin and, therefore it is a better immunogen. Allergen M has good thermal stability, whereas actomyosin is easily denatured under a variety of conditions (Hamoir, 1955). The difference in stability may account for the higher immunogenicity of allergen M, as fish is usually eaten after cooking. Allergen M belongs to a group of fish muscle sarcoplasmic proteins that have an undefined physiological role except for binding of calcium ion. These calcium-binding proteins are also known as parvalbumin. The structures of parvalbumin from whiting, hake, pike, and carp have been determined (cited by Elsayed and Bennich, 1975).

2. Bovine Milk Allergens Allergy to milk is a frequent problem in children. Bovine milk contains a number of proteins, namely, oaseias, p-laotoglobulins, lactalbumin, serum albumin, immunoglobulins, and various enzymes ( McKenzie, 1970, 1971 ) . Caseins, /&lactoglobulin,p-lactalbumin, and serum albumin are among the principal proteins and their concentrations in skim milk

90

T. P. KING

are respectively 24, 3, 0.7, and 0.3 mg/ml. The importance of these proteins in milk allergy was assessed by oral challenge (Goldman et d . , 1963a) and by skin tests (Goldinan et d.,1963b; Bleumink and Young, 1968). All sensitive persons reacted to one or more of these four antigens. Bleumink and Young reported that p-lactoglobulin is a more reactive allergen than the other three proteins. The presence of reaginic antibody in milk-sensitive individuals for p-lactoglobulin was also demonstrated by the chopped lung technique (Parish, 1967). Serum antibodies to bovine serum albumin are found more frequently among children (75%)than among young adults of 16 to 40 years of age (25%)or among older age groups (8%);incidence of antibodies to lactalbumin had the same distribution but only half the frequency of antibodies to serum albumin (Rothberg and Fan, 1965). The above-mentioned allergenic milk proteins are well characterized proteins with PIS in the range of pH P 6 . There are three different caseins, as,p, and K , in the molecular weight range of 18,000 to 24,000 daltons; in milk they occur as a colloidal complex with calcium phosphate. ,8-Lactoglobulin is a protein of two identical polypeptide chains, each with molecular weight of 18,363 daltons. a-Lactalbumin and serum albumin are proteins of a single polypeptide chain with molecular weights of 14,200 and 67,000 daltons, respectively. With the exception of K-casein, the amino acid sequences of different genetic variants of as and p caseins, p-lactoglobulins and a-lactalbumin (cf. Dayhoff, 1972), and that of bovine serum albumin (Brown, 1975) have been reported.

3. Chicken Egg White Allergens Egg-sensitive individuals are usually reactive only to the egg white. Chicken egg white contains several well-characterized proteins. Four of these proteins, ovalbumin, ovomucoid, ovomucin, and lysozyme, were shown to be allergenic when skin tested on egg-sensitive persons (Miller and Campbell, 1950; Bleumink and Young, 1969). Bleumink and Young indicated that ovomucoid was the most potent allergen of the four tested. Miller and Campbell had indicated that lysozyme was more reactive than ovomucoid, but this was later attributed to a nonspecific irritant effect ( Bleumink and Young, 1969; Stanworth, 1973). Allergenicity of egg white proteins was also determined by histamine release from human lung tissue sensitized with egg-allergic sera (Virtue and Wittig, 1970). The conclusion was that ovomucoid was much more active than ovalbumin and lysozyme, whereas conalbumin, another egg white protein, was apparently inactive. Ovomucoid is a minor component (48) of egg white proteins as compared to ovalbumin (658), conalbumin (178), and lysozyme (11%)

PROPERTIES OF SOME ATOPIC ALLERGENS

91

(Warner, 1954). Ovomucoid is a glycoprotein with molecular weight of about 27,000 daltons and pI of 3.9. Ovalbumin and lysozyme are proteins with molecular weights of 44,000 and 14,800, respectively. Ovalbumin is easily denaturable, but ovomucoid is highly stable. Exposure to 75"-100" heat resulted in no obvious change of ovomucoid, and this was the early procedure used for its isolation after heat coagulation of other egg proteins. These properties may play important roles in the apparent higher immunogenicity of ovomucoid.

G. HONEYBEE VENOMALLERCENS Allergy to insect stings, such as bees, yellow jackets, wasps, and hornets of the Hymenoptera order, is a common problem. Surveys have shown that about 0.4%of our population is allergic to bee stings (Chafee, 1970; Settipane, 1970). A number of studies on the purification of Hymenoptera venom allergens has lieen carried out, chiefly by Arbesman and his colleagues, and these studies have been reviewed ( Shulmaii, 1968). The venom components of honeybee ( Apis mellifera) have been extensively characterized by Habermann and his collaborators ( Habermann, 1972). Among the identified venom components are a hyaluronidase, a phospholipase A?, a hemolytic peptide (melittin) and a neurotoxic peptide (apamin), accounting, respectively, for 2, 12, 50, and 2% of the dry weight of venom. The complete amino acid sequences of phospholipase (Shipolini et al., 1974), melittin, and apamin are known and their respective molecular weights are 15,800, 2840, and 2038 daltons. Bee venom hyaluronidase has a molecular weight of about 50,000 daltons ( King et al., 1976). Both phospholipase and hyaluronidase are basic glycoproteins. The allergenic activities of the above four known components of bee venom have been tested by the leukocyte histainine release assay (King et al., 1976). For the majority of allergic persons tested ( 6 out of 7 ) , phospholipase is the most active, and hjduronidase is second; on a weight basis, their average activities are about 8 and 2 times those of the whole venom. It is appropriate to note that, on a molar basis, these two proteins are about equally active as allergens. In 1 of the 7 persons tested, a reverse order was found: Hvaluronidase and phospholipase were respectively 90 and 0.5 times as active as the venom; melittin was found to be a weak allergen, thc average being about 0.1 of the activity of the> venom, but apamin was not active. A detailed inspection of thc allergenic activity data showed that the relative activities of phospholipasc to hyaluronidase or melittin are different in different individuals. This variability indicates that allergcnicity is dependent on the nature of the allergens as well as on the individual responses to the allergen. Other

92

T. P. KING

reports have also indicated that phospholipase is the major bee venom allergen (Hoffman and Shipman, 1975; Ilea et al., 1975; Sobotka et al., 1975) and that melittin is a weak allergen (Mackler, 1972). Chemical modifications of bee venom phospholipase (King et al., 1976) showed that its allergenic determinants depend on the charge, the conformation, and the amino acid sequence of the molecule, just as has been indicated in the preceding sections on ragweed and ryegrass pollen allergens. The chemical modifications of phospholipase include succinylation of 8 of its 11 amino groups, cyanogen bromide cleavage of its three methionyl bonds, or reduction and carboxymethylation of its four disulfide bonds. All the derivatives showed reduced allergenic activities. The reduced and carboxymethylated enzyme showed a large decrease in its allergenic activity, about that of the native enzyme. The relative activity of the succinylated enzyme compared to that of the native enzyme covered a range of 0.003 to 0.7 in 5 allergic persons tested; the activity of the cyanogen bromide-cleaved enzyme also showed a wide range, The widely differing decreases in activities of these two modified enzymes in different persons tested is indicative of the fact that the test subjects recognize different antigenic determinants of the native enzyme; that is, the specificities of their IgEs differ so that they will show different activities with different modified enzymes. Melittin is immunogenic, but apamin is not. Melittin with a formula weight of 2840 is strongly associated in solution to form a complex with a molecular weight of about 12,000, but apamin with a formula weight of 2038 does not associate. This association property may be of importance as to why one is an immunogen and the other is not. IV. General Observations on Allergens

Under natural conditions of exposure, especially inhalants, atopic persons become sensitized on repeated challenges with extremely small doses of allergens. As an example, the estimated annual dosage of ragweed antigen E for a person is in the microgram range (Marsh, 1975). Prior to the development of RAST, the presence of allergen-specific IgEs could be detected only by biological tests because of their very low concentrations. The very high sensitivity of biological tests, coupled with the very low concentrations of allergens required for sensitization and of the specific IgEs formed, led a number of workers to consider the possibility of allergens being different from antigens. For example, Berrens (1974) has suggested that the allergenic activity of a molecule is dependent primarily on the proportion of N-glycosidic protein-sugar residues and their resulting decomposition products, and Stanworth

PROPERTIES OF SOME ATOPIC ALLERGENS

93

(1973) has stressed the importance of dimeric structure for protein allergens. However, chemical characterizations of allergens and immunological studies in the past decade have provided solid indic at'lolls that allergens are not different from antigens. Some of these studies will be reviewed in this section. The inhalants, foods, or insect venoms are all multiallergen systems as indicated in the preceding section. Each source has one or more major allergens to which the majority (>go%) of allergic persons are highly reactive and several minor allergens to which a smaller group (10-50%) of persons is also highly reactive. People show varying responses to the different allergens. This individual variation indicates that the allergenicity depends both on the chemical nature of allergen and on the response of the host. The major allergens from the inhalants and foods are all acidic proteins in the molecular weight range of 20,000 to 40,000 daltons with PIS in the range of 4 to 6. The minor allergens can be acidic or basic proteins (polypeptides), with molecular weights as large as 65,000 daltons for serum albumin and as small as 2800 for the bee venom allergen, melittin. The major allergens from bee venom, hyaluronidase and phospholipase, are exceptions, being basic proteins. However, the route of sensitization to bee venom is clearly different from that for inhalants or foods. All the known allergens from the inhalants and foods are globular proteins, as indicated by their chromatographic properties on porous gels and by their diffusion coefficients. Some allergens contain carbohydrates and some do not. Some allergens are proteins of a single polypeptide chain (ragweed antigens Ra3, 4, and 5; codfish allergen M; serum albumins) and some are proteins of two nonidentical polypeptide chains (ragweed antigen E ) or of two identical polypeptide chains (p-lactoglobulin). Some allergens (ragweed antigen E and mite allergen) in their native state are strongly resistant to proteolytic digestion, but others like grass and tree pollen allergens are readily digested. The amino acid compositions of most of the allergens described in Section I11 are known and they do not show any unusual chemical features. The complete amino acid sequences of several allergens are known: ragweed antigen Ra5, codfish allergen M, p-lactoglobulin, bovine serum albumin, and bee venom phospholipase and melittin. Simple inspection of these amino acid sequences does not reveal any obvious unusual structural features. Melittin is an exception, having an unusual distribution of the hydrophobic and hydrophilic amino acid residues. We cannot state unequivocally that allergens do not contain special features, since we do not know what combinations of amino acid residues participate in forming the antigenic determinants.

94

T. P. KING

The dominant antigenic determinants of the major allergens of ragweed and ryegrass pollens, codfish, and bee venom are dependent on both the primary structure and the conformation of the molecule, a property in common with other globular protein antigens. This was shown by the loss of their allergenic activities on denaturation, chemical modification, or proteolysis. Attempts to isolate haptenic peptides from proteolytic digests of modified ragweed antigen E or rye Group I antigen were unsuccessful. However, there are reports claiming the isolation of low molecular weight haptenic materials from timothy pollen ( Malley and Harris, 1967; Malley et nl., 1973) and from ragweed pollen ( Attalah and Sehon, 1969). Studies on the immunogenicities of the major allergens of ragweed and ryegrass pollens showed that the pollen allergens are no more potent than other protein antigens in eliciting IgG and IgE responses in mice ( Chang and Marsh, 1974). In addition to the “natural” allergens described in Section 111, many other diverse proteins are known to stimulate IgE responses in atopic or normal man, for example, bovine pancreatic ribonuclease ( Salvaggio et d.,1964), diphtheria and tetanus toxoids (Kuhns, 1962), insulins (Berson and Yallow, 1963), keyhole limpet hemocyanin ( Salvaggio et al., 1969), and the laundry detergent additives, proteases and amylases from Bacillus subtilis (Pepys et al., 1969). The biological functions of some of these protein allergens are known, and they are quite varied including enzymes, hormones, and blood ‘transport proteins. The widely different biological functions as well as origins of these protein allergens also argue for the unlikeliness that they may have certain structural features particularly suited for the induction of IgE production, such as the presence of hydrophobic surfaces of the molecule for attachment to immunocompetent cells. Most allergens are absorbed on inhalation or ingestion, and the allergens have to be absorbed through the respiratory or gastrointestinal tracts. These are the very sites where the IgE-forming plasma cells are known to be predominantly located (Tada and Ishizaka, 1970). Obviously there will be an upper size limit as to the solutes that can be rapidly absorbed through the mucosal membranes. The size of albumin molecule of about 65,000 daltons may well represent the upper limit of permeability of mucosal membrane, as serum albumins are the largest known allergens. There is one known exception to this size limit, namely keyhole limpet hemocyanin, an associating protein with estimated molecular weight of about 1 X lo6 daltons at pH 8.6 (Barte and Campbell, 1959). Intranasal immunization of normal or atopic persons with this protein as an aerosol led to the production of specific IgE ( Salvaggio et nl., 1969). Hemocyanins are known to be composed of associating subunits in the

PROPERTIES OF SOME ATOPIC ALLERGENS

95

size range of about 30,000 daltons (Pickett et al., 1966), so it is possible that the dissociated subunits are being absorbed. Studies have shown that atopic persons do not differ from normal persons by having greater permeability of the nasal mucosal membrane ( Kontou-Karakitsos et al., 1975). The immunogenic potency of a protein is dependent on its molecular complexity, its degree of foreignness to the host ( Landsteiner, 1936; Crumpton, 1974), and the genetic makeup of the immunized host (Benacerraf and McDevitt, 1972). Complex proteins with a large number of antigenic determinants are more likely to meet the genetic requirements of the immunized host than small proteins or peptides with fewer antigenic determinants. For example, keyhole limpet hemocyanin is a larger and more foreign protein to man than bovine pancreatic ribonuclease, and hemocyanin induced specific IgE formation in atopic and normal persons more frequently and readily than ribonuclease did (Salvaggio et al., 1964,1969); with either immunogen a greater percentage of atopic persons developed specific IgE than the percentage of normal persons. This combination of permeability and immunogenicity requirements may very well be the explanation as to why proteins of molecular weights of 20,000 to 40,000 are better allergens than proteins of other sizes. They are not too large, so that they cannot be readily absorbed through the mucosal membranes of atopic individuals. Yet they are not too small so that they can be sufficiently complex to be good immunogens. Studies on the effects of inbred strain and H-2 type, antigen, and antigen dose on immune responsiveness in the mouse (Levine and Vaz, 1970; Vaz and Levine, 1970) have provided several findings applicable to human atopy. They suggest the following: ( a ) Atopic persons differ from normal persons in their ease to respond immunologically to minute doses of antigen; ( 1 ) ) low-dose immunization favors persistent and boosterable IgE production or IgG production, and the reverse is true for high-dose immunization; and ( c ) both IgE and IgG responses to specific antigens are under the control of histocompatibility linked immune response genes. The last area has prompted a number of studies to establish the association of a specific HL-A type and IgE and IgG responses to a specific allergen. To date such studies have been carried out with the following purified allergens: ragweed antigen E (Levine et al., 1972; Blumenthal et al., 1974; Bias and Marsh, 1975), ragweed antigen Ra5 (Marsh et al., 1973) and rye Group I antigen ( Marsh, 1974). The presence of an IgE-regulating gene in man has been postulated (Hamburger et al., 1973; Marsh et al., 1974), and this gene in some way regulates the serum level of IgE. The IgE-regulating gene is believed to

96

T. P. KING

play an important role together with the immune response gene in the expression of atopic allergy in man ( Marsh et al., 1974). V. Uses of Purified Allergens

One important use for purified allergens is in the standardization of allergen extracts. These extracts are important reagents for the diagnosis and treatment of a number of allergic disorders. Since the beginning of clinical allergy practice, extracts have been standardized on the basis of their protein nitrogen contents, and they are not sbandardized on the basis of their biological activities. It is not uncommon to have two commercial extracts having the same protein nitrogen content but differing markedly in their biological activities (Baer et al., 1970). The difficulty in standardizing allergen extracts on the basis of their biological activities lies in the variability of individual responses to allergens. This problem can be partly overcome by doing RAST or RAST inhibition with a reference pool of reaginic sera. In cases where the major allergens( s ) are known and well-characterized, a simpler approach will be to standardize the extracts on the basis of their allergen contents. The allergen contents can be established by a number of simple methods independent of human reaginic sera. Such methods may include various forms of simple or electroimmunodiffusion using animal antisera specific for the allergens or, in cases where known, enzyme assays. Two other related uses of purified #allergensare as reagents for studying the immunological mechanism of immunotherapy and for the development of chemically modified derivatives better sui'ted for immunotherapy. Immunotherapy is one well-accepted form of treatment of allergic diseases since its iniitial description by Noon in 1911. The effectiveness of properly manlaged immunotherapy for hay fever has been reported in several studies, but symptom relief is not complete for most patients (Norman and Lichtenstein, 1971; Lichtenstein et al., 1974). On immunotherapy there is usually an initial rise in the serum level of allergenspecific IgE followed by a slow decrease, but the serum level of specific IgG continues to rise during therapy, usually reaching a plateau. The best clinical improvement has been associated with high doses of allergen extracts, high serum levels of allergen-specific IgG, and lowered levels of specific IgE, but it is not clear whether the clinical improvement is a consequence of changes in IgG, IgE, or both. Another serological change for hay fever patients is that immunotherapy prevents the seasonal rise in specific IgE following each pollinating season (reviewed by Irons et al., 1975). The immunological mechanism of these changes in IgG and IgE levels on immunotherapy is under study in animal models by several

PROPERTIES OF SOME ATOPIC ALLERGENS

97

laboratories. Some workers have suggested these changes to be a form of feedback suppression of IgE antibody production by IgG antibody (Tada and Okumura, 1971). Others have indicated these changes to be a form of immunological tolerance specific for the IgE class antibody (Ishizaka and Ishizaka, 1973; Gleich and Yunginger, 1974). Whatever the mechanism may be, the results of immunotherapy indicate that the best symptom relief is obtained on treatment with high doses of allergens. The only difficulty is that many persons cannot tolerate large doses of allergens because of allergic reactions. Investigators have sought to circumvent this problem by using adjuvants to give slow release of allergens into tissues and to enhance their immunogenicity. The repository treatment using allergen extracts incorporated as a mineral oil emulsion (Loveless, 1957) was used, but this practice was discontinued because of undesirable side effects, Alum-precipitated allergen extract is also used, but the amount that can be used safely is not significantly greater than that for the aqueous allergen extract ( Norman et al., 1972). For ragweed pollen extract, the mean total amount of pollen proteins to be injected into a person after a season’s treatment is on the order of 1.5 mg (or 25,000 PNU) ; less commonly, the total dose may be as high as 15 mg of pollen proteins for some individuals ( Melam et al., 1971; Yunginger and Gleich, 1973). Another approach, which has been tried (Stull et al., 1940; Fuchs and Strauss, 1959; Naterman, 1957, 1965), is to use chemically modified allergens. Marsh et al. (1970b) had proposed the term allergoids for modified allergens having reduced allergenic activities but retaining the immunogenic properties of the native molecule. The desired immunogenic property of the modified allergen may involve not only the induction of allergen-specific IgG antibody production but also IgE class-specific tolerance. However, this may not be the only important property, as it is indicated in the preceding paragraph that immunotherapy may be a form of IgE class-specific tolerance. The term allergoid was presumably chosen by analogy to the reduction of toxicity of a toxin to yield immunogenic tosoid. I t is appropriate to note here the difference in the toxic actions of allergens and toxins. The action of a toxin such as diphtheria or cholera depends on two welldefined reactive sites, one having enzymatic activity necessary for toxicity and the other for interaction with cell receptor (Gill et al., 1973; van Heyningen, 1974). Chemical modification of the single enzymatically active site of a toxin is sufficient for its conversion to toxoid. An allergen with molecular weight of about 40,000 daltons will have at least four or five antigenic determinants, by analogy to other known protein antigens, and its toxic action depends on the interaction of at least two of these antigenic determinants with cell-bound IgE ( Ishizaka, 1973 ) . Therefore,

98

T. P. KING

for the conversion of an allergen to an allergoid, chemical modifications of different combinations of antigenic determinants of an allergen are required so that the resulting allergoid is a mixture of derivatives. In the idealized case, each derivabive retains one of the determinants of the native allergen, and the derivatives together retain all the specificities of the native allergen. Marsh prepared allergoids of rye Group I antigen by formaldehyde treatment under conditions similar to those used for preparing bacterial toxoids (Marsh et al., 1970b). The product so prepared was a mixture of components of different molecular sizes, and the major components (about 70%) were dimers and trimers of the rye Group I antigen (Marsh, 1971) , The chemical reactions involved mainly inter- and intramolecular cross-linking of the €-amino groups of the protein through methylene bridges. The allergenic activity of the resulting allergoid was 10'-109 times less than that of the native antigen, and the large range of activity decrease is a reflection of the variability of allergic individuals in recognizing different antigenic determinants. The rye Group Iallergoid was immunogenic in normal persons in inducing IgG production specific for the native allergen. Following the pilot studies, allergoids of whole ragweed and grass pollen extracts were also prepared, and these allergoids were 102-104 times less allergenic than the unmodified extracts. Controlled studies on immunotherapies with the whole ragweed extract and the ragweed allergoid indicated that the allergoid was somewhat more effective clinically that the extract. The average dose of ragweed allergoid used was about 40 times larger than that of the extract, and the higher dose of allergoid was more effective than the limited dose of extract in inducing allergen-specific IgG production. However, complete symptom relief was still not attained ( Norman et al., 1975). Patterson and co-workers ( 1973a,b) have modified ragweed antigen E by treatment with glutaraldehyde. The resulting cross-linked preparation was separated into high and low molecular weight fractions with apparent molecular weight ranges of 0.2 to 4 million and of 4 to 20 million daltons, respectively. There was a greater reduction of the allergenic activity of the high molecular weight fraction than of the low molecular weight fraction in accord with their molecular size differences. Both fractions were as immunogenic in rabbits as the native antigen in inducing specific IgG antibody production. Interestingly, they differed in their induction of IgE antibody production: The native antigen produced a tnansient IgE response; the low molecular weight fraction gave an IgE response weaker than the native antigen, but the high molecular weight fradion gave a stronger and longer lasting IgE response than that of the native antigen ( Patterson and Suszko, 1974 ).

PROPERTIES OF SOME ATOPIC ALLERGENS

99

The allergenicity of a molecule can also be reduced by attaching to it large molecular weight groups. The bulky large groups can prevent or reduce the accessibility of antigenic determinants for combination with antibodies, as illustrated with the following two derivatives of ragweed antigen E. Antigen E attached to 2-5 molecules of devtran of molecular weight of about 20,000 was prepared by reductive coupling with periodate-oxidized dextran (King et al., 1975). Also, antigen E attached to about 8 molecules of methovypolyethylene glycol, of molecular weight of about 2,000, was prepared by reaction with a monochlorotriazine derivative of methoxypolyethylene glycol ( T. P. King, unpublished work; Abuchowski et al., 1974). On a molar basis, the allergenic activity of the dextran and the methoxJplyethylene glycol derivatives of antigen E were, respectively, about one-eighth and one-fiftieth of the native antigen. The methoxypolyethylene glycol antigen E did not form immune precipitates with sheep anttiantigen E sera, whereas the dextran-antigen E still did. The polyethylene glycol-antigen E has all the antigenic determinants of antigen E as shown by its inhibition of hemagglutination of antigen E-coated cells with anti-E sera, and, on a molar basis, its inhibitory activity was one-fortieth of the native antigen. Rabbits immunized wicth the polyethylene glycol derivative gave a lower antibody titer (about one-sixth ) than those with the native antigen. Although the polyethylene glycoliantigen E was less immunogenic than antigen E, it was just as effective as antigen E in priming the rabbiits for a secondlary response. The immunogenic property of the polyethylene glycol-antigen E for IgE response has yet to be studied. The modifications described in this paragraph differ from the formaldehyde and glutanaldehyde treatments in one significant aspect, namely that the dextran, or the polyethylene glycol, derivative is more homogeneous in size than are the formaldehydeor glutaraldehyde-treated allergens and that it does not contain large molecular weight aggregates. This chemical difference of these derivatives may be of importance for their capacities to alter IgE responses in man or animal, and this is indicated by the work of Patterson and Suszko ( 1974 ) on the immunogenic property of glutaraldehyde-treated antigen E. If the clinical effcct of immunotherapy is a result of depression of the IgE antibody response, then denatured or extensively modified allergens can be useful therapeutic reagents. This suggestion results from the following observations: As indicated in Section 111, the urea-denatured antigen E and its dissociated polypeptide chains are essentially inactive as allergens, that is, they have less than 0.0001 of the originla1 activity. Studies in mice indicated that denatured and native antigen E do not share common B-cell determinants, but they do share common T-cell determinants (Ishizaka et al., 1974, 1975). Mice immunized with the

100

T. P. KING

denatured antigen E did not produce IgG or IgE specific for the native antigen, but the denatured antigen E primed the animals for a secondary response to the native antigen E. More interestingly, repeated injections of the denatured antigen E into antigen E-primed mice depressed the ongoing IgE antibody response, whereas injections of the same dose of native antigen E showed little or no depression. The animals treated with the denatured antigen showed only a small rise in their IgG antibody specific for the native antigen E, whereas those treated with the native antigen E showed a very large rise. These findings on the treatment of mice with the denatured antigen E are believed to result from a suppression of the helper T-cell function as well as from decreases of antigen E-specific IgG and IgE B-cell populations ( Takatsu et al., 1976). These observations on the denatured antigen E are of general applicability and they may prove to be of practical importance. Another example on the use of chemically modified protein antigen to depress an ongoing IgE response specific for the native antigen in the mouse model has been reported (Bach and Brashler, 1975). The protein antigen studied was ovalbumin, and chemical modifications of the lysyl, the arginyl, and tryptophan$ residues were carried out. I t was found that the primary structural requirement for efficacy of the modified antigen was reduced antigenicity to prevent anaphylaxis. Since the antigenicity of globular proteins depends primarily on the conformation of the molecule, the findings can be taken to indicate that the more altered the antigen is conformationally, the better it is for suppression of the IgE response. VI. Concluding Remarks

Many different proteins from animal, plant, or microbial sources can be allergens in man. The known allergens are usually in the molecular size range of 20,000 to 40,000 daltons, and they do not have any distinguishing chemical features. The purified allergens are useful reagents as they can help us to understand better the heightened responsiveness of susceptible persons to minute doses of immunizing allergens for IgE and IgG productions. Continued chemical and immunological studies of the purified allergens will indicate the types of chemical modification required to prepare derivatives useful for altering the immune responsiveness of allergic man. The findings will be of general applicability, as most environmental allergens are proteins. ACKNOWEDGMENTS I would like to thank Dr. David Marsh for sending me a copy of his review article on allergens prior to its publication. My own studies on allergens have been supported in part by a grant from the National Institutes of Health (AI-08445).

PROPERTIES OF SOME ATOPIC ALLERGENS

101

REFERENCES Aaronson, T., and Wide, L. (1974). Int. Arch. Allergy Appl. Immunol. 47, 224. Aas, K. ( 1966). Int. Arch. Allergy Appl. Immunol. 30, 257. Aas, K. (1967). Int. Arch. Allergy Appl. Immunol. 31, 239. Am, K., and Elsayed, S. M. (1969). J . Allergy 44, 111. Aas, K., and Jebsen, J. W. (1967). Int. Arch. Allergy Appl. Immunol. 32, 1. Abuchowski, A., van Es, T., Palnuk, N. C., and Davis, F. F. ( 1974). Fed. Proc., Fed. Am. SOC. Exp. Biol. 33, 1317. Attallah, N. A., and Sehon, A. H. (1969). Immunochemistry 6, 609. Augustin, R., and Hayward, B. J. (1962). Immunology 5, 424. Austen, K. F. (1974). Allergol., Proc. Int. Congr., 8th, 1973 pp. 306-315. Bach, M. K., and Brashler, J. R. (1975). J . Immunol. 114, 1799. Baer, H., Codfrey, H., Maloney, C. J., Norman, P. S., and Lichtenstein, L. M. ( 1970). J . Allergy 45, 347. Barte, A. H., and Campbell, D. H. (1959). Arch. Biochem. Biophys. 82, 232. Belin, L. (1972). Int. Arch. Allergy Appl. Zmmunol. 42, 329. Benacerraf, B., and McDevitt, H. 0. ( 1972). Science 175, 273. Bennich, H., and Johansson, S. G. 0. (1971). Adu. Immunol. 13, 1. Berrens, L. (1971). “The Chemistry of Atopic Allergens.” Karger, Basel. Berrens, L. (1974). Ann. N.Y. Acad. Sci. 221, 183. Berson, S. A., and Yallow, R. S. ( 1963). Science 139, 844. Bias, W. A., and Marsh, D. G. (1975). Science 188,375. Bleumink, E., and Young, E. (1968). Int. Arch. Allergy Appl. Immunol. 34, 521. Bleumink, E., and Young, E. (1969). Int. Arch. Allergy Appl. Immunol. 35, 1. Blumenthal, M. N., Amos, D. B., Noreen, H., Mendell, N. R., and Yunis, E. J. (1974). Science 184, 1301. Brandt, R., Ponterius, G., and Yman, L. (1973). Int. Arch. Allergy Appl. Immunol. 45, 447. Brown, J. R. (1975). Fed. Proc., Fed. Am. SOC. Exp. B i d . 34, 591. Callaghan, 0. H, and Coldfarb, A. R. (1962). J. Immunol. 89, 612. Cebra, J. J. ( 1957). J . Immunol. 79, 118. Ceska, M. (1972). Int. Arch. Allergy Appl. Immunol. 43, 419. Ceska, M., and HultBn, E. (1972). Int. Arch. Allergy Appl. Immunol. 43, 427. Ceska, M., and Ponterius, C. (1973). Int. Arch. Allergy Appl. Immunol. 45, 808. Ceska, M., Eriksson, R., and Varga, J. M. (1972a). J. Allergy Clin. Immunol. 49, 1. Ceska, M., Eriksson, R., and Varga, J. M. (197213). Int. Arch. Allergy Appl. Immunol. 42, 430, Chafee, F. H. (1970). Acta Allergol. 25, 292. Chang, E. B., and Marsh, D. C. (1974). J. Allergy Clin. Imrnunol. 53, 65 (abstr.) Cooke, R. A. ( 1974). “Allergy in Theory and Practice.” Saunders, Philadelphia, Pennsylvania. Crumpton, M. J. (1974). In “The Antigens” (M. Sela, ed.), Vol. 2, pp. 1-72. Academic Press, New York. Dasgupta, A,, and Cunliffe, A. C. (1970). Clin. Exp. Immunol. 6, 891. Dayhoff, M. 0. ed. (1972). “Atlas of Protein Sequence and Structure,” Vol. 5. Nat. Biomed. Res. Found., Washington, D.C. De Weck, A. L. (1974). AUergol., Proc. Int. Congr., 8th, 1973 pp. 347-355. Elsayed, S., and Aas, K. (1971a). Int. Arch. Allergy Appl. Immunol. 40, 428. Elsayed, S., and Aas, K. (197113). J . Allergy 47,283. Elsayed, S , and Bennich, H. (1975). Scand. J. Immunol. 4, 203. Elsayed, S., Aas, K., Sletten, K., and Johansson, S. (1972). lmmunochemistry 9, 647.

102

T. P. KIXG

Elsayed, S., von Bahr-Lindstriini, H., and Bennich, H. ( 1971). Scond. J. Zmtnunol. 3, 683. Fink, J. N., Patterson, R., and Pruzansky, J. J. (1966). J. Allergy 38, 81. Fontana, B. J., Wittig, H., and &It, L. E, Jr. (1963). J. Allergy 34, 348. Foucard, T., Johansson, S. G. O., Bennich, H., and Berg, T. (1972). Znt. Arch. Allergy Appl. Zmmrinol. 43, 360. Frankland, A. W. (1974). Allergol., Proc. I n t . Cotigr., 8th, 1973 p. 123. Fuchs, A. M., and Strauss, hl. B. ( 1959). J. Allergy 30, GG. Cell, P. G. H., and Benacerraf, B. (1961). Ado. Immtcnol. 1, 319. Gill, D. hl., Pappenheimer, A. M., Jr., and Uchida, T. (1973). Fed. Proc., Fed. A m . SOC. Exp. Biol. 32, 1508. Gleich, G. J., and Jacob, G. L. (1975). Science 190, 1106. Gleich, G. J., and Yunginger, J. W. (1974). Allergol., Proc. Znt. Congr., 8th, 1973 pp. 44-53. Gleich, G. J., Larson, L. N., Jones, R. T., and Baer, R. ( 1974). J. Allergy Clin. lmniunol. 53, 158. Glovsky, hf., Ghekiere, L., Roth, R., and Caniplicll, D. (1975). 1. Allergy Clin. Ztnmtrnol. 55, 76 (abstr.). Goldfarb, A. R. ( 1968). J. Immtrnol. 100, 902. Goldfarb, A. R., Bhattacharya, A. K., and Koerner, S. K. (1958). J. Zmmrrnol. 81, 302. Goldnian, A. S., Anderson, D. W., Sellars, W. A., Saperstcin, S., Kniker, W. T., and Halpern, S. R. ( 1 9 6 3 ~ )Pediutrics . 32, 425. Goldnian, A. S., Sellars, W. A., Halpem, S. R., Anderson, D. W., Furlow, T. E., and Johnson, C. €1. (1963b). Pediatrics 32, 572. Goodfriend, L., and Lapkoff, C. (1974). Int. Arch. Allergy Appl. lmniunol. 46, 215. Goodman, J. W. ( 1975). In “The Antigens’’ (11. Sela, etl.), 1’01. 3, p. 127. Academic Press, New York. Griffiths, B. W. (1972). J. Chroniotogr. 69, 391. Griffiths, B. W. (1973). Can. J. Biochem. 51, 1275. Griffiths, B. W., and Brunet, R. (1971). Can. J. Biochem. 49, 396. Gutman, A. A., and hliyanioto, T. ( 1972). In “Allergic Diseases” ( R . Patterson, cd. ), pp. 87-123. Lippincott, Philadelphia, Pennsylvania. Haberniann, E. (1972). Sciencc 177, 314. Hanibrirger, R. N., Orgel, H. A., and Bazaral, hl. (1973). I n “hlechanisms in Allergy” ( L . Goodfriend, A. H. Sehon, and R. P. Orange, eds.), pp. 131-139. Dekker, New York. Hanioir, G. (1955). Ado. Protein Chcm. 10, 227. Herbeston, S., Porath, J., and Colldahl, H. (1958). Acto Chctn. Scond. 12, 737. Hoffman, D. R., and Shipman, W. H. (1975). J. Allergy Clin. Ztnmrrnol. 55, 74. Ilea, V., Okazaki, T., Wypek, J. I., Reisnian, R. E., and Arl)esrnan, C. E. ( 1975). J . Allergy Clin. Znitntrtiol. 55, 71 ( a h t r . ) . Irons, J. S., Pnizansky, J. J., and Patterson, R. (1975). J. Allergy Clin. Immrrtiol. 56, 64. Ishizaka, K. (1973). In “The Antigens” ( A l . Sela, e d . ) , Vol. I, pp. 479-528. Academic Press, New York. Ishizaka, K., and Ishizaka, T. ( 1973 ). In “Asthma: Physiology, Immimopharmacology, and Treatment” ( K . F. Austm and L. 51. Lichtcnstcin, c d s . ) , pp. 5568. Acadcniic Press, New York. Ishizaka, K., Kishinioto, T., Delc.spcssc, G., and King, T. P. ( 1074). J. I r t t t w d 113, 70.

PROPERTIES OF SOME ATOPIC ALLERGENS

103

Ishizaka, K., Okudaira, H., and King, T. P. (1975). J. Zmmunol. 114, 110. Johnson, P., and Marsh, D. G. (1965). Eur. Polym. J. 1, 63. Johnson, P., and Marsh, D. G. (1966a). Immunochemistry 3, 91. Johnson, P., and Marsh, D. G. (1966b). Immunochemistry 3, 101. Kabat, E. A., Turino, C. M., Tarrow, A. B., and Maurer, P. H. (1957). J. Clin. Inoest. 36, 1160. Katz, D. H., and Benacerraf, B. (1972). Adu. Zmmrcnol. 15, 1. Kawai, T., Marsh, D. G., Lichtenstein, L. M., and Norman, P. S. (1972). J. Allergy Clin. Immunol. 50, 117. King, T. P. (1972). Biochemistry 11, 367. King, T. P., and Norman, P. S. (1962). Biochemistry 1, 709. King, T. P., Norman, P. S., and Connell, J. T. (1964). Biochemistry 3, 458. King, T. P., Norman, P. S., and Lichtenstein, L. M. (1967a). Biochemistry 6, 1992. King, T. P., Norman, P. S., and Lichtenstein, L. M. (1967b). Ann. Allergy 25, 541. King, T. P., Norman, P. S., and Tao, N. (1974). Immunochemistry 11, 83. King, T. P., Kochoumian, L., Ishizaka, K., Lichtenstein, L. M., and Norman, P. S. (1975). Arch. Biochem. Biophys. 169, 464. King, T. P., Sobotka, A. K., Kochoumian, L., and Lichtenstein, L. M. (1976). Arch. Biochem. Biophys. 172, 661. Kontou-Karakitsos, K., Salvaggio, J. E., and Mathews, K. P. (1975). J . Allergy Clin. Zmmunol. 55, 241. Kuhns, W . J. (1962). J. Zmmunol. 89, 652. Landsteiner, K. ( 1936). “The Specificity of Serological Reactions” (rev. ed. Dover Publications, New York, 1962 ). Leskowitz, S., and Lowell, F. C. (1961). J . Allergy 32, 152. Levine, B. B. (1965). J . Zmmunol. 94, 111. Levine, B. B. (1966). N . Engl. J. Med. 275, 1115. Levine, B. B., and Vaz, N. M. (1970). Int. Arch. Allergy Appl. Immunol. 39, 156. Levine, B. B., Stember, R. H., and Jotino, M. (1972). Science 178, 1201. Levy, D. A. ( 1975). Int. Arch. Allergy Appl. Immunol. 49,219. Lichtenstein, L. M. (1974). Allergol., Proc. Int. Congr., 8th, 1973 pp. 294-305. Lichtenstein, L. M., King, T. P., and Osler, A. G. (1966). J . Allergy 38, 174. Lichtenstein, L. M., Marsh, D. G., and Campbell, D. H. (1969). J . Allergy 44, 307. Lichtenstein, L. M., Roebber, M., and Goodfriend, L. (1973). J . Allergy Clin. Immunol. 51, 285. Lichtenstein, L. M., Norman, P. S., and Ishizaka, K. (1974). Allergol., Proc. Int. Congr., 8th, 1973 pp. 61-74. Loveless, M. H. ( 1957). J. Immunol. 79, 68. LZwenstein, H., Weeke, B., and Nielsen, L. (1974). Acta Allergol. 29, 418. McKenzie, H. A., ed. ( 1970). “Milk Proteins: Chemistry and Molecular Biology,” Vol. 1. Academic Press, New York. McKenzie, H. A., ed. ( 1971). “Milk Proteins: Chemistry and Molecular Biology,” Vol. 2. Academic Prew, New York. Mackler, B. F. (1972). Clin. Allergy 2, 317. Malley, A,, and Harris, R. L., Jr. ( 1967). 1. Zmmunol. 99, 825. Malley, A., Crossley, G., Baecher, L., Wilson, B. J., Perlman, F., and Burger, D. ( 1973). I n “Mechanisms in Allergy” ( L . Goodfriend, A. H. Sehon, and R. P. Orange, eds.), pp. 83-96. Dekker, New York. Marsh, D. G. ( 1971). Int. Arch. Allergy Appl. Zmmunol. 41, 199. Marsh, D. G. (1974). Allergol., Proc. Znt. Congr., 8th, 1973 pp. 381-393.

104

T. P. KING

Marsh, D. G. (1975). In “The Antigens” (M. Sela, ed.), Vol. 3, pp. 271459. Academic Press, New York. Marsh, D. G., and Haddad, 2. H. (1968). Fed. Proc., Fed. Am. SOC. Erp. Biol. 27, 368. Marsh, D. G., Haddad, Z. H., and Campbell, D. H. (1970a). J . Allergy 46, 107. Marsh, D. G., Lichtenstein, L. M., and Campbell, D. H. (1970b). Immunology 18, 705. Marsh, D. G., Bias, W. B., Hsu, S. H., and Goodfriend, L. (1973). Science 179, 691. Marsh, D. G., Bias, W. B., and Ishizaka, K. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 3588. Melam, H., Pruzansky, J., Patterson, R., and Zinger, S. (1971). J . Allergy 47, 262. Miller, H., and Campbell, D. H. (1950). 1. Allergy 21, 522. Miyamoto, T. (1974). Allergol., Proc. Int. Congr., 8th, 1973 pp. 403-410. Mole, L. E., Goodfriend, L. Lapkoff, C. B., Kehoe, J. M., and Capra, J. D. (1975). Biochemistry 14,1216. Nakashima, S., Sugimura, K., Sawada, T., and Mazaki, M. (1974). J. Biochem. (Tokyo) 76, 349. Naterman, H. L. (1957). J. Allergy 28,76. Naterman, H. L. ( 1965). J . Allergy 36,226. Noon, L. (1911). Lancet 1, 1572. Norman, P. S. (1971). In “Immunological Diseases” ( M . Samter, ed.), 2nd ed., pp. 775-783. Little, Brown, Boston, Massachusetts. Norman, P. S., and Lichtenstein, L. M. (1971). In “Immunological Diseases” (M. Samter, ed.), 2nd ed., pp. 840-858. Little, Brown, Boston, Massachusetts. Norman, P. S., Winkenwerder, W. L., and Lichtenstein, L. M. (1971). J. Allergy 47, 273. Norman, P. S., Winkenwerder, W. L., and Lichtenstein, L. M. (1972). J. Allergy Clin. Immunol. 50, 31. Norman, P. S., Lichtenstein, L. M., and Ishizaka, K. (1973). J. Allergy Clin. Immunol. 52, 210. Norman, P. S., Marsh, D. G., Lichtenstein, L. M., and Ishizaka, K. (1975). J. Allergy Clin. Immunol. 55, 78. Ohman, J. L., Jr., Lowell, F. C., and Bloch, K. J. ( 1973). J . Allergy Clin. I m m u d . 52, 231. Ohman, J. L., Jr., Lowell, F. C., and Bloch, K. J. (1974). J . Immunol. 113, 1668. Ohman, J. L., Jr., Lowell, F. C., Bloch, K. J., and Kendall, S. ( 1975). J . Allergy Clin. Immunol. 55, 77 (abstr). Osler, A. G., Lichtenstein, L. M., and Levy, D. (1968). Ado. Immunol. 8, 183. Parish, W. E. (1967). Nature (London) 215,738. Patterson, R., and Suszko, I. M. (1974). J . Immunol. 112, 1855. Patterson, R. Suszko, I. M., and McIntyre, F. C. (1973a). J. Immunol. 110, 1402. Patterson, R., Suszko, I. M., Pruzansky, J. J., and Zeiss, C. R. ( 1973b). J . Immunol. 110, 1413. Pepys, J., Hargreaves, F. E., Longbottom, J. L., and Faux, J. (1969). Lancet 1, 1181. Pickett, S. M., Riggs, A. F., and Larimer, J. L. (1966). Science 151, 1005. Ponterius, G., Brandt, R., Hultbn, E., and Yman, L. (1973). Int. Arch. Allergy Appl. Immunol. 44, 679. Prausnitz, C., and Kustner, H. ( 1921). Zentralbl. Bakteriol., Pumsitenk. Infektionskr., Abt. 86, 160. Ricci, M., Romagnani, S., and Bilotti, G. (1974). Allergol., Proc. Int. Congr., 8th, 1973 pp. 411-416.

PROPERTIES OF SOME ATOPIC ALLERGENS

105

Richter, M., and Sehon, A. H. (1960). J. Allergy 31, 111. Robbins, K. C., Wu, H., and Hsieh, B. (1966). Immunochemistry 3, 71. Roebber, M. ( 1975). Ph.D. Thesis, McGill University, Montreal. Roebber, M., Marsh, D. G., and Goodfriend, L. (1975). J. Immunol. 115, 303. Rothberg, R. M., and Farr, R. S. ( 1965). Pediatrics 35, 571. Salvaggio, J. E., Cavanaugh, J. J. A., Lowell, F. C., and Leskowitz, S. (1964). J . Allergy 35, 62. Salvaggio, J. E., Castro-Murillo, E., and Kundur, V. ( 1969). 1. Allergy 44, 344. Santilli, J., Marsh, D. G., Bias, W., Schachter, B., and Goodfriend, L. (1975). J. Allergy Clin. Immunol. 55, 108. Settipane, G. A. (1970). Acta Allergol. 25, 286. Shipolini, R. A., Doonan, S., and Vernon, C. A. ( 1974). Eur. J. Biochem. 48, 477. Shulman, S. (1968). Prog. Alleigy 12,246. Sobotka, A., Franklin, R., Valentine, M., and Lichtenstein, L. M. (1976). J. Allergy Clin. Immunol. 55, 29. Squire, J. R. (1950). Clin. Sci. 9, 127. Stanworth, D. R. (1957a). Biochem. J. 65,582. Stanworth, D. R. (1957b). Int. Arch. Allergy Appl. Immunol. 11, 170. Stanworth, D. R. ( 1973). “Immediate Hypersensitivity.” North-Holland Publ., Amsterdam. Stull, A,, Cooke, R. A., Sherman, W. B., Hebald, S., and Hampton, S. F. (1940). J . Allergy 11, 439. Tada, T., and Ishizaka, K. (1970). J. Immunol. 104, 377. Tada, T., and Okumura, K. (1971). J. Immunol. 106, 1002. Takatsu, K., Ishizaka, K., and King, T. P. (1975). J. Immunol. 115, 1469. Underdown, B. J., and Goodfriend, L. (1969). Biochemistry 8, 980. Van Heyningen, J. M. ( 1974). Science 183, 656. Varga, J. M. (1972). Int. Arch. Allergy Appl. Immunol. 42, 438. Vaz, N. M., and Levine, B. B. (1970). Science 168,852. Virtue, C. M., and Wittig, H. J. (1970). Fed. Proc., Fed. Am. SOC. Erp. Biol. 29, p. 576 ( abstr. ). Warner, R. C. (1954). In “The Proteins” (H. Neurath and K. Bailey, eds.), 1st ed., Vol. 2A, pp. 435-484. Academic Press, New York. Weeke, B., and L@wenstein, H. (1975). Int. Arch. Allergy Appl. Immunol. 49, 74. Wide, L., Bennich, H., and Johansson, S. G. 0. (1967). Lancet 2, 1105. Wodehouse, R. P. (1954). Int. Arch. Allergy Appl. Immunol. 5, 425. Wodehouse, R. P. (1971). “Hayfever Plants,” 2nd ed. Hafner, New York. Yman, L. ( 1975). Int. Arch. Allergy A w l . Immunol. 49, 55. Yman, L., Brandt, R., and Ponterius, G. (1973). Int. Arch. Allergy Appl. Immunol. 44, 358. Yunginger, J. W., and Gleich, G. J. (1972). J. Allergy Clin. Immunol. 50, 326. Yunginger, J. W., and Gleich, G. J. ( 1973). J. Clin. Inoest. 52, 1268. Zeiss, L. R., Pruzansky, J. J., Patterson, R., and Roberts, M. (1973). J. Immunol 110, 414.

This Page Intentionally Left Blank

Human Mixed-Lymphocyte Culture Reaction: Genetics, Specificity, and Biological Implications BO DUPONT AND JOHN A. HANSEN Tissue Typing laboratory, Sloan-Kettering lnrtitute far Cancer Research, New York, New Yark

AND EDMOND J. YUNlS Department of Pathology and laboratory Medicine, University o f Minnesota Hospitals, Minneapolis, Minnesota

I. Introduction: Major Histocompatibility System in Man . . . . 11. Serology of Human Leukocyte Alloantigens (HLA-A,B,C) . . . A. Introduction . . . . . . . . . . . , B. One Genetic System of Several Closely Linked Loci . . . . C. Cross-reactivity . . . . . . . . . . . D. Genetic Linkage Disequilibrium . . . . . . . . 111. Cell-Mediated Allogeneic Reactions in Vitro. . . . . . . A. Mixed-Lymphocyte Culture Reaction , , . . . . . B. Induction of Cytotoxic Effector Cells in Mixed-Lymphocyte Culture C. Induction of Immunological Memory Cells in Mixed-Lymphocyte . . . . . . . . . . . . . Culture IV. Measurement of Antigenic Differences in Mixed-Lymphocyte Culture Reaction . . . . . . . . . . . . . . V. Single-Locus Concept for Mixed-Lymphocyte Culture Stimulation (HLA-D LOCUS) . . . . . . . . . . . . A. Family Studies . . . . . . . . . . . . B. Mixed-Lymphocyte Cultures between Unrelated Individuals . . VI. Mixed-Lymphocyte Culture ( HLA-D ) Specificities Defined by HLA-D. . . . . . . . . Homozygous Typing Cells . A. Families with Shared Parental Histocompatibility Haplotypes . . R . Identification of HLA-D-Homozygous Typing Cells . . . . C. Sources of HLA-D-Homozygous Typing Cells . . . . . D. Definition of Typing Responses . . . . . . . . E. Characterization of HLA-D Specificities . . . . . . F. Complexity of the HLA-D Locus: Cross-reacting Specificities versus Multiple Subloci . . . . . . . . . . . G . Family Studies with HLA-D-Homozygous Typing Cells . . . H. HLA-D Typing of Families with Recombinations within the HLA . . . . . . . . . . . . Complex . I. Population Studies with HLA-D-Homozygous Typing Cells . . J. Serological Identification of Alloantigens with Restricted Tissue . . . . . . . . . . . . Distribution K. Role of Lymphocyte Subpopulations in Mixed-Lymphocyte Culture . . . . . . . . . . . . . Reaction 107

108 110 110 114 116 117 119 119 120 123 124 130 130 132 135 136 138 140 142 146 148 154 154 162 164 168

108

DUPONT, HANSEN, AND YUNIS

VII. Genetic Control of Immune Response Related to Histocompatibility VIII. Mixed-Lymphocyte Culture As a Histocompatibility Test for Clinical . . . . . . . . . . . . Transplantation IX. Genetic Mapping of the HLA Complex on Chromosome C-6 . . . X. Conclusions . . . . . . . . . . . . . References . . . . . . . . . . . . .

169 177 183 185 187

I. Introduction: Maior Histocompatibility System in M a n

In all vertebrates studied thus far, the major histocompatibility complex ( M H C ) has been described as a genetic system of closely linked genes. The MHC controls alloantigens that are the predominant determinants in allograft or transplantation reactions. In addition to the control of individual diversity, this system also controls determinants of tissue diversity. Interest in the MHC goes beyond transplantation to include a number of biologically important functions. For example, genes linked to the MHC control immune responses to some specific antigens and production of some components of the complement system. The MHC in mouse ( H-2) controls certain factors influencing resistance or susceptibility to virus infections, autoimmune disease, and neoplastic disease. The MHC in man (HLA) is at present defined by at least four distinct loci (Fig. 1) : HLA-A, B, C, and D. Three of these loci code for alloantigens readily detectable by serological methods (HLA-A, B, and C ) . The fourth locus ( HLA-D ) controls lymphocyte responses in the in uitro mixed-lymphocyte culture reaction ( MLR) , Three important characteristics are shared by the genes or gene products of each locus: ( 1 ) each locus represents an allelic system with considerable polymorphism; ( 2) the determinants of each locus demonstrate cross-reactivity; and ( 3) genetic linkage disequilibrium exists among the alleles of the different loci. In addition to these four HLA loci, several different genetic traits have been linked to HLA, although the actual mapping of bhe genes in relation to HLA is still largely unknown (Fig. 1 ) . This review will summarize the major achievements in the clarification of the genetic control and polymorphism of the serologically deteotable HLA-A, B, and C determinants. The emphasis will, however, be placed on the recent developments in the identification of the HLA-D locus and HLA-D determinants. The mixture of lymphocytes from two different individuals usually results in a blastogenic response with cell proliferation. During the MLR, a series of events, in addition to cell proliferation, takes place, including the production and secretion of biologically active molecules called lymphokines, which mediate certain immunological readions [e.g., production of migration inhibitory factor ( MIF), lymphocytotoxic factors,

109

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION Location of Cantrornara

n

-8

=0.0074

--

* O ; - , Jm - -

.

Lorua Symbol‘

’OM3

= 0.0017-

-e

n . I n 1 . 1

1 . 1 n

u u

U

0

Chido

C

A

1 - 1 1-1

FACIOR I C-2 DEFICIENCY C-4 DEFlCllNCY I-CELL ALLOANTIGENS

SIRONG MLR GvH REACIIONS1

1

1

=

1 WEAK MLR

I

t

1

CML REAClIONS FIRS1 SET SKIN REJECTION

I

DISEASE ASSOCIATIONS1

-

RAGWEED ALLERGY

1

1

I

IMMUNE RESPONSE GENES (lr)

FIG. 1. Schematic diagram of the major histocompatibility complex (HLA) in man and associated genetic traits. The four presently recognized HLA loci are designated by the letters A, B, C, and D following the Sixth International Histocompatibility Workshop 1975. The A locus was previously known as LA or first locus; B locus as four or second locus; C locus as AJ or third locus; and D locus as MLR-S, MLC locus, LD-I, or LAD. The bars indicate associated genetic traits. Open bars indicate that no definite position for a locus can be determined. A filled bar underneath a region of HLA indicates that there is strong evidence for localizing the trait within this region. The arrows indicate a specific localization of a trait based on study of at least one recombinant family. ( e ) Recombination fraction between two loci; ( PGMI ) phosphoglucomutase-3; ( Chido) erythrocyte antigen, Chido; ( Factor B ) glycine-rich &glycoprotein (GBC), Bf, complement C3 proactivator; ( C2) serum complement C2 deficiency; ( C 4 ) serum complement C4 deficiency; (CML) cell-mediated lympholysis; ( MLR ) mixed-lymphocyte culture reaction; ( GvH ) graft-versus-host reactions.

and mitogenic factors]. Following 6-7 days of in uitro mixed-lymphocyte culture, specific cytotoxic “killer” lymphocytes may appear, and after 10-14 days of in uitro culture, small lymphocytes with specific memory for secondary responses, both proliferative and cytotoxic, appear. The MLR is of special interest, since this in uitro test reflects the individual’s capacity to identify foreign cells and to initiate specific immunological reactions against the foreign cells. The study of genetic control of human MLR has advanced rapidly during the last few years as a result of the development of new methods for identifying HLA-D specificities. Using

1

110

DUPONT, HANSEN,

AND YUNIS

lymphocytes from HLA-D-homozygous individuals as stimulators in MLR, it is possible to identify the HLA-D specificities in the popul at'ion. As a result of these studies, six to eight different provisional groups of HLA-D specificities have been defined, accounting for approximately 50%of the determinants in random Caucasians. The study of MLR has ramifications in cellular immunology, immunogenetics, and clinical transplantation. Although many immunological traits and allogeneic phenomena are clearly associated with HLA, the precise region or determinant involved in a specific effect is sometimes difficult to define. The MLR typing now allows direct analysis of the HLA-D region, with the possibility of looking at HLA-D determinants either as markers or as factors in many different immune reactions and in susceptibility to disease. The biological significance of the determinants controlling the MLR in man is largely unknown as is also the case for the other presently known HLA determinants. The conclusions that can be made from the study of genetic control of immune response and from clinical transplantations emphasize, however, the significance of the HLA-D segment of the HLA complex. II. Serology of Human leukocyte Alloantigens (HLA-A,

B, C)

Several authors have furnished extensive reviews of the major histocompatibility system in man: Amos (1969), Amos and Yunis (1973), Amos and Ward ( 1975), Bodmer ( 1973), Ceppellini (1971), Ceppellini and van Rood ( 1974), Dausset ( 1971, 1973), Kissmeyer-Nielsen and Thorsby ( 1970), McDevitt and Bodmer ( 1974), Morris ( 1974), Terasaki and Singal (1969), Thorsby (1974), and Walford et al. (1970). The early history of the detection of antibodies has been reviewed by Walford et al. (1969), Killman (1960), and van Rood (1962). For reviews of the mouse histocompatibility system, consult Stimpfling ( 1971), Shreffler and David (1975), and Klein ( 1975). Reviews focusing on human mixed-lymphocyte culture have been published by SGrensen ( 1973), Bach and Bach (1974), Eijsvoogel (1974), and Thorsby (1974). The following section summarizes the developments in HLA serology and includes information obtained during the Sixth International Histocompatibility Workshop, Aarhus, Denmark, 1975. In Table I are given the presently recognized HLA specificities (World Health Organization Terminology Committee, 1975; Kissmeyer-Nielsen, 1975). A. INTRODUCTION Knowledge of the serology and inheritance of human leukocyte alloantigens (HLA-A, B, and C ) has resulted from the studies of many dif-

TABLE I LISTINGO F CUBRENTLT ItECOGNIZED HLA

A locus N em

Previous

c locus

B locus New

Previous

sPECIFICITIi,S

Sew

Previous

(1975)" 1) locus

New

Previous

HLA-I>Wl

L1) 101-Pf, J, Lad 27a, L1) SVIII, L1) iV5a LI) 102-L1)7a, Pi, S , Ld V L1) 10:3-L1>8~. L1) X I LI) 104-LI) \\l.;a, li, L, L1) X l r l l l LI) lI).?-Ll) I V LI) 106-L1>-pin, L1) SIL' Ll>ll'a, LI) XI1 L 1Lae

~

HLA-A1 HLA-A2 HLL4-A3 HLA-A9 HLA-A10 HLA-A11 HL.4-ABX HLA-A29

HL-A1 HL-A2 HL-A:3 HL-A!) HL-A10 HL-A11 WJ28 W29

HLA-B.5 HL.4-J%7 HLA-BX HLA-Bl2 HLA-Bl3 HLA-B14 HLA-B1X HLA-B27

HL-A,; HL-.47 HL-AH HL-Al2 HL-A1:3 W14 \?'I8 W27

H1.A-AWBB - .--

W'L3

HLA-AW24 HLA-AW24 HLA-AWB6 HLA-AWS1 HLA-AWS2

W'24 W2.i W26 L$rS1 W32

HT, 4.- . A-W 3 3 -

U'l9.6

HLA-AW34 HLA-AWX HLA-AW43

Malay 2 No* BK

HLh-BW1.i HL.4-BIV16 HLA-BW17 HLA4-B\V21 HLA-BW22 HL&B\V3S HL.4-B\V37 I-ILA-BW3H HLA-BbV3'3 HLA-B\\'40 HLA-B\V41 HLA-BW42

W15 LV16 W17 W21 lV22 W5 TY W16.1 W16.2 tV 10 Sahell 11i\-A4

HLA-CW1 HLA-CLV2 HLA-C\V:I HLA-CW4 HL.4-CW5

TI-AJ Ti-Sa532, 170 T3-UPS T4-315 TS

HL.4-I>W2 HL.4-I)jVS HLA-I)W4 HLA-I)\V5 HLA4-I)\V6 LI) 107 LI) 10%

a The genetic region containing the genetic determinants of the rnajor histocompatil~ilitycomplex in man is called HL.4 following the Sixth International Histocompatibility IVorksIrop 1'375. The locus previously called LA (or first) is now designated the letlrr A ; four locus (or second) is now designated 13, and the AJ locus (or third) is now designated C. The genetic locus controlling hlLR is now designated the letter 1) (previously MLIt-S, LI)-l, hlLC locus). T h e alleles of the A and B loci are nuinliered jointly (for historical reasons). The alleles of the C and I) loci are nuinbered separately for each locus starting in sequence from 1. Individual alleles of each locus are designated a numI)er following the locus symbol. Provisionally identified specificities will in addition carry the letter \V, inserted between the locus letter and the specificity number for the allele. An HLA ohenotvoe will lie written as HLA-.41, 2. B7. 12, CLVl, C\V2, I)\Vl, lllV2, and the genotype could be HLA-A2, B12, CW1, I>Wl)Al, B7: CW2, l)\V2 No locus symhol has been designated to the alloantigen systenr coding for antigens n i t h restricted tissue distrihution (Ia equiralents?).

8 5 2 E

3

?

w

4

m

2 m

3

5

! c1

112

DUPONT, HANSEN, AM, WNIS

ferent investigators over the last 18 years. The development of this body of knowledge is an unprecedented example of international collaboration, taking place during six histocompatibility workshops: Histocompatibility Testing, Duke University, North Carolina 1964 (Russell et al., 1965); Histocompatibility Testing, Leiden, The Netherlands, 1965 ( Balner et al., 1965); Histocompatibility Testing, Torino, Italy, 1967 ( Curtoni et al., 1967) ; Histocompatibility Testing, Los Angeles, California, 1970 ( Terasaki, 1970) ; Histocompatibility Testing, Evian, France, 1972 ( Dausset and Colombani, 1973); and Histocompatibility Testing, Aarhus, Denmark, 1975 ( Kissmeyer-Nielsen, 1975). In the beginning of this collaboration, two developments, among many, contributed most to the rapid progress in the field: ( 1 ) the miniaturization and standardization of the lymphocytotoxic technique (Terasaki and McClelland, 1964) and (2) the introduction of rabbit complement for lymphocytotoxic reactions ( Walford et aZ., 1965). The characterization of a human leukocyte alloantigen, Mac, was furnished by Dausset (1958). Following this, van Rood and van Leeuwen (1963) described the first leukocyte antigen system, which consisted of two alleles, 4a and 4b, controlled by one locus. Van Rood and van Leeuwen also introduced the use of two-by-two comparison analysis to identify similar patterns of serological reactions of leukocyte alloantibodies. This has been an essential tool in the subsequent identification of HLA specificities. A second group of leukocyte antigens, labeled LA 1 to 4, was identified by Payne et al. (1964). These antigens belonged to the same allelic system and were shown by family studies to be mutually exclusive. Dausset et al. (1965) studied ten identifiable antigens in the population and postulated a unified concept of human leukocyte antigens, describing a single complex genetic system, which was then termed Hu-1, and later changed to HL-A (Nomenclature Committee, 1968). Assumption of the control of human leukocyte antigens by a single genetic system of closely linked allelic groups was proven in family studies (Dausset et al., 1967; Ceppellini et al., 1967; Amos, 1967). According to this concept, one genetic system controls a large number of alleles. The segregation of paternal and maternal leukocyte antigens can be identified by serological reactions of the cells from family members, and only six different genotypes can be found in a given family. The paternal haplotypes were usually designated a and b, with the genotype ab; the maternal haplotypes were designated c and d, with the genotype cd. Four genotypes could thus occur in the family among the children: ac,ad, bc, bd (Fig. 2 ) .

113

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

a b

I c!l

1

a c

Cornbina tion

Pdrent-Child

0 Haplotype Diff.

none

c d

b

b

I

L

a d

b c

b d

I Haplotype D i f f .

2 Haplo type

Diff.

ab

or

cd

ab

or

cd

vs

ad

b

or

cd

vs

bc

ab

or

cd

vs

bd

d C

VS

ad

ac

vs

bc

dc

vs

bd

ad

vs

bd

ad

vs

bc

bc

vs

bd

d

vs

ac

none

FIG.2. Segregation of HLA haplotypes in a family. Letters a, b, c, and d denote HLA haplotypes; a/b, c/d, a/c, a/d, b/c, and b / d denote HLA genotypes. The probability for genotypically identical siblings in each subsequent mating is 0.25. Parent-child combinations will always share one HLA haplotype. The MLC between HLA-identical siblings ( 0 haplotype difference) will cause no response. The MLC between parent-child and some sibling combinations will cause one-haplotype response; MLC between other sibling combinations will cause two-haplotype response. If the parents share a HLA-D specificity (e.g., the D specificity of the a and c haplotype is cross-reacting ), the disparity in MLR combinations involving this shared specificity (e.g., a/d vs c / d ) will not represent full one-haplotype stimulation. The concept of a single genetic system in control of human leukocyte antigens received further support when Bach and Amos (1967) observed that the in uitm MLR between leukocytes from family members corresponded with the results obtained by serological identification of the leukocyte antigens. klixed-lymphocyte cultures between cells from siblings genotypically identical for the serologically defined leukocyte antigens failed to stimulate. It was also demonstrated that the genes controlling the production of these leukocyte antigens were codominant, since one paternal and one maternal haplotype was expressed in each child.

114

DUPONT, HANSEN, AND YUNIS

B. ONE GENETICSYSTEMOF SEVERAL CLOSELY LINKEDLOCI Based on population and family studies (Ceppellini et al., 1967; Dausset et al., 1970), it was concluded that the HLA specificities belong genetically to the same system but are probably the product of at least two different genetic loci, or regions, now known as the HLA-A locus (previously LA, first locus), and the HLA-B locus (previously four or second locus). It was also concluded that the two loci, each with multiple alleles, were closely linked on the same chromosome. This concept was proven when crossing over was found to take place between the two segregant series ( Kissmeyer-Nielsen et d., 1969; Ward et al., 1969). From analysis of the haplotypes of Scandinavians, it was estimated that the recombinant fraction between the HLA-A and the HLA-B loci was 0.0056 (Svejgaard et al., 1970). In a later report, the same investigators analyzed information on 1362 parental meiotic divisions; in 11 cases, recombinations had occurred and the recombinant fraction was calculated to be 0,0081 & 0.0024, with the maternal recombination frequency slightly higher than the paternal one (Svejgaard et al., 1971). Combined data obtained from six European laboratories have recently been analyzed, and the recombination fraction between the HLA-A and HLA-B locus has been determined to be 0.00874 0.00136. These data consisted of 4614 parental meiotic divisions, with forty recombinations (Belvedere et al., 1975). Within a family, the alleles of one locus cannot be transmitted with the same gamete, and cannot be brought together by crossover (recombinations). By contrast, the alleles of the A and B loci are inherited together as a genetic unit through the gamete, and may be changed by crossover. The combination of the two alleles, transmitted together within the same chromosome, is called a haplotype, and a complete genofype is formed b y two haplotypes. Mattiuz ei al. (1970) used computer analysis of the HLA specificities in families and populations. First, they demonstrated that HLA specificities could be distributed into two nonoverlaping series, corresponding to the two HLA loci, HLA-A and HLA-B. Second, they calculated the haplotype frequencies from the phenotype frequency in the population. This was performed by using gene frequencies found in the population with a correction factor-the delta value-compensating for the nonrandom association between different alleles on the two loci. Third, they studied the zygotic assortment of the four parental haplotypes in individual sibships. The observed random segregation of the parental haplotypes suggested that genetic linkage disequilibrium between the HLA-A and the HLA-B locus was not based on selection at

*

HUMAN MIXED-LYMPHOCYTE CULlTJRE REACXION

115

the level of gametic assortment. These findings were confirmed during the Fourth International Histocompatibility Workshop 1970 ( Allen et al., 1970). It was concluded that the alleles from each of the two series did not occur with more than two specificities in the phenotypes (very few triplets were observed) and that the phenotype frequencies for the two antigen series fitted the Hardy-Weinberg distribution. One additional locus controlling HLA antigens, now considered to be well established, is the HLA-C locus ( AJ locus or third locus). Sandberg et al. (1970) described an HLA antiserum that reacted with about 10% of Scandinavians; the specificity detected by this antiserum did not seem to fit with any of the known antigens of the HLA-A or HLA-B locus. Thorsby et aZ. (1971) further defined two additional antisera that seemed to belong to the same series. These sera described specificities that were clearly shown to segregate with the HLA complex. A number of antigens that segregate as true alleles and fit the Hardy-Weinberg distribution have been assigned to this series (Svejgaard et nl., 1973). Formal proof of a separate third locus has come from the observation of recombinations between the HLA-A locus and the HLA-C locus, and between the HLA-C locus and the HLA-B locus (Mayr et al., 1973; Low et d., 1974; H. E. Hansen et al., 197s). Only two families, however, have been described with a recombination between the HLA-C locus and the HLA-I3 locus (Low et al., 1974; H. E. Hansen et al., 1975), but several families have been found with rccoinbination between the HLA-A locus and the HLA-C locus (Pierres et al., 1975). The HLA-C locus is thus presumably closely linked to the HLA-B locus, but the exact position of the C locus within the HLA complex is still uncertain. Only five specificitics are at present defined for the HLA-C locus (CW1-CW5) ( World Health Organization Tcmninology Committee, 1975; KissmeyerNielsen, 1975). The existence of these three different series of antigens has also been demonstrated by the phenomenon of determinants capping on the cell membrane. Because the detcrminants of HLA-A, 13, and C were shown to be present on the cell surface as separate molecules, they had to be controllcd by different cistrons (Bcrnoco et d.,1973; Solheim et al., 1973). Two important factors have complicated the definition of polymorphism at the three HLA loci: ( 1 ) serological cross-reactivity between different HLA antigens, and ( 2 ) the linkage disequilibrium between the determinants at different loci. The present concept of three HLA loci controlling serologically defined HLA antigens could well be an oversimplification. Most, if not all, HLA antisera are multispecific, and definition of the specificities is only obtained by careful control of

116

DUPONT, HANSEN, AND

methodology in terms of length of incubation, temperature of incubation, quality and quantity of complement, dilution of the reagents (antisera), and/or addition of antihuman globulin, Therefore, the antisera against the HLA antigens are only operationally monospecific. It is possible that the HLA antigens of the different loci have considerable homology at the molecular level, which could explain the cross-reactivity observed among them. Such homologies have recently been suggested from studies of tyrosine residues within different HLA antigens (C. CunninghamRundles et al., 1975). C. CROSS-REACTIVITY Leukocyte antibodies may be mono- or multispecific. In general, it may be assumed that a leukocyte antibody is monospecific when it can be exhaustively absorbed by the immunizing cells and by cells of several other individuals possessing the same specificity. There may be common determinant parts between allelic products of one locus, but each allele must have individual or unique determinant. Monospecific antibodies may be more or less cross-reactive among the different alleles of one locus or they may be directed toward a unique antigenic determinant. Antibodies reacting against a panel of different lymphocytes can define at least two different types of specificities: ( 1 ) narrow HLA specificities, which are defined when the antibody reacts with one known specificity, and can only be absorbed by cells giving positive reactions for that specificity and ( 2 ) cross-reacting groups ( CREG), which are complex and defined by antibodies reacting with cells expressing specificities for two or more different alleles. These antisera can be completely absorbed by cells of any one of these specificities. The shared specificities are said to be included in the products of each of the different alleles in the cross-reacting group. Analogous to the H-2 system of the mouse, this phenomenon has been explained on the basis of private (narrow) or public (broad) specificities. The crossreacting antibodies are of at least two different types: ( I ) antibodies defined by antigens of relatively restricted specificity, for example, one specificity of the HLA-A locus was originally designated as A9 but was later found to include two cross-reacting specificities, AW23 and AW24, and ( 2 ) the broad reacting antibodies, termed supertypic by Ceppellini (1971), such as anti-4a (W4) and anti-4b ( W 6 ) . These broad antibodies include several specificities of the HLA-B locus: 4a includes HLA-B5, 12, 13, W35, W17, W21, 27; and 4h includes B7, 8, W40, 14, 15, 16 and 22 (van Rood, 1973). The cytotoxicity negative and absorption positive ( CYNAP ) phenomenon was introduced by Ceppellini et al. (1965) and is illustrated in the following example. One serum

HUMAN MIXED-LYMPHOCYTE CULTLTRE REACTION

117

( B C ) was shown to cross-react with HLA-AS, A l l , and A1 and could be absorbed by cells phcnotyped as A3, All , or Al, although at some dilutions it was cytotoxic only for A3 (Yunis et al., 1970). This may be due to different avidity of some antibodies to some restricted antigenic specificities, and could explain the many examples of the CYNAP phenomenon. The foregoing description is oversimplified, since there are many paradoxes that cannot at present be explained. For instance, many sera fail to differentiate HLA-B13 and W40, and yet HLA-B13 is included in W4 and W40 in W6 supertypic specificities. Most women, responding to the conceptus, recognize more readily the supertypic specificities dctected by CREG than the type-specific HLA antigens, and therefore the shared portion or the molecule must be highly immunogenic (Nielscn and Svejgaard, 1972; Ceppellini, 1971). The knowledge of CREG has practical applications, since matching for traiisplantation finds its greatest limitation in the nonavailability of fully matched donors. Matching at cross-reactive groups may be important for kidney allograft survival (Dausset et al., 1975).

D. GENETICLINKAGE DISEQUILIBRIUM Antigens of the HLA-A and HLA-R loci can be in linkage disequilibrium. This means that some alleles of the two loci can be found together on the same haplotype with a inuch higher or lower frequency than expected from the genc frequency of the alleles in the population. This association is measured by a constant ( A ) that is defined as the difference between the observed haplotype frequency and that expected from the product of the gene frequencies (Mattiuz et al., 1970). The delta values for the different combinations of alleles vary from population to population. For instance, in European Caucasoids, the highest delta value is found in the HLA-A1, B8 haplotype; in Japanese, HLA-All-B22; and in Asian Indians, HLA-A1-Bwl7 ( Dausset and Colombani, 1973). The reason for these associations is unknown. Chance inbreeding in genetic isolatcs, natural selection, and genetic drift, however, may be factors in the occurrencc of this phenomenon. Two important basic mechanisms may I,c involved: ( I) the selective advantage derived from maintaining closely linked irninune response genes within the HLA complex, and ( 2 ) control of recombination frequency within the HLA complex by other genes. Dunn and Gluecksohn-Waelsch ( 1953) observed that different strains of mice had different recombination frcquency within the H-2 complex. Bennet et al. (1972) have recently noted that the suppression of recombinations within the H-2 complex was controlled by genes at another

118

DUI’ONT, HANSEN, AND YUNIS

locus, the T locus, within the ninth linkage group of chromosome 17 in the mouse. Similarly, it is possible that a T locus equivalent, which would control the recombination frequency between the different loci of the HLA complex, may exist in man (Amos and Ward, 1975). In summary, HLA is a system of several closely linked codominant genes. Three different loci are assumed to exist for the serologically defined HLA antigens, namely HLA-A, B, and C. Only the HLA-A and HLA-B loci, however, fulfill formal genetic requirements, since they control genetically the expression of different mutually exclusive alleles, and since several proven recombinations between the determinants of the two loci have been described. The HLA-C locus has tentatively been assigned five specificities identified by lymphocytotoxic antibodies, but for many individuals the HLA-C determinants cannot yet be defined due to lack of specific antibodies for identification of other possible specificities, Independent capping of cell-surface determinants, however, confirms the existence of all three loci. The exact mapping of the HLA-C locus cannot be made at present, since only two recombinant families have separated the HLA-C determinant from the HLA-B determinant. It is characteristic of the HLA antigens that the identification of different specificities by antisera is complicated by the serological crossreactivity that exists between some antigens. These serological reactions also demonstrate inclusions between different specificities which may represent public and private determinants. Genetic linkage disequilibrium between the alleles of the different loci further complicates the evaluation of the reactions of the HLA antisera. The pattern of reactions obtained with an antiserum when tested against a panel of lymphocytes may indicate that one HLA specificity is included in another specificity. This phenomenon would, however, also be observed if the two specificities defined by the antiserum were in strong positive linkage disequilibrium. A selected panel that minimizes the influence of genetically determined associations between HLA antigens of different loci and family studies make it possible to discriminate antibodies identifying HLA inclusions from antibodies directed against multiple determinants controlled by closely linked genes. Very little is presently known about the biological significance of the serologically defined HLA antigens, These antigens occur as cell surface antigens on all tissue cells and are, thus, different from the alloantigens that have restricted tissue distribution, as described in Section VI,J. The most direct proof that HLA antigens are important in transplantation comes from the observation that allografts are rejected in an accelerated manner when performed in the presence of alloantibody specific

HUMAN MIXED-LYMPHOCYTE CULTURE REACXION

119

to antigen(s) of the donor (detected in cross-match testings). As will be discussed later, the reports of longer allograft survival in HLA-matched unrelated donor-recipient combinations could be explained by assuming that the HLA-A and B antigens themselves elicit the allograft immunity. Another possibility is that there are loci independent from but closely linked to HLA-A and B that are responsible for allograft immunity. Ill. Cell-Mediated Allogeneic Reactions

in Vifro

A. MIXED-LYMHPOCYTE CULTURE REAC~ION The MLR is an in vitro test of lymphocyte recognition and proliferation. Bain et al. (1963) were the first to show that the mixture of lyniphocytes from different individuals in vitro resulted in the production of blastlike cells, the appearance of some cells in mitosis; and the appearance of cells labeled with thymidine- 'H in radioautograph smears. The control and specificity of this reaction and its relation to certain allogeneic determinants is still an area of intense investigation. The very early studies, however, established that the strength of the MLR was related to certain genetic differences. Bain et nl. (1964) showed that there was no reaction in mixed-lymphocyte cultures between three pairs of monozygotic twins, but reactions between four pairs of dizygotic twins were variable. Two pairs of dizygotic twins showed rcactions comparable to those seen bctwcen unrelated individuals, and two pairs showed no reaction. Hirschhorn et al. (1963) confirmed that the mixture of lymphocytes from unrelated individuals led to cell enlargement and division. Bach and Hirschhorn (1964) suggcsted that the degree of transformation in mixed-lymphocyte culture might prove to be useful as a quantitative measure of histocompatibility. Although the reactions between the majority of normal unrelated persons were found to be positive, there was great variation between sibling pairs, ranging from a large numbcr of blast cells in some combinations to no reaction in others ( Bain and Lowenstein, 1964; Chalniers et al., 1966). These studies suggested that some genetic factor( s ) was segregating in families and that this factor( s ) determined certain antigenic differences on leukocytes that could be detected by the MLR. Important innovations in mixed-lymphocvtc culture methodology have included: ( I) the use of isotopically labeled thymidine to determine the synthetic ratc of DNA in reactive cultures (Bain et a]., 1964); ( 2 ) the development of microculture systems (Hartznian et al., 1971); ( 3 ) the development of multisamplc harvesting machines ( Hartzmm et al., 1972); and ( 4 ) the one-way mixed-lymphocyte culture technique. The

120

DUPONT,

HANSEN, AND WNIS

one-way mixed-lymphocyte culture technique makes possible the measurement of proliferation of a single responder in a given combination. In the two-way mixed-lymphocyte culture, lymphocytes from two diffcrent donors are allowed to respond against each other so that the final reactivity measured reflects the total proliferation of both populations. A unidirectional response is obtained by metabolically inhibiting the proliferative capacity of one of the cells in the mixture while leaving its stimulating capacity intact. Stimulation in the mixed-lymphocyte culture requires a metabolically active and viable lymphocyte. Heat-inactivated cells, freeze-thawed cells, or disrupted cells are not able to stimulate allogeneic lymphocytes, even though the serologically detected HLA antigens may still be present (Hardy and Ling, 1969; Schellekens and Eijsvoogel, 1970). The one-way reaction can be achieved by pretreatment of aliquots of the designated stimulating cells with X-rays (Kasakura and Lowenstein, 1965) or mitomycin C (Bach and Voynow, 1966). By different mechanisms these agents block lymphocyte proliferation by inhibition of DNA synthesis, Measurement of DNA synthetic rate has become the standard method to measure lymphocyte proliferation in MLR. Deoxyribonucleic acid synthesis in unstimulated lymphocytes cultured for 3 to 6 days is relatively low and the rate of DNA synthesis in stimulated cultures can be readily determined by measuring radioactive labeled thymidine-14C or -3H incorporation. To make quantitative comparisons between weak and strong stimulation, pulse labeling with thymidine is performed during the log phase of proliferation (Sgrensen et al., 1969). Depending on the culture conditions, the log phase of proliferation for stimulated lymphocytes cultured in a microsystem is 96-144 hours (DuBois et al., 1974; Bondevik et al., 1974; Jorgensen and Lamm, 1974). Different aspects of MLR methodology have recently been analyzed in a cooperative study between several laboratories and reported by Thorsby et al. (1974b). In addition to the proliferative response in the mixed-lymphocyte culture, other cell-mediated allogeneic reactions may occur, The induction of cytotoxic effector cells and immunological memory, as shown by accelerated responses in secondary culture, will be briefly summarized.

B. INDUCTION OF CYTOTOXIC EFFECTOR CELLS IN MIXED-LYMPHOCYTE CULTURE Following stimulation by allogeneic cells or tumor cells, lymphocytes proliferate and acquire a capacity for specific cytotoxicity against the stimulating cells (Cerottini and Brunner, 1974). Some degree of proliferative response requiring several days of incubation is generally necessary before a cytotoxic effect of cells activated in vitro can be de-

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

121

tected. This cell-mediated cytotoxicity is independent of antibody or complement. The first in oitro system showing some specific sensitization against transplantation antigens was developed by Ginsburg and associates, utilizing the xenogeneic combination of rat lymphocytes sensitizcd on mouse cell monolayers (Ginsburg, 1968; Berke et al., 1969). They showed that sensitized rat cells had their greatest cytotoxicity when tested against mouse cells syngeneic with the immunizing monolayer. Hiiyry and Defcndi ( 1970) used the one-way mixed-lymphocyte culture to sensitize mouse lymphocytes against allogeneic mouse lymphocytes and showed cytotoxicity by chromium-51 release against lymphoid cell lines isogeneic with thc sensitizing cells. Solliday and Bach (1970) showed that human cells could also generate cytotoxicity in mixedlymphocyte culturc and that cytotoxicity was greatest against the same lymphoid cell linc that was used to sensitize. However the necessity of using lymphoid cell lines that were more sensitive to the cytotoxic effect of sensitized lymphocytcs than normal cells severely restricted the possibilities for studying the specificity of the cytotoxic reaction. Lightbody et al. (1971) and associates (Miggiano et al., 1972) introduced an innovation that made it possible to use normal lymphocytes as targets in the cytoxicity reaction. Normal lymphocytcs treated with phytohemagglutinin ( PHA ) produce blasts that function as targets just as well as lymphoid cell lines in the chromium-51 release assay. Target lysis of PHA blasts is maximum for the PHA blasts syngeneic with the cells used to sensitize in mixed-lymphocyte culture. Usually no cytotoxicity occurs against PHA blasts syngeneic with the sensitized effector cells, This test for cell-mediated lympholysis, following mixed lymphocyte culture and using PHA blasts as target, is called the CML test. During the mixed-lymphocyte culture cytotoxic lymphoid cells are produced when the responder and stimulator cells differ at the HLA chromosomal region (Eijsvoogel et al., 1972b, 1973a,b,c; Miggiano et al., 1972; Trinchieri et al., 1973; Bonnard et al., 1973; Mawas et al., 1973; Alter and Bach, 1974). Studying the CML test in families, Miggiano et al. (1972) showed that a greater cytotoxicity could be achieved between two-haplotype different siblings as compared to one-haplotype different siblings. Eijsvoogel et al. (1972b, 1973c) studied families with recombination within the HLA region. When the CML test was performed between combinations differing at one or two HLA haplotypes, clear cytotoxicity always resulted, Cytotoxicity did not occur, however, between MLR-positive combinations that were seroidentical. In three combinations that were HLAD-identical but serologically incompatible, no proliferation occurred and

122

DUPONT, HANSEN, AND W N I S

no cytotoxicity could be detected. The conclusion from these studies was that disparity at HLA-D is necessary to generate effector cells, but measurable cytotoxicity in CML depends on disparities at the serologically defined HLA-A or B loci or at some other closely linked determinants that are responsible for the specificity of effector and target cell interaction ( Eijsvoogel, 1973a,b). More recently, it has been suggested that the HLA-D determinants are not esssential for the generation of cytotoxic cells (Long et al., 1975). Studies with CML both in recombinant families and between unrelated individuals, matched or mismatched for various serologically defined HLA antigens, indicate that HLA-A antigens tend to be weaker immunogens in CML than most HLA-B antigens (Eijsvoogel et al., 1973a,b). The HLA-A antigen A2, however, has been shown to be as strong a target as the HLA-B antigens (Trinchieri et al., 1973; Eijsvoogel, 1974; Grunnet et al., 1974). The specificity of the killer cells generated is directed toward the stimulator cell, but another cell may be a target cell if it shares determinants with the stimulator cell. These determinants are either HLA-A or HLA-B, or products of closely linked genes. In the mouse, it has been demonstrated that cytotoxic cells can be generated in MLC combinations where the responder and the stimulator cells differ at the H-2K locus alone. Conversely, Sprrensen, using congeneic strains carrying recombinant alleles of the H-2 complex, found that the specificity of the effector phase is determined by serologically detectable antigens or by products of genes that map together with H2-D or H-2K and that incompatibility at MLC is not an absolute requirement for sensitization to occur (Sgrensen and Hawkes, 1973; Sorensen, 1973). More recently, Fcstenstein and Dkmant ( 1975) have presented evidence for the existence of a mouse CML-stimulating locus, ECS, mapping in thc I-B and I-C subrcgions of thc H-2 complex. In man, the HLA antigens A, B, and C may be important as target determinants. Whether or not these antigens are the exclusive target determinants has been studied by testing the CML between HLA-identical individuals. Bach et al. (1973) investigated 5 HLA-A, B-identical unrelated individuals and found that one combination exhibited positive CML. Schapira and Jeannet (1974) found several positive CML reactions in HLA-identical unrelated individuals. The CML effectors educated between HLA-A, B-identical unrelated individuals cxhibited a higher mean chromium release than combinations involving HLA-identical siblings (Grunnet et al., 1974; Grunnet and Kristensen, 1975). Kristensen and Grunnet (1975) used two CML effectors educated against different HLA-A, B, C antigens on two different stimulators and found

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

123

that they gave a similar pattcrn of cytotoxicity against a panel of target cells. The “CML determinant” prcmmed to be the target for cytotoxicity was found to be in association with HLA-B8, both in population and family studies. This study suggests the possibility of a separate CML locus within HLA and also suggests :in approach for using hlLC-activatcd effector cells in “CML typing.”

c. INDUCTION OF IN

IMMUiXOLOGICAL MEhfonY CELLS MIXED-LYMPHOCYTE CULTURE

Following the first 6-7 days of mixed-lymphocyte culture, mouse cells revert to small nondividing cells with loss of cytotoxic effect against the primary target cells (Andersson and Hayry, 1973, 1974). When primed mouse cells are rechallenged between 14-21 days of culture with fresh primary target cells, they regain their cytotoxic effect within 24 hours, and the rate of secondary proliferation is accelerated (Hayry and Andersson, 1973; Cerottini et al., 1974; MacDonald et al., 1974; Wagner et al., 1972). Similarly, the proliferative response of human lymphocytes primed in oitro and restimulated on day 14 is accelerated (Fradelizi and Dausset, 1975; Sheehy et al., 1975a). Studying an HLA-B/D recombinant family, Mawas et al. (1975a) have shown that disparity for HLA-D alone is sufficient for the secondary response of primed cells, and disparity for HLA-A, B, or C alone does not initiate secondary proliferation. An accelerated secondary response can occur against a new cell donor if the donor shares an HLA-D determinant with the stimulator in primary culture (Fradelizi et al., 1975; Sheehy et al., 1975a). If a haploidentical cell donor is used to stimulate in primary MLC, the primed cells will identify the same HLA-D determinant in secondary culture by an accelerated response. The use of primed lymphocytes as typing cells against unrelated individuals has been proposed as a new approach to HLA-D typing ( Sheehy et al., 1975a,b; Fradelizi et al., 1975). In addition to the acceleration of proliferation following restimulation, lymphocytes primed in vitro show an increased efficiency as effector cells when rechallenged with the same target cell (Wagner et al., 1972; Andersson and Hayry, 1973; Hiiyry and Andersson, 1974; Cerottini et al., 1974; MacDonald et al., 1974). The reinduction of specific effector cells was found to result from activation by HLA-D products, which need not be identical to those present on the primary cells (Charmot et al., 1975a,b). Even nonspecific niitogens were shown to be capable of reinducing specific effector cells. However, HLA-A, B, or C differences alone could not restimulate primed effector cells. Mawas et al. (1975b,c) have studied two families with HLA-B/D recombinations using the more sensitive secondary CML test. In one combination, an HLA-D difference

124

DUPONT, HANSEN, AND W N I S

did not elicit killer cells after secondary challenge, but in the other combination an HLA-D difference did elicit killer cells after secondary challenge. These preliminary observations are consistent with the possibility of a locus between HLA-B and HLA-D that controls a CML determinant. A similar locus ( H D R ) had been previously postulated to map between HLA-B and D, based on skin graft survival in HLA-A,B-identical HLA-D-different siblings (Yunis et al., 1973). In addition, Mawas et al. (1975b,c), using the secondary CML test to prime and rechallenge HLA-AS, B7 heterozygous responders with a homozygous HLA-AS, B7,DW2 stimulator showed that the secondary effector cells were cytotoxic against several unrelated target cells, most of them carrying the HLA-DW2 determinant, These experiments, as well as those reported by Kristensen and Grunnet (1975), using primary CML, suggest that CML is directed toward antigen( s ) in strong linkage disequilibrium with other antigens of the HLA-B, D region. In summary, the in vitro triggering of allogeneic responses and the resulting cellular differentiation might enable us to study the cellular events leading to rejection of a graft or a graft-versus-host reaction. Mixture of allogeneic lymphocytes in vitro results in proliferation in the MLR. The generation of killer lymphocytes is measured in the CML test. Both MLC and CML have immunological specificity, and MLC-primed cells display specific immunological memory ( secondary MLC and CML). The proliferative phase in primary MLC occurs without previous immunization and is initiated by differences a t determinants genetically controlled by the HLA-D locus. The determinants controlling the specific lysis of target cells are not known but are probably genetically controlled by genes different from, but closely linked to HLA-B. IV. Measurement of Antigenic Differences in Mixed-lymphocyte Culture Reaction

Quantitative expression of MLR data for man has been an important feature in the deductive analysis of the genetic factors controlling the mixed-lymphocyte culture reaction. Responses in MLR (one-way) have been expressed using increments or net counts per minute, stimulation ratios ( S R ) , or relative responses ( R R ) . Stimulation ratios are calculated by dividing the counts per minute in stimulated cultures by the counts per minute in autostimulated or nonstimulated cultures:

SR

=

cpm st,imulated culture (test) cpm autostimulated culture (cont,rol)

Relative responses are calculated by dividing the net counts per minute

125

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

in a test combination by the net counts per minute in an average maximally stimulated combination (which is said to define the 100%reference value ) :

RR

=

cpm stimulated - cpm autostimulated cpm reference response - cpm autostimulated

x

100%

The reference value for a given responder is derived from the average response against a panel of lymphocytes from unrelated individuals (usually 5 different donors) or from the response against a standardized pool of stimulator cells containing lymphocytes from three to five HLAdisparate unrelated donors. Osoba and Falk (1974) suggested that a pool of cells from 3 different individuals representing the major cross-reacting groups of serologically defined HLA antigens be used to compose a standardized pool. We have used the cells from 4 different individuals who are HLA heterozygous and who are known not to carry any of the common HLA-D determinants. The calculation of stimulation ratio corrects for interexperimental variation in autostimulated cultures but does not allow for correction in the variation of maximal response capacity. Small variations in counts per minute in unstimulated cultures can have a great influence on the value for SR. The use of relative responses, with a reference value included to stabilize the maximum response, has been shown by Jorgensen et al. (1973) to allow better reproducibility of interexperimental results than the use of stimulation ratios. Studies on the kinetics of lymphocyte responses in weak and strong stimulation indicate that the doubling time of DNA-synthesizing cells is constant and independent of the antigenic strength of the stimulating cells (Wilson et al., 1968; Bach et al., 1969). The number of cells initially responding in MLC seems to be related to the antigenic strength of the stimulating cells. It has been shown through experiments involving sequential elimination that diff erent clones of responder cells react to specific stimulating cells (Salmon et al., 1971; Zoschke and Bach, 1971). The depletion of a specific antigen-reactive clone by broinodeoxyuridine (BUDR) incorporation and UV irradiation inhibits the response of that population of cells to repeated challenge by the same stimulating cell but not the response to other allogeneic cells. Assuming that the lag phase before onset of proliferation is the same for strong and weak combinations, proliferation, which is measured by the DNA synthetic rate, should be proportional to the nuniber of antigen-reactive cells. Empirically, the measurement of proliferation in MLC does reflect differences in antigenic strength and allows the ranking of MLR for combinations representing strong, intermediate weak, and zero stimulation.

126

DUPONT, HANSEN, AND YUNIS

In MLC family studies, it has repeatedly been shown that quantitative differences can be measured between combinations representing HLAseroidentical siblings ( zero haplotype differences, Oh), discrepancy at one haplotype ( 111. differences), or discrepancy at two haplotypes (212 differences ) ( Albertini and Bach, 1968; Sprrensen and Kissmeyer-Nielsen, 1969; Schellekens and Eijsvoogel, 1970; Seigler et al., 1971). The average response for one haplotype-different combinations is approximately half the average response for two haplotype-different combinations ( Schellekens and Eijsvoogel, 1970); Sgrensen, 1970; Mempel et al., 1973a; Dupont et al., 1974a; Jiirgensen and Lamm, 1974). No overlap is seen between the responses of HLA-seroidentical siblings and the responses of family members different for two haplotypes. Sprrensen (1970) found no significant difference in the average responses between two haplotypedifferent family members as compared to the average responses between randomly chosen unrelated individuals. It has been shown by several investigators that the stimulation between HLA-seroidentical sibling pairs is essentially zero, as measured by standard MLC techniques. Thus, it is concluded that the histocompatibility system controlled by the HLA region is the only significant contributor to stimulation in the MLR in vitro (Bach and Amos, 1967; Amos and Bach, 1968; Albertini and Bach, 1968; Eijsvoogel et al., 1970; S@rensen,1970). A quantitative analysis of a large number of MLC tests, accumulated from two different laboratories and representing 983 different one-way combinations, has been evaluated with respect to average stimulation between HLA seroidentical siblings (Oh), family members discrepant for one haplotype ( l h ) , and combinations discrepant for two haplotypes (2h) ( Dupont et al., 1974a; Thomsen et al., 1974). When the stimulation ratios for Oh, lh, and 2h combinations were converted to log values (log SR ) and then plotted against the cumulative percentage of observations in each group, three straight-line curves were generated, indicating that the log-converted stimulation ratios in each group of combinations are normally distributed, From the log-converted SRs, the mean and 95% confidence interval for each group was calculated. These values, retransformed into SRs, are given in Table IIa. The mean SR, derived from ninety-six different combinations between Oh sibling pairs, was 0.95 (0.45-2.0). This range defines the mean SR and 95%confidence limits for MLC identity. The mean for combinations differing by l h was 6.5 (2.0-2l.O), and the mean SR for combinations differing by 2h was 12.0 (3.7-38.0). Although the mean SR for 2h differences is nearly twice the mean for l h differences, there is a large overlap between the two groups. This overlap between I h and 2h groups has been demonstrated in several different studies ( Albertini and Bach,

127

H U M AN MIXED-LYMPHOCYTE CULTURE REACTION

TABLE I1

Q u INTIT.I~IYON O F MLC I)Isr \ n i w a. STIMUI, I T I O N I1 \ T I O S ( S Z 2 ) u . b HLA (haplotype) disparity:c Parameter

Oh

lh

2h

No. of combinations Mean (95 % Confidence range)

96 0.9-5 (0.4.5-2.0)

266 6..5 (2.0-21 , 0)

62 1 12.0 (3.7-38.0)

HLA (haplotype) disparity :c Parameter

No. of combinations Mean (net rpm) Mean (relative response) (9.5 % Confidence range)

Oh 41 92 0 7% (0-3 6 %)

lh

2h

117 7489 66 % (16-1 16 %)

130 924,; 90 % (48-132 %)

Reference response to pooled cells 78 10,826 100 % -

D a t a from Dupont et al. (1974a) and Thornsen et al. (1975a) Table I I a represents 983 different one-way MLC combinations. The mean and 95 % confidence limits are obtained from log converted stimulation ratios. (Oh) IILA-identical sibling conibinations; ( 1 h) one HLA haplotype-different family combinations; (2h) two HLA haplotype-different family or unrelated combinations. Unpublished data, J. A . Hansen and B. Dupont (1973). Table IIb represents 288 different one-way MLC combinations. The reference response is the response to a pool of 4 different unrelated, IILA-heterozygous individuals. The relative response ( K R ) is clalculated as the percentage of the net counts per minute in stimulated cultures relative to the net reference response. a

1968; Schellekens and Eijsvoogel, 1970; SGrensen, 1970). It may be the result of some technical factors causing interexperimental variation or it may be the result of genetic heterogeneity within each group. The assumption that antigenic disparity between combinations within the l h group or 2h group is always of the same strength may not be true. Genetic uniformity in the strength of l h stimulation can be expected only if the antigenic disparity between the four haplotypes inherited within a family is of equal strength. If parents share a haplotype with cross-reacting specificities, then the disparity in l h combinations involving the haplotypes will not represent full l h stimulation, and the response in MLR will be less than the average l h response (Fig. 3 ) . The

128

DUPONT, HANSEN, AND YUNIS

argument that shared specificities on different haplotypes may came less stimulation than expected is also relevant to the occasional observation of weak MLR between some unrelated combinations. Jorgensen and Lamm ( 1974) have evaluated relative responses among fifty-four random unrelated 2h combinations and found a good fit with the normal distribution, The analysis of our own data for the responses in Oh, lh, and 2h combinations is shown in Fig. 3. The average response in Oh combinations, expressed as relative responses, is 0.7%,with a range of 0 to 3.6%(Table IIb). There is no overlap between the Oh responses and the responses in the 2h groups. Although the mode of distribution of l h and 2h combinations approaches normality, there are a number of outliers, particularly in the low range of distribution. These outliers probably represent combinations that share more than the expected number of determinants. Regardless of whether the data are expressed in stimulation ratios or relative responses, the same quantitative distinction between the different groups of responses (Oh, lh, and 2h) can be made. Lack of MLR between randomly selected unrelated individuals is an extremely rare event. If, however, unrelated combinations are selected on the basis of HLA seroidentity, the number of weak MLC responses increases greatly (van Rood and Eijsvoogel, 1970; Kissmeyer-Nielsen et al., 1970; Sgrensen and Nielsen, 1970; Singal et al., 1975). In more recent studies, it has been shown that approximately 10%of HLA seroidentical combinations may show no response or very weak response ( Mempel et al., 1973a,b), Even unrelated combinations,which are selected for identity at only the HLA-B locus, may show some very weak reactions (Mempel et al., 1973b; Thomsen et al., 1974; L'Espkrance et al., 1975). Repeated testing of 5 HLA-B locus-identical individuals, selected as relatively MLR identical from a group of 16 HLA-B locus-identical, unrelated normal blood donors, indicated that the mutual MLR between these 5 unrelated individuals constituted a group of very weak responses that were clearly separable from both Oh- and lh-different combinations ( Thomsen et al., 1974). As shown in Section VI, the degree of stimulation produced by HLA-D-homozygous cells against heterozygous cells that share the same HLA-D specificity belongs to this same weak intermediate range of responses. These weak responses are known as typing responses. In conclusion, MLR can be ranked into four different groups based on antigenic differences between responder and stimulator. A clear separation exists between HLA-seroidentical combinations (Oh) and combinations that differ by one or two haplotypes ( l h and 211). The mean for one haplotype responses is approximately half the mean for two haplotype

HUMAN MIXED-LYMPHOCYTE

CULTURE =(;TION

129

15

Differences

10 5

HLA Identical Siblings

I

0

N=77

o;

' 40

60 ti0 ' 160 Relative Response (%)

o1;

140

FIG.3. Frequency distribution of relative responses of 305 heterozygous responders against one HLA-D homozygous typing cell ( DW2 specificity). Frequency distribution of relative responses for 77 combinations representing HLA-identical siblings, 287 combinations representing one-haplotype stimulation, and 328 combinations representing two-haplotype stimulation.

130

DUPONT, HANSEN, AND YUNIS

responses. Considerable overlapping, however, exists between these two groups. Evaluation of this intermediate group in terms of HLA genetic differences is discussed further in Sections V and VI. V. Single-Locus Concept for Mixed-Lymphocyte Culture Stimulation (HLA-D locus)

A. FAMILY STUDIES The genetic control of stimulation in the MLR was clearly assigned to the major histocompatibility system (HLA) in man by Bach and Amos (1967; Amos and Bach, 1968). They demonstrated that a significant correlation existed between lack of response in MLR and identity between siblings, with respect to the histocompatibility antigens detected by the lymphocytotoxicity test. Of a total of 209 siblings, 29.2% showed no response in reciprocal MLR, whereas 282 unrelated cell combinations all showed significant stimulation. It was concluded that the genetic system controlling MLR is identical with the major histocompatibility locus (HLA). Similar observations had been obtained in the mouse where strains differing at H-2 showed stimulation, whereas H-2-identical cell combinations did not react (Dutton, 1966). Bach and Amos (1967), however, observed one MLR-positive combination of cells from siblings in one family where the siblings were identical for the serologically defined histocompatibility antigens. Also, two combinations between parent and child did not show a positive MLR. Similar observations were subsequently obtained. Albertini and Bach (1968) confirmed the findings by Bach and Amos and further demonstrated the quantitative differences in MLR elicited by one and two HLA-haplotype-different family combinations. These investigators, as well as Bach et al. (1969), presented a few combinations of HLA-seroidentical siblings with positive MLR, and single, one haplotype-different combination with lack of positive MLR, or very weak response. Plate et aZ. (1970) described another family in which one pair of HLA-A, B-seroidentical siblings showed strong positive MLR in mutual MLC testings. Seigler et al. (1971) observed some parent-child combinations with very weak mutual responses in MLR, or with positive MLR in only one direction. The possibility that MLR might be controlled by a separate gene closely linked to the serologically detectable HLA antigens was suggested by Yunis et aZ. (1971), based on studies in recombinant families. Two different phenomena were observed: ( I ) The MLR in a known recombinant family, with recombination occurring between the HLA-A locus and the HLA-B locus, was controlled by genetic factors closely linked

131

CULTURE REACTION

HUMAN MIXED-LYMPHOCYTE

to the HLA-B locus; ( 2 ) MLR in another family with two pairs of HLA seroidentical siblings did not correspond to the results of the HLA serotyping. In this family, the sibling combination with the HLA genotypes a l c was mutually nonresponsive in MLC, whereas the sibling combination with genotypes a l d was mutually responsive. One of the siblings, however, with the a l d genotype was mutually nonresponsive with the two a / c haplotype siblings (Fig. 4).This suggested that the child's d haplotype was a recombinant haplotype with a maternal recombination occurring between the HLA-B locus and the hypothetical locus for the control of MLR (MLR-S), now designated HLA-D. These findings, together with findings from studies of HLA seroidentical unrelated combinations, led to the hypothesis of a separate locus for the genetic control of MLR. This locus was tentatively placed outside the do/b

F

I Responder Cells

Family

d

Stimulating

Ch 1

Chi

7

ac

Ch3

ad

-

-

Ch4

ad

+

+

Yunis & A m a s ,

-

-

Ch2

by

~~

Ch3

-

ac

-

described

Cells

ChZ

1971

.

+

I

Chd

+ + +

a = AI,B8 b:

A9.612

c

6

A3.BW17

d

= A2,BIZ

FIG. 4. Informative families with indirect evidence of HLA-B,D recombination. Pedigree indicating the serologically defined HLA haplotypes in a family: a, b, c, and d. The lower part of the figure gives the results of the MLC testings between the siblings. An MLR identity is found between u/c ( C h l ) = u / c (Ch2) = a / d (Ch3). Positive MLR is seen between the two HLA-seroidentical siblings u/d (Ch3) and n / d (Ch4). The HLA-B/D locus recombination can, therefore, be assigned to the d haplotype of Ch3. ( - ) MLC identity (or very weak responses); ( + ) strong positive MLR.

132

DUPONT,

HANSEN,

AND W N I S

segment of the chromosome controlling the serologically defined HLA-A and HLA-B antigens (Yunis and Amos, 1971). A series of reports of mixed-lymphocyte cultures in families with recombination between the HLA-A and the HLA-B locus followed. These studies consistently demonstrated that sibling combinations that differ for only one HLA-A antigen did not generate strong stimulation in MLR. In combinations where the siblings differed, however, for a single HLA-B locus antigen, the MLR was strongly positive (Dupont et al., 1971; Gatti et al., 1971; Lebrun et al., 1971; Eijsvoogel et al., 1972a,b; Yunis et al., 1974, 1975). This confirmed that the genetic determinants responsible for strong MLR were located close to the HLA-B locus, or that they were identical to the HLA-B determinants. Several other family studies of MLR demonstrated HLA-seroidentical sibling combinations that were mutually strongly responsive (Eijsvoogel et al., 1972a,b,c; Mempel et al., 1972; Sasportes et al., 1972, 1973; Thorsby et al., 1973b; Keuning et al., 1975b). In other families, it was shown that some siblings differing for one HLA haplotype did not generate positive MLR( Eijsvoogel et al., 1972a,b; Sasportes et al., 1972; Dupont et al., 1972, 1974a). This information further supported the hypothesis that a separate gene, different from the serologically defined HLA determinants, was responsible for MLR. B. MIXED-LYMPHOCYTE CULTURES BETWEEN UNRELATED INDIVIDUALS The MLR between cells from randomly selected unrelated individuals is practically always positive. Even when HLA-seroidentical combinations are selected, it has been shown that most of these pairs develop strong positive responses. Approximately 10%of these combinations, however, are MLR identical or show only weak positive responses (van Rood and Eijsvoogel, 1970; Kissmeyer-Nielsen et al., 1970; Sgrensen and Nielsen, 1970; Eijsvoogel et al., 1970; 1971; Sengar et al., 1971; Mempel et al., 1973a,b; Jorgensen and Kissmeyer-Nielsen, 1973; Segall et al., 1973). This discrepancy demonstrated that typing for the serologically defined HLA-A and HLA-B determinants cannot be used to predict the results obtained in mutual mixed-lymphocyte cultures and that the HLA-A and HLA-B determinants cannot be the major genetic factors controlling MLR. The observation that randomly selected unrelated pairs practically always show positive MLR, whereas 101%of the HLA-seroidentical combinations are clearly weakly responsive or MLR-identical, indicates that HLA-A antigens and/or the HLA-B antigens must have some relation to the specificity of the MLR controlling determinants. This concept is further supported by the observation of some (less than 10%)MLR-identical or weakly responsive MLR combinations among individuals carrying the same HLA-B specificities. The majority of pairs in this group will

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

133

show strong mutual MLR (Mempel et nl., 1973a,b; Dupont et al., 1973c; L’Espkrance et al., 1975) . A few reports have described MLR identity, or weak MLR, among unrelated HLA-A, B-different combinations, even with differences between the two cells on the HLA-B locus (Dupont et d., 1972; Pentycross et al., 1972; Mempel et ul., 1973a,b). Some of these combinations have later been reevaluated and shown to represent HLA-A, B-heterozygous, but HLA-D locus-homozygous cells (Dupont et al., 1973c, 1974a; Mempel et aZ., 1975). The results of family studies and studies of unrelated combinations can be summarized: The genetic control of strong MLR in man derives from a ringle gene (or genetic region) closely linked to HLA-B. It is located outside the HLA-A, B segment of the HLA complex. This genetic region controls an allelic system, wherein the specificity of determinants to some extent is related to the specificity of the determinant on the HLA-B locus. A schematic representation of the position of the four, presently defined genetic determinants of the msjor histocompatibility complex (HLA) in man is given in Fig. 1. The locus controlling strong MLR has been named HLA-D, following nonieiiclature established by the Sixth international Histocompatibility Workshop, 1975. As shown in the figure, the recombination fraction between the HLA-A and the HLA-€3 locus has been estimated to 0.008 (Svcjgaard et id., 1971; Belvedere et nl., 1975). Thorsby et al. ( 1975a) estimated the recombination fraction between HLA-B/D to be 0.014. The rcwmibination fraction between the HLA-B and HLA-D has recently been estimated to be 0.00’74 by Keuning et al. (1975b). These authors studied 39 familics, including 8 families in which parents did not share an HLA-B, D hapIotype (incIuding the HLA-D allele), 24 families in which the parents shared one haplotype, and 12 sets of HLA seroidentical siblings from 10 families in which parents or grandparents were not available for study ( 3 of which also belong to the second group). Positive MLR was found in two pairs of HLA-seroidentical siblings in two different families (the material represents 167 children). The families selected were, in most instances, studied for possible HLA-D-homozygous offspring from first-cousin marriages. The material contains a high fraction of HLAidentical siblings and a high fraction of mating pairs sharing at least one HLA-A, B haplotype. The studies were also based on the assumption that there was no more than one recombination between the HLA-B locus and thc HLA-D locus per family. Corrections were made to compensate for parental homozygosity, but the influence of genetic linkage disequilibrium, which exists between some HLA-B alleles and HLA-D alleles, was ignored.

134

DUPONT, HANSEN, AND YTJNIS

This indirect approach for determining the recombination fraction between the HLA-B and the HLA-D locus is, however, the first attempt to obtain an estimation of the distance between these two loci. Although the HLA-B determinants are identified readily in the lymphocytotoxicity test, the HLA-D determinants were not characterized by typing but indirectly assumed by evaluating the results of MLC between HLA-A, B-seroidentical siblings and parent-child combinations. Further development of typing for the HLA-D determinant will supply the tools for direct determinantion of the recombination fraction between the HLA-B and the HLA-D locus ( see Section VI, H ) . The genetic control of the mixed-lymphocyte culture reaction in mouse developed in parallel with the studies in man. Rychlikovl et al. (1970, 1971) demonstrated that differences at the K end of the H-2 gene complex in general provoked much stronger MLR than H-2D end differences. Studies of MLR in pairs of congenic strains were subsequently performed with combinations carrying H-2 recombinant haplotypes or with mutants within the H-2 gene complex (see Shreffler and David, 1975, for review). Bach et a2. (1972a,b) and Meo et al. (1973a,b) established that strong MLR was controlled by genes within the I region of the H - 2 gene complex and not by the H-2K region or the H-2D region. It was, however, demonstrated that genetic differences within other areas of the €€-2 gene complex would generate definite but weak MLR ( Abbasi et al., 1973; Widmer et al., 1973a,b; Meo et al., 1973a,b; Plate, 1973). Festenstein et al. (1970, 1974) and Huber et n2. (1973) demonstrated another genetic system ( M locus) separate from H-2 which can provoke moderate MLR. The MLR in the mouse is, thus, controlled by several genes, most of which are located within the H-2 gene complex. Strong MLR is controlled by genes within the I region, whereas weak-to-moderate MLR can be induced by all other regions of the H-2 gene complex, as well as by some non-H-2 genes. It is concluded that strong MLR in man can be caused by disparity on a single genetic locus (or genetic region) HLA-D. It is, however, not known if disparities within the HLA-D locus can cause weak MLR and if weak MLR can be induced by disparities on other genetic factors, other than HLA-D, within the HLA complex. A few family studies have been reported involving children with recombination between the HLA-A and the HLA-B locus, or between the HLA-B and the HLA-D locus, in which the single HLA-A locus disparity or disparity for the entire HLA-A, B segment of the HLA complex induced weak MLR (Eijsvoogel et al., 1972a; Mempel et al., 1973a; Thorsby et al., 1973b; Dupont et al., 1974a). It was postulated that another locus within the HLA-A, B segment of

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

135

HLA, possibly close to the HLA-A locus, could be responsible for weak MLR. Proof for the existence of such a locus has not been presented. The main difficulties involved in resolving the question of genetic control in MLR are closely related to the problem of quantitation of the responses. The present conclusions are based on the classification of responses into three groups: (1) MLR identity, as seen in most cases of HLA-A, B-seroidentical siblings; ( 2 ) weak MLR, as seen between some HLA-seroidentical unrelated or HLA-B locus-identical individuals; and ( 3 ) strong MLR obtained in combinations differing for one or two HLA haplotypes. The conclusions made in this section are based on the assumption that there is a qualitative difference between strong and weak MLR. Strong MLR can result only from disparities between two cells at the HLA-D locus (or HLA-D region). A few recombinant families demonstrate weak responses in MLR, which may be causcd by disparities at genetic determinants (located close to HLA-A or HLA-B) other than HLA-D, although this has not been proven. In man, MLR has never been shown to result from genetic determinants outside of HLA. VI. Mixed-Lymphocyte Culture (HLA-D) Specificities Defined by HLA-D-Hornozygous Typing Cells

When it was recognized that strong MLR is controlled by a separate gentic locus (HLA-D) different from the genes controlling the serologically defined HLA-A, B, C antigens, attempts were made to develop methodologies for typing the HLA-D determinants. The breakthrough in the development of HLA-D typing was a consequence of the following observations ( I ) The weak MLR obtained from 10%of HLA-seroidentical unrelated individuals indicated a genetic linkage disequilibrium between the serologically defined HLA antigens and the determinants controlling strong MLR; ( 2 ) since weak MLR is observed in 10% of HLA-seroidentical unrelated individuals, it could be assumed that the polymorphism of the strong MLR determinants is relatively restricted and that the strong MLR specificities could be defined not only within families but in populations. Two different approaches were used to identify HLA-D specificities : the study of serological methods for identification of lymphocyte membrane components shared by individuals identical for the HLA-D determinants and the study of MLR, using well-defined HLA-D-homozygous or HLA-D-heterozygous cells as stimulators. The use of HLA-D-homozygous cells as stimulators in typing experiments has been the major advance in the identification and study of HLA-D specificities.

136

DUPONT,

HANSEN, AND WNIS

A. FAMILIES WITH SHARED PARENTAL HISTOCOMPATIBILITY HAPLOTYPES In the early reports of MLR in families, Bach and Amos (1967; Amos and Bach, 1968) mentioned that a few parent-child combinations were nonresponsive in MLR. Seigler et al. (1971) demonstrated that some HLA-seroidentical parent-child combinations, and some family combinations differing for only one HLA-A or B allele, were nonresponsive in MLR or showed very low levels of response. In addition, they described a few combinations with stimulation in only one direction. Nonresponsive HLA-seroidentical parent-child combinations were also observed by Eijsvoogel et al. (1971) and Thompson et al. (1972). These authors presented data from families in which the parents had one HLA-A, B haplotype in common. In such families a child could inherit the same HLA haplotype from each parent and thus be genotypically homozygous. Eijsvoogel et al. (1971) observed that in mixed-lymphocyte cultures some of these HLA-A, B-homozygous cells generated very low response when used as stimulators against cells from parents or siblings who were heterozygous for the shared parental haplotype (e.g., aJ/bvs U J / U , ) . The low levels of MLR obtained in these unusual combinations, however, were clearly different quantitatively from the MLR identity responses obtained from HLA-identical sibling combinations. The first suggestion of a restricted polymorphism in determinants responsible for strong MLR was derived from further studies of outbred families, in which HLA-seroidentical parent-child combinations were mutually nonresponsive in MLR. When parents share one HLA haplotype, as defined by serological typing, there is a 50%chance that a child will be genotypically HLA-seroidentical with at least one of the parents (Fig. 5). Children who are seroidentical with a parent will inherit one haplotype from that parent and the second haplotype, the shared haplotype, from the other parent. The MLR between the seroidentical parent-child combinations will test a l h difference, represented by the two parentally shared haplotypes. Identity response in this combination will indicate that the MLR dkterminants on these two serotypically identical haplotypes aretalsa identical. As illustrated in Fig. 5, the two parental haplotypes HLA-AS, B7 must carry the same HLA-D determinant, because the MLR between the father and child 3 is mutually nonresponsive. These observations indicate that certain HLA-A, B haplotypes are associated with specific HLA-D specificities. The MLR-identical parentchild combinations are most frequently observed in families in which the parents share one of the more common HLA haplotypes, e.g., HLA-A1,

HUMAN MIXED-LYMPHOCYTE CULTURE REACI'ION

A3,B7/A3,#7

'm

'f

Ch I

A3, B7

A2, B12

A l l , B W 4 0 /A3,B7 b am

af i h 2

Ch 3 Ch 1 ChZ Ch3

HLA-D HLA-D HLA-D

137

All,BW40 /A2>#12 b C

Ch

4

Homozygous i f a f = a m Identical w i t h mother i f d f = d m Identical H i t h father i f a f = a m

FIG. 5. Segregation of HLA haplotypes in a family with one shared haplotype. The a,, b, a,, and c denote the HLA haplotypes in the family. The a, = a, for the serologically defined HLA-A, B antigens. Child 1 is HLA-A, B-homozygous. If the a, and a , haplotype carry the same HLA-D determinant, the a,/a, cell is a Dhomozygous typing cell. In other families, the parents may share a HLA-B antigen (e.g., B7), but differ for the HLA-A antigens (e.g., a, = A3, B7 and a m = Al, B7). If these two haplotypes carry the same HLA-D determinant, the a,/a, cell is a D-homozygous typing cell.

B8; A3, B7; or A2, B12 (Mempel et al., 1973a,b; Dupont et al., 1973c; Dausset et al., 1973; Lebrun et aZ., 1973; Sasportes et al., 1973). The association of different HLA-D determinants with these common HLA-A, B, haplotypes suggests that the HLA-D determinants are in linkage disequilibrium with certain HLA antigens and that some HLA-D specificities are associated with common HLA-A, B haplotypes. Therefore, these HLA-D specificities should also be relatively frequent determinants in the population. Additional family studies have demonstrated that the association between HLA-D specificities and the HLA-A, B antigens is strongest between HLA-D and the HLA-B determinants. Some parent-child combinations differing for one HLA-A antigen but having the same HLA-B antigen on the shared parental haplotype, were MLR-identical (Fig. 5 ) (Seigler et al., 1971; Thompson et al., 1972; Dupont et al., 1973c; Dausset et al., 1973). One intriguing puzzle in MLR has been the finding of HLA-heterozygous unrelated combinations with lack of MLR response in one direction, but with positive response in the opposite direction (Dupont et al., 1972, 1973b; Mempel et al., 1973a, b; Pentycross, 1972). Some of these HLA-A, B-heterozygous individuals have been shown to be HLA-D-homozygous (Mempel et al., 1973b; Dupont et al., 1973a b, 1974a; J. A. Hansen et al., 1975). The results between some other of these heterozygous combinations remain unexplained. Nevertheless, the observation that an HLA-A,

138

DUPONT, HANSEN,

AND WNIS

B-heterozygous cell can be HLA-D-homozygous ( demonstrating that one HLA-D determinant may occur together with more than one HLA-B determinant further suggests that polymorphism of the HLA-D system is relatively restricted. In summary, MLR family studies have shown that up to 50%of HLAseroidentical parent-child combinations may be mutually nonresponsive or only very weakly responsive. These nonresponsive combinations are usually seen when the shared HLA haplotype involves serologically identified HLA antigens that are common in the population e.g., Al, B8; A3, B7; A2, B12; A2, BW15; A l l , BW35. The association between certain HLA-D and HLA-B determinants is much stronger than the association between the HLA-D and the HLA-A determinants. The experiments that led to these conclusions are based on the following assumptions: ( I ) that disparities between HLA-D determinants induce strong MLR and ( 2 ) that the weak MLR seen in some combinations is caused either by disparities at other loci or near identity at the HLA-D locus. The frequent observation of either MLR identity or weak MLR between HLA-seroidentical parent-child combinations in the outbred population indicates that the polymorphism of the HLA-D determinants is relatively restricted.

B. IDENTIFICATION OF HLA-D-HOMOZYGOUS TYPINGCELLS Family studies of MLR constitute the basis for identification of HLA-D-homozygous typing cells. Most frequently such homozygous cells are obtained from families in which the parents are known to share one HLA haplotype. This is based on serological typing for the HLA-A and B antigens. An example of such a family is given in Fig. 5 and Table 111. The shared parental haplotype in this family is designated a. The paternal haplotype af and the maternal haplotype a, carry the same HLA-D determinant. The two remaining HLA haplotypes, b and c, are different. The two HLA-seroidentical parent-child combinations ( q b X u,b) and ( GC X q c ) should be mutually nonresponsive in MLR, and the cells of the homozygous a,af child should not stimulate the cells from either parent or from siblings who are heterozygous for the a haplotype. The use of HLA-D-homozygous cells for typing of HLA-D determinants in the population was first formulated by Mempel et al. ( 1973a). The concept was built on the basis of substantial MLR data obtained in both related and unrelated combinations. Mempel et al. divided their data into three different groups of responses representing Oh-, lh-, and 2h-different combinations. They further identified an intermediate group of weak MLR which consisted of some combinations of HLA-seroidentical

139

H U M A N MIXED-LYMPHOCYTE CULTURE MACXION

TABLE I11 MIXED-LYMPHOCYTE CULTURE RE.\CTIONI N A FAMILY WITH A N HLA-I>-IIOMOZYGOUS CHILD(PARENTS SHARET H E HLA HAPLOTYPE A3,B7,DW2)a Stimulating cells* HLA haplotypes Subject

A

Father Mother Child1 Child 2 Child3 Child 4

3 3 3 3

B 7 7

7

7 1 1 W40 11 W40

D

A

DW2 DW2 DW2 DW2 X X

11 2 3 2 3 2

B

D

W40 X 12 Y 7 I>W2 12 Y 7 1lW2 12 Y

Father Mother AS

+ + + +

Child Child Child Child 1 2 3 4

+

-

AS + -

A

+

+

-

+

+ S + A S + A + +

- + + + + + S A

+ + + S

a All family members who are heterozygous for the same haplotype that is homozygous in Child 1 give negative or weak MLIt, which is characteristic of a typing response against homozygous cells. Child 1 is homozygous HLA-DW2. Father-Child 3 and Mother-Child 2 represent seroidentical parent-child combinations. They are mutually nonresponsive in MLR, indicating that the maternal DW2 and paternal DW2 are identical (+) Positive MLR; ( - ) negative or weak MLR; (AS) autostimulated control cultures.

or HLA-B-identical unrelated individuals. Combinations belonging to this intermediate group repeatedly gave responses midway between the zero and one haplotype responses. Mempel et al. also identified 2 unrelated individuals, homozygous for HLA-AS, B7, who were mutually MLR identical. When cells from these 2 individuals were used as stimulators against several unrelated cells heterozygous for the same haplotype, it was observed that some of the responses fell into the intermediate group of weak responses. This weak reaction against a homozygous stimulator cell is called a typing response. It was suggested that HLA-D-homozygous cells could be used for the typing of HLA-D specificities of unrelated cell donors. The authors discussed the possibility that perhaps the HLA-AS, B7 haplotype in these two unrelated families carried the same HLA-D specificity as a result of some common genetic origin, since these two families lived in the same geographically isolated region. These first two typing cells were assumed to be HLA-D-homozygous on the basis of family studies in which the homozygous cells did not stimulate the two parental cells. The two unrelated typing cells appeared to be homozygous for the same HLA-D specificity because of mutual MLR identity. Similar families were studied by Lebrun et al. (1973). In spite of the observation that the HLA-seroidentical parent-child combinations were mutually MLR identical, the HLA homozygous cells generated a weak

140

DUPONT, HANSEN,

AND YUNIS

but definitely positive response against the parental cells. This could not be explained by assuming a single locus was responsible for strong activation in MLR, and these reaction patterns are still not conceivable with our present concept of the genetic control of MLR. It was suggested that stimulation induced by HLA-D-homozygous typing cells in responders heterozygous for this HLA-D specificity is caused by allelic interaction. The problems of defining HLA-D-homozygous typing cells and typing responses will be discussed later in this section. It can, however, be concluded that even typing cells that elicit relatively high typing responses within the family can be useful in the population for identification of HLA-D specificities. Operationally, a typing cell is useful if it gives a clear bimodal distribution of responses.

C. SOURCEOF HLA-D-HOMOZYGOUS TYPINGCELLS Two different approaches have been used to find HLA-D-homozygous typing cells. One approach is based on the concept that some HLA-D specificities are in strong positive linkage disequilibrium with certain common HLA-B antigens. Therefore, some HLA-D-homozygous individuals should be found even in the outbred population by searching for HLA-A, B-homozygous children in families in which the parents share one of the common HLA-A, B haplotypes (Mempel et al., 1973a; Dupont et al., 1973a,b). The second approach, the study of inbred families, is genetically cleaner. In most inbred families studied, the parents have been first cousins (van den Tweel et al., 1973; Jorgensen et al., 1973). In a first-cousin marriage, or incest family, the HLA-homozygous child will be a true HLA homozygote by descent. The segregation of HLA haplotypes in such a family is shown in Fig. 6. I

FIG.6. Pedigree of a family with a first-cousin marriage in generation 111 illustrating the segregation of HLA haplotypes a, b, c, d, n, m, p, and 9 in the family. Child IV.1 is HLA-homozygous ( a / a ) by descent. The probability of an HLAhomozygous child in any first-cousin marriage is 0.0625.

HUMAN MIXED-LYMPHOCYTE CULTURE REACIlON

141

As will be described later, there does not seem to be any special advantage in using HLA-D-homozygous typing cells obtained from either one or the other source. It should be noted that the weak response obtained when typing cells stimulate parental cells is seen with both homozygous cells from inbred families and with homozygous cells from the outbred population (Eijsvoogel et al., 1971; du Toit et al., 1973; JIdrgensen et al., 1973; Thorsby and Piazza, 1975). Proof that a suspected HLA-D-homozygous cell is truly homozygous may be difficult to obtain: Critical family members may not be available for study, and the prcsumed HLA-D-homozygous typing cell may induce weak or even a moderate degree of stimulation in family members heterozygous for the same HLA-D specificity. When testing a family for a possible HLA-D-hamozygous cell, the most informative combinations to study are those of the HLA-seroidentical parent-child. If this combination is mutually nonresponsive, the HLA-homozygous child must be HLA-D-homozygous, assuming that a recombination has not occurred. The HLA-D-homozygous typing cells almost always induce some weak stimulation of cells from heterozygous siblings and parents. This weak response presents the main difficulty in evaluating HLA-D typing experiments. Since the discrimination between weak responses and the group of responses representing lh-different-sibling or child-parent combinations is relatively poorly defined, the limits of a typing response cannot be clearly defined. The identification of a typing cell is, in practice, based on how the cell functions in a typing experiment. If all of the responses against a homozygous cell can be separated into a bimodal distribution representing a group of weak responses, on the one hand, and a group of I h or greater responses on the other, then empirically the cell is functioning as a typing cell. Some HLA-D-homozygous cells that are heterozygous for the serologically defined HLA-A,B antigens have been identified (Dupont et al., 1972, 1973b,c, 1974a; Mempel et al., 1973b, 1975; Suciu-Foca and Dausset, 1975; J. A. Hansen et al., 1975; F u et al., 1975a,b). Such cells are identified from family studies by the same criteria as used for identification of HLA-A, B-homozygous cells. Family members heterozygous for any of the HLA haplotypes of the typing cell should show typing responses when stimulated by the typing cell. When the two parental haplotypes, assumed to have the same HLA-D determinant, are tested against each other in a parent-child combination, the MLR should be nonresponsive. In summary, HLA-D-homozygous typing cells can be obtained from random matings of the outbred population (Mempel et al., 1973a,b; Dupont et al., 1973c) and from inbred populations, primarily from children of first-cousin marriages (van den Tweel, 1973; Jpgensen et al.,

142

DUPONT, HANSEN,

AND YUNIS

1973; Keuning et al., 1975a), but also from geographical isolates (Layrisse et al., 1975). The main source for HLA-D-homozygous typing cells are families in which the parents share one serologically defined HLA haplotype. Homozygosity for the HLA-D determinant can be demonstrated if the HLA-homozygous cell induces typing responses in family members heterozygous for the haplotype of the homozygous cell and if the HLAseroidentical parent-child combinations are mutually nonresponsive. This approach is applicable to outbred and inbred families. If family members of the HLA-homozygous individual are not available for study, the potential HLA-D-homozygous typing cell may be identified by mutual MLC testing against a panel of defined HLA-D-homozygous typing cells. If none of these approaches can answer the question, a potential typing cell may be operationally identified by using it as a stimulating cell against a large responder cell panel. If the potential typing cell gives a bimodal stimulation pattern that shows a discrimination between weak responses and strong responses, the cell may be said to behave as a typing cell. A typing cell that represents a very rare HLA-D specificity will have to be tested against a large number of responders before it can be definitely identified as a typing cell.

D. DEFINITION OF TYPING RESPONSES A typing response is defined as the weak MLC response obtained when a responder cell reacts to an HLA-D-homozygous typing cell that has at least one D specificity identical to that of the responder. As discussed previously (Section IV), it is only possible to discriminate clearly MLC responses obtained between HLA-identical sibling combinations, on the one hand, and family members or unrelated combinations that differ for one or more HLA-D specificities or haplotypes, on the other. The discrimination between identity responses and positive responses in MLR is a quantitative one. Evaluation of response differences in the weak intermediate range, however, remains a major problem in MLR interpretation. To minimize technical variation in experimental results, many attempts have been made to normalize the data, particularly so that comparisons can be made between data obtained in different experiments and in different laboratories. Variation in the results obtained in any mixed-lymphocyte culture experiment may be the result of: ( I ) various technical factors, ( 2 ) the status of the responding cells, (3) the status of the stimulating cells, and ( 4 ) genetic disparity between the responder and stimulator cells. As discussed previously, responder cell function can be normalized by using relative responses; stimulator cell function is normalized by using the methods of Ryder and Thomsen (Thomsen et al., 1975a; Ryder et al., 1975).

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

143

Figure 3 shows the frequency distributions of relative responses obtained when one HLA-D-homozygous typing cell was used as a stimulating cell against a panel of unrelated responders, For comparison, the frequency distribution of relative responses is shown for family combinations of HLA-identical siblings (Oh), Ih-different combinations, and 2h-different family or unrelated combinations, Some of the responses elicited with the typing cell are clearly lower than the responses for lh-different combinations. There is, however, obvious overlap. The problem in evaluating rcsponses of unrelated individuals is the uncertainty in discriminating between 111 stimulators ( i.e., disparity for at least one strong HLA-D determinant) and typing responses. At present, no definition for the upper limit of a typing response that gives statistical limits of confidence is available. From our empirical analysis of the relative responses of each homozygous cell in family studies, and from the frequency distribution of the relative responses of a large donor panel, we have selected an arbitrary limit of 35% for the upper level of a typing response. The definition and reproducibility of typing responses were evaluated in the Joint Report of the collaborative MLC experiments a t the Sixth International Histocompatibility Workshop ( Thorsby and Piazza, 1975) . In the combined analysis of typing experiments, which included a reproducibility experiment, it was clear that the assignment of a positive MLR against a typing cell could be based o n a single testing. The assignment of typing response, however, could not be reproduced in a large proportion of combinations if the experiment was repeated. The number of combinations that gave discordant results between typing versus nontyping responses in repeated experiments varied from 25 to 70%, depending on the method used to define a typing response. Different HLA-D-homozygous cells, independent of the specificity that they represented, had different discriminating capacities. For selected homozygous cells used as typing cells in a standardized procedure in a single laboratory, the reproducibility of typing responses can be quite good (see Fig. 7 ) . Why a weak MLR response results when a heterozygous cell reacts against a typing cell that is homozygous for one of the same HLA-D determinants is unknown. Stimulating cells are pretreated with irradiation of mitomycin C to block DNA synthesis, but it is possible that certain lymphokines may be secreted by the stimulating cell in response to the allogeneic responder cell. This is a postulated reaction that has been called back stimulation. Mitogenic factors or blastogenic factors have been reported in the supernatants of mixed-lymphocyte cultures (Janis et al., 1970). Preliminary studies on the production of blastogenic factor by homozygous

144

DUPONT, HANSEN, AND YUNIS 130 120 110 100

-

-

-

90,

t; 60 80

70

N YI

c

50

-

30 -

J’

40

Dw2

20 A

FIG. 7. Reproducibility testing of responses to two HLA-D-homozygous typing cells. The lymphocytes of 20 unrelated individuals were tested against the DW2 typing cell and 21 were tested against the DW3 typing cell. The correlation coefficient of identity between test 1 and test 2 for each test cell was r = 0.921 (DW2) and r = 0.921 ( DW 3 ) at a significance level of p 0.001.

<

stimulating cells in MLC typing suggest that such a mechanism may be at work, but conclusive evidence has not yet been obtained (Jorgensen et al., 1974a). Mempel et al. (1973a) performed chromosomal analysis of “two-way” MLC between an HLA-D-homozygous brother and his heterozygous sister. The analysis revealed 28 mitoses: 21 male and 7 female. This corresponded to the stimulation observed in the two directions tested and indicates that the typing responses reflect cell proliferation of the responder cells rather than insufficient blocking of DNA synthesis in the stimulating homozygous cells. Weak reactions might occur in the typing responses between a homozygous cell and an unrelated cell: ( 1 )if the typing cells share an HLA-D specificity with the responder cells and ( 2 ) differ at other genetic loci within the HLA complex. Studies of HLA recombinant families have indicated that weak MLRstimulating determinants may be present with the HLA-A,B segment of the HLA complex. These so-called weak MLC determinants have not

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

145

been conclusively identified. When typing unrelated individuals, it might also be possible that the HLA-D determinant, represented by the homozygous stimulating cell, has a specificity that is only cross-reacting and not identical to the specificity of the responder cell. In a first-cousin marriage family, an HLA-homozygous child should be genotypically homozygous by descent. The typing responses elicited by this homozygous cell against heterozygous family members should be free of noise that might result from possible disparities at HLA-D or at other HLA loci. Nevertheless, even within an inbred family, the typing responses are of the same level of magnitude as typing responses between unrelated individuals. Therefore, the weak responses characteristic of typing responses cannot be completely explained by weak disparities at other HLA loci or by crossreactions between similar HLA-D specificities, Such disparities, however, may indeed cause weak MLR. Experimental animals provide a useful model for the study of interaction between homozygous and heterozygous cells in MLC. Parental and F, hybrid cell combinations are considered to represent unidirectional MLR, because either parental cell donor is homozygous for the histocompatibility haplotype for which the F, cell is heterozygous. Wilson et al. (1967) observed, by karyotypic analysis of combinations of peripheral blood lymphocytes, that no proliferation of F, cells occurred, and Jones ( 1972) confirmed this by identification of histocompatibility antigens on the blast cells in parental-F, cell combinations. When spleen cells were used in such studies, however, it was found that some F, cell proliferation did occur, and, under some conditions, the degree of F, proliferation could equal the proliferation of parental cells ( Huemer et aZ., 1968; Alder at al., 1970; Harrison and Paul, 1973; Piquet et al., 1975). The proliferation of F, cells in these combinations was dependent on the presence of parental T cells (Harrison and Paul, 1973; von Boehmer, 1974). It was shown that proliferating F, cells consisted primarily of B cells ( Piquet et al., 1975). Similar observations have been made in MLC between the spleen cells of nude mice, and mitomycin C-treated or irradiated, normal, allogenic spleen cells, indicating that cell proliferation may occur among B cells (Wagner, 1972; Croy and Osoba, 1973). It is possible that secretion of mitogenic factors may activate cells from nude mice against allogenic stimulator cells. Although the stimulator cells have been treated with mitomycin C or irradiation to block DNA synthesis, they may be capable of reacting against the nude cells or against Ft cells by secretion of lymphokines. The B cells of the nude mouse or the cells of the F, animal may be induced by these lymphokines into cellular proliferation. This phenomenon has been demonstrated using spleen cells or lymph

146

DUPONT,

HANSEN, AND YUNIS

node cells as stimulator cells; it has not been seen when peripheral blood lymphocytes are used. The weak cell proliferations responsible for the typing responses in HLA-D typing experiments have not yet been satisfactorily explained.

E. CHARACTERIZATION OF HLA-D SPECIFICITIES The HLA-D specificities defined today are based on a collaboration organized by the Sixth International Histocompatibility Workshop; a Joint Report (Thorsby and Piazza, 1975) from the Workshop appears in Kissmeyer-Nielsen ( 1975). The combined experiments of the Workshop led to definition of eight different groups of HLA-D specificity. These results were, in general, consistent with previous work from individual laboratories in which some of the specificities were originally identified (Table I ) . Jgrgensen et al. (1973) described a first-cousin marriage family in which two children who were HLA-A2, B27-homozygous could also be shown to be HLA-D-homozygous. When the cells from the homozygous child were used as typing cells against a panel of responders, it clearly gave a bimodal distribution of responses. The HLA-D determinant for which this homozygous cell was typing was named the J determinant. After the Sixth International Histocompatibility Workshop, the J determinant was assigned to the HLA-DW1 group of specificities. Jgrgensen et al. (1974b) showed that this J determinant existed in the population in linkage disequilibrium with HLA-B35 ( W5). A second HLA-D specificity was defined independently in two different laboratories. This specificity was represented by HLA-homozygous cells obtained from different families of nonconsanguineous parents. These parents shared the HLA-A3, B7 haplotype. The MLR family studies showed that the children homozygous for the A3, B7 haplotypes gave typing responses against family members heterozygous for A3, B7. The determinant representing this specificity was called Pi by Mempel et al. ( 1973a,b) and LD-7a by Dupont et al. (1973a,b). This specificity is now known as HLA-DW2. More than one unrelated individual was found to be homozygous for (HLA-DW2, Pi, or LD-7a), and these different typing cells were mutually nonresponsive in MLR. In typing experiments against a large panel of responders, the specificity represented by these homozygous HLA-DW2 cells was strongly associated with HLA-B7. The gene frequency in the Caucasian population was estimated to be 0.07-0.10 (Mempel et al., 1973a,b; Dupont et nl., 1973a; Jersild et al., 1973b; Albert et al., 1973). Several different D-homozygous cells representing these and additional specificities were subsequently identified in a number of laboratories :

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

147

HLA-DW1 by Sachs et al. (1974), Mempel et al. ( 1973b), and Keuning et al. (1975a); HLA-DW2 by JGrgensen et al. (1974b), Keuning et al. (1975a); and Sasportes et al. (1973); HLA-DW3 by Keuning et al. (1975a), Thomsen et al. (1975a), L’Espkrance et al. (1975), and Page et al. (1975); HLA-DW4 by JGrgensen et al. ( 1974b), Thomsen et al. (1975a), Keuning et al. (1975a), Stastny ( 1974), and L’Espkrance et al. ( 1975). During the Sixth Histocompatibility Workshop, definition of the different HLA-D specificities was based on a computerized analysis of the typing response pattern against a large panel of responder cells and by mutual MLC testing between all typing cells. A two-by-two comparison was made between each typing cell, and if there was a similarity in pattern of response against them, they were assigned to the same group of specificities. Sixty-two typing cells were included against 741 different individuals on the responder panel. Eight different groups of HLA-D specificities were defined (Table I ) . These were provisionally designated as DW1, DW2, DW3, DW4, DW5, DW6, LD-107, and LD-108. The correlation in typing response between cells believed to be typing for groups DW1 through DW6 was statistically good. The specificities representing HLA-D groups LD-107 and LD-108 were less well-defined.Thirtyeight of the sixty-two typing cells could be assigned to one of these eight groups. Each group comprised two to six different cells. The twenty-four cells that could not be assigned to a group might possibly each represent a distinct specificity. Alternatively, a typing cell might not be assigned a specificity because of poor discriminatory power, because of technical failure, or because they might not be HLA-D-homozygous. The association between typing cells in each specificity group is only partial. No pair of typing cells within the same specificity group was found to type in exactly the same way, and, because of this, the provisional assignment of HLA-D specificities should be considered as groups of relatively similar but not identical HLA-D determinants. Mixed-lymphocyte culture testing between D-homozygous cells assigned to the same group usually shows mutual typing responses and sometimes identity responses. There are exceptions, however, in which mutual MLR is strongly positive between two cells typing for the same specificity (J. A. Hansen, et al., 1975). Sometimes MLR is positive in ony one direction, suggesting an inclusion phenomenon. Some HLA-D specificities are clearly associated with certain HLA-B antigens (Table IV) : DW1 ( J , Pf ) shows significant linkage disequilibrium with BW35( W5); DW2 (LD-7a) is strongly associated with B7; DW3 (LD-8a) is strongly associated with B8; DW4 shows significant association with BW15; and DW5 shows significant association with BW16.

148

DUPONT,

HANSEN, AND YUNIS

TABLE I V FREQUENCY OF HLA-D SPECIFICITIES A s DEFINED BY T H E SIXTHINTERNATIONAL HISTOCOMPATIBILITY WORKSHOP^ Gene HLA-I) Antigen specificity groupb frequency ( %) frequency ( p ) “ ~~

DW1 DW2 DW3 DW4 DW5 DW6 “Blank”

19.3 15.2 16.4 15.6 14.6 10.5 -

~

Most significant HLA-B association

Delta valued

~

0.102 0.078 0.085 0.082 0.075 0.054 0.524

BW35(W5) B7 B8 BW15 BW16 BW16c -

0.021 0.031

0.044 0.017 0.013 0.007 -

Frequencies for the six provisional HLA-D specificity groups calculated from the pooled data of the Sixth International Histocompatibility Workshop. The responses of 171 random Caucasian donors were available for analysis. A specificity was assigned to a responder if a t least 50% of the typing cells belonging to a specific DW group elicited typing responses from that responder. Only 4 % of the responders gave triplets. b (DW1-DW6) provisional HLA-D specificities; (“Blank”) number of haplotypes that were not assigned a DW specificity. 5 Calculated using the method of maximum likelihood. d The coefficient of linkage disequilibrium (delta value) was calculated using the formula of Mattiue et al. (1970). d An association of borderline significance. Q

F. COMPLEXITY OF THE HLA-D Locus : CROSS-REACTING SPECIFICITIES VERSUS MULTIPLE SUBLOCI Population studies and segregation studies of HLA-D specificities in families have indicated that the DW1-DWG determinants represent mutually exclusive and non-cross-reacting groups of D specificity (Thorsby and Piazza, 1975). Earlier, MLR between different homozygous typing cells representing the same D-specificity group indicated that the more common specificities might be quite restricted. As new D-homozygous cells were identified, it became obvious that even the common D specificities were more heterogeneous than originally assumed. No two typing cells in any specificity group elicit exactly the same pattern of responses from a panel. Mutual MLC testing between typing cells belonging to the same specificity group can show variable MLR. Some combinations are nearly nonresponsive, some give mutuaI typing responses, some combinations are positive in one direction, and some combinations are strongly positive in both directions. The MLR responsiveness between homozygous typing cells from the same DW specificity group reflects some disparity between the cells. The question to be resolved is whether these disparities in each

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

149

DW group reflect differences at the D locus or at other genetic determinants (J. A. Hansen, et al., 1975; Reinsmoen et al., 1975; Thomsen et al., 1975b; Keuning et al., 1975a,b). This complexity of the D specificities is also evident in the Workshop material. The behavior of typing cells representing the same DW group against a responder panel and against each other has led to several descriptive phenomena. Sometimes, when two typing cells from the same DW group are used to type a panel, one of the cells will give a larger number of typing responses than the other. This has led to the concept of broad and narrow D specificities. A typing cell that elicits the most frequent number of typing responses is said to represent a narrow specificity. If there is complete overlap in the typing response patterns between a broad and narrow cell, the specificity of the narrow cell is said to be included in the specificity of the broad cell. This inclusion phenomenon implies that the identification of a D specificity on one haplotype by a single typing cell does not necessarily define the complete D specificity of that haplotype. A situation has been described, for example, in which some unrelated individuals who share an HLA-A3, B7, DW2 haplotype, but differ at their second haplotype, may show no MLR in one direction while being positive in the other direction. This has been interpreted as resulting from inclusion of the same D determinant on each of the two nonidentical haplotypes ( Dupont et al., 1973b). A possible model for the interaction between broad and narrow D specificities is given in Table Va. Responding cells that carry the broad specificity would give typing responses to both broad and narrow typing cells. The determinant of the narrow cell is included in the determinant of the broad cell. In mutual MLR, the broad cell will stimulate the narrow cell, but the narrow cell elicits a typing response from the broad cell. These observations of broad and narrow D specificities may be analogous to the inclusion phenomenon described in HLA serology. A cross-reacting group of antigens belonging to the same allelic system react with a broad HLA antiserum directed against a shared determinant common to each member of the group. A narrow or monospecific antiserum reacts with that part of the allelic product which is unique for that specificity. The possible explanations of broad and narrow D specificities include: ( I ) Broad and narrow D specificities may represent different allelic products that share a common cross-reacting determinant; ( 2 ) broad and narrow D specificities may represent the complex products of subloci or pseudoalleles within the D locus; ( 3 ) they may represent disparities at other genetic determinants. A second kind of inclusion phenomenon has been observed between

150

DUPONT, HANSEN, AND WNIS

TABLE V HLA-D INCLUSIONS a. INCLUSION PHENOMENA TYPEI (BROADA N D NARROW SPECIFICITIES W I T H I N T H E SAMEHLA-D GROUP)= Responder cellsb

HLA-11-homozygous stimulator cellsc ab/ab a/a

Homozygous

ab/ab

-

-

Heterozygous It1 _Heterozygous R,

ab/ -

-

-

-/ -

+

+

b. INCLUSION PHENOMENA TYPIC 11 (HLA-D SPECIFICITIES CONTROLLED BY A GENETIC SYSTEMO F SEVERAL CLOSELYLINKEDGENES)^ HLA-D-homozygous Responder cellsb Homozygous Homozygous Heterozygous Heterozygous Heterozygous Heterozygous Heterozygous

y/y

z/z

R, Y Z / RZ Y / Z R, Z/R4 Y / Rs -/ -

Stimulator cellsc

y/y

z/z

-

+

1 -+ -

+

-

+ +

0 Narrow and broad D specificities are designated a and ab. This pattern of typing responses is consistent with the concept of broad (ab) and narrow ( a ) D specificities representing different allelic products that share cross-reacting determinants. * R1,Rz,R3, R4,and R, are HLA-D-heterozygous responder cells. c (-) Typing response or weak MLR; (+) positive MLR (no typing response). d The Y and Z represent the specificities of two HLA-D-homozygous typing cells that belong to the DW4 specificity group. The typing cells ( Y and Z) are mutually MLR-positive. The reaction pattern of typing responses illustrated in this table indicate that the two typing cells define different D specificities, The frequent observation of similar typing response patterns for the two cells is consistent with the hypothesis that the D specificity is defined by a genetic system of closely linked genes. These two typing cells would then identify D Specificities that are in strong, positive linkage disequilibrium in the population.

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

151

two cells within the DW4 specificity group. These two cells, which frequently elicited similar typing responses from a panel, react strongly in mutual MLR (J. A. Hansen, et al., 1975). A proposed model for the inclusion phenomenon demonstrated by these two cells is given in Table Vb. The two typing cells for which this phenomena is observed are HLAA2,B12,DW4-homozygous and HLA-A2,BW15,DW4-homozygous. The two cells frequently give typing responses with the same responder cell, but some responders will only be typed with one cell whereas others will be typed with the other cell. It is characteristic that the BW15,DW4homozygous cells most frequently show isolated typing responses against responder cells with the BW15 antigen, whereas B12, DW4 typing cells sometimes show isolated typing responses with Bl2-heterozygous responder cells. The two typing cells belong, however, to the DW4 specificity group. One interpretation of these data might be that these typing cells share a common determinant that identifies them as belonging to the DW4 group. The mutual positive MLR between them, however, indicates that each cell must represent some unique D specificities in addition. Another interpretation of these data is possible. The D specificity on one haplotype could be determined by a system of closely linked genes. The determinants on the subloci within this system could be in strong positive genetic linkage disequilibrium in specific combinations. The two mutually MLR-responsive DW4 typing cells would then represent different D determinants at two different subloci. These two different determinants would be found in the population in strong positive linkage disequilibrium. Another possible explanation for the lack of response to a typing cell is that the responder cell may lack the appropriate gene necessary for response against a specific HLA-D determinant. Yunis and Amos (1971) suggested that such response genes could be similar to the immune response (Zr) genes that are postulated to exist within the HLA complex. The genetic control of MLR would then be determined by genes coding for stimulating determinants ( HLA-D ) and genes coding for responsiveness to specific HLA-D determinants. The presence of both stimulation and response genes would explain why some HLA-D-heterozygous cell combinations show positive MLR in only one direction (Dupont et al., 197313). They would also provide an alternative explanation to the typing response patterns presently interpreted as representing HLA-D inclusions. Support for this concept comes from the study of one family (Dupont et al., 1975a). In this family, it was shown that one maternal HLA-haplotype ( c ) induced strong stimulation of one paternal HLAhaplotype ( b ) in some combinations, but the same c haplotype did not induce stimulation of the b haplotype in other combinations (no MLR

152

DUPONT, HANSEN, AND YUNIS

response in combination cld stimulating b l d ; positive MLR in combination alc stimulating a / b ) (Table V I ) . In the same family, the paternal HLA-D specificity on the b haplotype could not stimulate in any comTABLE VI HLA GENOTYPES OF FAMILY SCH Subject

F M

First haplotype

Al,B8, A3, BW17, Al,B8, A9, B12, Al,B8, Al,B8, Al,B8,

Ch 1 Ch 2 Ch 3 Ch 4 Ch 5

DW3 D-17 DW3 D-Y DW3 DW3 DW3

Second haplotype

IILA genotype

AS, B12, I)-Y A2, B12, I)-I*,II** A2, B12, 11-17 A2, B12, I)-1,II A2, B12, I)-1,II A3, BW17, D-17 A3, B17, D-17

alb cld

ale X

d

bld ald alc ale

MIXED-LYMPHOCYTE CULTURE REACTION I N SELECTED COMBINATIONS

MLR

W%)

Responder cell

Stimulator cell

Haplotype tested

(cpm)

Ch 2 bld Father alb

Mother cld Ch 4 or Ch 5 alc

c-+ h

776

4

c--, b

7701 767 1

51 51

a/c X d

cxd-tb

2631

16

Father alb

a This HLA-B/D recombinant family was originally presented by Yunis et al. (1971) and Yunis and Amos (1971) as evidence for a separate MLC-stimulating locus. By HLA-D typing it has been demonstrated that the paternal Al,B8 haplotype carries the DW3 specificity. The maternal A2,B12 haplotype carries the D specificity (I*) identified by typing cell SFN-2 (J.A. Hansen et al., 1975; Suciu-Foca and Dausset , 1975) and in addition a D determinant belonging to the DW3 specificity group (II**) (J.A. Hansen et al., 1975). The maternal haplotype A3,BW17 carries the D specificity D-17 (defined by Grosse-Wilde et al., 1975, typing cell EI). The B / D recombination is thus proven by direct identification of the B and D specificities on the recombinant haplotype. The MLR demonstrates that the c haplotype stimulates the b haplotype in two combinations (a/b vs a/c), but does not stimulate in the b/d vs c/d combination. The recombinant haplotype (c X d ) has lost the capacity to generate strong MLR against the a / b cells (Dupont et al., 1975~). * (a, b, c, and d ) Denote HLA haplotypes; (c X d ) denotes HLA-B/D recombinant haplotype; (+) denotes direction of stimulation in MLR. Uncharacterized D specificities are designated X and Y.

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

153

bination the maternal c haplotype. It could, therefore, be concluded that the HLA-D determinant of the c haplotype demonstrated the inclusion of the HLA-D specificity of the b haplotype. The MLC testing in this family demonstrates an asymmetrical response pattern consistent with the concept of inclusion of multiple HLA-D specificities on a single HLA haplotype. One of these haplotypes expresses HLA-D inclusions that only induce MLR activation in some combinations. The responder cell in the MLR-positive combination was HLA-B locus-heterozygous, whereas the MLR-nonreactive combination was HLA-B12-homozygous. This suggests that MLR activation may also depend on the immune response capacity of the responder cell. In this family, there was also an HLA-BID recombinant child. The HLA-B/D recombination occurred in the mother and gave rise to a recombinant HLA haplotype with the HLA-D determinant from the maternal c haplotype and the HLA-A,B determinants of the maternal d haplotype. This recombinant haplotype did not generate positive M t R in the cell combination showing positive MLR to the original c haplotype (e.g., no MLR in combination a l c x d stimulating alb, but positive MLR in combination alc stimulating a / b ) , The recombinational event between HLA-B and D may, therefore, have separated the HLA-D inclusion from the remaining part of the D specificity of the c haplotype (e.g., the maternal recombination occurred between A3, B17, D-17, X and A2, B12, D-Z, leading to the recombinant haplotype A2, B12, D-17) (Table V ) . During the Sixth International Histocompatibility Workshop, eight provisional groups of HLA-D specificities have been identified (DW1DW6, LD-107, and LD-108). Each group has been defined by two or more D-homozygous typing cells, None of these homozygous cells, however, behaved in the typing experiments in a completely identical way. Differences between cells presumably belonging to the same group could also be shown in mutual MLC. There are at least three different possibilities for explaining these findings: ( I ) Each provisional group of D-locus specificity represents broadly cross-reacting clusters of alleles of the same locus; ( 2 ) the cross-reacting D specificities represent determinants of closely linked genes or subloci within the same genetic system, and the clusters represent determinants in strong positive linkage disequilibrium in the population; or ( 3 ) the clusters represent typing cells with different HLA-D specificities for which the responder cells cannot generate a proliferative response because of restrictions in immune response genes. This last hypothesis is based on the assumption that MLR activation involves immune response genes in the responder cells in addition to disparities on the HLA-D determinants between stimulator and responder. This hypothesis is consistent with the findings in the mouse

154

DUPONT, HANSEN,

AND YUNIS

that many .cell combinations only show stimulation in one direction [e.g., BlO.A(4R) responds to BlO.A(BR), whereas BlO.A( 2R) fails to respond to BlO.A(4R)] (Bach et al., 1972b). Such asymmetrical response patterns in human MLR need very careful analysis including studies of segregation patterns within the family. The further analysis of inclusion phenomena and the question of possible involvement of immune response genes in MLR may come from the study of families in which one family member shows typing responses to typing cells from more than two of the presently defined HLA-D specificity groups. G. FAMILY STUDIESWITH HLA-D-HOMOZYGOUS TYPINGCELLS The segregation of the HLA-D specificities in families can now be studied directly in typing experiments with HLA-D-homozygous cells. Only one HLA-D specificity was included in each of the early segregation studies. Jgrgensen et al. (1973) studied the segregation of a DW1 specificity. Jersild et al. (1973b) and Dupont et al. (1974a) studied the segregation of a DW2 specificity. It was clearly shown that the HLA-D specificity defined by a typing cell segregated in the family with one HLA haplotype and was the first direct evidence of genetic linkage between HLA and the MLC-stimulating determinants. Grosse-Wilde et d. (1975) have recently reported a large series of family segregation studies consisting of 74 families with 205 children. The families were analyzed with typing cells representing the DW1, DW2, and DW3 specificities. Forty-two informative backcross matings could b e analyzed. The distribution of positive and negative haplotypes seemed to be balanced. For these three groups of HLA-D specificities, no more than one D determinant on a single haplotype was ever identified. The observation that the DW1, DW2, and DW3 determinants did not occur together on one haplotype reaffirms that typing cells belonging to the DWI, DW2, and DW3 specificity groups are mutually exclusive. Family studies, however, cannot distinguish between the segregation of true alleles or the segregation of determinants controlled by two closely linked series of subloci or pseudoalleles. The DW specificity groups, as they are now defined, could represent a complex of closely linked genes with a high degree of positive linkage disequilibrium in the population. WITH RECOMBINATIONS H. HLA-D TYPINGOF FAMILIES WITHIN THE HLA COMPLEX

The concept of one locus, independent of the serologically defined HLA antigens, controlling MLR activation in man came from the study of just a few informative families exhibiting recombination within the HLA system (Section IV). The identification of the MLC gene was indirectly

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

155

proven by comparing the results of serological typing with the results obtained in MLC testing between the family members. The finding that most unrelated seroidentical combinations are MLR positive also supported the concept of a separate locus for MLC. The accepted proof of a separate locus rests on observed HLA recombinations, which are presumed to separate HLA-B from HLA-D. Recombinations between the HLA-B and the HLA-D have primarily been claimed in three types of families : 1. The MLR-positive sibling combinations that are genotypically identical for the serologically defined HLA antigens. If other sibling combinations are not available for study, the recombination might occur within any of the two haplotypes present in the children (see Table VII, Group I ) . 2. The MLR-identical sibling combinations, in which siblings are genotypically different for one of the two serologically defined HLA haplotypes. In such families the recombination could take place on either of the two HLA haplotypes. If there is a third child available who is serologically identical with one of the other children, and these serologically identical siblings are MLC identical, then it can be claimed that the HLA-incompatible haplotype carries the recombination ( see Table VII, Group 11). 3. Families with two different pairs of HLA-seroidentical siblings. The two pairs differ for one HLA haplotype. One pair, with haplotypes a and c, is MLR identical; the other pair, with haplotypes a and d, is MLR responsive, but one of these siblings is MLR identical with the two siblings of the a / c haplotypes. The HLA-B/D recombination is then assigned to the d haplotype of the child who was MLR identical with the a / c siblings (Fig. 4 and Table VIII). The last situation provides the best, although indirect, evidence for a recombination between the B and D locus and assigns the recombination to one specific haplotype. Estimation on the basis of family studies of the recombinant fraction between the HLA-B locus and the HLA-D locus, which at best gives only indirect evidence for identification of recombination, is difficult. All instances of recombination will not be appreciated if informative siblings are not available for testing. Only 6 families with well-established indirect proof of HLA-B/HLA-D recombination have been reported (Table VIII). One additional family has been reported by Mempel et a2. (1972), but in that family the recombination could have occurred between the HLA-A and HLA-B locus. The recombination involved the maternal HLA haplotypes, which were both carrying the HLA-BW35 determinant.

156

DUPONT, HANSEN, AND W N I S

ATYPICALMLR

IN

TABLE VII FAMILIES PRESUMED TO REPRESENT IILA-B/D RECOMBINATION^

Group*

MLR resultse

References

I

Positive MLR between one pair of genotypically HLA-seroidentical siblings:

Bach and Amos (1967), Amos and Bach (1968), Bach el al. (1969), Plate el al. (1970)

a/c

I1

+ a/c

MLR identity between one pair of HLA-seroidentical siblings and one HLA serohaplotype-different sibling, MLR-identical to the others:

Dupont et al. (1972), Dupont el al. (1974a)

a / c - a/c - a/d MLR positivity between one pair of HLA-sero- Sasportes et al. (1972) identical siblings and one additional sibling differing from the others on one HLA-A, B haplotype (this last sibling is MLR-identical to one of the other siblings):

alb--rcld + -

a/c a/d Three genotypically HLA-seroidentical siblings (two siblings are MLR-identical, and the third sibling is MLR-positive with the other two) : a/c

a/c

Dupont et al. (1974a)

- a / c ia/c

a A summary of the published family .studies originally interpreted as representing possible HLA-B/D recombinations. These family studies and the families presented in Table VIII represent the presently published families with evidence for HLA-B/D recombination. A few additional family studies may occur in “Histocompatibility Testing 1975” but have not been available to the authors a t the time of writing. * Group I represents only HLA-seroidentical sibling combinations with positive MLR. Group I1 consists of families with three siblings. Two families have one pair of HLA-identical, MLIt-nonresponsive siblings and a third sibling differing for one HLA-A, B haplotype, but MLR-identical to each of the other siblings. Another family has one pair of HLA-seroidentical but MLR-positive siblings and a third sibling differing for one HLA-A, B haplotype. This last sibling is MLR-identical with one of the other siblings. The last family has three HLA-seroidentical siblings of which one is MLR-positive with each of the two other siblings. (a, b, c, and d ) HLA serohaplotypes; a plus (+) between two cell combinations denotes positive MLH; and a minus ( - ) between two cell combinations denotes MLR identity or very weak MLR.

P.\RENT.%L

TABLE 1.111 HLA H.\PLOTYPE;S A N D HLA-B/D RECOMBINANT H.\PLwrYPm I N FAMILIES WITH I N D I R E C T P R O O F O F KECOMBINATIONa

Paternal HLA haplotypes A9, B12, 11-Y A l , B8, D-X All,WYS, D-Y A2, B8, D-X A3, B7, D-Y A2, g 2 , D-X A3, I37, B Y ) (A2, BWlT,, I>-X* AW26, BW22, 1)-X i l l , B8, D-Y A2, B13, D-Y) ( A l , B8, D-X

Maternal HLA haplotypes A.3, B17, D-Z * A3, B7, D-Z* AW24, B?, D Z * A3, B7, 1)-Z ilW26, BW22, D-Z* A41,BX, D-Z*

z

SIX

INFORMATIVE

Recombinant haplotypes

A2, B12, D-Q A2, B12, I)-Z A l , B8, D-Z .41, A 3 , 8 7 , D-Q A3, B7, D-Z A428,BW40, D-Q A3, B7, D-X AW24, BW17, D-Q AW26, BWB'L, D-Q A2, B7, D-Q Al, B8, D-Q

KD-Q

C

5

2

References Yunis et al. (1971) Eijsvoogel et al. (1972b) Sasportes el al. (1972) Thorsby et a / . (1973b) Keuning ~t a/. (1975b) Keuning el al. (1975b)

= The HLA-D locus specificities are indicated as X and Y for the paternal specificities and Z and Q for the maternal specificities. The asterisk (*) indicates whether the HLA-B/D recombination has occurred between maternal or paternal haplotypes. If the parents share one HLA-B locus antigen or share the HLA-A, B haplotype, this is indicated by underlining the shared antigens or haplotype. Parentheses indicate that the IILA genotypes of these individuals have been deduced from the family study since the individuals were unavailable for studies.

n

2

158

DUPONT, HANSEN, AND YUNIS

Among the 6 informative families with evidence of recombination, only

1 involved a paternal recombination. In 4 of the families the parents shared one HLA-A, B haplotype, and in the other families the parents shared the HLA-B12 or the B8 determinant, This sharing of parental haplotypes cannot influence the occurrence of the recombination, but it could have implications as to alternative explanations for the results obtained in the family MLC tests. The development of HLA-D typing with homozygous cells has made it possible to evaluate the possible HLA-B/ HLA-D locus recombinants by a direct method. Formal proof of a recombination should be based on the positive identification of one allele on each of the two genetic loci between which the recombination has occurred, and the products of each locus should be identified on the recombinant haplotype. This has only been shown in two HLA-B/D recombinant families. The father, however, was not available for study in one family (Goulmy et al., 1975) (Tablc V I ) . In view of newly acquired knowledge about the determinants of the HLA-D locus, some “atypical” or “abnormal” MLC family studies previously reported as showing unexpected lack of MLR between different family members can now be reinterpreted. This can be illustrated by using as an example the MLC family study previously reported by Dupont et al. (1972, 1973a, 1974a). It was originally concluded that this family demonstrated an HLA-B/ HLA-D locus recombination in one child. This was based on the observation that one pair of HLA-seroidentical, MLC-identical siblings and their 1 h-different sister were MLR identical. The pedigree for key family members of this family ( K J ) is given in Fig. 8. The propositus, 111-2, is a child with severe combined immunodeficiency disease. The parents, 11-3 and 11-4, are unrelated husband and wife. The father 11-4 is DW2-homozygous, in spite of the fact that he is HLA-A, B-heterozygous (Dupont et al., 1973a,b,c, 1974a). Individuals 11-2, 11-3, and 111-1 are DW2-heterozygous, but the paternal grandfather, 1-1, does not carry DW2. The family MLC study is summarized in Table IX. It can be shown that 1-1 is behaving as a D-locus-homozygous typing cell. Grandfather 1-1is HLA-A, B-heterozygous, but his cells are incapable of stimulating the cells of his children or the cells of the patient’s sibling 111-1, who has inherited the A2, BW35 haplotype from 1-1. Since 111-1 has inherited the paternal A3, B7 DW2 haplotype, it is possible to show that the A2, BW35 haplotype carries the same D-locus specificity as the A2, B12 haplotype by testing combinations 11-2 vs 11-3and 11-2 vs 111-3.The MLC identity in these combinations confirms the hypothesis that 1-1 is HLA-D-homozygous. This example illustrates that the observation of MLC identity between HLA-haploidentical siblings does not constitute proof for HLA-B/ HLA-D recombinations.

HUMAN MIXED-LYMPHOCYTE CULTURE REACI'ION FAMILY

159

KJ

I

A2, B W 5 5 . 0 - X A1,BI2,D-X

11-1

~ ~ , m i D-x a, A a, ~ 7 D,w a

(AS,B7,DWD Al,BO,O-Y)

11-2

A2.Bl2,D-X ~ a , ~ owa 7 ,

11-4

Aa,B55,0-X A2,B7 , O W 1 111-1

AS,B7,DWP AlO,DlO, O W 2 Ill?

Ill A3,B7, D W 2 A2,BWSI,D-X

AIO, mia, o m AS,BW55,0-X

FIG. 8. Pedigree of family KJ previously described by Dupont et al. (1972, 1973a, 1974a) and C. Koch et al. (1973). The propositus (111-2) has severe combined immunodeficiency disease. The HLA genotypes are given as HLA-A, B, D specificities. The genotypes of 1-2 are deduced from HLA serotyping of additional family members. Individuals 11-3 and 11-4 are unrelated; 11-4 is HLA-A, B-heterozygous but HLA-D-homozygous for the DW2 specificity (Dupont ct al., 1974a). The DW2 specificity has been identified in 11-1, 11-2, 11-3, and 111-1 in typing experiments which included both 11-4 and other DW2-homozygous typing cells. The D specificity D-X has not been identified in typing experiments, but the family MLR indicates that 1-1 is homozygous for D-X. The segregation of the D-X specificity in the family is deduced from the family MLC testings summarized in Table IX.

Only a few families with HLA recombinations have been studied with HLA-D-homozygous typing cells. Rittner et al. (1975a) reported on 2 families with HLA recombination in which D-locus typing showed that the D specificity segregated with the HLA-B determinants. In both families, however, the recombination occurred between the two maternal haplotypes, both of which had the same HLA-B determinant. Therefore, the precise site of recombination could not be determined. Sasazuki et al. (1975) studied an HLA-A/B recombinant family, but the typing cells used could not identify the recombinant haplotype. Similarly, 5 families studied by Suciu-Foca et al. (1975) were uninformative with respect to the segregation of an identifiable D-locus determinant. Informative studies in HLA-B/ D recombinant families have been demonstrated by Keuning et al. (1975b,c), Mawas et al. (1975b,c), and Dupont et al. ( 1 9 7 5 ~ ) .In the family studied by Mawas et al. (1975a), the recombinant haplotype acquired the DW2 specificity, but the HLA-B specificity could not be determined. In the family studied by Dupont et al. (1975c),

160

DUPONT, HANSEN, AND YUNIS

TABLE I X MIXED-LYMPHOCYTE CULTIJRlC ItEACTION HLA-A, B,

L)

B

1-1 2W35 11-1 2 12 11-2 2 12 11-3 2 W35 111-1 3 7 111-2' 10 18 II-4b 10 18 DW2 cellb 3 7

Second haplotype L)

X X

X X W2 W2 W2

FAMILY KJ

genotypes

First haplotype Responder A

IN

A

B

1)

Stimulating cellsa

1-1 11-1

+

2 12 X A S 2 7 W2 - AS 2 7 W2 2 7 W2 X - n.d. 2 W35 2 W35 X 3 7 W2

W2 3

7 W2

+ + + n.d.

11-2

+ -

AS -

+

+

11-3 111-1 111-2 II-4b

+

+

+- +-

DW2 cellb

+

AS -

n.d. AS

-

-

n.d.

-

-

+ +

+ +

+ +

AS

-

-

AS

-

n.d. -

-

a (n.d.) Not done; (+) positive MLR; ( - ) very weak MLR; (AS) autostimulated control culture. 11-4 and DW2 cell: HLA-DW2-homozygous cells. c Individual 111-2 has severe combined immunodeficiency disease and is, therefore, not included as a responder cell. Immunological reconstitution of the lymphoid T-cell system and partial reconstitution of the B-cell system has resulted following bone marrow transplantation with maternal uncle 11-1 as bone marrow donor. According to this scheme the donor and recipient have inherited unrelated EILA haplotypes, which, however, carry the same HLA-D specificities, DW2 and D-X.

the D specificity carried by the HLA-A2, B12 haplotype of the mother could be defined by two HLA-D-homozygous cells. In the recombinant child, the involved haplotype, A2, B12, no longer carried one of these D specificities and acquired the D-17 specificity (see Table VI). The segregation of D specificities in a presumed recombinant family reported by Keuning et al. (1975b) is also in agreement with the concept of recombination between HLA-B and a separate locus for MLC (HLA-D). Only the family presented in Table VI and the family studied by Goulmy et al. (1975) proves the HLA-B/D recombination. It has previously been suggested that some genetic factors between HLA-A and HLA-B may contribute to at least weak MLR activation (Eijsvoogel et al., 197213; Thorsby et al., 1973b; Dupont et al., 1974a). The contribution of this additional genetic factor( s ) to the MLR activation has not been proven. An additional family suggesting the existence of an MLR-stimulating determinant associated with the HLA-A locus has recently been reported by Johnson et al. (1975). This family study demonstrates a unique finding that an HLA-A/ B recombinant child (11-1,Table X and Fig. 9 ) is strongly responsive to an HLA-B locusidentical sibling. The HLA-D typing in this family indicates that the

161

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

TABLE X MIXED-LYMPHOCYTE CUI.TURER.EACTION IN FAMILY HuR HLA-A, B genotypes First haplotype

Second haplotype

Stimulating cells*

Re- - sponder A B A B 1-1 1-2 11-1 11-2 11-3 11-4

2 11 2 2 9 2

18 9 W35 28 7 11 18 28 7 11 18 28

1-1

7 D W 2 W35 w35 W35 W35 DW2 W35

AS

+ + + + +

1-2

11-1

11-2

11-3

+ + + + + AS + 2 + AS + + AS 2 + + +

+ + + +

+

AS

-

11-4

DW2 cell'

+ + + +

-

+

2

+ +

AS

The HLA-A/HLA-B recombinant child (11-1) is HLA-B-identical to sibling 11-3. However, the MLIt between 11-1 and 11-3is positive. The l)-lorus typing with I)W2 typing cell identifies the paternal haplotype IILA-AS, 13, 7, 11W2 (in the fathrr 1-1 and rhild 11-3). Recombinant child 1-1 shows a strong positive response to the DW2 typing rrll. (15. Yunis, unpublishcd data.) * (+) Positive MLR; (-) MLC identity; (2) denotes the critical MLR-positive combinations; (AS) autostimulated control culture.
A2,87/All,BW35* not DWl

A2, B18/A28;8W35

HU

A9, B7/AIl,BW35* DW2

A2,BIa /A28,

BW35

FIG. 9. Pedigree of family HU. Child 11-1 has a paternal recombinant haplotype A2, B7. This recombinant haplotype is also transmitted to the child of 11-1. The HLA-B antigen W35" was defined as a subspecificity of BW35. Four antisera could differentiate the W35 antigen on the A l l , BW35" from the W35 antigen on the A28, BW35 haplotype. A maternal recombination involving the BW35' and BW35 can be ruled out (Johnson et al., 1975). Workshop serum No. 165 (Anner1)Histocompatibility Workshop 1975 also identified the BW35O as different from the BW35 in this family.

162

DUPONT,

HANSEN, AND WNIS

paternal haplotype A9, B7 carries DW2 and this is transmitted to one child, along with A9, B7 (11-3).The recombinant child (11-l), who is A2, B7, does not carry the DW2 specificity (Table X ) ( E . J. Yunis, unpublished data). Thus, HLA-D typing in this family seems to indicate that child 11-1is a double recombinant, with recombination between the HLA-A/B and between the HLA-B/D. Formal proof, however, by identification of the D specificity on the recombinant haplotype is not established. Two different studies that demonstrate more than one recombination in a single family have been reported. In both families, the second recombination appeared to involve the same maternal haplotype. In one family, both recombinations were assumed to have occurred between HLA-B and HLA-D, giving rise to 2 children with HLA recombinant haplotypes (Dupont et nl., 1974a). The other family also involved two recombinational events between maternal haplotypes. The location of the recombination, however, could not be established because of HLA-B locus homozygosity (Rittner et al., 1975a). This unusual finding in two different families might reflect a genetic “hot spot” within the involved haplotypes. In summary, very few informative families giving proof (although indirect) of HLA-B/D locus recombination have been described. The complete genotyping of the HLA determinants in an HLA-B/D recombinant family, especially for the determinants on the recombinant haplotype, has only been described in two families. Only such a study would constitute formal proof of recombination between the B and D loci. The reported occurrence of double recombinants within one HLA haplotype and the reports of two recombinational events in two different families add further complication to the interpretation of “unexpected” or “abnormal’’ MLC family studies. The mapping of the D locus was, until HLA-D typing became available, based on the analysis of MLC family studies in HLA-A/B and HLA-B/D recombinant families. The HLA-D typing in recombinant families seems thus to confirm the previous mapping of the HLA-D locus close to the B locus but outside the HLA-A, B segment of the HLA complex. The direct identification of the B and D specificity on the B/D recombinant haplotype has, however, only been presented in two families.

I. POPULATION STUDIESWITH HLA-D-HOMOZYGOUS TYPINGCELLS The HLA-D-homozygous typing cells that have been obtained from the outbred population seem in most instances to identify common HLA-D specificities (Mempel et al., 1973a,b; Albert et al., 1973; Dupont et al., 1973b,c; Jersild et al., 1973b). Some of the homozygous cells ob-

H U M A N MIXED-LYMPHOCYTE CULTURE REACTION

163

tained from the inbred population, however, seem to represent some relatively infrequent HLA-D specificities ( Kcuning et al., 1975a,b). The gene frequencies of the six D-specificity groups (DWl-DW6) were calculated from the combined Workshop material. There were 174 random Caucasian donors available for analysis. One of the DW specificities was said to be present if a responder gave typing responses to at Ieast 50%of the typing cells belonging to that group. By this method of assignment, the DW1-DW6 groups appeared to represent distinct and non-cross-reacting specificities. Only 4% of the responders gave triplets. These individuals were excluded from the calculation of gene frequencies. The accumulated gene frequencies for these six specificities was 0.524 (Table IV). The gene frequencies for the DW1-DW6 specificities (from the combined data of two Scandinavian laboratories) was 0.561 (Thomsen et al., 197513; Thorsby et al., 1 9 7 5 ~ )If. the gene frequencies for the two less well-defined HLA-D specificities, LD-107 and LD-108, are included, the accumulated gene frequency for eight HLA-D specificities in the Scandinavian material is 0.652. These preliminary data indicate that the first eight HLA-D-specificity groups, defined in the Workshop by 38 out of 62 submitted D-homozygous cells, account for approximately 65% of the HLA-D specificities in the Caucasian population. The Workshop data €or the frequencies of the DW1-DW6 specificities were tested in the Hardy- Weinberg equation, and the distribution of observed phenotypes fit the expected distribution. This supports the concept that the HLA-D specificities as defined (DW1-DW6) represent a single segregating system of alleles or a system of closely linked genes in strong positive linkage disequilibrium. Several observations indicate that the DW1-DW6 specificities behave more like a pseudoallelic rather than a single allelic system. These preliminary data seem to indicate that the HLA-D specificities identified by the presently described eight D-specificity groups account for approximately 50-70% of the D specificities in the Caucasian population. In order to add the gene frequencies for the individual D specificities, it must be assumed that they represent alleles within the same genetic system. There are, however, as previously discussed, a number of observations indicating that identification of one D specificity on one HLA haplotype does not necessarily describe the complete D specificity for that haplotype. It was, for example, observed that some unrelated individuals who shared one HLA haplotype, e.g., HLA-AS, B7, DW2, but differed on the other haplotype, would have strong positive MLR in one direction and no MLR in the opposite direction (Dupont et al., 197313).

164

DUPONT, HANSEN, AND YUNIS

One of the approaches to clarify the question of how well the already defined non-cross-reacting DW specificities identify the D specificity of a responder cell is to select unrelated individuals who show typing responses with two different DW specificities. This is assumed to identify the phenotype of the individuals. If such unrelated pairs are mutually nonresponsive, the DW1-DW6 must identify the most important components of the MLC-stimulating gene product. If, however, these combinations frequently are mutually responsive or show responses in one direction, it means that the HLA-D genetic region is more complex and could be composed of genetic subloci. This would also indicate that the presently defined specificities would have to be split into subspecificities and, consequently, indicate a higher degree of polymorphism of the D alleles. Jgrgensen et al. (1975,) observed that twelve combinations consisting of six pairs of unrelated individuals with typing responses to two different DW specificities showed relative responses varying from 5 to 45%. This indicated that HLA-D typing to some extent could predict the results obtained in mutual MLC testing between individuals characterized with identical DW specificities. Grosse-Wilde et al. (1975) studied forty-eight combinations of phenotypically HLA-D-identical individuals and found that only 66% showed weak MLR (defined as relative responses below 20%).It should be remembered in this context that the MLR observed in mutual testing between unrelated individuals should be compared with the three MLR response groups consisting of zero, one, and two haplotype-different combinations ( Fig. 3). The initial studies seem to indicate that the DW2 specificity is relatively homogeneous and defines a unique D determinant when it occurs on an HLA-AS, B7 haplotype or on a B7 haplotype. The DW2 specificity may, however, be more complex when associated with other B antigens, e.g., B18 together with DW2 ( F u et al., 1975a,b; J. A. Hansen et al., 1975). The same may account for the other DW specificities when they occur together with the B specificities for which a strong positive linkage disequilibrium is observed, e.g., DW1 with BW35, DW3 with B8, DW4 with BW15, and DW5 with BW16. J. SEROLOGICAL IDENTIFICATION OF ALLOANTICENSWITH RESTRICTED TISSUEDISTRIBUTION Following the observation that strong MLC activation in the mouse was elicited by genetic determinants within the I region of the H-2 complex, attempts were made to produce antibodies against determinants controlled by the I region. It was assumed that MLC-stimulating genes would code for cell surface structures that should be antigenic. The I-region associated antigens ( Ia antigens) were initially described

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

165

by David et al. (1973), Hauptfield et al. (1973), Gotze et al. ( 1973), and Sachs and Cone ( 1973). Different Ir specificities have been mapped with all three subregions of the I region (Ir-lA, Ir-B, and I-C) (Shreffler and David, 1975). The relationship between the Ia antigens and thc MLR-stimulating determinants in the mouse is not clear. There is some evidence that Ia determinants may stimulate in MLR. Meo et al. (1975) have shown that specific Ia antibodies directed against the Ia antigens of the stimulating cell will specifically inhibit MLC activation, whereas anti-H-2D or anti-H-2K antisera will not inhibit. The Ia antigens are strongIy expressed on lymphocytes from lymph node and spleen and are only rarely identified on thymic lymphocytes. Moreover, Ia antigens are absent from erythrocytes, kidney and liver cells, and from brain tissue (David et al., 1973). I t was initially concluded that Ia antigens are absent from T lymphocytes (Sachs and Cone, 1973; Hammerling et al., 1974). Later studies, however, have shown that some antisera may contain antibodies against both B cells and T cells (Frelinger et al., 1974). It is possible that some I a antisera are directed exclusively against T cells, whereas others are B-cell specific. Inhibition of human MLC with antisera containing HLA antibodies was first described by Ceppellini et al. (1969). This was subsequently confirmed by several others (Grumet and Leventhal, 1970; Ceppellini, 1971; Buckley et al., 1972; Revillard et al., 1972). I t has never been clear, however, whether the antibodies responsible for MLC inhibition were truly anti-HLA or directed against some other determinants. Revillard et al. (1972) showed that MLC-blocking factors could be absorbed with leukocytes and to a lesser extent with platelets. Greenberg et al. (1973) described several HLA antisera that specifically blocked MLC stimulation or response. The inhibition was not related to the specific HLA antigens on the lymphocytes involved in the MLC test. Gatti et al. (1973) described an antiserum with a broad MLC-blocking capacity unrelated to HLA specificity. Van Leeuwen et al. (1973) used an MLC inhibition protocol in which the responding cells were obtained from the antibody producer and the stimulating cells were from unrelated individuals who were HLA-seroidentical with the antibody producer. Some MLC combinations were strongly inhibited, suggesting that non-HLA antibodies are involved in MLC blocking. Wernet and Kunkel (1973) studied sera S ( SLE ). Some of these from patients with systemic ~ U ~ L Ierythematosus sera wcre shown to inhibit MLC. Dorf et al. (1973) tested sera from planned immunizations between HLA-seroidentical, MLC-positive, unrelated individuals, but was unable to show any MLC-blocking antibodies, Thorsby et al. (1973a) demonstrated that an anti-HLA-Bl2 anti-

166

DUPONT, HANSEN, AND YUNIS

serum blocked the stimulating and responding capacity of HLA-B12-positive lymphocytes. Lymphocytes from HLA-Bl2-negative donors, however, were also blocked. These studies demonstrated a complex pattern of MLC inhibition which generally seemed to be independent of HLA-A or B specificity. Some antisera seemed to act on the responding cell, whereas other antisera seemed to act on stimulating cells. In 1973, van Leeuwen et al. combined studies of MLC inhibition with investigation of indirect immunofluorescence of responding and stimulating lymphocytes reacting with the same antisera. Three antisera were described for which the specificity of the immunofluorescence reaction against the stimulating cell to some extent corresponded with MLC inhibition. One of these antisera (Pl) was further evaluated in family and population studies (van Rood et al., 19754. The occurrence of positive immunofluorescent staining in four families corresponded to the segregation of HLA. Population studies demonstrated that the PI specificity is associated with HLA-BS and HLA-DW3. When the PI antiserum was used to stain peripheral blood lymphocytes, a positive reaction gave only 15-2S% staining with a background of 4-7% The recent identification of several different human alloantigens that differ from the previously described HLA-A, B, and C antigens, but probably are coded for by genes within the HLA complex, have been made because of three discoveries: ( I ) Some antisera identify alloantigens that occur primarily on B lymphocytes and probably not on T lymphocytes (Winchester et al., 1975a,b, 1976; Wernet et al., 1975; Mann et al., 1975a,b, 1976); ( 2 ) the use of F ( a b ) , fragments prepared from isolated IgG of rabbit antisera specific for human IgG and conjugated with rhodamine made possible a much lower background staining in the indirect immunofluorescence technique (Winchester et al., 1975a; Fu et al., 1975~);( 3 ) the antisera directed against the B-lymphocyte alloantigens are cytotoxic for B lymphocytes and can be used in a cytotoxicity assay if purified B lymphocytes or B-cell-enriched cell suspensions are used as target (Walford et al., 1971, 1975; Winchester et al., 1975a,b,c; van Rood et al., 1975a; Terasaki et al., 1975; Bodmer et al., 1975; Ferrara et al., 1975; Solheim et al., 1975). Most of the antisera specific for B lymphocyte alloantigens have been obtained from pregnancy sera. Some of these contain, in addition to the B-cell antibody, conventional HLA antibodies directed against T lymphocytes, B lymphocytes, and platelets. Antibodies directed against the HLA-A, €3, or C specificities can be removed by absorption with platelets or T-cell lines, leaving only the antibodies against B cells (Winchester et al., 1975 a,b).

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

167

Preliminary family studies of segregation of B-cell alloantigens have shown that the determinants are associated with the HLA complex (van Leeuwen et al., 1973; van Rood et al., 1975a-d; Wernet et al., 1975; Winchester et al., 1975b). In one family with an HLA-B/D locus recombinant child, the B-cell alloantigen segregated with the HLA-D determinant (van Rood et al., 1975a-d; Wernet et al., 1975). Comparison of HLA-D-typing experiments using homozygous typing cells with B-cell alloantigen determination have shown that some B-cell specificities are strongly associated with specific HLA-D determinants. This has been shown for DW1, DW2, and DW3 by van Rood et al. (1975c,d) and for DW2, DW3, and DW4 by Winchester et al. (1975b). There are some data, however, indicating that the B-cell lymphocyte alloantigen is not coded for by the same genes as the HLA-D determinants. This is shown by the discordant segregation of HLA-D specificities and B-cell alloantigens in some families (van Rood et al., 1975d; Winchester et al., 1975b,c). Kovithavongs et al. ( 1975) have recently used the antibody-dependent cell-mediated cytotoxicity test ( ADCC ) to identify antisera directed against MLC-stimulating determinants. In a continuation of this work Dossetor et al. (1975) have demonstrated a significant correlation between HLA-D typing with one homozygous typing cell and positivity in the ADCC test with one antiserum. Ferrara et al. (1975) have raised antisera by planned immunizations in three different donor-recipient groups: ( I ) HLA-identical siblings identical for the A, B, and D determinants); ( 2 ) HLA-identical unrelated individuals (identical for the A, B, and D determinants); ( 3 ) HLA-A, B-compatible unrelated individuals. They studied these antibodies by complement-dependent cytotoxicity ( CDC ), ADCC, MLC inhibition studies, and cytotoxicity against a panel of B lymphocytes obtained from patients with chronic lymphocytic leukemia ( CLL) . In the first group of immunizations ( HLA-identical siblings), they could not demonstrate any antibodies. In the second group (HLA-identical, unrelated), they demonstrated an antibody directed against four of thirty different donor-recipient combinations. In the last group ( HLA-A, B compatible, D incompatible, unrelated), they demonstrated B-lymphocyte alloantibodies, blocking of MLC stimulation, but no blocking of MLC response. In another set of planned immunizations, they produced antibodies against HLA-B antigens or HLA-C antigens. These antisera blocked both MLC stimulation and response. In summary, the B-cell alloantigen system represents a new HLA determinant that has a restricted tissue distribution. The B-cell alloantigens are expressed on peripheral blood B lymphocytes, on B lymphoblasts,

168

DUPONT, HANSEN,

AND YUNIS

on R-lymphoid-cell lines, and on chronic lymphatic leukemia cells of B-cell origin. Prcliminnry studies indicate that the genetic determinants controlling some B-cell alloantigens are closely linked to the HLA-D locus. Othcr specificities may be controlled by genes in other segments of the HLA complcx (Mann et al., 1976). Some B-ccll alloantigens may be controlled by genetic dcterminnnts transmitted independently of the HLA complex ( Legrand and Dausset, 1975b; Winchester et al., 1975b). The antibodies dirccted against some B-cell alloantigens will specifically block MLC stimulation. It has not becn proven, however, that some B alloantigens are identical with the HLA-D determinants. There arc thus considerable similarities between some B-cell alloantigens in man and the Ia antigens in the mouse.

K. ROLEOF LYMPHOCYTE SUBPOPULATIONS IN MIXED-LYMPHOCYTE CULTUREREACTION Participation of lymphocyte subpopulations in stimulation and response in MLR have been studied primarily in mouse and man. The contribution of T and B lymphocytes in MLR have been evaluated by selective elimination of one cell population by specific antibodies or by different separation proccdurcs. Conflicting results havc been obtained by different investigators. Von Boehmer (1974), Chess et al. (1974), and Sondel et al. (1975a) have shown that T and B cells stimulate equally well. Plate and McKenzie (1973) and Lohrmann et al. (1974), however, demonstrated that B cells stimulate better than T cells. It seems, however, well established that non-B cells containing lymphocyte populations are effective in MLC stimulation in man and mouse. Also pure T-cell suspensions may stimulate (Sondel et al., 1975a; Chess et al., 1974; Rocklin et al., 1974; Lonai and McDcvitt, 1974; Frelinger et al., 1974). Alter and Bach (1970) demonstrated that monocytes are required to induce MLC stimulation, and stimulation in MLC can be induced by adherent cells alone (Alter and Bach, 1970; Rode and Gordon, 1974). Reconstitution of a monocyte-depleted culture, however, requires only a small number of adherent cells. Recently, Sondel et nl. (1975b) have shown reconstitution of normal MLR in monocyte-depleted cultures of purified human T cells by addition of monocytes from the responder. This supports the idea that T cells can stimulate and that the origin of the monocytes is not critical for MLC activation. Cell response in MLC and generation of the cytotoxic killer cells in CML are solely a T-cell function (Hayry et al., 1972; Wagner et al., 1972; MacLaurin, 1972; Sondel et al., 1975a,b). The in vitro interaction in MLC is probably mediated primarily by B cells in the stimulation phase and by T cells in the proliferative and

H U M AN MIXED-LYMPHOCYTE CULTURE REACTION

169

cytotoxic effector phase of the reaction. B cells and possibly T cells express the genetic determinants responsible for MLK activation. I t is, however, not known if all T cells and B cells express these determinants. It has recently been shown by Lonai and Gruniet (1975) that different subregions within the I region in the mouse code for the expression of MLR determinants either 011 T or B cells, but not on both. VII. Genetic Control of Immune Response Related to Histocompatibility

Two studies appcar to have provided the basis for present dcvelopinent in the understanding of genetic control of resistance to infections and immune response. First, Gorer and Schiitze (1938) demonstrated a correlation between antibody production and resistance to Salmonella infections in genetically resistant mouse strains. Second, Scheibel ( 1943) immunized randomly bred guinea pigs with diphtheria toxoid and divided the animals into good and poor responders. After several generations of selective breeding, it was possible to obtain groups of guinea pigs that were homogeneously good or poor responders to diphtheria antigen. Similar observations were obtained by Fink and Quinn (1953), Ipsen (1959), Stern et al. (1956), Dineen ( 1964), and Playfair (1968) in inbred strains of mice. Selective breeding of outbred mice over several generations was required to obtain high or low responders to specific antigens, indicating that multiple genes were involved in the genetic control of these immune responses ( Biozzi et al., 1968). The genetic control of immune response in experimental animals has been extensively reviewed ( McDevitt and Benacerraf, 1969; Benacerraf and McDevitt, 1972; McDevitt and Landy, 1973; Hildemann, 1973; Gasser and Silvers, 1974). Histocompatibility-linked genetic control of specific immune responses to viral infections has been reviewed by McDevitt et al. (1974), and HLA and disease associations have been reviewed by McDevitt and Bodmer (1974), Svejgaard et al. (1975), Vladutiu and Rose (1974), and by several different authors in Transplantation Reviews, Vol. 22 ( 1975). Milestones in the current knowledge about histocompatibility-linked immune response genes can be summarized by four sets of observations : ( 1 ) Immune response to well-defined synthetic polypeptides is genetically controlled and has been observed in inbred mice, guiea pigs, rats, and monkeys (Kantor et al., 1963; Levine et al., 1963; Levine and Benacerraf, 1964, 1965); ( 2 ) these immune response genes are inherited as autosomal dominant genes (Pinchuck and Maurer, 1968; McDevitt and Sela, 1965), the majority of these gcnes being associated with the major histocompatibility complex in several species ( McDevitt and Chinitz,

170

DUPONT, HANSEN, AND YUNIS

1969; McDevitt et al., 1969); ( 3 ) the immune responses controlled by these genes are dependent on thymus-derived T lymphocytes (Grumet et al., 1971); and ( 4 ) the susceptibility to a number of different autoimmune, virus-induced or spontaneous diseases in experimental animals is associated with histocompatibility determinants ( Lilly et al., 1964). The majority of the data in man related to the question of genetic control of immune response and histocompatibility determinants have focused on the association between disease susceptibility and the HLA specificities. Experimental animal studies on genetic control of disease susceptibility were initiated when Lilly et al. (1964) observed that inbred strains of mice varied in susceptibility to infection with murine leukemia virus. Mice homozygous for H-2" were practically all susceptible to the leukemia virus, whereas F-1 hybrids between susceptible and resistant strains were resistant to infection. Subsequently, a number of other diseases in experimental animals have been shown to be associated with histocompatibility determinants. Muhlbock and Dux ( 1971) demonstrated that susceptibility to virus-induced murine mammary tumor was related to certain H-2 determinants. Autoimmune and spontaneous thyroiditis in chicken occurs more frequently in relation to certain histocompatibility antigens (Rose et al., 1973; Bacon et al., 1973). Experimental allergic encephalomyelitis (Williams and More, 1973) and lymphocytic choriomeningitis ( LCM ) are also associated with certain histocompatibility determinants ( Oldstone et al., 1973). Analysis of viral leukemogenesis in mouse has provided the best model for study of the multigenic control of disease development. The leukemogenic process includes several steps, each of which is genetically controlled: ( I ) Viral expression depends on the Aku genes (Rowe et al., 1972); ( 2 ) viral replication is controlled by another gene Fu-I (Lilly, 1967); Pincus et al., 1971); and ( 3 ) resistance to virus infection is controlled by the Rgu-I gene (Lilly el al., 1964). The Rgu-I locus seems to influence the degree of immune response to virus-induced changes in cell surface antigens and does not affect virus replication (Lilly and Pincus, 1973). Only the Rgu-I locus is linked to the H-2 complex. Genes Aku and Fu-I, the two other genetic systems of importance in murine leukemogenesis, are not associated with the major histocompatibility system. This experimental model may suggest why the associations between certain human diseases and HLA is often partial or incomplete. Even in family studies, there is frequently discordant segregation between disease susceptibility and suspected HLA haplotypes. In human disease, the strongest association with an HLA antigen exists between HLA-B27 and ankylosing spondylitis (Brewerton et al., 1973; Schlosstein et al., 1973). The HLA-B27 is increased in these patients from

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

171

the expected 8% to 70-90% This antigen is also increased in reactive arthritis, Heiter’s syndrome, and acute anterior uveitis. The relative risk for developing one of these disease is increased thirty- to eighty-fold for anyone with the B27 antigen. In patients with psoriasis, there is an increase in HLA-B13, B17 (White et al., 1972; Russell et al., 1972) and BW37 (Tiilikainen, 1974). The relative risk of psoriasis for individuals with these antigens is increased four- to eightfold. Sevcral different diseases have been associated with the antigen HLA-B8: ( I ) dermatitis herpetiformis, ( 2 ) celiac disease (Stokes et al., 1972; Falchuk et al., 1972), ( 3 ) chronic active hepatitis, ( 4 ) myasthenia gravis, (5) juvenile diabetes (also associated with HLA-BW15), ( 6 ) Grave’s disease, and ( 7) idiopathic Addison’s disease. Multiple sclerosis and optic ncuritis are associated with HLA-B7. Some diseases have been shown to bc more frequently associated with certain HLA-D determinants: ( I ) multiple sclerosis (Jekild et al., 1973a,b) and optic neuritis with DW2 (LD7a) ( Sandberg-Wollheim et al., 1975); ( 2 ) rheumatoid arthritis with DW4 (LDW15a) (Stastny, 1974); and ( 3 ) dermatitis herpetiforms with DW3 (LD8a) (Solheim et al., 1975; Thomsen et al., 1975a). Other diseases have been shown to be equally associated with HLA-B and HLA-D specificities: ( I ) juvenile diabetes is associated with HLA-B8, BW15, and DW3 (LD8a) and DW4 (LDW15a); ( 2 ) idiopathic Addison’s disease with HLA-B8 and DW3 (Thomsen et al., 1975a); ( 3 ) myasthenia gravis with HLA-B8 and DW3 in two studies (Kaakinen et al., 1975; Moller et al., 1975) (Table XI). The disease associations with HLA antigens described so far have indicated that the genetic determinants of the HLA system are unequally involved, Because of genetic linkage disequilibrium within HLA, some antigens may be increased secondary to an increase in an antigen at another locus. In multiple sclerosis, the relative risk of disease for individuals with DW2 (LD7a) is higher than for B7 or A3, indicating that the increase in A 3 and B7 is probably secondary to the increase in DW2 (Jersild et al., 1975). Attempts have been made to see if certain HLA haplotypes are increased in any disease. I t has not bcen possible to show that a complete haplotype (HLA-A, B, C , D ) is selectively increased or decreased in any disease (Jersild et al., 1975). It has becn proposed that the chromosomal region between the T locus and the H-2 complex in the mouse represents a “supergcne” and that some combinations of determinants within the supergene would carry a survival advantage ( Snell, 1968). It is usually assumed that disease association with HLA antigens reflect a disease susceptibility based on the presence or absence of some specific immune response gene. It must also be assumed that disease

172

DUPONT, HANSEN, AND WNIS

I)ISEASE

TABLE XI HLA DETERMINANTS~

ASSOCIATIONS TO

Most significant HLA antigen increase Disease Rheumatic diseases: Ankylosing spondylitis Reiter’s syndrome Reactive arthritis Acute anterior uveitis Rheumatoid arthritis Neurological diseases : Multiple sclerosis Optic neuritis Myasthenia gravis Endocrine diseases: Juvenile diabetes Idiopathic Addison’s disease Graves’ disease Gastrointestinal diseases: Chronic hepatitis Celiac disease Skin diseases: Psoriasis

B locus

D locus

nw4

IlW2 IlW2 DW3 DW3, DW4 DW3 -

DW3

C locus A locus

B27 B27 B27 B27

-

B8 B8, BW15 B8 B8 B8 B8 B13, BW17, BW37

T 7

0 Disease associations to HLA antigens have only been included if more than one study has demonstrated significant associations. An exception to this is juvenile diabetes and idiopathic Addison’s disease, where the association to the D determinants have only been reported in one study. They are, however, included since the association appears to be strong. Significant association is based on calculation of relative risk (Svejgaard et al., 197.5).

susceptibility genes are dominant, since most patients are heterozygous for the HLA determinant characteristic for that disease. In experimental animals, however, it has not yet been shown that disease susceptibility is caused by a dominant gene. Disease susceptibility is usually a recessive trait. Nor has it been shown that disease susceptibility is caused by an abnormal immune response gene. Proof of involvement of histocompatibility-linked Zr genes in disease susceptibility would necessitate demonstration of an association between histocompatibility antigens, specific immunological responses, and disease development. In multiple sclerosis ( M S ) , it has been shown that antimeasles antibody titers in serum are slightly increased among patients and that the increased antibody titers occur primarily among patients with HLA-B7 ( Jersild et al., 1 9 7 3 ~ )The . response of MS patients’ leukocytes to paramyxovirus antigens (measles, parainfluenza, mumps virus) in the direct leukocyte migration inhibition

HUMAN MIXED-LYMPHOCYTE CULTURE REACXION

173

assay was found to be decreased (Ciongoli et al., 1973; Platz et al., 1974a), whereas Utermohlen and Zabriskie (1973) in a similar assay systcm found lack of response to the measles antigen. This phenomenon was, however, observed in all patients and not limited to the patients with specific histocompatibility antigens. Recently, we have demonstrated that lymphocytes of MS patients and normal controls respond equally well with in uitro blast transformation to paramyxovirus antigens (S. Cunningham-Rundles et al., 1975). A selective lack of cell-mediated immune response of MS patients to paramyxovirus antigens has thus not been clearly established, and the mechanism of specific histocompatibility antigen involvement in the development of MS is still unknown. In idiopathic Addison’s diseasc, it has been shown that antiadrenal antibodies are primarily present in patients with HLA-B8 and/or DW3 (Thomsen et al., 1975a). The mechanism, however, by which the HLA complex influences disease development is yet unknown. Involvement of Zr genes may be only one of many possibilities. Molecular mimicry between HLA antigens and disease-causing agents could also explain the associations between HLA and disease susceptibility. Only in ankylosing spondylitis is there nearly complete association between one HLA antigen and a disease. It is possible that this disease represents a situation in which the antigen, B27, is directly involved in disease development because of molecuIar mimicry with the diseasecausing agent. Molecular mimicry would suggest that the individual is unable to respond to a foreign agent because it mimics the host antigen or the immune response cross-reacts with the host antigen as a target. The HLA antigens might also serve as specific receptors for diseasecausing agents. This possibility, however, does not explain situations in which the patient lacks the HLA antigen characteristic of the disease. Ankylosing spondylitis, which is primarily associated with the antigen B27, has not been associated with any of the presently defined HLA-D specificities nor is it primarily associated with specific HLA-C locus specificities (Sachs et al., 1974; Moller et al., 1975; Truog et al., 1975). This is in contrast to some diseases that are primarily associated with the D-locus specificities or with C-locus specificities. For example, multiple sclerosis is clearly more strongly associated with the HLA-D specificity DW2 than with antigen B7. Recently, it has been shown in psoriasis that the increase in the three antigens, B13, B17, and BW37, seems to be secondary to a primary increase in one HLA-C antigen, T7 (Table XI) (Svejgaard, 1976). It thus seems that there are several different regions within the HLA complex that can be related to disease susceptibility. The mechanism of this susceptibility may, however, be different for different diseases.

174

DUPONT, HANSEN, AND YUNIS

Recently, Fu et al. (1974) have demonstrated in family studics that deficiency of complement C2 synthesis is genetically controlled and linked with the HLA system. This abnormality has been associated with SLE and Hodgkin’s disease. Five out of 6 patients homozygous for C2 deficiency were found also to be homozygous for the HLA-DW2 (LD7a) specificity. Rittner et al. (1975b) have recently shown in a single family segregation study that deficiency of synthesis of complement C4 is associated with HLA. The homozygous C4-deficient patient has renal disease. These findings suggest another possible mechanism by which histocompatibility-linked determinants may influence disease susceptibility. The disease susceptibilities that have so far been associated with HLA determinants cannot be considered as evidence for the existence of Zr genes within the HLA complex. There are some circumstantial indications that immune responsiveness may play a role in certain histocompatibility-linked disease susceptibilities, but several other possibilities are also likely. Studies of the genetic control of immune response in man have been very difficult. This is primarily due to the outbred nature of human populations, ethical considerations in human experimentation, the lack of pure antigens with restricted specificity, and the lack of an exact methodology to assess first molecular events in the immune response in man. Nevertheless, Levine et al. (1972) and Marsh et al. (1973) have reported associations between HLA and responsiveness to ragweed antigen E in family studies. These findings were supported by the study of a large family with one HLA-A and B recombinant child, which indicated that the inheritance of the ragweed sensitivity segregated in the family, together with the HLA-B determinant ( Blumenthal et al., 1974). By mixed-lymphocyte cultures, it was suggested that the locus ( I r E ) controlling the production of reagenic antibody was placed outside the HLA-B, D loci (Yunis et al., 1975). These studies were based on evaluation of the immediate skin reaction to ragweed antigen. Although the antigen E preparation used had a high degree of purity, it produced responses with different limiting doses, and the degree of exposure to the antigen could not be controlled. I t was also impossible to rule out the presence of lesser amounts of other antigens that might cause a positive immune response. Furthermore, a negative response in an individual does not indicate absence of the genetic capacity to produce reagenic antibody, since young individuals may not have been adequately exposed, and older persons may lose the ability to respond. Although these criticisms cannot be resolved at present, the families with ragweed sensitivity have been useful in demonstrating the association between the HLA-B, D region and ragweed sensitivity. I t is remarkable that a presumably polygenic

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

175

disease with varying degrees of penetrance and age of onset should demonstrate such a degree of association between the HLA-B determinant and the reactions to antigen E. Recently, Greenberg e t al. (1975) have studied the association of in vitro lymphocyte response to purified antigens and HLA among unrelated individuals. They used purified streptococcal nuclease A as the antigen, and by selecting limiting doses, it was possible to define “high” and “low” responders. The study indicates that there may be an association between responsiveness to this antigen and the presence of HLA-B5. It is not proven, however, that differences in responsiveness to streptococcal nuclease A are genetically controlled. Haverkorn et al. (1975) studied 71 monozygotic and 72 dizygotic twin pairs for serum antibody titers against poliomyelitis virus I, 11, and I11 and diphtheria antigen. These antigens are used in routine immunizations. They also compared antibody titers against common infections that were assumed to be endemic in the community, such as, rubella, measles, and influenza. It was shown that there is significant evidence for genetic control of ‘antibody response to poliomyelitis I, diphtheria toxoid, and measles. Only the response to measles virus, however, seemed to be associated with HLA. The association between histocompatibility and disease susceptibility may be related to the sustained genetic polymorphism of the HLA complex. It is assumed that a polyallelic system will be maintained only if it represents a survival benefit for the species ( Ford, 1965). The persistence of HLA polymorphism may be attributed to an advantage of heterozygotes, as indicated by Degos et al. (1974) in a study of inbred populations. It is interesting in this respect that the presently described disease associations to HLA have involved some of the most common HLA determinants present in the population, e.g., B7, B8, DW2, DW3, DW4. A few studies in experimental animals have shown that offspring heterozygous for the major histocompatibility system may have a survival advantage over homozygote offspring. Morton e t al. (1965) observed in chickens that B-antigen heterozygotes had a lower mortality rate during the embryonic period than homozygotes. Gorer and Mikulska (1959) showed the same phenomenon in some H-2 homozygote and heterozygote mice. Palm (1970) analyzed this question in rats homozygous and heterozygous for Ag-B. Homozygous females were mated with F, heterozygous males (F, backcross) and the survival of homozygous and hetcrozygous progeny were compared. The study demonstrated that twice as many Ag-B homozygotes died from runt disease in the neonatal period. Further analysis indicated that runt disease developed on the basis of non-Ag-B-immune reactions. It was suggested that Ag-B incomptibility would protect the fetus against the occurrence of immune reactions from

176

DUPONT, HANSEN, AND YUNIS

the mother, which would dcvelop more easily in situations where mother and fetus were Ag-B compatible. Similar phenomena may be obscrved in human disease. The study of MLRs in families of patients with certain hemopoietic and immunological disorders have demonstrated an unexpected number of identical, or only weakly responsive, MLR between mother and child. These nonresponsive combinations are predominantly observed when the child or the mother suffers from the disease. This lack of response is specific because the cells from mother and child will stimulate lymphocytes from other donors. We have observed this phenomenon in the families of 2 children with severe combined immunodeficiency disease (SCID) (Dupont et al., 1972, 1975b). The same has been observed by several other investigators ( R. H. Buckley, personal communication, 1975; S. Goldmann, personal communication, 1975; P. Ernst, personal communication, 1975; D. Gunther, personal communication, 1975). The MLR identity between mother and patient with Fanconi’s congenital aplastic anemia has been reported in two different families (Dupont et al., 1975b). Similar findings have been obtained in the families of some patients with idiopathic aplastic anemia, congenital neutropenia, and acute leukemia ( Opelz, 1975; Dupont et al., 1975b). I t has been suggested that HLA-D compatibility between mother and fetus may sometimes allow fetal engraftment with maternal immunocompetent cells and that these cells might induce damage to the developing hemopoietic system of the fetus. The possible role of fetal graft-versus-host ( GVH ) reaction as a pathogenic factor in the development of progressive bone marrow aplasia and other hemopoietic disease is supported by the finding of Lafferty et al. (1972) that in the chick embryo the hemopoietic stem cell can be a primary target for GVH reaction. A reaction induced on the chorioallantoic membrane will produce different lesions in the hematopoietic system of the chick, depending on the stage of embryonation. Development of fatal GVH reactions after intrauterine grafting of maternal immunocompetent lymphoid cells in patients with SCID has been shown in 2 cases (Kadowaki et al., 1965; O’Reilly et al., 1973). There is substantial evidence of the same in 2 additional cases ( Biggar et al., 1975; Naiman et al., 1969). It has not been demonstrated that MLR identity between mother and child is more frequent in the above-mentioned diseases than in the normal population. Most of these diseases are relatively rare, and appropriate control families have not been studied. It seems established, however, that homozygosity for the serologically defined HLA-A and B determinants in children of first-cousin marriages does not result in decreased survival. Keuning et al. (1975a) found 45 HLA-A, B-homozy-

HUMAN MIXED-LYMPHOCYTE CULTURE

177

REACTION

gous children among 593 offspring of first-cousin marriages, and this does not differ significantly from the 51 homozygous children that were expected. VIII. Mixed-Lymphocyte Culture As a Histocompatibility Test for Clinical Transplantation

The major histocompatibility region in man (HLA) and mouse (H-2) play dominant roles in transplantation. The H2 disparity has a quantitative effect on the rate of allogeneic skin rejection in the mouse, where such grafts are rejected within 9 to 10 days. Disparity at non-H2 (minor histocompatibility loci) can lead to variable rejection times, depending on the accumulative strength of the minor loci (Eichwald and Weissman, 1966; Klein, 1967). In experimental animals, several observations comparing major and minor histocompatibility loci have been made. The strength of minor loci differences depends on the accumulated effect of multiple allelic systems. In mouse, multiple minor loci differences can have as great an effect on skin graft rejection as a major locus difference. It is, howevcr, relatively easier to cause prolongation of skin graft survival using immunosuppression across multiple minor loci than across a major locus difference (Silvers et al., 1967). An H-2 identity leads to significant prolongation of skin grafts, even though minor histocompatibility differences eventually cause rejection. In man, it has been shown that disparity at HLA usually results in skin graft rejection within 10 to 13 days, whereas skin graft survival between HLA-seroidentical siblings is prolonged from 19 to 25 days (Amos et al., 1966; Ceppellini et al., 1965; Dausset et al., 1970; C. T. Koch et al., 1973). HLA disparity within a family, as measured by O h , Ih, and 2h differences, has a quantitative effect on the survival of skin grafts transplanted within the family. Amos et al. (1969) showed a mean of survival of 24.9 k 1.1 days ( N = 43) for HLA-seroidrntical siblings, a mean of 14.4 0.3 days ( N = 115) for l h combinations, and a mean survival of 11.6 0.5 days ( N = 18) for 21e combinations. Dausset et al. (1970) performed 238 skin grafts between Ih-diffcrent family members representing 164 different HLA haplotypes (mean survival time = 12.7 -+ 2.4 days). Several recipients received grafts from different donors. Each graft represented a challenge with the same A and B locus antigens. These grafts all had similar survivals, in spite of presumed incompatibility for minor histocompatibility antigens, indicating that HLA is the dominant histocompatibility system. The inimunogenicity of A and B locus antigens appeared to be similar. Seigler et al. (1971) evaluated MLR and skin graft survival in selected pairs from thirty different families. They found

* *

178

DUPONT, HANSEN, A N D YUNIS

no correlation within families between the degree of MLC reactivity and skin graft survival. Skin graft survival between unrelated, but HLAseroidentical pairs was evaluated by Koch et al. (1971).The mean graft survival for 11 unrelated HLA-A, B-identical donor-recipient pairs was 11.4 t 0.2 days. This was significantly longer than the mean graft survival (10.0 & 1.3 days) between 20 HLA-nonidentical unrelated pairs. One recipient was MLC-identical with each of two unrelated, but HLAseroidentical donors (HLA Al, B8/A3, B7). Both of these HLA-A, B, D identical grafts were rejected on day 12. This study seemed to indicate that serological matching for HLA-A and B alone resulted in prolonged survival and that additional MLC matching did not contribute significantly to the skin graft survival. In a second group of unrelated HLAseroidentical pairs, these investigators correlated stimulation ratios for two-way MLC between donor and recipient. Combinations with a low stimulation ratio (SR-2) tended to have a longer graft survival (C. T. Koch et al., 1973; van Rood et al., 1973). However, even with matching at the D locus, the survival of serologically identical unrelated grafts is much less than the survival of grafts between HLA-identical siblings. Sasportes et al. (1972, 1973) have shown evidence suggesting a correlation between MLC identity and prolongation of skin graft survival in haploidentical family combinations. Mutual MLR nonresponsiveness was observed in a few family combinations, possibly as a result of recombination or sharing of HLA-D determinants between parents. The mean time for rejection in haploidentical, but MLR-positive combinations was 11.3 days (9-17.5 days). Some prolongation of graft survival occurs even though the MLR is positive. There was a tendency for slightly longer graft survival in haploidentical MLR-negative combinations (rejection times: 14, 15, 16.5, 18, 19, and 190 days; mean = 28.4 days). Skin grafting in HLA recombinant' families has generally failed to show that D locus disparity alone can cause early skin graft rejection (Ward and Seigler, 1973; Yunis et al., 1973). Amos et al. ( 1976) sensitized a normal volunteer with skin from an HLA-A, B-identical unrelated donor. Repeated skin grafting 10 days later from the same donor and from 6 additional HLA-A, B-identical unrelated donors resulted in hyperacute or second-set graft rejections, suggesting that factors other than the HLA-A, B alloantigens are responsible for accelerated skin graft rejection. Amos and Yunis (1973) postulated that some genetic factor( s ) other than HLA-A, B, D is responsible for eliciting delayed hypersensitivity reactions (HDR) and that they might also be responsible for accelerated rejection of skin grafts between unrelated individuals (Yunis and Amos, 1971; Amos and Yunis, 1973). The gene( s ) for HDR segregate with the

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

179

HLA haplotype in family studies and may be linked more to the A and B locus than to the D locus, There is similarity between the postulated HDR locus and the CML locus, which seems to be closely linked to HLA-B (Section 111), suggesting that the CML test might be an in vitro model for the in vivo function of HDR. Further studies, particularly in informative recombinant families, comparing CML and skin graft survival are indicated. The effect of the major histocompatibility system ( H L A ) in clinical organ transplantation is clearly established, but the significance of the individual components has been difficult to determine. In kidney transplantation, the best graft survival is obtained with grafts from HLAidentical sibling donors (Singal et al., 1969; Hors et al., 1971; Seiglcr et al., 1972; Opelz et al., 1974). In kidney recipients receiving immunosuppressive therapy, the ranking of matches within families shows a significant decrease in graft survival when comparing O h , lh, and 2h matches ( Dausset and Hors, 1972; Opelz et al., 1974). The probability of finding an HLA-seroidentical cadaveric kidney for a patient awaiting kidney transplantation depends on the size of the donor and recipient pool. To maximize the chance of providing a “good” match whenever a cadaveric kidney becomes available, cooperative groups have been organizccl. This has made possible a number of cadaveric kidney transplants in which donor and recipient are seroidentical (four antigen matches). Although there are now a large number of transplant patients in several different series the effect of HLA-A, B matching for cadaveric kidney grafts remains controversial. Several questions remain: Does the number of antigens matched affect graft survival? What is the relative importance of matching for HLA-A versus HLA-B? What is the effect of matching for certain HLA haplotypes? And what is the effect of lymphocytotoxins in the recipient? Joint analysis by Dausset et al. (1974) of data from both the France transplant and the London transplant groups strongly supports a positive correlation between graft survival and the number of HLA antigens matched between recipient and donor. This is confirmed by the Scandiatransplant Report ( 1975), whcrein a significant difference is shown between four and three antigen niatchcs. The improvement, however, in survival for four antigen matches is not as striking as reported by Dausset et al. (1974). Eurotransplant found no significant effect of HLA matching in their overall material, yet analysis of only the recipients with lymphocytotoxic antibodies showed decreasing graft survival between combinations sharing fewcr HLA antigens (van Hooff et al., 1972). From San Francisco, Belzer et al. (1974) showed no significant correlation between matching for three, two, or one antigen and graft survival.

180

DUPONT, HANSEN, AND YUNIS

NO patient received a four antigen match in this series. In a large combined analysis from transplant centers in the United States and Canada representing 2172 cadaver grafts, Opelz et al. (1974) reported no significant difference in graft survival for one, two, three, or four HLA antigen matches. Eurotransplant and the London Transplant group reported that matching for antigens of the second locus (HLA-B) was more important than matching for antigens of the first locus ( HLA-A) (van Hooff et al., 1972; Oliver et al., 1972). More recent analysis, however, has not confirmed this (Dausset et al., 1974; Belzer et al., 1974; Opelz et al., 1974; Scandiatransplant Report, 1975). A later analysis of Eurotransplant, reported by van Hooff et al. (1974) showed that matching for certain haplotypes significantly influenced graft survival. Matching for haplotypes, however, was not found to be significant in the Scandiatransplant Report ( 1975). The presence of lymphocytotoxic antibodies, an indication of presensitization, has been shown in some studies to have an adverse effect on graft survival (Terasaki et al., 1971; van Hooff et al., 1972; Oliver et al., 1972); other studies show that the presence of lymphocytotoxic antibodies has no significant effect (Belzer et al., 1974; Dausset et al., 1974; Scandiatransplant Report, 1975). Opelz and Terasaki ( 1972) and Opelz et al. (1974) reported that patients who did not develop cytotoxins after more than 1 year on hemodialysis had a marked improvement in survival at 1 year. The London Transplant group and Eurotransplant reported that HLA serotyping had a more prominent effect on graft survival in patients with cytotoxins (van Hooff et al., 1972; Oliver et d., 1972). This was not found to be so in the joint analysis of combined data from London Transplant and France Transplant. In patients who previously rejected a cadaveric graft and are, therefore, presumably sensitized, the effect of HLA-A, B typing seems more significant than for patients receiving their first graft (Oliver et al., 1972; Dausset et al., 1974; Scandiatransplant Report, 1975). Cochrum et al. (1973) studied graft survival in 59 cadaveric kidney transplants. They were specifically interested in MLC reactivity between donor and recipient. All the combinations were divided into groups of high and low MLC stimulation (stimulation ratio <8 or stimulation ratio >8), and they found that high MLC reactivity correlated with relatively poor graft survival, and low MLC reactivity with relatively good graft survival. The disparity for HLA-A and B, however, could not be controlled in this experiment. Because of the linkage disequilibrium between several of the more common products of HLA-B and HLA-D, the selection of a group of donor and recipients on the basis of less

H U M A N MIXED-LYMPHOCYTE CULTURE REACTION

181

disparity at HLA-D would likely result in the inadvertent selection of certain HLA-B antigens that are found in linkage disequilibrium with certain HLA-D determinants. The low MLC rcactivity group would, therefore, have less disparity at the HLA-B locus antigens as well. The relative importance of serologically dcfined antigens vs MLC determinants in transplantation may not be discernible from such genetically complex clinical material. The most revealing study of MLR in terms of transplantation has been the correlation of MLR with the GVH reaction in congenic strains of mice. Using strains identical for the serologically determined H2-K and H2-D regions, it has been shown that the determinant responsible for splenic enlargement in neonatal mice which have been given immunocompetent cells is very closely associated with the determinants causing MLC (Livnat et al., 1973; Klein and Park, 1973). Using a lymph node enlargement assay, Oppltovi and Dkmant (1973) obtained similar results. The I-region differences give strong MLR and GVH reaction; K-and D-region differences give moderate-to-weak MLR and GVH reaction; and S-region differences probably produce no significant effect. Studying the proliferative response of donor thymocytes received from the spleens of lethally irradiated F, hosts, a technique that isolates the donor response in the GVH reaction from the host response, Elkins et al. (1973) found that I-region differences were primarily responsible for the proliferation of donor cells, at least during the first 5 days after transplantation. This in vivo model of MLR gives results consistent with the concept that certain histocompatibility determinants, located in the I region of the mouse, control the potential proliferation of immunocompetent cells. Rodcy et al. (1974) used congenic strains of mice to test the relation betwecn degree of MLR and chronic GVH reaction, as measured by mortality from secondary disease. There was a remarkable correlation between the strength of MLR, as measured by stimulation ratios, and death from secondary disease, suggesting a significant relation betwecn I-region differences, in vitro lymphocyte proliferation ( M L R ) , and intensity of the GVH reaction. Allogeneic bone marrow transplantation in the treatment of children with SCID has been very informative concerning the role of histocompatibility in transplantation. Successful transplantation and immunological reconstitution in this disease has depcnded largely on the use of HLAidentical sibling donors (Buckley, 1971; Dupont et al., 1974b). Bone marrow from a family donor mismatchcd by a single haplotype has invariably resulted in fatal GVH disease (van Bekkum, 1972; Park et al., 1973). Recombination within the HLA system allowed successful bone

182

DUPONT, HANSEN, AND YUNIS

marrow transplantation of a child with SCID when marrow cells from an HLA-A-nonidentical, but MLR-nonresponsive, sibling donor was used (Gatti et al., 1968). The HLA recombination had presumably led to HLA-B, D identity between donor and recipient, but nonidentity for the HLA-A determinant. Family MLR in the case of a second patient with SCID suggested again a recombination in such a way that the maternal uncle and patient were nonresponsive in MLR, even though they were HLA-A, B-nonidenticaI ( Dupont et al., 1972, 19734. Bone marrow transplantation from the maternal uncle resulted in successful immunological reconstitution of the child with proved chimerism (C. Koch et al., 1973). Retrospective analysis of this family (KJ), however, suggests there was probably no recombination (see Section V,H). Although the father and maternal grandfather were both HLA-A, B heterozygous, they appear to both have been HLA-D homozygous. The patient and maternal uncle were nonresponsive in MLR because they each carried the same two HLA-D specificities, although on different and unrelated haplotypes. This is an instructive case because it suggested that matching for bone marrow transplantation could be achieved even for nonsibling donors by matching for HLA-D. An attempt was made to find an unrelated'donor for a patient with aplastic anemia by matching for HLA-A, B, and D (Speck et al., 1973). An HLA-A, B, and D identical unrelated donor was found, but bone marrow transplantation was unsuccessful because of failure to achieve engraftment. This trial must be considered inadequate, however, because the recipient was preconditioned with antilymphocyte serum only, making likely the rejection of a marrow graft from even an HLA-identical sibling. Bone marrow from an unrelated donor, MLC-identical but incompatible for one HLA-A antigen, has been used in repeated graftings of a child with SCID ( L'EspArance et al., 1975). Chimerism has been achieved with complete immunological reconstitution. A second patient with SCID has been grafted with an unrelated, HLA-Dmatched, but one HLA-A antigen-nonidentical donor. The patient died on day 30 from cytomegalovirus (CMV) pneumonia, but here was some evidence of immunological reconstitution (Horowitz et ul., 1975). The duration of the course, however, was probably too short to evaluate the potential for significant or serious GVH disease. Gale et a2. (1975a,b) reported on bone marrow transplantation in a young male with acute myelogenous leukemia, in which case marrow from a seroidentical, but MLR-positive, female sibling donor was used. There was early evidence of engraftment, but leukemia recurred, along with predominantly host markers, and the patient died on day +38. The family MLR was in-

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

183

terpreted as representing an HLA recombination between HLA-B and HLA-D, but formal proof of recombination is not available. Because of the early recurrence of leukemia and the apparent loss of the graft, the potential for GVH reaction in this MLH-positive situation could not be adequately evaluated. Clinical bone marrow transplantation in patients with aplastic anemia or leukemia is more complex than bone marrow transplantation in children with SCID. In the first place, these patients are immunocompetent and it is necessary to treat them before transplantation with immunosuppressive drugs (usually cyclophosphamide ) . These patients have frequently received multiplc transfusions of blood products, which in the dog has becn shown to increase the chance of rejecting even a well matched graft ( Storb et al., 1971 ). Even with pregraft immunosuppression, graft rejcction has been reported in up to 25%of transplants between HLA-seroidentical, MLR-nonreactive, sibling pairs ( Storb et al., 1974). Following bone marrow transplantation, the incidence of clinically significant GVH reaction is much greater in patients with aplastic anemia or leukemia than in patients with SCID. The incidence of clinically significant GVH reaction exceeds 50%in some combined series (Thomas et al., 197s) and may bc as high as 80%in patients with acute myelogenesis leukemia (Thomas et al., 1973). The reason for the increased intensity of GVH reaction in these patients is unknown. Important differences between patients with aplastic anemia or leukemia and infants with SCID may include age factors, differences engendered by the primary disease process, presensitization caused by blood transfusions, and complications introduced by immunosuppressive or antileukemic therapy. In summary, identity in HLA is of primary importance in matching within families for transplantation of skin, kidney, and bone marrow. However, HLA seroidentity alone is an inadequate predictor of skin or kidney allograft behavior between unrelated individuals. Matching for MLR (HLA-D), or some other closely associated determinants, seems to be the primary factor in determining the intensity of the GVH reaction, although significant GVH reaction may occur even between HLAidentical siblings. IX. Genetic Mapping of the HLA Complex on Chromosome C-6

Somatic cell hybridization between cells of man and Chinese hamster have given evidence for synteny between HLA, phosphoglucomutase-3 (PGM,), indophenol oxidase B (IPO-B), and cytoplasmic malic enzyme

184

DUPONT, HANSEN, AND YUNIS

( M E , ) (van Someren et al., 1974). The PGM., had been shown to be in genctic linkage with HLA by Lamm et al. (1971), and assigned to chromosome C-6 by Jongsma et al. ( 1973). Recently Middleton et al. (1974) demonstrated that the erythrocyte blood group Chido is linked to the HLA complex. The ME, was also assigned to the same chromosome by Chcn et al. (1973). Family studies have shown that the PGM, locus is linked at a distance of about 0.22 morgan with HLA and the sequence is likely to be PGM,- -HLA-D- -HLA-B- -HLA-C- -HLA-A. The HLA complex was directly assigned to chromosome number 6 in a family study by Lamm et al. (1974) who found an individual with a large periccntric inversion that segregated in the family with one HLA haplotype. Recently, several components of the serum complement system have been shown to be in linkage with the HLA complex. Allen (1974) demonstrated close genetic linkage between the glycine-rich p,-glycoprotein (GBG, Bf, or Factor B ) . Fu et al. (1974) presented evidence for linkage between HLA and the genes involved in the synthesis of the second component of complement ( C2), which was subsequently confirmed by Day et al. (1975) and Wolski et al. (1975). Recently, Rittner et al. (1975b) demonstrated that the genes controlling the synthesis of the fourth component of complement ( C 4 ) segregated in one family with one HLA haplotype, and the C4-homozygous-deficient child was also HLA-homozygous. The presence of genes in the human HLA complex that control synthesis of complement components is another example of homology with the H-2 complex in the mouse. Two serum proteins, the Ss protein, the serum serological variant, and the sex-limited protein ( S l p ) have been linked to H-2 and map between H-2K and the H-2D region (Shreffler, 1965; Passmore and Shreffler, 1971). The genes coding for the Ss and the S l p traits are located within the S region of the H-2 complex between the I region and the D region. Dkmant et al. (1973) and Capkovh and Dkmant (1974) reported that serum levels of hemolytic complement were controlled by the H-2 complex and were coded for by genes within the S region. Ferreira and Nussenzweig (1975) demonstrated linkage between the synthesis of complement C3 and the H-2 complex, and Goldman and Goldman (1975) gave evidence for association between C1, C2, C4, and the H-2 complex. Genes controlling the synthesis of several complement components are thus identified within the H-2 complex. The initial study linking C2 deficiency with HLA ( F u et al., 1974) demonstrated that C2 deficiency was associated with HLA-B18. Subsequent studies have shown C2 deficiency to occur with HLA-B18, HLABA2*, HLA-B5, and other HLA-€3 locus antigens in other families. Com-

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

185

plement C2 deficiency was associated with HLA-DW2 in six families studied Iiy Fu et al. (1975a,b) and in one additional family described by Friend et al. (1975). In these seven familics, there were six C2homozygous-deficient propositii. They were all HLA-DW2-homozygous and behaved in MLC testings as HLA-D-homozygous typing cells for the DW2 (LD7a) spccificity ( FU et al., 1975a,b; J. A. Hansen et ul., 1975; Friend ct al., 1975). In one family, the C2-deficiency genc was associated with an HLA-A1O,B18 haplotypc but without thc DW2 determinant ( F u et al., 1975b). It is an interesting obscrvation that homozygous C2 deficiency was found together with HLA-DW2 and HLA-B18 homozygosity in three familics, and in two other familics the C2-deficient propositus was HLA-DW2 homozygous but HLA-R18 heterozygous. In the normal population, the DW2 specificity is in strong positive genetic linkage with HLA-B7 and not with B18 (J. A. Hansen et al., 1975; Thorsby and Piazza, 1975). One family reported may demonstrate a recombination betwecn the HLA-B18 determinant and the C2-deficicncy genc. Although HLA-D typing of the recombinant child is not conclusive, the C2-deficiency gene can be niapped tentatively outside the HLA-B locus, probably close to the HLA-D locus (Friend et nl., 1975). Mapping of the Bf locus in the HLA region has been attempted by Rittner et al. (1975a), Netzel et d. (1975), and Teisberg et al. (1975). In summary, these data are conflicting, and mapping of the B f gene in relation to the different loci of HLA cannot yet be made. The presently described associations between the HLA complex and different genetic traits are shown in Fig. 1 together with the few welldefined genes that can be shown to be mapped close to HLA. The proper mapping of the Factor B locus, the C2-deficiency locus, and thc C4deficiency locus may be very important in future studies of the HLA region especially in the analysis of HLA recombin at'ions.

x.

Conclusions

At present four distinct segregant-series of HLA determinants can be identified. HLA-A, B, C, and D. The determinants coding for control of MLR (HLA-D) appear to belong to a chromosomal region that maps outside the A, B, C segment of HLA. The concept of a single locus controlling MLR, and the identification of some individuals homozygous for the HLA-D determinants, has made possible the definition of HLA-D specificities. The current methodology for HLA-D typing utilizes lymphocytes from HLA-D-homozygous individuals as stimulating cells in MLR. When an HLA-D-homozygous cell stimulates an individual heterozygous

186

DUPONT, HANSEN, AND W N I S

for the same HLA-D determinant, the weak response elicited is called a typing response. Accuracy in HLA-D typing depends on the discriminating power of the individual typing cell. If the responses elicited by a typing cell against a large panel has a clear bimodal distribution, then limits defining the cut-off level for a typing response can be made. The presently recognized HLA-D specificities have been defined in population studies and mutual MLR between different homozygous cells. As a result of collaborative work in the Sixth International Histocompatibility Workshop, six relatively well-defined provisional groups ( HLADW1 to DW6) and two less well-defined groups of specificity (LD 107 and LD 108) now exist. These groups appear to be mutually exclusive and non-cross-reacting. No two typing cells in any provisional group, however, type in exactly the same way. Some typing cells of the same group may even show mutual responses in MLR. The determinants represented by the different HLA-D-homozygous cells assigned to each group represent either cross-reacting specificities of the same allelic system or they may represent the products of subloci that exist in the population in strong positive linkage disequilibrium. As is generally true for the antigens of HLA-A, B, and C, the determinants of HLA-D, as defined by the provisional specificity groups, show strong linkage disequilibrium with certain HLA-B antigens. The HLA-DW1 is frequently associated with HLA-B35 ( WS), HLA-DW2 is very frequently associated with HLA-B7, and HLA-DW3 is frequently associated with HLA-B8. There is considerable homology between the H-2 complex in mouse and HLA in man. The mapping of HLA-D outside the segment coding for the serologically defined HLA antigens represents, however, a major exception. Yet, the region controlling HLA-D in man may be homologous to the I region in mouse. The I region controls MLR, certain immune response genes, and alloantigens with restricted tissue distribution ( Ia) . There is some suggestion that some of the B-cell alloantigens in man are coded for by determinants close to HLA-B and D. The homology between Ia in mouse and B-cell alloantigens in man has, however, not been proven. Disease association with HLA are primarily linked to HLA-B and D. The immune response to ragweed antigen also seems to be linked to this part of the HLA segment. It is likely, therefore, that well-defined specific immune response genes will be found to be closely linked to the HLA-B and D region. Some examples of unexplained MLR in families may be elucidated by assuming that the HLA-D region is composed of subloci and that control of MLR is the result of both stimulating determinants (MLR-S) and response determinants ( MLR-R). The possible existence of immune

H U M A N MIXED-LYMPHOCYTE CULTURE REACTION

187

response genes close to the H L A - B , D region is also supported by the recent findings suggesting a separate locus controlling cytotoxic killer cells in CML. The relative importance of the different loci of HLA in clinical transplantation is yet poorly understood. The HLA-D determinants, however, appear to be of major importance in controlling the GVH reaction and in determining the outcome of bone inarrow transplantation. ACKNOWLEDGMENTS

Original work cited was supported by grants from the American Cancer Society, U.S. Public Hcalth Grants IIL-06314, 1-R01 HD 08145, N01-CB-43853, and CA08748-0851, and by NCI Program Project Grant NCI-CA 17401-01, National Foundation March of Dimes, and by The Special Fund for the Advanced Study of Cancer. One of the authors ( J . A. H . ) is a Special Fellow, Leukemia Society of America. Also grants from Danish Multiple Sclerosis Foundation, the Foundation for Medical Research in Copenhagen, Greenland and Faroe Islands, the P. Carl Petersen Foundation, and the Danish Medical Research Council have supported the work.

REFERENCES Abbasi, K., DCmant, P., Festenstein, H., Holmes, J., Huber, B., and RychlikovA, M. ( 1973). Transplant. Proc. 5, 1329. Adler, W. H., Takiguchi, T., Marsh, B., and Smith, R. T. (1970). J. Zrnmunol. 105, 984. Albert, E. D., Mempel, W., and Grosse-Wilde, H. (1973). Transplant. Proc. 5, 1551. Albertini, R. J., and Bach, F. H. (1968). J. Exp. Med. 128, 639. Allen, F., Amos, D. B., Batchelor, R., Bodmer, W., Ceppellini, R., Dausset, J., Engelfriet, C., Jeannet, M., Kissmeyer-Nielsen, F., Morris, P., Payne, R., Terasaki, P., van Rood, J. J., Walford, R., Zmijewski, C., Albert, E., Mattiuz, P., Mickey, M. R., and Piazza, A. (1970). Zn “Histocompatibility Testing 1970” ( P . I. Terasaki, ed. ), pp. 17-44. Munksgaard, Copenhagen. Allen, F. H., Jr. ( 1974). Vor Sang. 27, 382. Alter, B. J., and Bach, F. H. (1970). Cell. Immtrnol. 1,207. Alter, B. J., and Bach, F. H. (1974). J. Exp. Mcd. 140, 1410. Amos, D. B. (1967). Transplantation 5, 1015. Amos, D. B. (1969). Ado. Immtrnol. 10, 251. Amos, D. B., and Bach, F. H. ( 1968). J. Exp. Med. 128,623. Amos, D. B., and Ward, F. E. ( 1975). Physiol. Reu. 55,206. Amos, D. B., and Yunis, E. J. (1973). In “Tissue Typing and Organ Transplantation” ( E . J. Yunis, R. A. Gatti, and D. B. Amos, eds.), pp. 117-133. Academic Press, New York. Amos, D. B., Hattler, B. G., Hutchin, P., McCloskey, R., and Zmijewski, C. M. (1966). Lancet 1, 300. Amos, D. B., Seigler, H. F., Southworth, J. G., and Ward, F. E. (1969). Transplant. Proc. 1, 342. Amos, D. B., Johnson, A. H., Rudeman, R. J., Mendell, N., and Yunis, E. J. ( 1976 ). Proc. M . D. Anderson Symp. (in press). Anderson, L. C., and Hlyry, P. (1973). Eur. J. Irnmunol. 3, 595. Anderson, L. C., and Hlyry, P. (1974). Scand. J. Imrnrrnol. 3, 461.

188

DUPONT, HANSEN, AND W N I S

Bach, F. H., and Amos, D. B. (1967). Science 156, 1506. Bach, F. H., and Hirschhorn, K. (1964). Science 143, 813. Bach, F. H., and Voynow, N. K. (1966). Science 153,545. Bach, F. H., Albertini, R. J., Amos, D. B., Ceppellini, R., Mattiuz, P. L., and Miggiano, V. C. (1969). Transplant. Proc. 1, 339. Bach, F. H., Bach, M. L., and Klein, J. (1972a). Science 176, 1024. Bach, F. H., Widmer, M. B., Bach, M. L., and Klein, J. (197213). J. Exp. Med. 136, 1430. Bach, F. H., Segall, M., Zier, K. S., Sonde], P. M., Alter, B. J., and Bach, M. L. (1973). Science 180, 403. Bach, M. L., and Bach, F. H. (1974). Semin. Hematol. 11, 291. Bacon, L. D., Kite, J. H., Jr., and Rose, N. R. (1973). Transplantation 16, 591. Bain, B., and Lowenstein, L. (1964). Science 145, 1315. Bain, B., Vas, M., and Lowenstein, L. ( 1963). Fed. Proc., Fed. Am. SOC.Exp. Biol. 22, 428 (abstr. ). Bain, B., Vas, M. R., and Lowenstein, L. (1964). Blood 23, 108. Balner, H., Cleton, F. J., and Eemisse, J. G., eds. (1965). “Histocompatibility Testing 1965.” Munksgaard, Copenhagen. Belvedere, M. C., Curtoni, E. S., Dausset, J., Lamm, L. U., Mayr, W., van Rood, J. J., Svejgaard, A., and Piazza, A. (1975). Tissue Antigens 5, 99. Belzer, F. O., Perkins, H. A., Fortmann, J. L., Kountz, S. L., Salvatierra, O., Cochrum. K. C?. and Payne. R. .f 1974,). Lancet 1. 774. Benacerraf, B., and McDevitt, H. 0. (1972). Science 175, 273. Bennett, D., Boyse, E. A., and Old, L. J. (1972). In “Cell Interactions” ( L . G. Silvestri, ed. ), pp. 247-263. North-Holland Publ., Amsterdam. Berke, G., Ginsburg, W. A,, and Feldman, M. (1969). Immunology 16, 643. Bernoco, D., Cullen, S., Scudeller, G., Trinchieri, G., and Ceppellini, R. (1973). In “Histocompatibility Testing 1972” (J. Dausset and J. Colombani, eds.), pp. 527-537. Munksgaard, Copenhagen. Biggar, W. D., Park, B. H., and Good, R. A. (1975). In “Immunodeficiency in Man and Animals” ( D . Bergsma et al., eds.), pp. 385-390. Sinauer, Sunderland, Massachusetts. Biozzi, G., Stiffel, C., Mouton, D., Bouthillier, Y., and Decreusefond, C. (1968). Ann. lnst. Pasteur, Paris 115, 965. Blumenthal, M. N., Amos, D. B., Noreen, H., Mendell, N. R., and Yunis, E. J. (1974). Science 184, 1301. Bodmer, W. ( 1973). Biochem., Ser. One 10,295-328. Bodmer, W. F., Jones, E. A., Young, D., Goodfellow, P. N., Bodmer, J. G., Dick, H. M., and Steel, C. M. (1975). In “Histocompatibility Testing 1975” ( F . Kissmeyer-Nielsen, ed. ), pp. 677-684. Munksgaard, Copenhagen. Bondevik, H., Helgesen, A., Thoresen, A. B., and Thorsby, E. (1974). Tissue Antigens 4, 469, Bonnard, G. D., Chapuis, M., Glauser, A., Mempel, W., Baumann, P., GrosseWilde, H., and Albert, E. D. ( 1973). Transplant. Proc. 5, 1679. Brewerton, D. A., Caffrey, M., Hart, F. D., James, D. C. O., Nicholls, A., and Sturrock, R. D. (1973). Lancet 1, 904. Buckley, R. H. (1971). Prog. lmmunol. lnt. Congr. lmmunol., l s t , 1971 pp. 106 1-1 080. Buckley, R. H., SchiE, R. I., and Amos, D. B. (1972). J. lmmunol. 108, 34. Capkovi, J., and DAmant, P. (1974). Folia Biol. (Prague) 20, 101.

H U M A N MIXED-LYMPHOCYTE CULTURE REACTION

189

Ceppelhni, R. (1965). In “Histocompatibility Testing 1965” ( H . Balner, F. J. Cleton, and J. G. Eernisse, eds.), p. 158. Munksgaard, Copenhagen. Ceppelhni, R. (1971). Prog. Immunol. Int. Congr. Immttnol., lst, 1971 pp. 973-1025. Ceppellini, R., and van Rood, J. J. ( 1974). Scmin. Hematol. 11, 233. Ceppellini, R., Curtoni, E. S., Leigheb, G., Mattiuz, P. L., Miggiano, V. C., and Visetti, M. ( 1965). In “Histocompatibility Testing 1965” ( H . Balner, F. J. Cleton, and J. G. Eernisse, eds. ), pp. 13-22. Munksgaard, Copenhagen. Ceppellini, R., Curtoni, E. S., Mattiuz, P. L., hliggiano, V., Scudeller, G., and Serra, A. (1967). In “Histocompatibility Testing 1967” ( E . S. Curtoni, P. L. Mattiuz, and R. M. Tosi, eds. ), pp. 149-185. hlunksgaard, Copenhagen. Ceppellini, R., Bigliani, S., Curtoni, E. S., and Leigheb, G. (1969). Transplant. Proc. 1, 390. Cerottini, J. C., and Brunner, K. T. (1974). Adv. Immnnol. 18, 67. Cerottini, J. C., Engers, H. D., MacDonald, H. R., and Brunner, K. T. (1974). J. Erp. Med. 140, 703. Chalmers, D. G., Coulson, A. S., Evans, C., and Yealland, S. (1966). Int. Arch. Allergy Appl. Immunol. 30, 177. Charmot, D., Mawas, C. E., and Sasportes, hl. (1975a). Immnnogenetics 2, 465. Charmot, D., Mawas, C. E., Sasportes, M., Legrand, L., and Dausset, J. (1975b). In Histocompatibility Testing 1975” ( F. Kissnieyer-Nielsen, ed. ) pp. 867-870. Munksgaard, Copenhagen. Chen, T. R., McMorris, F. A., Creagan, R., Ricciuti, F., Tischfield, J., and Ruddle, F. ( 1973). Am. J . Hum. Genet. 25, 200. Chess, L., MacDermott, R. P., and Schlossman, S. F. (1974). J. Immtrnol. 113, 1113. Ciongoli, A. K., Platz, P., Dupont, B., Svejgaard, A., Fog, T., and Jersild, C. (1973). Lancet 2, 1147. Cochrum, K. C., Perkins, H. A., Payne, R. O., Kountz, S. L., and Belzer, F. 0. (1973). Transplant. Proc. 5, 391. Croy, B. A., and Osoba, D. (1973). Transplant. Proc. 5, 1721. Cunningham-Rundles, C., Jersild, C., Svcjgaard, A., Dupont, B., and Good, R. A. (1975). Proc. Natl. Acad. Sci. U.S.A. 72,5081. Cunningham-Rundles, S., Dupont, B., Posner, J. B., Hansen, J. A., and Good, R. A. (1975). Lan.cet 2, 1201. Curtoni, E. S., Mattiuz, P. L., and Tosi, R. M., eds. ( 1967). “Histocompatibility Testing 1967.” Munksgaard, Copenhagen. Dausset, J. (1958). Acta Haematol. 20, 156. Dausset, J. (1971). Transplant. Prm. 3, 8. Dausset, J. (1973). In “Tissue Typing and Organ Transplantation” (E. J. Yunis, R. A. Gatti, and D. B. Amos, eds.), pp. 13-47. Academic Press, New York. Dausset, J. and Colombani, J., eds. ( 1973 ). “Histocompatibility Testing 1972.” Munksgaard, Copenhagen. Dausset, J,, and Hors, J. ( 1972). Nature (London),New Biol. 238, 150. Dausset, J., Ivanyi, P., and Ivanyi, D. ( 1965). In “Histocompatibility Testing 1965” ( H . Balner, F. J. Cleton, and J. G. Eernisse, eds.), pp. 51-62. Munksgaard, Copenhagen. Dausset, J., Ivanyi, P., Colombani, J., Feingold, N., and Legrand, L. ( 1967). In “Histocompatibility Testing 1967” ( E . S. Curtoni, P. L. Mattiuz, and R. M. Toni, eds. ), pp. 188-201. Munksgaard, Copenhagen. Dausset, J., Colombani, J., Legrand, L., and Fellous, M. (1970). In “Histocompatibility Testing 1970” ( P. I. Terasaki, ed.), pp. 53-75. Munksgaard, Copenhagen.

190

DUPONT, HANSEN, AND YUNIS

Dausset, J., Sasportes, M., and Lebrun, A. (1973). Transplant. Proc. 5, 1511. Dausset, J , , Hors, J., Busson, M., Festenstein, H., Oliver, R. T. D., Chir, B., Paris, A. M. I., and Sachs, J. A. (1974). N . Engl. J. Med. 290, 979. Dausset, J., Hors, J,, and Busson, M. ( 1976). In press. David, C. S., Shreffler, D. C., and Frelinger, J. A. (1973). Proc. Natl. Acad. Sci. U . S. A . 70, 2509. Day, N. K., L’EspCrance, P., Good, R. A,, Michael, A. F., Hansen, J. A., Dupont, B., and Jersild, C. (1975). J. Exp. Med. 141, 1464. Degos, L., and Dausset, J. (1974). Zrnrntrnogenetics 1, 195. Degos, L., Colombani, J., Chaventre, A,, Bengtson, B., and Jacquard, A. (1974). Nature ( L o n d o n ) 249, 62. DCmant, P., Capkovi, J., Hinzovd, E., and Vordcovd, B. (1973). Proc. Natl. Acad. Sci. U . S. A. 70, 863. Dineen, J. K. ( 1964). Nature (London) 202,101. Dorf, M. E., Eguro, S. Y., Flye, M. W., and Amos, D. B. (1973). Transplantation 15, 253. Dossetor, J. B., Kovithavongs, T., Marchuk, L., Butcher, D., St. Louis, G., and Schlaut, J. ( 1975). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 620-624. Munksgaard, Copenhagen. duBois, R., Meinesz, A,, Bierhorst-Eijlander, A., Groenewoud, M., Schellekens, P. T. A., and Eijsvoogel, V. P. ( 1974). Tissue Antigens 4,458. Dunn, L. C., and Gluecksohn-Waelsch, S. (1953). Genetics 38, 512. Dupont, B., Nielsen, L. S., and Svejgaard, A. (1971). Lancet 1, 1336. Dupont, B., Andersen, V., Hansen, G. S., and Svejgaard, A. (1972). Transplantation 14, 557.

Dupont, B., Andersen, V., Ernst, P., Faber, V., Good, R. A., Hansen, G. S., Henriksen, K., Jensen, K., Juhl, F., Killmann, S. Aa., Koch, C., Muller-Berat, N., Park, B. H., Svejgaard, A., Thomsen, M., and Wiik, A. (1973a). Transplant. Proc. 5. 905. Dupont, B., Jersild, C., Hansen, G. S., Nielsen, L. S., Thomsen, M., and Svejgaard, A. (1973b). Transplant. Proc. 5, 1481. Dupont, B., Jersild, C., Hansen, G. S., Nielsen, L. S., Thomsen, M., and Svejgaard, A. ( 1 9 7 3 ~ )Transplant. . Proc. 5, 1543. Dupont, B., Good, R. A,, Hansen, G. S., Jersild, C., Nielsen, L. S., Park, B. H., Svejgaard, A., Thomsen, M., and Yunis, E. J. (1974a). Proc. Natl. Acacl. Sci. U . S. A. 71, 52. Dupont, B., O’Reilly, R. J., Jersild, C., and Good, R. A. (197413). In “Progress in Immunology II” (L. Brent and J. Holborow, eds. ), Vol. 5, pp. 203-214. NorthHolland Publ., Amsterdam. Dupont, B., Hansen, J. A., and Good, R. A. (1975a). Ado. Biosci. 165-190. Dupont, B., Hansen, J. A., and Good, R. A. (1975b). Transplantation 20, 255. Dupont, B., Yunis, E. J., Hansen, J. A., Reinsmoen, N., Suciu-Foca, N., Mickelson, E., and Amos, D. B. ( 1 9 7 5 ~ ) .In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 547-551. Munksgaard, Copenhagen. du Toit, E. D. (1973). Quoted by Eijsvoogel (1974). Dutton, R. W. (1966). J . Exp. Med. 123, 665. Eichwald, E. J., and Weissman, I. L. (1966). Ann. N.Y. Acad. Sci. 129, 94. Eijsvoogel, V. P. (1974). Sernin. Hematol. 11, 305. Eijsvoogel, V. P., Schellekens, P. T. A., Breur-Vriesendorp, B., van Leeuwen, A., Koch, C., and van Rood, J. J. (1970). In “Histocompatibility Testing 1970”

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

191

( P. I. Terasaki, ed. ), pp. 523-529. Munksgaard, Copenhagen. Eijsvoogel, V. P., Schellekens, P. T. A., Breur-Vriesendorp, B., Koning, L., Koch, C., van Leeuwen, A,, and van Rood, J. J. ( 1971). Transplant. Proc. 3, 85. Eijsvoogel, V. P., Koning, L., de Groot-Kooy, L., Huismans, L., van Rood, J. J., van Leeuwen, A., and du Toit, E. D. (19721). Transplant. Proc. 4, 199. Eijsvoogel, V. P., van Rood, J. J., du Toit, E. D., and Schellekens, P. T. A. (1972b). Eur. J . Immrrnol. 2, 413. Eijsvoogel, V. P., du Bois, R., Melief, C. J. M., Zeyleniaker, W. P., Raat-Kooning, L., and de Groot-Kooy, L. (1973a). Transplant. Proc. 5, 415. Eijsvoogel, V. P., du Bois, M. J. G. J., Meinesz, A,, Bierhorsh-Eijlander, A., Zeylemaker, W. P., and Schellekens, P. T. A. (1973b). Transplant. Proc. 5, 1675. Eijsvoogel, V. P., du Bois, M. J. G. J., Melief, C. J. M., de Groot-Kooy, M. L., Koning, C., van Rood, J. J., van Leeuwen, A,, du Toit, E., and Schellekens, P. T. A. ( 1 9 7 3 ~ ) .In “Histocompatibility Testing 1972” ( J . Dausset and J . Colombani, eds. ), pp. 501-508. Munksgaard, Copenhagen. Elkins, W. L., Kavathas, P., and Bach, F. H. (1973). Transplant. Proc. 5, 1759. Falchuk, Z. M., Rogentine, G. N., and Strober, W. (1972). J . Clin. Inoest. 51, 1602. Feldman, M. ( 1973). Tronsplant. Proc. 5, 1803. Ferrara, G. B., Tosi, R. M., Antonelli, P., and Longo, A . (1975). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 608-619. Munksgaard, Copenhagen. Ferreira, A,, and Nussenzweig, V. (1975). I. Erp. Med. 141, 513. Festenstein, H. (1970). Tronsplant. Reo. 3,74. Festenstein, H. ( 1974). Transplantation 18, 555. Festenstein, H., and Dkniant, P. ( 1975). In “Histocompatibility Testing 1975” ( F. Kissnieyer-Nielsen, ed. ) pp. 824-826. Munksgaard, Copenhagen. Festenstein, H., Abbasi, K., and DBmant, P. (1974). J . Immunogenet. 1, 47. Fink, M. A., and Quinn, V. A. (1953). J . Immunol. 70, 61. Ford, E. B. (1965). In “Genetic Polymorphism.” Faber & Faber, London. Fradelizi, D., and Dausset, J. ( 1975). Eur. J . Immunol. 5, 295. Fradelizi, D., Mawas, C . , Charmot, D., and Sasportes, M. (1975). In “Histocompatibility Testing 1975” ( F. Kissnieyer-Nielsen, ed. ), pp. 584-587. Munksgaard, Copenhagen. Frelinger, J. A., Niederhuber, J. E., David, C. S., and ShrefRer, D. C. (1974). J . Exp. Med. 140, 1273. Friend, P., Kim, Y., Handwerger, B., Reinsmoen, N., Michael, A., and Yunis, E. ( 1975). In “Histocompatibility Testing 1975” (F. Kissmeyer-Nielsen, ed. ), pp. 928-932. Munksgaard, Copenhagen. Fu, S. M., Kunkel, H. G., Brusman, H. P., Allen, F. H., Jr., and Fotino, M. (1974). J . Exp. Med. 140, 1108. Fu, S. M., Stern, R., Kunkel, H. G., Dupont, B., Hansen, J. A., Day, N. K., Good, R. A., Jersild, C., and Fotino, M. (1975a). J . Exp. Med. 142,495. Fu, S. M., Stern, R., Kunkel, H. G., Dupont, B., Hansen, J. A., Day, N. K., Good, R. A., Jersild, C., and Fotino M. (1975b). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 933-936. Munksgaard, Copenhagen. Fu, S. M., Winchester, R. J., and Kunkel, H. G. ( 1 9 7 5 ~ ) J. . Immrrnol. 114, 250. Gale, R. P., Opelz, G., Sparkes, R., Zighelboim, J., Fahey, J., Cline, M., Craddock, C., Young, L., Feig, S., Smith, G . , Terasaki, P. I., and Rupard, N. (1975a). Transplant. Proc. 7 , Suppl. 1, 833. Gale, R. P., Opelz, G., Sparkes, R., and The UCLA Bone Marrow Transplantation Group. ( 197513). Transplantation 20, 194.

192

DUPONT, HANSEN, AND YUNIS

Gasser, D. L., and Silvers, W. K. ( 1974). Ado. Immunol. 18, 1. Gatti, R. A., Meuwissen, H. J., Allen, H. D., Hong, R., and Good, R. A. (1968). Lancet 2, 1366. Gatti, R. A., Meuwissen, H. J., Terasaki, P. I., and Good, R. A. (1971). Lancet 1, 239. Gatti, R. A,, Yunis, E. J., and Good, R. A. (1973). Clin. Exp. Immunol. 13, 427. Ginsburg, H. (1968). Immunology 14, 621. Goldman, M. B., and Goldman, J. N. (1975). Fed. PTOC., Fed. Am. SOC. E x p . B i d . 34, 979 (abstr.). Corer, P. A., and Mikulska, Z. B. (1959). Proc. R. SOC. London, Ser. B 151, 57. Corer, P. A., and Schiitze, H. (1938). J. Hyg. 38, 647. Gotze, D., Reisfeld, R. A., and Klein, J. (1973). J. Exp. Med. 138, 1003. Goulmy, E., Tennijtelen, A., Keuning, J. J., and van Rood, J. J. (1975). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 845-848. Munksgaard, Copenhagen. Greenberg, L. J., Reinsmoen, N., and Yunis, E. J. (1973). Transplantation 16, 520. Greenberg, L. J., Gray, E. D., and Yunis, E. J. (1975). J. Exp. Med. 141, 935. Grosse-Wilde, H., Netzel, B., Mempel, W., Ruppelt, W., Brehm, G., Bertrams, J., Ewald, R., Lenhard, V., Rittner, C., Scholz, S., and Albert, E. D. (1975). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 526-532. Munksgaard, Copenhagen. Grumet, F. C., and Leventhal, B. G. (1970). Transplantation 9, 405. Grumet, F. C., Mitchell, G. F., and McDevitt, H. 0. (1971). Ann. N.Y. Acad. Sci. 190, 170. Grumet, F. C., Konishi, J., Payne, R. O., and Kriss, J. P. (1973). Clin. Res. 21, 493 (abstr.). Grunnet, N., and Kristensen, T. ( 1975). In “Histocompatibility Testing 1975” (F. Kissmeyer-Nielsen, ed. ), pp. 862-866. Munksgaard, Copenhagen. Grunnet, N., Kristensen, T., JGrgensen, F., and Kissmeyer-Nielsen, F. ( 1974). Tissue Antigens 4, 218. Hammerling, G. J., Deak, B. D., Mauve, G., Hammerling, U., and McDevitt, H. 0. (1974). Immunogenetics 1, 68. Hansen, H. E., Ryder, L. P., and Nielsen, L. S. ( 1975). Tissue Antigens 6, 275. Hansen, J. A., Dupont, B., Rubinstein, P., Suciu-Foca, N., Fu, S. M., Mickelson, E., Whitsett, C., Jersild, C., Kunkel, H. G., Day, N. K., Good, R. A., Thomas, E. D., Reemtsma, K., Allen, F. H., Jr., and Fotino, M. (1975). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 470-478. Munksgaard, Cqpenhasen. Hardy, D. A., and Ling, N. R. (1969). Nature ( L o n d o n ) 221,545. Harrison, M. R., and Paul, W. E. (1973). J. E x p . Med. 138, 1602. Hartzman, R. J., Segall, M., Bach, M. L., and Bach, F. H. (1971). Transplantation 11, 268. Hartzman, R. J., Bach, M. L., and Bach, F. H. (1972). Cell. Immunol. 4, 182. Hauptfeld, V., Klein, D., and Klein, J. (1973). Science 181, 167. Haverkorn, M. J., Hofman, B., Masurel, N., and van Rood, J. J. (1975). Transplant. Rev. 22, 120. Hayry, P., and Andersson, L. C. ( 1974). Scand. J. Immunol. 3, 823. Hayry, P., and Defendi, V. (1970). Science 168, 133. Hayry, P., Andersson, L. C., Nordling, S., and Virolainen, M. (1972). Transplant. Reo. 12, 91.

H U M A N MIXED-LYMPHOCYTE CULTURE REACTION

193

Hildemann, W. H. ( 1973). Annu. Ret;. Genet. 7, 19. Hirschhorn, K., Bach, F., Kolodny, R. L., Firschein, I. L., and Hasheni, N. (1963). Science 142, 1185. Horowitz, S. D., Groshong, T., Bach, F. H., Hong, R., and Yunis, E. J. (1975). Lancet 2, 431. Hors, J., Feingold, N., Fradelizi, D., and Dausset, J. (1971). Lancet 1, 609. Huber, B., Pelia-Martinez, J., and Festenstein, H. ( 1973). Transplant. Proc. 5, 1373. Huemer, R. P., Keller, L. S., and Lee, K. D. (1968). Transplantation 6, 706. Ipsen, J. (1959). J. Immunol. 83, 448. Janis, M., Hartzmann, R. J., and Bach, F. H. (1970). In “Histocompatibility Testing 1970” ( P. I. Terasaki, ed.), pp. 517-521. Munksgaard, Copenhagen. Jersild, C., Dupont, B., Fog, T., Hansen, G. S., Nielsen, L. S., Thomsen, M., and Svejgaard, A. (1973a). Transplant. Proc. 5, 1791. Jersild, C., Fog, T., Hansen, G. S., Thomsen, M., Svejgaard, A,, and Dupont, B. (1973b). Lancet 2, 1221. Jersild, C., Svejgaard, A., Fog, T., and Ammitzboll, T. ( 1 9 7 3 ~ )Tissue . Antigens 3, 243. Jersild, C., Dupont, B., Fog, T., Platz, P. J., and Svejgaard, A. (1975). Transplant. Reu. 22, 148. Johnson, A . H., Amos, D. B., Noreen, H., and Yunis, E. J. (1975). Transplantation 20, 291. Jones, G. (1972). Clin. Exp. Irnrnunol. 12, 391. Jongsma, A., Van Someren, H., Westerveld, A., Hagemeijer, A., and Pearson, P. (1973). Hurnangenetik 20, 195. JBrgensen, F., and Kissineyer-Nielsen, F. ( 1973). Symp. Ser. Imrnunobiol. Stand. 18, 135-140. JZrgensen, F., and Lamm, L. U. ( 1974). Tissue Antigens 4,482. JBrgensen, F., Lamm, L. U., and Kissmeyer-Nielsen, F. ( 1973). Tissue Antigens 3, 323. JZrgensen, F., Lamm, L. U., and Kissmeyer-Nielsen, F. ( 1974a). Tissue Antigens 4, 404. JBrgensen, F., Lamni, L. U., and Kissmeyer-Nielsen, F. (1974b). Tissue Antigens 4, 419. J@rgensen, F., Lamm, L. U., and Kissmeyer-Nielsen, F. (1975a). Tissue Antigens 5, 262. JBrgensen, F., Lamm, L. U., Ferrara, G. B., Ipsen, S., JZrgensen, H., and KissmeyerNielsen, F. ( 1975b ). In “Histocompatibility Testing 1975” ( F. KissnieyerNielsen, ed. ), pp. 519-525. Munksgaard, Copenhagen. Kaakinen, A., Pirskanen, R., and Tiilikainen, A. (1975). Tissue Antigens 6, 175. Kadowaki, J., Thompson, R. I., Zuelzer, W. W., Woolley, P. V., Jr., Brough, A. J., and Gruber, 0. ( 1965 ) . Lancet 2, 1152. Kantor, F. S., Ojeda, A., and Benacerraf, B. (1963). J . Exp. Med. 117, 55. Kasakura, S., and Lowenstein, L. ( 1965). In “Histocompatibility Testing 1965” ( H . Balner, F. J. Cleton, and J. G. Eernisse, eds.), pp. 211-212. Munksgaard, Copenhagen. Keuning, J. J., Termijtelen, A., BloussB van Oud Alblas, A., van den Tweel, J. G., Schreuder, I., and van Rood, J. J. (1975a). Transplant. Proc. 7, Suppl. 1, 35. Keuning, J. J., van den Tweel, J. G., Gabb, B. W., Termijtelen, A,, Goulmy, E., Blokland, E., Elferink, B. G., and van Rood, J. J. (1975b). Tissue Antigens 6, 107.

194

DUPONT, HANSEN, A N D YUNIS

Keuning, J. J., Termijtelen, A., Bloussk van Oud Alblas, A., Gabb, B. W., D ’ h a r o , J., and van Rood, J. J. ( 1 9 7 5 ~ ) .In “Histocompatibility Testing 1975” (F. Kissmeyer-Nielsen, ed. ), pp. 533-543. Munksgaard, Copenhagen. Killman, S. A. ( 1960). “Leukocyte Agglutinins.” Blackwell, Oxford. Kissmeyer-Nielsen, F., ed. ( 1975). “Histocompatibility Testing 1975” Munksgaard, Copenhagen. Kissmeyer-Nielsen, F., and Thorsby, E. ( 1970). Transplant. Reo. 4, 11. Kissmeyer-Nielsen, F., Svejgaard, A., Ahrons, S., and Nielsen, L. S. (1969). Nature (London) 224,75. Kissmeyer-Nielsen, F., Svejgaard, A., Sflrensen, S. F., Nielsen, L. S., and Thorsby, E. ( 1970). Nature (London) 228,63. Klein, J. ( 1967). In “Histocompatibility Testing 1967” ( E . S. Curtoni, P. L. Mattiuz, and R. M. Tosi, eds.), pp. 21-28. Munksgaard, Copenhagen. Klein, J. ( 1975). “Biology of the Mouse Histocompatibility-2 Complex.” SpringerVerlag, Berlin and New York. Klein, J., and Park, J. M. (1973). J. Erp. Med. 137, 1213. Koch, C., Henriksen, K., Juhl, F., Wiik, A., Faber, V., Andersen, V., Dupont, B., Hansen, G. S., Svejgaard, A., Thomsen, M., Ernst, P., Killmann, S. A., Good, R. A:,Jensen, K., and Muller-Berat, N. (1973). Lancet 1, 1146. Koch, C. T., Frederiks, E., Eijsvoogel, V. P., and van Rood, J. J. (1971). Lancet 2, 1334. Koch, C. T., van Hooff, J. P., van Leeuwen, A., van den Tweel, J. G., Frederiks, E., van der Steen, G. J., Schippers, H. M. A., and van Rood, J. J. ( 1973). In “Histocompatibility Testing 1972” ( J . Dausset and J. Colombani, eds. ), pp. 521-524. Munksgaard, Copenhagen. Kovithavongs, T., Hyshka, L., McConnachie, P. R., and Dosseter, J. B. (1975). Tissue Antigens 5, 165. Kristensen, T., and Grunnet, N. ( 1975). In “Histocompatibility Testing 1975” ( E’. Kissmeyer-Nielsen, ed. ), pp. 835-844. Munksgaard, Copenhagen. Lafferty, K. J., Walter, K. Z., Scollay, R. G., and Killby, V. A. A. (1972). Transplant. Rev. 12, 198. Lamm, L. U., Svejgaard, A., and Kissmeyer-Nielsen, F. (1971). Nature (London), New Biol. 231. 109. Lamm, L. U., Friedrich, U., Petersen, G. B., Jflrgensen, J., Nielsen, J., Therkelsen, A. J., and Kissmeyer-Nielsen, F. ( 1974). Hum. Hered. 24,273. Layrisse, Z., Heinen, H. D., Malavk, I., Pulido, M., and Layrisse, M. (1975). Transplant. Proc. 7, Suppl. 1, 45. Lebrun, A., Sasportes, M., Lebrun, D., and Dausset, J. (1971). C. R. Hebd. Seances Acad. Sci. 273, 2130. Lebrun, A,, Sasportes, M., and Dausset, J. (1973). Transplant. Proc. 5, 363. Legrand, L., and Dausset, J. (1975a). In “Histocompatibility Testing 1975” (F. Kissmeyer-Nielsen, ed. ), pp. 665-670. Munksgaard, Copenhagen. Legrand, L., and Dausset, J. (1975b). Transplant Proc. 7, 5. L’Espkrance, P., Hansen, J. A., Jersild, C., O’Reilly, R., Good, R. A., Thomsen, M., Nielsen, L. S., Svejgaard, A., and Dupont, B. (1975). Transplant. Proc. 7, Suppl. 1, 823. Levine, B. B., and Benacerraf, B. ( 1964). J. Exp. Med. 120,955. Levine, B. B., and Benacerraf, B. (1965). Science 147, 517. Levine, B. B., Ojeda, A., and Benacerraf, B. ( 1963). 1. E r p . Med. 118, 953. Levine, B. B., Stember, R. H., and Fotino, M. (1972). Science 178, 1201.

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

195

Lightbody, J. J., Bernoco, D., Miggiano, V. C., and Ceppellini, R. (1971). G . Batteiiol., Virol., Immunol. Ann. Osp. Maria Vittoria Torino. 64, 243. Lilly, F. (1967). Science 155, 461. Lilly, F., and Pincus, T. (1973). Adu. Cancer Res. 17,231. Lilly, F., Boyse, E. A., and Old, L. J. ( 1964). Lancet 2, 1207. Livnat, S., Klein, J., and Bach, F. H. (1973). Nature (London), New Biol. 243, 42. Lohrmann, H. P., Novikovs, L., and Craw, R. G., Jr. (1974). Nature (London) 250, 144. Lonai, P., and Grumet, F. C. ( 1975). Transplant. Proc. 7, 149. Lonai, P., and McDevitt, H. 0. (1974). J . Exp. Med. 140, 1317. Long, M. A., Handwerger, B., and Yunis, E. J. (1975). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 849-854. Munksgaard, Copenhagen. Low, B., Messeter, L., Maansson, S., and Lindholm, T. (1974). Tissue Antigens 4, 405 ( abstr. ). McDevitt, H. O., and Benacerraf, B. (1969). Adu. Immunol. 11, 31. McDevitt, H. O., and Bodmer, W. F. ( 1974). Lancet 1, 1269. McDevitt, H. O., and Chinitz, A. (1969). Science 163, 1207. McDevitt, H. O., and Landy, M., eds. (1973). “Genetic Control of Immune Responsiveness.” Academic Press, New York. McDevitt, H. O., and Sela, M. (1965). J . Exp. Med. 122, 517. McDevitt, H. O., Shreffler, D. C., and Stimpfling, J. H. (1969). J . Clin. Inuest. 48, 57a ( abstr. ) . McDevitt, H. O., Oldstone, M. B. A., and Pincus, T. (1974). Transplant. Rev. 19, 209. MacDonald, H. R., Engers, H. D., Cerottini, J. C., and Brunner, K. T. (1974). J . Exp. Med. 140, 718. MacLaurin, B. P. (1972). Clin. Exp. lmmunol. 10, 649. Mann, D. L., Abelson, L., Harris, S., and Amos, D. B. (1975a). J . E x p . Med. 142, 84. Mann, D. L., Abelson, L., Henkart, P., Harris, S., and Amos, D. B. (197513). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 705-707. Munksgaard, Copenhagen. Mann, D. L., Abelson, L., Harris, S., and Amos, D. B. (1976). A’ature (London) 259, 145. Marsh, D. G., Bias, W. B., and Hsu, S. H. (1973). Science 179, 691. Mattiuz, P. L., Ihde, D., Piazza, A., Ceppellini, R., and Bodmer, W. F. (1970). In “Histocompatibility Testing 1970” ( P. I. Terasaki, ed. ), pp. 193-205. Munksgaard, Copenhagen. Mawas, C. E., Christen, Y., Legrand, L., and Dausset, J. (1973). Transplant. Proc. 5, 1691. Mawas, C. E., Charmot, D., and Sasportes, M. (1975a). Immunogenetics 2, 449. Mawas, C. E., Charmot, D., and Sasportes, M. (1975b). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 855-857. Munksgaard, Copenhagen. Mawas, C. E., Charmot, D., and Sasportes, M. ( 1 9 7 5 ~ ) .In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 858-861. Munksgaard, Copenhagen. Mayr, W. R., Bernoco, D., deMarchi, M., and Ceppellini, R. (1973). Transplant. Proc. 5, 1581.

196

DUPONT, HANSEN, AND YUNLS

Mempel, W., Albert, E., and Burger, A. ( 1972). Tissue Antigens 2, 250. Mempel, W., Grosse-Wilde, H., Albert, E., and Thierfelder, S. ( 1973a). Transplant. Proc. 5, 401. Mempel, W., Grosse-Wilde, H., Baumann, P., Netzel, B., Steinbauer-Rosenthal, I., Scholz, S., Bertrams, J., and Albert, E. D. (1973b). Transplant Proc. 5, 1529. Mempel, W., Grosse-Wilde, H., Netzel, B., Luboldt, W., Bertrams, J., and Albert, E. D. (1975). Transplant. Proc. 7, Sup$ 1, 81. Meo, T., Vives, J., Miggiano, V., and Shreffler, D. C. (1973a). Transplant. P ~ o c . 5, 377. Meo, T., Vives, G., Rijnbeek, A. M., Miggiano, V. C., Nabholz, M., and Shreffler, D. c. (1973b). TTU?Isplant.Pfoc. 5, 1339. Meo, T., David, C. S., Rijnbeek, A. M., Nabholz, M., Miggiano, V. C., and Shreffler, D. C. ( 1975). Transplant. Proc. 7, Suppl. 1, 127. Middleton, J., Crookston, M. C., Falk, J. A,, Robson, E. B., Cook, P. J. L., Batchelor, J. R., Bodmer, J., Ferrara, G. B., Festenstein, H., Harris, R., Kissmeyer-Nielsen, F., Lawler, S. D., Sachs, J. A,, and Wolf, E. (1974). Tissue Antigens 4, 366. Miggiano, V. C., Bernoco, D., Lightbody, J., Trinchieri, C., and Ceppellini, R. (1972). Transplant. Proc. 4,231. Moller, E., Link, H., Matell, G., Olhagen, B., and Stendahl, L. (1975). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 778-781. Munksgaard, Copenhagen. Mdler, E., Hammarstrorn, L., Smith, E., and Matell, G. (1976). Tissue Antigens 7, 39. Morris, P. (1974). Contemp. Top. Immunobiol. 3, 141. Morton, J. R., Gilmour, D. G., McDermid, E. M., and Odgen, A. L. (1965). Genetics 51, 97. Miihlbock, O., and Dux, A. (1971). Transplant. Proc. 3, 1247. Naiman, J. L., Punnett, H. H., Lischner, H. W., DestinB, M. L., and Arey, J. B. (1969). N . Engl. J . Med. 281, 697. Nerup, J., Platz, P., Ortved-Andersen, O., Christy, M., LyngsZe, J., Poulsen, J. E., Ryder, L. P., Nielsen, L. S., Thomsen, M., and Svejgaard, A. (1974). Lancet 2, 864. Netzel, B., Grosse-Wilde, H., Rittner, C., Pretorius, A. M. G., Scholz, S., and Albert, E. D. ( 1975). In “Histocompatibility Testing 1975” (F. Kissnieyer-Nielsen, ed. ), pp. 955-959. Munksgaard, Copenhagen. Nielsen, L. S., and Svejgaard, A. ( 1972). Tissue Antigens 2, 316. Nomenclature Committee. (1968). Bull. W. H . 0 . 39, 483. Oldstone, M. B. A., Dixon, F. I., Mitchell, G . F., and McDevitt, H. 0. (1973). J . Erp. Med. 137, 1201. Oliver, R. T. D., Sachs, J. A., and Festenstein, H. (1972). Lancet 2, 1381. Opelz, G . (1975). Transplantation 19, 48. Opelz, G., and Terasaki, P. I. (1972). Transplant. Proc. 4, 433. Opelz, G., Mickey, M. R., and Terasaki, P. I. (1974). Transplantation 17, 371. Oppltovh, L., and Dkmant, P. ( 1973). Transplant. Proc. 5, 1367. OReilly, R. J.. Patterson, J. H., Bach, F. H., Bach, M. L., Hong, R., KissmeyerNielsen, F., and Therkelsen, A. J. (1973). Transplantation 15, 505. Osoba, D., and Falk, 1. ( 1974). Cell. Immunol. 10, 117. Page, A. R., Sharp, H. L., Greenberg, L. J., and Yunis, E. J. (1975). J . Clin. Inoest. 56. 530. Palm, J: (1970). Transplant. Proc. 2, 162.

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

197

Park, B. H., Biggar, W. D., Yunis, E. J., and Good, R. A. (1973). Transplant. Proc. 5, 899. Passmore, H. C., and Shreffler, D. C. (1971). Biochem. Genet. 5,201. Payne, R., Tripp, M., Weigle, J., Bodmer, W., and Bodmer, J. (1964). Cold Spring Harbor Symp. Quant. Biol. 29, 285. Pentycross, C. R., Klouda. P. T., and Lawler, S. D. (1972). Transplantation 14, 657. Persson, I., Ryder, L. P., Nielsen, L. S., and Svejgaard, A. (1975). Tissue Antigens 6, 50. Pierres, M., Fradelizi, D., Neauport-Sautes, C., and Dausset, J. (1975). Tissue Antigens 5, 266. Pinchuck, P., and Maurer, P. H. (1968). In “Regulation of the Antibody Response” (B. Cinder, ed. ), pp. 97-113. Thomas, Springfield, Illinois. Pincus, T., Rowe, W. P., and Lilly, F. (1971). J. Exp. Med. 133, 1234. Piquet, P. F., Dewey, H. K., and Vassalli, P. (1975). J. Erp. Med. 141, 775. Plate, J. M. D. (1973). Transplant. Proc. 5, 1351. Plate, J. M. D., and McKenzie, I. F. C. (1973). Nature (London), New B i d . 245, 247. Plate, J. M. D., Ward, F. E., and Amos, D. B. (1970). In “Histocompatibility Testing 1970” ( P . I. Terasaki, ed.), pp. 531-535. Munksgaard, Copenhagen. Platz, P., Dupont, B., Fog, T., Ryder, L., Thomsen, M., Svejgaard, A., and Jersild, C. (1974a.).Proc. R . Soc. Med. 67, 1133. Platz, P., Ryder, L., Nielsen, L. S., Svejgaard, A., Thomsen, M., Nerup, J., and Christy, M. (1974b). Lancet 2,289. Playfair, J. H. L. (1968). Immunology 15, 35. Reinsmoen, N., Stewart, M., Emme, L., Hanrahan, L. A., Dupont, B., Hansen, J. A., Friend, P., Amos, D. B., and Yunis, E. J. (1975). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nieken, ed. ), pp. 439-463. Munksgaard, Copenhagen. Revillard, J. P., Robert, M., Betuel, H., Latour, M., Bonneau, M., Brochier, J., and Traeger, J. (1972). Transplant. Proc. 4, 173. Rittner, C., Grosse-Wilde, H., Rittner, B., Netzel, B., Scholz, S., Lorenz, G., and Albert. E. D. 1975a ). Humangenetik 27, 173. Rittner, C., Hauptmann, G., Grosse-Wilde, H., Grosshans, E., Tongio, M. M., and Mayer, S. ( 1975b). In “Histocompatjbiljty Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 945-954. Munksgaard, Copenhagen. Rocklin, R. E., MacDerniott, R. P., Chess, L., Schlossman, S. F., and David, J. R. (1974). J. E x p . Med. 140, 1303. Rode, H. N., and Gordon, J. (1974). Cell. Immunol. 13, 87. Rodey, G. E., Bortin, M. M., Bach, F. H., and Rinim, A. A. (1974). Transplantation 17, 84. Rose, N. R., Kite, J. H., Jr., Vladutio, A. O., Tomazic, V., and Bacon, L. D. (1973). Int. Arch. Allergy Appl. Zmmrtnol. 45, 138. Rowe, W. P., Hartley, J. W., and Bremner, T. (1972). Science 178, 860. Russell, P. S., Amos, D. B., and Winn, H. J., eds. (1965). “Histocompatibility Testing 1964,” Publ. No. 1229, Nat. Acad. Sci.-Nat. Res. Counc., Washington, D.C. Russell, T. J., Schultes, L. M., and Kuban, D. J. (1972). N . Engl. J. Med. 287, 738. RychlikovQ, M., D&mant, P., and Ivinyi, P. (1970). Folia Biol. (Prague) 16, 218. RychlikovQ, M., Dkniant, P., and Ivinyi, P. (1971). Nature ( L o n d o n ) , Ncw Biol. 230, 271. Ryder, L. P., Thomsen, M., Platz, P., and Svejgaard, A. (1975). In “Histocom-

198

DUPONT, HANSEN, AND YUNIS

patibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 557-562. Munksgaard, Copenhagen. Sachs, D. H., and Cone, J. L. (1973). J. E r p . Med. 138, 1289. Sachs, J. A., Sterioff, S., Robinette, M., Wolf, E., Curry, H. L. F., and Festenstein, H. (1974). Tissue Antigens 5, 383. Salmon, S. E., Krakauer, R. S., and Whitmore, W. F. (1971). Science 172, 490. Sandberg, L., Thorsby, E., Kissmeyer-Nielsen, F., and Lindholm, A. (1970). In “Histocompatibility Testing 1970” ( P. I. Terasaki, ed. ), pp. 165-169. Munksgaard, Copenhagen. Sandberg-Wollheim, M., Platz, P., Ryder, L. P., Nielsen, L. S., and Thomsen, M. (1975). Actu Neurol. Scund. 52, 161. Sasazuki, T., McMichael, A., Payne, R., Grumet, F. C., Radvany, R., Colombe, B., McDevitt, H., Stastny, P., Marchuk, L., and Dosseter, J. (1975). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 464-469. Munksgaard, Copenhagen. Sasportes, M., Lebrun, A., Rapaport, F. T., and Dausset, J. (1972). Transplant. Proc. 4, 209. Sasportes, M., Lebrun, A., Rapaport, F. T., and Dausset, J. (1973). Transplant. Proc. 5, 353. Scanditransplant Report. ( 1975). Lancet 1,240. Schapira, M., and Jeannet, M. (1974). Tissue Antigens 4, 178. Scheibel, I. F. ( 1943). Acta Pathol. Microbiol. Scand. 20, 464. Schellekens, P. T. A,, and Eijsvoogel, V. P. (1970). Clin. Exp. Immunol. 7, 229. Schendel, D. J., and Bach, F. H. (1974). J. Erp. Med. 140, 1534. Schlosstein, L., Terasaki, P. I., Bluestone, R., and Pearson, C. M. (1973). N . Engl. J. Med. 288, 704. Segall, M., Omodei-Zorini, C., Bach, F. H., Jorgensen, F., and Kissmeyer-Nielsen, F. ( 1973). Transplant. Proc. 5, 383. Seigler, H. F., Ward, F. E., Amos, D. B., Phaup, M. B., and Stickel, D. L. (1971). 1. Erp. Med. 133, 411. Seigler, H. F., Stickel, D. L., Gunnells, J. C., Jr., McCoy, R., Ward, F. E., Anderson, E. E., and Amos, D. B. (1972). Transplant. Proc. 4,439. Sengar, D. P. S., Mickey, M. R., Myhre, B. A., Chen, H. H., and Terasaki, P. I. (1971). Transfusion 11,251. Sheehy, M. J., Sondel, P. M., Bach, M. L., Wank, R., and Bach, F. H. (1975a). Science 188, 1308. Sheehy, M. J., Sondel, P. M., Bach, F. H., Sapori, M. L., and Bach, M. L. (197513). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 569575. Munksgaard, Copenhagen. Shreffler, D. C. (1965). In “Isoantigens and Cell Interactions” (J. Palm, ed.), pp. 11-19. Wistar Inst. Press, Philadelphia, Pennsylvania. Shreffler, D. C., and David, C. S. (1975). Adu. Immunol. 20, 125. Silvers, W. K., Wilson, D. B., and Palm, J. (1967). Transplantation 5, 1053. Singal, D. P., Mickey, M. R., and Terasaki, P. I. (1969). Transplantation 7, 246. Singal, D. P., Villanueva, R., and Naipaul, N. (1975). Tissue Antigens 6, 1. Snell, G.D. ( 1968). Folia Bid. (Prague) 14, 335. Solheim, B. G., Bratlie, A,, Sandberg, L., Nielsen, L. S., and Thorsby, E. (1973). Tissue Antigens 3, 439. Solheim, B. G., Bratlie, A., Winther, N., and Thorsby, E. (1975). I n “Histocom-

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

199

patibility Testing 1975” (F. Kissmeyer-Nielsen, ed. ), pp. 713-718. Munksgaard, Copenhagen. Solheim, B. G., Ek, J., Thune, P. O., Baklien, K., Bratlie, A,, Rankin, B., Thoresen, A. B., and Thorsby, E. ( 1976). Tissue Antigens 7, 57. Solliday, S., and Bach, F. H. (1970). Science 170, 1406. Sondel, P. M., Chess, L., MacDermott, R. P., and Schlossman, S. F. (1975a). J. Immunol. 114, 982. Sondel, P. M., Chess, L., and Schlossman, S. F. (1975b). Cell. Immunol. 18, 351. Sgirensen, S. F. (1970). Acta Pathol. Microbiol. Scand., Sect. B 78, 711. Serensen, S. F. ( 1973). Transplant. Proc. 5, 1657. Sgirensen, S. F., and Hawkes, S. P. (1973). Transplant. Proc. 5, 1361. Sgirensen, S. F., and Kissnieyer-Nielsen, F. (1969). Acta Pathol. Microbiol. Scand. 77, 117. Sgirensen, S. F., and Nielsen, L. S. (1970). Acta Pathol. Microbiol. Scand., Sect. B 78, 719. Serensen, S. F., Andersen, V., and Giese, J. (1969). Acta Pathol. Microbiol. Scand. 76, 259. Speck, B., Zwaan, F. E., van Rood, J. J., and Eernisse, J. G. (1973). Transplantation 16, 24. Stastny, P. (1974). Tissue Antigens 4, 571. Stem, K., Brown, K. S., and Davidsohn, I. (1956). Genetics 41, 517. Stimpfling, J. H. (1971). Annu. Rev. Genet. 5, 121. Stokes, P. L., Asquith, P., Holmes, G. K. T., Mackintosh, P., and Cooke, W. T. ( 1972). Lancet 2, 162. Storb, R., Rudolph, R. H., Graham, T. C., and Thomas, E. D. (1971). J . Immunol. 107, 409. Storb, R., Thomas, E. D., Buckner, C. D., Clift, R. A,, Johnson, F. L., Fefer, A., Glucksberg, H., Giblett, E. R., Lerner, K. G . , and Neiman, P. (1974). Blood 43, 157. Suciu-Foca, N., and Dausset, J. ( 1975). Tissue Antigens 5, 137. Suciu-Foca, N., Hansen, J. A., Dupont, B., Rubinstein, P., Fu, S. M., Yunis, E. J., Mickelson, E., Thomas, D. L., Fotino, M., Whitsett, C., Jersild, C., Kunkel, H. G., Good, R. A., Dausset, J., and Reemtsma, K. (1975). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 937-940. Munksgaard, Copenhagen. Svejgaard, A. ( 1976). In “Manual of Immunological Methods” (in press). Svejgaard, A., Thorsby, E., Hauge, M., and Kissmeyer-Nielsen, F. (1970). Vox Sang. 18, 97. Svejgaard, A., Bratlie, A., Hedin, P. J., Hognian, C., Jersild, C., Kissmeyer-Nielsen, F., Lindblom, B., Lindholm, A,, Low, B., Messeter, L., Moller, E., Sandberg, L., Nielsen, L. S., and Thorsby, E. (1971). Tissue Antigens 1, 81. Svejgaard, A., Nielsen, L. S., Ryder, L. P., Kissmeyer-Nielsen, F., Sandberg, L., Lindholni, A., and Thorsby, E. ( 1973). In “Histoconipatibility Testing 1972” (J. Dausset and J. Colombani, eds.), pp. 465-473. Munksgaard, Copenhagen. Svejgaard, A., Platz, P., Ryder, L. P., Nielsen, L. S., and Thomsen, M. (1975). Transplant. Rev. 22, 3. Teisberg, P., Olaisen, B., Gedde-Dahl, T., Jr., and Thorsby, E. (1975). Tissue Antigens 5, 257. Terasaki, P. I., ed. ( 1970). “Histocompatibility Testing 1970.” Munksgaard, Copenhagen.

200

DUPONT, HANSEN, AND W N I S

Terasaki, P. I., and McClelland, J. D. (1964). Nature (London) 204, 998. Terasaki, P. I., and Singal, D. P. (1969). Annu. Reu. Med. 20, 175. Terasaki, P. I., Mickey, M. R., and Kreisler, M. (1971). Postgrad. Med. J.-47, 89. Terasaki, P. I., Opelz, G., Park, M., Billing, R., Ting, A., and Mickey, M. R. (1975). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 657-664. Munksgaard, Copenhagen. Thomas, E. D., Buckner, C. D., Clift, R. A., Fass, L., Fefer, A., Lerner, K. G., Neiman, P., Rowley, N., and Storb, R. (1973). Transplant. Proc. 5, 917. Thomas, E. D., Storb, R., Clift, R. A,, Fefer, A,, Johnson, F. L., Neiman, P. E., Lerner, K. G., Glucksberg, H., and Buckner, C. D. (1975). N . Engl. J . Med. 292, Part 1, 832; Part 2, 895. Thompson, J. S., Parmely, M. J,, Flink, R. J., Canady, M. S., and Severson, C. D. ( 1972). J . Exp. Med. 135,596. Thomsen, M., Hansen, G. S., Svejgaard, A., Jersild, C., Hansen, J. A., Good, R. A., and Dupont, B. ( 1974). Tissue Antigens 4, 495. Thomsen, M., Platz, P., Ortved Andersen, O., Christy, M., LyngsZe, J., Nerup, J., Rasmussen, K., Ryder, L. P., Nielsen, L. S., and Svejgaard, A. (1975a). Transplant. Rev. 22, 125. Thomsen, M., Jacobsen, B., Platz, P., Ryder, L. P., Nielsen, L. S., and Svejgaard, A. ( 1975b). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 509-518. Munksgaard, Copenhagen. Thomsen, M., Platz, P., Marks, J., Ryder, L. P., Shuster, S., Svejgaard, A., and . Antigens 7,60. Young, S. H. ( 1 9 7 5 ~ )Tissue Thorsby, E. (1974). Transplant. Reu. 18, 51. Thorsby, E., and Piazza, A. ( 1975). In “Histocompatibility Testing 1975” (F. Kissmeyer-Nielsen, ed. ), pp. 414-458. Munksgaard, Copenhagen. Thorsby, E., Lindholm, A., Sandberg, L., and Nielsen, L. S. (1971). VOXSang. 21, 69. Thorsby, E., Bondevik, H., and Solheim, B. G. (1973a). Transplant. Proc. 5, 343. Thorsby, E., Hirschberg, H., and Helgesen, A. (1973b). Transplant. Proc. 5, 1523. Thorsby, E., Dupont, B., Eijsvoogel, V. P., and J@rgensen, F. (1974a). Tissue Antigens 4,455. Thorsby, E., duBois, R., Bondevik, H., Dupont, B., Eijsvoogel, V. P., Hansen, J. A., Jersild, C., J@rgensen, F., Kissmeyer-Nielsen, F., Lamm, L. U., Schellekens, P. T. A., Svejgaard, A., and Thomsen, M. (1974b). Tissue Antigens 4, 507. Thorsby, E., Bondevik, H., Helgesen, A., and Hirschberg, H. (1975a). Transplant. Proc. 7, Suppl. 1, 87. Thorsby, E., Helgesen, A,, Rankin, B., Moller, E., and Kaakinen, A. (1975b). Tissue Antigens 6 , 147. Thorsby, E., Bratlie, A., Helgesen, A., Rankin, B., Solheim, B. G., Kaakinen, A., and . “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, Moller, E. ( 1 9 7 5 ~ )In ed. ), pp. 502-508. Munksgaard, Copenhagen. Tiilikainen, A. ( 1974). Quoted in Svejgaard et al. ( 1975). Trinchieri, G., Bernoco, D., Curtoni, S. E., Miggiano, V. C., and Ceppellini, R. ( 1973). In “Histocompatibility Testing 1972” (J. Dausset and J. Colombani, eds. ), pp. 509-519. Munksgaard, Copenhagen. Truog, P., Steiger, U., Contu, L., GalfrB, G., Trucco, M., Bemoco, D., Bemoco, M., Birgen, I., Dolivo, P., and Ceppellini, R. (1975). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen. ed. ), pp. 788-796. Munksgaard, Copenhagen. Utermohlen, V., and Zabriskie, J. B. (1973). Lancet 2, 1147. van Bekkum, D. W. ( 1972). Transplant. Reu. 9, 9. van den Tweel, J. G., BlussB van Oud Alblas, A,, Keuning, J. J., Goulmy, E.,

HUMAN MIXED-LYMPHOCYTE CULTURE REACTION

201

Temiijtelen, A., Bach, M. L., and van Rood, J. J. (1973). Transplant. Proc. 5, 1535. van Hooff, J. P., Schippers, H. hl. A., van der Steen, G. J., and van Rood, J. J. (1972). Lancet 2, 1385. van Leeuwen, A., Schuit, H. R. E., and van Rood, J. J. (1973). Transplant. Proc. 5, 1539. van Rood, J. J. (1962). Thesis, University of Leiden, Leiden. van Rood, J. J. (1973). I n “Histocompatibility Testing 1972” ( J . Dausset and J. Colombani, eds. ), p. 645. Munksgaard, Copenhagen. van Rood, J. J., and Eijsvoogel, V. P. (1970). Lancet 1, 698. van Rood, J. J., and van Leeuwen, A. (1963). J. Clin. Inuest. 42, 1382. van Rood, J. J., Koch, C. T., van Hooff, J. P., van Leeuwen, A., van den Tweel, J. G., Frederiks, E., Schippers, H. h4. A., Hendriks, G., and van der Steen, G . J. (1973). Transplant. Proc. 5, 409. van Rood, J. J., van Leeuwen, A., Keuning, J. J., and BloussC van Oud Alblas, A. ( 1975a). Tissue Antigens 5, 73. van Rood, J. J., van Hooff, J. P., and Keuning, J. J. (1975b). Transplant. Rev. 22, 75. van Rood, J. J., BloussC van Oud Alblas, A., Keuning, J. J., Frederiks, E., Termijtelen, . A., van Hooff, J. P., Peria, A. S., and van Leeuwen, A. ( 1 9 7 5 ~ ) Transplant. Proc. 7, Suppl. 1, 25. van Rood, J. J., van Leeuwen, A., Parlevliet, J., Termijtelen, A., and Keuning, J. J. ( 1975d). In “Histocompatibility Testing 1975” ( F. Kissmeyer-Nielsen, ed. ), pp. 629-636. Munksgaard, Copenhagen. van Sorneren, H., Westerveld, A,, Hagemeijer, A., Mies, J. R., Khan, P. M., and Zaalberg, 0. B. (1974). Proc. Natl. Acad. Sci. U.S.A. 71, 962. Vladutiu, A. O., and Rose, N. R. (1974). Immunogenetics 1, 305. von Boehmer, H. (1974). 1. Immunol. 112, 70. Wagner, H. (1972). 1. Immunol. 109, 630. Wagner, H., Harris, A. W., and Feldman, M. (1972). Cell. Immunol. 4, 39. Walford, R. L., Gallagher, R., and Troup, G. M. (1965). Transplantation 3, 387. Walford, R. L., Finkelstein, S., Hanna, C., and Collins, Z. ( 1969). Nature ( L o n d o n ) 224, 74. Walford, R. L., Waters, H., and Smith, G. S. (1970). Fed. Proc., Fed. Am. SOC. Exp. Biol. 29, 2011. Walford, R. L., Zeller, E., Combs, L., and Konrad, P. (1971). Transplant. Proc. 3, 1297. Walford, R. L., Gossett, T., Smith, G. S., Zeller, E., and Wilkinson, J. (1975). Tissue Antigens 5, 196. Ward, F. E., and Seigler, H. F. (1973). Transplant. Proc. 5, 359. Ward, F. E., Southworth, J. G., and Amos, D. B. (1969). Transplant. Proc. 1, 352. Wernet, P., and Kunkel, H. G. (1973). Transplant. Proc. 5, 1875. Wernet, P., Winchester, R. J., Kunkel, H. G., Wernet, D., Giphart, M., van Leeuwen, A., and van Rood, J. J. (1975). Transplant. Proc. 7, Suppl. 1, 193. White, S. H., Newcomer, V. D., Mickey, M. R., and Terasaki, P. I. (1972). N . Engl. 1. Med. 287, 740. Widmer, M. B., Alter, B. J., Bach, F. H., and Bach, M. L. (1973a). Nature (Lond o n ) , New Biol. 242, 239. Widmer, M. B., Peck, A. B., and Bach, F. H. (197313). Transplant. Proc. 5, 1501. Williams, R. M., and Moore, M. J. ( 1973). J. Exp. Med. 138, 775. Wilson, D. B., Silvers, W. K., and Nowell, P. C. (1967). 1. Exp. Med. 126, 655. Wilson, D. B., Blyth, J. L., and Nowell, P. G. (1968). J. Exp. Med. 128, 1157.

202

DUPONT, HANSEN, AND Y U N I

Winchester, R. J., Fu, S. M., Wernet, P., Kunkel, H. G., Dupont, B., and Jersild, C. (1975a). J . Exp. Med. 141, 924. Winchester, R. J., Dupont, B., Wernet, P., Fu, S. M., Hansen, J. A., Laursen, N., and Kunkel, H. C. (197513). In “Histocompatibility Testing 1975” (F. KissmeyerNielsen, ed. ), pp. 651-656. Munksgaard, Copenhagen. Winchester, R. J., Dupont, B., Wernet, P., Hansen, J. A., Fu, S. M., Ojea, F., Laursen, N., and Kunkel, H. G. (1976). Proc. Leukocyte Cult. Conf. 10, (in press ) . Wolski, K. P., Schmid, F. R., and Mittal, K. K. (1975). Science 188, 1020. World-Health-Organization-International Union of Immunological Societies Terminology Committee. ( 1975). In “Histocompatibility Testing 1975” ( F. KissmeyerNielsen, ed. ), pp. 5-10. Munksgaard, Copenhagen. Yunis, E. J., and Amos, D. B. (1971). Proc. N d l . Acad. Sci. U.S.A. 68, 3031. Yunis, E. J., Ward, F. E., and Amos, D. B. (1970). In “Histocompatibility Testing 1970” (P. I. Terasaki, ed.), pp. 351-356. Munksgaard, Copenhagen. Yunis, E. J., Plate, J. M., Ward, F. E., Seigler, H. F., and Amos, D. B. (1971). Transplant. Proc. 3, 118 Yunis, E. J., Seigler, H. F., Simmons, R. L., and Amos, D. B. (1973). Transplantation 15, 435. Yunis, E. J., Krivit, W., Reinsmoen, N., and Amos, D. B. (1974). Nature (London) 248,517. Yunis, E. J., Amos, D. B., and Blurnenthal, M. N. (1975). Transplant. PTOC. 7, Suppl. 1, 49. Zoschke, D. C., and Bach, F. H. (1971). Science 172, 1350.

lmmunochemical Properties of Glycolipids and Phospholipids DONALD M. MARCUS1 AND GERALD A. SCHWARTING’ Departments o f Medicine, Microbiology and Immunology, Albert Einstein College o f Medicine, Bronx, N e w York

I. Introduction

.

.

11. Glycolipids . . A. Glycosphingolipids

B. Glycosyl Glycerides C. Lipoteichoic Acids D. Other Glycolipids 111. Phospholipids . . A. Cardiolipin . . B. Phosphatidyl Inositol C. Sphingomyelin . IV. Concluding Remarks . References . . .

.

. . .

. . . . . . . .

.

. . .

. . . . . . . .

.

. . .

. . . . . . . .

.

. . .

. . . . . . . .

.

. . .

. . . . . . . .

.

. . .

. . . . . . . .

.

. . .

. . . . . . . .

.

. . .

. . . . . . . .

.

. . .

. . . . . . . .

.

. . .

. . . . . . . .

.

. . .

. . . . . . . .

203 204 204 221 226 228 229 229 232 232 233 233

I . Introduction

Although several important observations on lipid antigens were made early in the century, this class of antigens received little attention from immunologists until quite recently. The lipoidal nature of many bacterial and animal cell antigens was suggested by their solubility in alcohol (reviewed by Landsteiner, 1962), and it was noted that lipid “haptens” were immunogenic only when mixed with an antigenic carrier protein or serum (Landsteiner and Simms, 1923). The first lipid antigen to be isolated was cardiolipin (Pangborn, 1942), which is the active principle of the alcoholic extract of beef heart used in serological tests for syphilis. The foundations of modern glycolipid immunology were laid by Yamakawa’s ( 1962) studies of erythrocyte glycosphingolipids (GSL), by Rapport’s ( 1961) observations on tumor GSL and methodology, and by Hakomori. Much of the current interest in GSL stems from Hakomori’s purification of complex GSL with blood group activity, and his demonstration of the alterations in GSL content and molecular arrangement in tumor cell membranes ( Hakomori, 1973,1975). Supported by Grant AI-05336 from the National Institutes of Allergy and Infectious Diseases. Recipient of a Research Fellowship, 1-F32 HL05162-01, from the Heart and Lung Institute. 203

204

DONALD M. MARCUS AND GERALD A. SCHWARTING

Immunologists have been deterred from working with lipid antigens by methodological problems associated with their isolation and with the preparation and characterization of antisera. Recent advances have resolved many of these difficulties, but lipid antigens do present certain unique problems, and they will be considered in detail in this review. We will also summarize recent data on the immunological properties of complex lipid antigens and indicate their growing importance in mammalian and microbial immunology. II. Glycolipids

A. GLYCOSPHINCOLIPIDS 1. Structure and Distribution Glycosphingolipids are the principal glycolipids of mammalian tissues. They are composed of a hydrophobic ceramide ( N-acylsphingosine ) moiety and one or more sugars attached to the terminal hydroxyl group of sphingosine (Fig. 1 ) . The nomenclature of GSL is confusing because many compounds were assigned trivial names when they were discovered, and new GSL described recently are not readily accommodated in previous schemes. We have divided GSL into three groups (Tables 1-111) based on their sugar composition and sequences. This is a selection of GSL based on their immunological interest and abundance and not a complete catalog. Table I includes galactosyl ceramide ( cerebroside ) and a family of compounds with related structures, the globoside series. The latter compounds are the most abundant GSL of extraneural tissues (Martensson, 1969; Ledeen and Yu, 1973)-The group of GSL containing N-acetylglucosamine (Table 11) are the most recently described and least abundant of the three groups, but it includes the ABH, Lewis, and P, blood group antigens. Gangliosides (Table 111), GSL containing sialic acid, are the principal glycolipids of the central nervous system. Gangliosides are also found in extraneural tissues but in much lower sphingosine

H0-CH2-CH-CH-CH=CH-(CH2),,-CH3

HA

ceramide - -- -

dH

____ _ _ _ ____ - -__ _ _ c=o I R

fatty acid

FIG.1. Structure of ceramide (N-acylsphingosine ).

205

GLYCOLIPIDS AND PHOSPHOLIPIDS

TABLE I NEUTRAL GLYCOSPHINCOLIPIDS~ Substance

Structure

Galactosyl ceramide Gal-Cer Glucosyl ceramide Glc-Cer Lactosyl ceramide Gal(@,144)Glc-Cer Trihexosyl ceramide Gal(a,l-+4)Gal(P,1+4)Glc-Cer (blood group Pk antigen) Globoside (blood GalNAc (p,1--1 3)Gal(a, 144)Gal (p,144)Glc-Cer group P antigen) Forssman GalNAc (a,I-. 3)GalNAc@, 1 4 3)Gal(a,1+4)Gal(P, 1 4 4 )Glc-Cer Abbreviations: Glc = o-glucose; galactosamine; Cer = ceramide.

Gal

=

D-galactose;

GalNAc

=

N-acetyi-o-

concentration than the globoside series of compounds. The same GSL are found in many mammalian tissues but each tissue and cell type has a distinctive glycolipid composition (reviewed by Martensson, 1969; Ledeen and Yu, 1973). For example, globoside is the major GSL of human erythrocytes and kidney (Makita et al., 1964; Rapport et al., 19644, but lactosyl ceramide predominates in spleen and liver ( Svennerholm and Svennerholm, 1963). Data on the subcellular distribution of GSL is limited, but most GSL are found in plasma membranes (Dod and Gray, 1968; Weinstein et al., 1970; Klenk and Choppin, 1970; Keenan et al., 1972; Critchley and Macpherson, 1973). The fatty acid and sphingosine components of GSL are quite heterogeneous. They vary in chain length, degree of unsaturation, and substitution by hydroxyl groups ( Martensson, 1969; Wiegandt, 1971; Ledeen and Yu, 1973). The oligosaccharide moiety of a GSL has immunological properties that differ from that of the same oligosaccharide free in solution (discussed in Sections II,A,2 and 5 ) , but there is no evidence that the microheterogeneity of lipid components of naturally occurring GSL is important in this respect. 2. Preparation of Antisera

Most pure GSL are poor immunogens, and it is necessary to mix them with immunogenic carrier substances in order to obtain good antisera. A list of the carrier substances employed includes heterologous whole serum (reviewed by Rapport and Graf, 1969), bovine serum albumin (Niedieck and Palacios, 1965; Tal et al., 1967; Koscielak et al., 1968), methylated bovine serum albumin (Inoue and Nojima, 1967), the major

TABLE I1 GLYCOSPHINGOLIPIDS CONTAINING N-ACETYLGLUCOSAMINE" Substance Paragloboside (lacto-N-neotetraosyl ceramide) Blood GOUP PI Ganglioside Blood group H Blood group A1

Structure Gal(p,l-+4)GlcNAc (p,143)Ga1((3,l-+g)Glc-Cer

Gal(a,1-14)Ga1(~,1~4)GlcNAc(~,1-+3)Ga1(~,1~4)Glc-Cer NeuNAc(a,2+3) Gal ( 0 , 1 4 4 )GlcNAc(p, 1-+3)Gal(p,1+4)Glc-Cer Fuc(a,l+2)Gal(~,1+4)GlcNAc(p, 1-+3)Gal(p,144)Glc-Cer GalNAc(~,l~3)Gal(~,l+4)Glc~Ac(~,l-+3)Gal(~,l+4)Glc-Cer r?

2 Ta

1 v

Blood group B (pancreas)

0

m

E

-

Fuc Gal(a,l~3)Ga1(p,1~3)GlcNAc(~,1-+3)Gal(~,1-,4)Glc-Cer 2

1 W

Blood group B (pancreas)

-

Fuc Gal(a,1-,3)Gal(~,1~4)GlcNAc(~,1-13)Ga.1(~,1+4)Glc-Cer

i

v

Fuc

-

Gal(&1+ 3)GlcNAc ( & + 3)Gal(&1+4)Glc-Cer

Blood group Lea (adenocarcinoma)

4

-

-

Gal(~,1+3)GlcNAc(~,1+3)Gal(~,1+4)Glc-Cer

Blood group Leb (adenocarcinoma)

2

4

Ta

1 v

Fuc Blood groups H and I

Fuc

Fuc (a,1+2)Gal(p, 1+4)GlcNAc (a, 1 4 3))Gal(j3,1-+4)GlcNAc(~,1+3)Gnl(p,l+4)(:lc-Cer Fuc(a, 142)Ga1(&1+4)GlcNAc(a, 1+6)

Abbreviations: Gal = D-galactose; GlcNAc = N-acetyl-o-glucosamine; Glc neuraminic acid: Fuc = L-fucose; GalNAc = A'-acetyl-D-galactosarnine. a

Ta

1 v

=

o-glucose; Cer

=

ceramide; NeuNAc

=

,v-acetyl-

E

208

DONALD M. MARCUS AND GERALD A. SCHWARTING

Substance GM3

Structure Gal(@,lb4)Glc-Cer

(hematoside)

n

3

Tff 2

v

-

NeuNAc GalNAc(P, 144)Gal(@,l+4)Glc-Cer

GM: (Tay-Sachs)

n

J

Tff 2

W

Asialo G M I GMI

NeuNAc Gal (@,1+3) GalNAc (@,1+4)Gal(@,1h4)Glc-Cer Gal(@,1+3) GalNAc (@,1+4)Gal@, lh4)Glc-Cer

3

Ta

2 v

NeuNAc

-

Gal(@,1+3)GalNAc(P,1+4)Gal(fl,1-+4)Glc-Cer 3

Tff 2

v

3

tff 2

v

-

NeuNAc NeuNAc Gal(@,1+3)GalNAc(@,1+4)Gal(@,1-,4)Gl~-Cer 3

Tff

2

W

NeuNAc (0,8+-2) NeuNAc Gal(& 1+3)GalNAc(P, 1+4)Gal(P, 1+4)Glc-Cer n

n

3

3

Tff

Tff

2

v

NeuNAc

2

v

NeuNAc(a,8+2)NeuNAc

Abbreviations: Gal = D-galactose; Glc = D-glucose; Cer = ceramide; GalNAc N-acetyl-u-galactosamine; NeuNAc = N-acetylneuraminic acid.

=

glycoprotein of human erythrocytes (glycophorin) (Marcus and Janis, 1970), and a protein extracted from mycoplasma membranes ( Razin et al., 1971a). Antisera have also been obtained to glycolipids incorporated into liposomes (Nagai and Ohsawa, 1974) or adsorbed to erythrocytes ( Yokoyama et al., 1963), but the antisera obtained by the latter

GLYCOLIPIDS AND PHOSPHOLIPIDS

209

method were weak. Alternatives to immunization with intact GSL are the use of deacylated glycolipids (Taketomi and Yamakawa, 1963) or synthetic glycosylsphingosine compounds coupled to carrier compounds ( Arnon et al., 1967) or immunization with lactosyl protein conjugates (Graf et al., 1965). In the latter study, only a fraction of the sera containing antilactoside antibodies reacted effectively with lactosyl ceramide. This procedure might be more successful with immunogens containing longer carbohydrate chains, where the “carrier specificity” (i.e., recognition of regions of the carrier adjacent to the carbohydrate chain) of the antisera would be less pronounced. Antisera to GSL may also be obtained by immunization with tissue homogenates (Rapport et al., 1959, 1964a,b) or cells, e.g., the traditional procedure for obtaining antibody to Forssman glycolipids is immunization with boiled sheep erythrocyte stroma (Kabat and Mayer, 1961). The other antibodies present in thcse sera present no problem in studying the reactions of the pure glycolipid with antibodies in vitro, but these sera could not be utilized for studies of intact cells or crude tissue extracts. Antiglycolipid antibodies can be purified by utilizing antigen affinity columns, as discussed in the following. 3. Detection and Measurement of Antibodies

Antibodies to GSL are directed primarily agninst the oligosaccharide chain ( Rapport and Graf, 1969). Individual glycolipid molecules are univalent, but GSL dispersed in aqueous solutions form large aggregates that behave like multivalent antigens in forming immune precipitates, fixing complement, etc. Dispersions of pure GSL react very poorly with antibodies. As found previously with cardiolipin ( Maltaner and Maltaner, 1945), auxiliary lipids are required for optimal reactivity of GSL with antibodies (Rapport et al., 1959). The auxiliary lipids, usually lecithin and cholesterol, must be mixed with the glycolipids before evaporation of the organic solvents and dispersion in buffcred saline. Sonication for several minutes is useful in obtaining uniform dispersions of glycolipidlecithin-cholesterol micelles ( Naiki et al., 1974). These preparations may be stored at 4” for several weeks but they should be sonicated immediately before each experiment. An example of the requirement for auxiliary lipids is provided by the data in Fig. 2, taken from a recent study (Naiki et al., 1974) of antibodies to G,, ganglioside and asialo GM]. No more than 10% complement fixation could be obtained with 200 ng of pure asialo GM,,whereas 50% fixation could he obtained with 2 116 of asialo GM, mixed with 3.5 parts each ( w / w ) of lecithin and cholesterol. The simplest and least sensitive technique for detection of antiglycolipid antibodies in double diffusion in agar gel. The antigen is used as an

210

DONALD M. MARCUS AND GERALD A. SCHWARTING

*b 2

02

40

80

120

160 200

2

4

6

8

1

0

ASIALO G M ~( n g )

FIG. 2. Effect of lecithin and cholesterol on complement fixation by asialo G%I1 and the IgC fraction of rabbit 519 antiasialo Garl (11 weeks); 5 x lo-‘ ml of antibody was employed. ( a l ) Asialo Ghrl alone; ( b l ) asialo G 4 e c i t h i n (1:1, w/w); ( c l ) asialo G d e c i t h i n (1:2, w/ w) ; ( d l ) asialo Ghfl-lecithin (1:3.5, w/w); ( e l ) asialo Gxl-lecithin ( 1:5, w/w ) ; ( f 1) asialo GM,-lecithin ( 1: 10, w/w ) ; ( a2 ) asialo G d e c i t h i n ( 1:3.5, w/w); ( b 2 ) asialo G~rl-lecithin-cholesterol ( 1:3.5:3.5, w/w); ( c2) asialo Gdecithin-cholesterol 1 :3.5:7.0, w/w); ( d 2 ) asialo GAecithincholesterol ( 1:3.5:17.5, w/w); and ( e 2 ) asialo Gh[l-lecithin-cholesterol ( 1:3.5:35.0, w/w). (Reproduced from Naiki et al., 1974, by permission of the Williams and Wilkins Company. )

aqueous dispersion and in general the glycolipid is mixed with twice its weight of lecithin or sodium taurocholate. The precipitin bands formed between antibodies and dispersions of pure glycolipids are initially sharp but they become diffuse and split in the absence of an auxiliary lipid (Graf and Rapport, 1965). The gel diffusion technique is useful as a screening procedure for detection of antibodies but it probably should not be used for analysis of cross-reactions between glycolipid antigens. We have noted spurious “reactions of identity,” i.e., extensive or complete fusion of precipitin bands between glycolipid antigens that cross-react weakly if at all (Naiki et al., 1974). We believe that the lipid particles containing the different antigens coalesce to form a lipid phase in the agar gel, which prevents the antigens from diffusing independently. Crossreactions can be evaluated by allowing antibody to react with a heterologous glycolipid in a gel diffusion plate that does not contain the homologous antigen. Two other simple, semiquantitative techniques that are useful are passive hemagglutination and a microtiter complement fixation technique. When erythrocytes are suspended in an aqueous dispersion of a glycolipid, the glycolipid is firmly adsorbed to the erythrocyte and is not re-

GLYCOLIPIDS AND PHOSPHOLIPIDS

211

moved by washing with aqueous buffers (Yokoyamo et al., 1963; Marcus and Cass, 1969). Antiglycolipid antibodies can then be detected by conventional hemagglutination techniques. Most erythrocytes are agglutinated by naturally occurring antibodies in heterologous antisera, and if the erythrocytes and antisera are from different species, the antisera should be absorbed with uncoated red cells prior to this assay. The hemagglutination reaction should be carried out a t 4"C, and treatment of the erythrocytes with proteolytic enzymes such as ficin or papain markedly increases the sensitivity of this technique ( Mollison, 1972, p. 401 ). Immunoglobulin M antibodies are much more efficient agglutinins than IgG and the latter class of antibodies may go almost undetected if other methods of assay are not employed (Mollison, 1972, p. 200). A simple, complement fixation assay can be carried out with a microtiter system in a total volume of 0.1 ml (Naiki and Taketomi, 1969). This test is more sensitive than gel diffusion and provides a semiquantitative estimate of the antibody titer. The microcomplement fixation technique of Wasserman and Levine (1961) is several orders of magnitude more sensitive than the microtiter assay and provides a means of quantitating cross-reactions and studying hapten inhibition. The optimal proportions of auxiliary lipids must be determined for each antigen-antibody system by the type of titration illustrated in Fig. 2. This assay is very sensitive because it employs a limited quantity of complement, but it is also very sensitive to nonspecific anticomplementary effects, and some experience is required to obtain reproducible results. Many antiglycolipid sera (Koscielak et al., 1968; Marcus and Janis, 1970; Naiki et al., 1974) have appreciable quantities of IgG and IgM antibodies. It should be noted that the optimal condition for complement fixation with IgG antibodies is incubation at 4°C overnight, whereas a l-hour incubation at 37°C is optimal for IgM (Stollar and Sandberg, 1966; Sandberg and Stollar, 1966). Rapport uses a different quantitative complement fixation technique (Rapport and Graf, 1957) that yields isofixation curves, i t . , plots of the optimal antigen-antibody ratio for fixation of complement at varying dilutions of both antigen and antibody. This technique provides more information than conventional complement fixation curves, but it requires a considerable amount of additional effort. Another useful technique is the liposomal model membrane system developed by Kinsky and collaborators. In this procedure, glycolipid antigens are incorporated into a model membrane containing lecithin, cholesterol, and a detergent. The membrane surrounds an aqueous compartment containing a marker substance such as a radioactive compound or an enzyme substrate (Kinsky et al., 1969). In the presence of specific antibody and complement, the permeability of the liposomal membrane

212

DONALD M. MARCUS AND GERALD A. SCHWARTING

is altered and the marker substance that is released is measured by sensitive enzymatic or isotopic techniques (Kinsky, 1972). Many of these studies were reviewed by Kinsky (1972). The technique was improved by development of a procedure for making single-compartment liposomes that can release all of their marker substances (Six et al., 1974). Some limitations of this technique have been noted recently. A human monoclonal IgM that lysed sheep erythrocytes in the presence of complement and released marker substances from liposomes containing Forssman glycolipid was found (Joseph et al., 1974; Alving et al., 1974). The macroglobulin did not react with globoside-containing liposomes, but reactivity with globoside was demonstrated by precipitation in agar gel, by agglutination of enzyme-treated human erythrocytes, and by hemagglutination inhibition (M. Naiki and D. M. Marcus, unpublished observations, 1974). The liposomal technique is very sensitive and flexible, but in view of the heterogeneous serological properties of immunoglobulins it is advisable not to depend solely on a single immunological technique in evaluating a new immune system. 4. Purification of Antibodies

Most antiglycolipid sera contain relatively small quantities of antibodies, and it is desirable to concentrate and purify the antibodies for certain applications, such as analysis of glycolipid distribution in tissues. Two types of glycolipid antigen immunoadsorbents have been developed recently: One involves the covalent linkage of a glycolipid fragment to an insoluble support, and the other technique consists of incorporation of glycolipids into a polyacrylamide gel. In the first procedure (Laine et al., 1974), the sphingosine double bond is subjected to ozonolysis and the fragment containing the carbohydrate moiety is coupled to aminoethyI agarose or derivatized glass beads containing an amino group. These adsorbents were used for purification of rabbit antibodies to globoside and hematoside ( G M J ) We . have adapted a method for incorporation of proteins into polyacrylamide gels (Carrel and Barandun, 1971) for use with glycolipids. The details of this procedure are described elsewhere (Marcus, 1976). The principle of the procedure is that an aqueous dispersion of 3H-labeled glycolipid, lecithin, and cholesterol is polymerized into a 5%acrylamide gel, and the gel is then ground and washed extensively to remove trapped glycolipids. The glycolipids are labeled by oxidation of the terminal galactose or N-acetylgaIactosamine residue with ( Suzuki galactose oxidase and reduction with sodium b~rohydride-~H and Suzuki, 1972), in order to estimate how much of the glycolipid is incorporated into the gel. We have used these immunoadsorbents to purify antibodies to asialo G,, , GMl,paragloboside, globoside, and trihexosyl ceramide.

I

1

1

I

GLYCOLIPIDS AND PHOSPHOLIPIDS

5. Specificity

of

213

Antiglycolipid Antibodies

The specificity of antisera to GSL has been determined by studying their cross-reactions with other glycolipids, and by inhibition of complement fixation by oligosaccharide haptens. The carbohydrate portion of GSL can be obtained for immunological studies by cleaving the glycolipid by ozonolysis (Wiegandt and Baschang, 1965) or periodate oxidation and alkaline hydrolysis ( Hakomori, 1966). Rapport first demonstrated that antibodies to GSL are directed against the carbohydrate portion of the molecule by inhibiting antibodies to lactosyl ceramide with lactose (Rapport et al., 1961) and numerous studies with antisera to other glycolipids have confirmed this principle ( Somers et al., 1964; Koscielak et al., 1968; Rapport and Graf, 1969; Marcus and Janis, 1970; Naiki et al., 1974). In general, antibodies raised against short oligosaccharide chains that do not contain repeating sequences react most strongly with the terminal nonreducing sugar residue, which has been termed the immunodominant residue ( Kabat, 1968). Alterations in the configuration, sequence, or sites of substitution of internal residues have much less effect on the interaction of antibodies with carbohydrate determinants. There is some evidence that the ceramide portion of GSL affects the immunological reactivity of these compounds. Rapport and Graf ( 1969) noted that synthetic lactosyl dihydroceramides containing short-chain fatty acids, 8 carbon atoms or less, were much less effective in complement fixation than compounds containing fatty acids of 10 carbon atoms or more. Hydrogenation of the sphingosine double bond or lactosyl ceramide did not affect its immunological activity, nor did configuration about the carbon-2 residue of sphingosine (Rapport and Graf, 1969). As mentioned previously, there is an incomplete cross-reaction between antibodies to lactosyl azophenyl haptens and lactosyl ceramide (Graf et al., 1965). We have also noticed discrepancies between the cross-reactions of antibodies with intact glycolipids and their isolated oligosaccharide moieties (Naiki et al., 1974). The ceramide moiety could alter the preferred conformation of the oligosaccharides or it could form part of the antigenic determinant of glycolipids with short oligosaccharide chains. It may not be possible to define the role of ceramide in the interaction of GSL with antibody solely on the basis of hapten inhibition studies. Arnon et al. (1967) found that antibodies to a lactosyl sphingosine conjugate were inhibited more effectively by lactosyl N-acetylsphingosine than by lactosyl sphingosine, and inferred that sphingosine was part of the antigenic determinant. An alternative explanation is that an unfavorable interaction between the positively charged amino group of lactosyl sphingosine and the antibody was responsible for the weaker binding of that hapten.

214

DONALD M. MARCUS AND GERALD A. SCHWARTING

In considering the specificity of antibodies to carbohydrate determinants, it should be noted that IgG and IgM antibodies in the same sera may differ significantly in specificity. For example, IgG antibodies to asialo GM, did not cross-react with GM,, but the IgM antibodies crossreacted extensively with GDlband G,, (Naiki et aZ., 1974). In view of the potential differences in specificity of IgM and IgG antiglycolipid antibodies and the known differences in their capacity to fix complement and agglutinate cells, it is advisable to separate the two classes of immunoglobulins before carrying out extensive studies of their specificity and reactions with cells. Antibodies directed against carbohydrate determinants may cross-react with compounds bearing the same nonreducing terminal sugar residues regardless of other differences in structure. For example, antibodies to Type XIV pneumococcal polysaccharide react with p-galactosyl residues on other polysaccharides, glycoproteins, or glycolipids (Heidelberger, 1956; Allen and Kabat, 1959; Siddiqui and Hakomori, 1973). Th'is crossreactivity limits the conclusions that one is entitled to draw from the cross-reaction of an anticarbohydrate antibody with a substance of unknown structure, a complex biological fluid, or a cell membrane. This concept is important because it has been assumed by some investigators (Graf and Rapport, 1970; Inoue et al., 1972) that if an antibody to a purified GSL reacts with a cell membrane, the specific GSL is present in the membrane. If the membrane is not known to contain that GSL, one may conclude only that the membrane contains one or more compounds with a terminal sugar sequence similar or identical to the immunizing antigen. Even if the GSL is known to be a component of the membrane, it may be only one of the receptors for that antibody. One might reasonably conclude that the GSL was the major receptor for the antibody if the cell membrane contained no cross-reacting glycolipids and if no immunological activity was found in an aqueous or detergent extract of the membrane. 6. Reactions of Antiglycolipid Sera with Intact Cells and Tissue Sections

A variety of techniques, including agglutination, cytotoxicity, and uptake of antibodies labeled with fluorochromes or enzymes are available for studying reactions of antiglycolipid sera with intact cells. It should be noted, however, that there is not a direct correlation between the quantity of glycolipid in a cell membrane and its reactivity with antibodies because some GSL in cell membranes are not readily accessible to antibodies (Hakomori, 1973). Globoside is the major glycolipid of human erythrocytes, but antigloboside sera react weakly with these cells.

CLYCOLIPIDS AND PHOSPHOLIPIDS

215

After treatment of the erythrocytes with trypsin, there is a marked increase in the agglutinability of the cells by antigloboside (Hakomori, 1969). The uptake of antigloboside was not studied in a quantitative manner, but the untreated erythrocytes were much less effective in absorbing antigloboside antibodies than the trypsinized cells, suggesting that most globoside molecules were not accessible to antibodies. By contrast, human fetal erythrocytes reacted well with antibodies to globoside, and trypsin treatment produced no enhancement of antibody uptake or agglutination. A substantial portion of the erythrocyte glycoprotein and about 50%of the sialic acid is cleaved from the red cell membrane by trypsin (Makela et al., 1960; Cook et al., 1960; Seaman and Heard, 1960; Eylar et al., 1962; Luner et al., 1975). Reduction in the net charge of the cell membrane renders the cells more agglutinable by antibodies, and removable of the glycoprotein chains apparently removes a steric impediment to the reaction of antibodies with the GSL (Hakomori, 1973). Similar observations were made on the cytotoxic action of antihematoside antibodies on mouse fibroblast 3T3 cells and baby hamster kidney cells. The cytotoxic activity of antihematoside sera on these cells was greatly increased by trypsinizing the cells or transforming them with oncogenic viruses, even though the transformed cells contained much less hematoside than the untransformed cells ( Hakomori et al., 1968, 1972). The reactivity of the transformed cells was not increased by trypsinization. A discrepancy between the immunological reactivity and abundance of a GSL in a hamster NIL cell line was found by Wolf and Robbins (1974). During 8 hours after mitosis, the quantity of Forssman GSL per cell increased two- to threefold, but the uptake of radiolabeled antiForssman antibodies by the cells increased only slightly. After trypsinization, however, the uptake of antibodies by the cells related directly to their content of Forssman GSL. These data demonstrate how antibodies to glycolipids may be used to reveal changes in the organization of cell membranes. Cell membrane glycolipids have also been demonstrated by immunofluorescence techniques. Purified antibodies to ganglioside G,, ( Naiki et al., 1974) react specifically with thymocytes and peripheral T cells of AKR/J, BALB/c, and A strains of mice (Stein-Douglas et al., 1976). Studies of the functions of GSL would be facilitated by data on their distribution in cells and organelles. Although there is considerable information on the glycolipid content of many mammalian organs, it would be more informative to determine the glycolipid composition of each type of cell and the distribution of glycolipids among cellular organelles. These studies are hindered by the technical problems involved in separa-

216

DONALD M . MARCUS AND GERALD A. SCHWARTINC

ting different types of cells and in obtaining pure membranes and organelles. Immunohistochemical techniques can complement biochemical analyses by localizing specific glycolipids in tissue sections. Globoside is the major neutral glycolipid of human kidney (Makita et al., 1964; Rapport et al., 1964a). Immunofluorescence studies of kidney with antigloboside sera revealed that this glycolipid is concentrated in the plasma membrane and cytoplasm of epithelial cells of the proximal convoluted tubule ( Marcus and Janis, 1970). Lactosyl ceramide, which is known to be present in the kidney (Makita, 1964), was not detectable by immunofluorescence. In the same study, globoside and lactosyl ceramide were detected only in the plasma membrane of reticuloendothelial cells of the red pulp of human spleen. These studies suggest that there is wide variation among cells in their content and expression of glycolipid, and that GSL may be essential for certain specialized functions. We have recently studied the localization of ganglioside in sections of rat brain by an immunohistochemical method that utilizes purified rabbit anti-GM, antibodies and peroxidase-conjugated sheep antirabbit IgG ( de Baecque et al., 1975). These antibodies react with ganglioside G,,, as well as GM, (Naiki et al., 1974). Specific staining was noted on white matter astrocytes, cerebellar granule cells, neurons in the deep cortical layer, and many areas of neuropil. The staining appeared to be localized on the plasma membrane. Purkinje cells and the cerebellar molecular layer were not stained. These studies were extended to the ultrastructural level by employing peroxidase-conjugated F ( ab’ ) fragments of sheep antirabbit IgG in order to facilitate penetration of the cells. With this technique, specific staining was noted on the plasmalemma and nuclear envelope of cerebellar granule cells.

7. Biological Aspects When Rapport and Graf surveyed the field in 1969, the number of well-characterized GSL with immunological properties was small enough to permit detailed analysis of each compound. Since that time the field has grown SO rapidly that this is no longer possible. Many individual GSL have been considered above, and it is evident that it should now be possible to elicit antibodies to any GSL. In this section we will review the role of GSL in some broad areas of biomedical interest. a. Cellular Antigens. Chemical analyses of the ABH and Lewis blood group antigens were performed initially on soluble glycoproteins, but a number of GSL have been isolated from erythrocytes recently (Ando and Yamakawa, 1973; Ando et al., 1973; reviewed by Hakomori and Kobata, 1974), and it appears that most of the erythrocyte ABH antigens

GLYCOLIPIDS AND PHOSPHOLIPIDS

217

are glycolipids and not glycoproteins. The major glycoprotein of human erythrocytes has a hydrophobic region (Marchesi et al., 1973; Javaid and Winzler, 1974a,b ) and binds lipids tenaciously. A single extraction with butanol or chloroform-methanol is not sufficient to extract all of the tightly bound glycolipid ( Zvilichovsky et al., 1971) . The sugar sequcnces of the GSL are identical to the glycoprotein determinants except that the former contain glucose, which is not a constituent of ABH glycoproteins. To date, ABH glycolipids isolated from erythrocytes contain only the Type I1 backbone structure found in glycoproteins, i.e., the subterminal disaccharide is Gal (p,1+4) GlcNAc (Table 11). Type I chains contain the two sugars linked (p,1+3). Glycolipids with Le" and Le" activities contain the Type I chain, as does a GSL with blood group B activity that was isolated rccently from the pancreas of a patient with Fabry's disease (Wherrett and Hakomori, 1973). Erythrocyte A and H GSL are remarkably polymorphic in size. Structures of GSL containing five to ten sugars have been determined (Hakomori and Kobata, 1974; Watanabe et nl., 1975), and larger compounds have been detected but not completely purified. A ceramide decasaccharide with a branched structure was found to possess blood group I and H activities ( Watanabe et al., 1975). Glycolipids with blood group activity have also been isolated from animal tissues, including hog gastric mucosa (B. L. Slomiany et al., 1973a; 1974, 1975; A Slomiany et al., 1974; Slomiany and Slomiany, 1975), dog intestine (Smith and McKibbin, 1972; Smith et al., 1973, 1975), and bovine serum ( Slomiany and Horowitz, 1972; Slomiany et al., 1973b). The GSL with blood group activity represent such a small fraction of total lipids that their immunological activity may not be detectable in a crude lipid extract. Fractions enriched in glycolipid may be obtained by alkaline hydrolysis of glycerophosphatides or by the acetylation technique of Saito and Hakomori ( 1971). Under certain circumstanccs, the antigenic activity of a GSL may be inhibited by other GSL. Glycolipids with blood group H activity were not detected for many ycars until it was realized that the H activity was inhibited by globoside, but readily demonstrable after separation of the H GSL from the morc abundant globoside (Stellncr et at., 1973; Koscielak et nl., 1973). The human erythrocyte P blood group system contains two common antigens, P, and P, and the rare PI' antigen (Race and Sanger, 1968). A GSL with P, activity was rccently isolated from erythrocyte stroma (Naiki et al., 1975). The P and PI' antigens were identified as the GSL globoside and trihexosyl ceramide, respectively, by hemagglutination inhibition studies (Naiki and Marcus, 1974, 1975) and by chemical analysis of rare erythrocytes lacking thesc antigens. Thc P: erythrocytes,

218

DONALD M. MARCUS AND GERALD A. SCHWARTING

which lack the P antigen, lack globoside, and the p erythrocytes, which lack all P antigens, do not have detectable quantities of trihexosyl ceramide or globoside (Marcus et al., 1976). The Pk and p erythrocytes appear to be normal structurally and have a normal life-span despite the absence of their major neutral GSL. The apparently rare Pk antigen is one of the principal neutral GSL of all normal red blood cells but it is so situated in the cell membrane that it is not reactive with antibodies, even after treatment of the cells with proteolytic enzymes. The Pk erythrocytes contain increased quantities of trihexosyl ceramide because they are unable to convert it to globoside, and the absence of globoside may remove a steric obstruction that prevents the accesss of antibodies to trihexosyl ceramide. The P antigens are not restricted to erythrocytes. All three antigens are found on fibroblasts, and P, and P are detectable on peripheral blood lymphocytes (Fellous et al., 1973, 1974). Antigen PI' was not detected on freshly isolated peripheral blood leukocytes, but it was found on lymphoid cells from the same individuals maintained in culture after stimulation by mitogens (Fellous et al., 1974). The P, and P antigens are linked to the human histocompatibility HL-A locus (Edwards et al., 1972; Fellous et al., 1973; Edwards, 1974) and they should be valuable markers in studying cell hybrids and in mapping the sixth chromosome. Antibodies that sensitize erythrocytes at 4" and then lyse them at higher temperatures in the presence of complement ( Donath-Landsteiner antibodies) occur in some patients with syphilis and following viral infections. These antibodies may cause hemolysis in vivo [paroxysmal cold hemoglobinuria ( PCH ) ] when patients are exposed to cold temperatures. Many of the PCH antibodies are directed against the P antigen (Levine et d.,1965; van der Hart et al., 1964; Worlledge and ROUSSO,1965). One of these PCH antisera studied previously by Hinz (1963) was recently shown to be inhibited by globoside, and purified antigloboside antibodies give a typical Donath-Landsteiner reaction ( Schwarting and Marcus, 1976). b. Acquired Cellular Antigens. Human erythrocyte Lewis antigens are GSL (Hakomori and Kobata, 1974) that are acquired by the red cells from lipoproteins (Marcus and Cass, 1969) and not synthesized in situ. The site( s ) of synthesis of these GSL is not known. Sheep, cattle, and hog erythrocytes possess blood group antigens that cross-react with human A and H antigens, but they resemble Lewis antigens in being acquired from the plasma (Stormont, 1949; Stone, 1962; Saison and Ingram, 1962; Rasmusen, 1962). The Lewis antigens were associated with both high- and low-density lipoprotein fractions (Marcus and Cass,

GLYCOLIPIDS AND PHOSPHOLIPIDS

219

1969), but bovine J (A-like) lipid antigens were found in the very lowdensity and low-density lipoproteins ( Radas and Thiele, 1974). Dawson and Sweeley (1970) found that glucosyl ceramide of pig erythrocytes was acquired from the plasma and not synthesized by the erythrocytes. Cellular antigens are generally assumed to be synthesized by the cells on which they are detected, but these data indicate clearly that GSL antigens may be acquired. This suggests that cell membrane antigens may be altered by pathological processes that affect other cells that synthesize either the GSL or the lipoproteins that transport these molecules. As noted previously (Section II,A,2), when cells are suspended in dispersions of GSL the GSL adhere firmly to the cell membrane, and the cells will react with antibodics directed against that glycolipid. This provides a method for introducing new antigens onto cell surfaces to explore the role of specific GSL in cellular interactions or in binding immunoglobulins, complement components, or regulatory molecules. c. Lymphoid Cells. The theta (thy-1) antigens are alloantigenic surface markers of mouse thymocytes, T lymphocytes, brain, and epithelial cells (Reif and Allen, 1964; Raff, 1971). Two alloantigens are recognized in this system: Thy-1.2 occurs in the prototype C3H/HeJ strain and most other strains of mice, and thy-1.1 occurs in AKR/J mice and a few other strains. Rabbit antisera to mouse brain can be absorbed so that they react selectively with mouse T cells [antibrain associated theta (BA theta)] (Golub, 1971), but anti-BA theta sera react equally well with C3H and AKR thymocytes, as well as with thymocytes and brain tissue of other species (Golub, 1972; Peter et al., 1973; Clagett et al., 1973). Esselman and Miller (1974) reported recently that the cytotoxic action of anti-BA theta sera on thymocytes was inhibited selectively by ganglioside G,,,, (Table 111) and subsequently noted weak inhibition by G,, as well (Miller and Esselman, 1975). Anti-thy-1.2 sera were found to be inhibited strongly by G,, and less effectively by GI),I,(Miller and Esselman, 1975). No studies of thc inhibition of anti-thy-1.1 by glycolipids were reported. Antibodies to purified ganglioside G,,, cross-react extensively with G,,, and asialo G,,, (Naiki et al., 1974). Purified antibodies to G,, produce specific immunofluorescent staining of AKR/ J and C3H thymocytes, and peripheral T cells of BALB/cJ, BALR/c NIH, and AKR/J mice ( Stein-Douglas et al., 1976). Monospecific anti-Ghrlantibodies do not react with thymocytes or lymphocytes. Purified antibodies to asialo Gar,also react with purified thymocytes and T cells. These data are compatible with the suggcstion that antibodies to are among the specificities present in anti-BA theta sera, but they do not support the identificat’ion

220

DONALD M. MARCUS AND GERALD A. SCHWARTING

of GM1 as the thy-1.2 antigen. In addition, the partial purification of thy-1.2 glycoprotein antigens was reported by Letarte-Muirhead et al. ( 1974) and Kucich et al. ( 1975). Liposomes containing GMl-lecithin-cholesterol were found to inhibit synthesis of antibody to sheep erythrocytes in an in vitro system (Miller and Esselman, 1975). The liposomes were thought to act on the B cells of this system. d . Toxin Receptors, Ganglimides are capable of binding and neutralizing a number of toxins. Tetanus toxin is bound by gangliosides GDlband GT1 (van Heyningen, 1963, 1974), and cholera toxin binds to GM1 (van Heyningen et al., 1971; Holmgren et al., 1973; Cuatrecasas, 1973). Botulinus toxin was inhibited by several gangliosides (Simpson and Rapport, 1971), and there appeared to be a direct relation between the number of sialic acid residues per molecule and the activity of the gangliosides. Staphylococcal a-toxin is neutralized by the N-acetylglucosamine ganglioside (Table 11) (Kato and Naiki, 1975). There is no definitive evidence that these gangliosides are actually the cellular receptors for the toxins, but tetanus and botulinus toxins are thought to act on presynaptic membranes, and plasma membranes of synaptosomes are enriched in gangliosides (Derry and Wolfe, 1967; Hamberger and Svennerholm, 1971; Avrova et al., 1973; Morgan et al., 1971). There is considerable evidence that GM1 is the cholera toxin receptor on adipose (Cuatrecasas, 1973) and intestinal epithelial cells (Holmgren et al., 1975). Cholera toxin stimulates adenyl cyclase and increases the levels of intracellular cyclic AMP (reviewed by Bennett and Cuatrecasas, 1975). These metabolic changes are detectable following the reaction of cholera toxin with GM1 molecules that are endogenous to the cell membrane or adsorbed to the cell ( Cuatrecasas, 1973). e . Neurological Diseases. Cerebroside ( galactosyl ceramide ) is the principal GSL and a major antigenic determinant of myelin (Rapport et al., 196413). Animals that develop experimental allergic encephalomyelitis (EAE) following immunization with brain or spinal cord produce antibodies to this GSL (Niedieck et al., 1965), and anticerebroside antibodies inhibit myelination or demyelinate explants of nervous tissue (Bornstein and Appel, 1961; Bornstein and Raine, 1970; Dubois-Dalcq et al., 1970; Fry et al., 1974). Antibodies to cerebroside are not, however, a significant factor in the pathogenesis of EAE (reviewed by Paterson, 1971) or in its prevention (Hughes and Leibowitz, 1975). Antibodies to asialoganglioside and gangliosides were found in the sera of some patients with neurological disorders (Yokoyama et al., 1962), but there have been no additional reports on this subject, and the significance of this observation is not clear.

GLYCOLIPIDS AND PHOSPHOLIPIDS

221

f. Cancer. Changes in the GSL content of spontaneous and induced tumors have been the subject of several recent reviews (Hakomori, 1973, 1975; Brady, 1975). In brief, normal growing cells are deficient in certain complex glycolipids that are present in larger quantities when the cells reach saturation density; these GSL have been termed densitydependent glycolipids. Tumor cells lack these complex GSL and exhibit no density-dependent alterations in GSL content. These data suggest that GSL may play a role in cellular interactions and adherence and that the abnormally low levels of certain GSL in tumor cells is causally related to their abnormal growth pattern ( reviewed by Hakomori, 1975). Exogenous GSL added to the culture media are incorporated into the plasma membrane of growing cells and reduce the growth rate and saturation density of transformed and nontransformed cells ( Laine and Hakomori, 1973; Keenan et al., 1975). Hamster fibroblast NIL cells transformed by polyoma virus contain paragloboside, which is not detectable in untransformed NIL cells (Gahmberg and Hakomori, 1975), and animals bearing these tumors make anti-paragloboside antibodies ( Sundsmo and Hakomori, 1976). The greater accessibility to antibodies of GSL in transformed cells has been noted in a previous section. Graf and Rapport (1960) noted that antisera to cancer tissues or extracts contain antibodies to GSL, particularly lactosyl ceramide (cytolipin H ) , much more frequently than antisera to normal tissues. This may result from the increase in simpler GSL in many tumors and/or the greater accessibility of the GSL on the cell surface. Tal (1965) detected an antibody to ceramide lactose in the sera of pregnant women and cancer patients. She reported subsequently that this antibody, termed T globulin, possessed unique antigenic determinants (Tal and Halperin, 1970). A test for the detection of T globulin in cancer sera was evaluated and found to be unsuitable for clinical use (Tal et al., 1973). In view of these alterations in GSL composition and molecular arrangement in tumors, there have been surprisingly few studies of GSL in cancer patients or in experimental tumor systems. B. GLYCOSYL GLYCERIDES

1. Structure and Distribution Glycosyl glycerides contain one or more sugar residues linked to the 3-(sn)-position of a diglyceride (Fig. 3 ) . They were first isolated from plants (Carter et al., 1956) and subsequently detected in microbial organisms and animal tissues. They are most abundant in plants, comprising about 50% of the total lipid in many plants and about 35-404: of the

222

DONALD M. MARCUS AND GERALD A. SCHWARTING CH,OH I 0

It

CH-0-C-R CH,-0-C-R OH

A. Monogalactosyl diglyceride ( I ,2-diacyl-3-o-p-ogalactopyranosyl-sn -glycerol)

OH

B. Digalactosyl diglyceride 1 1,2-dlacyl-3-o-(a-D -galactopyranosyl( I +6)-0 -p-o -galactopyranosyl)-sn -glycerol 1

0

H O P > & 0 OH

HO

O-CH,

I

0

I F I1 ;

CH-0-C-R CH,-

0--C -R

C. Galactosylglucosyl diglyceride [ I ,2-diacyl-3-o-(a-o -galactopyranosyl( 1+2)-O-p-O -glucopyranosyl)-sn-glycerol]

FIG.3. Structures of some glycosyl glycerides.

dry weight of the chloroplasts. Among microbial organisms, glycosyl glycerides are present in most gram-positive bacteria and mycoplasma (Brundish et al., 1965a,b), in treponemas, and in some gram-negative bacteria (Sastry, 1974). In animals they have been found only in the nervous system (Steim, 1967). Since many of the microbial and animal glycosyl glycerides contain the same sugar sequences as the plant compounds the latter are a useful source of material for chemical and immunological studies. Glycosyl glycerides containing up to five sugars have been described but diglycosyl glycerides are the most abundant compounds, and some

GLYCOLIPIDS AND PHOSPHOLIPIDS

223

TABLE I\' B.\cTI,: I

~ AI L

1) I G LY COSY I, ( ;L Y c I,:RII) E

S ~

(:lc(a,1+2)(;1c (a,1+3)-diglyccridc Glc(P, 1 ~ 6 ) ( ~ 1 c ( ~ , l + : 3 ) - d i ~ l y r e r i d ~ Gal@, l+G)C;al(~,l-t3)-di~lyrcridc Gal(a, 1+2)C;nl(a, 1+3)-diglyrcride hlan(a,l-t3)P\lan(a,l--r9)-diglyceridc Ahbrcviations: (>lC = D-glUCOSe;

(;a1 = u-galactose;

Man

=

D-mannosc.

of the most common disaccharide structures are listed in Table IV. Galactose, glucose, and mannose are the major constituents of microbial glycosyl glycerides, and the mono- and digalactosyl glycerides are the only compounds detected thus far in animals. Microbial glycosyl glycerides contain palmitic, stearic, and oleic acids, fatty acids of plant compounds are highly unsaturated, and bacteria contain branched fatty acids (Sastry, 1974). A comprehensive review of the structures of bacterial glycolipids and glycophospholipids was published recently by Shaw (197s). 2. Immunological Properties Antibodies to glycosyl glycerides are produced in the course of natural (Beckman and Kenny, 1968; Plackett et al., 1969) and experimental infections (Brunner et al., 1973) caused by mycoplasma. They may also be elicited by immunization with whole mycoplasma cell membranes or pure glycosyl glycerides aggregated with a membrane protein from Acholeplasma laidlawii (Razin et al., 1970, 1971b). The immunological specificity of these compounds is determined by their sugar sequence and linkages, and the nonreducing terminal residue is immunodominant. For optimal complement fixation, glycosyl glycerides must be mixed with lecithin and cholesterol (Plackett et al., 1969; Kenny and Newton, 1973). Cross-reactions betwecn glycosyl glycerides of a number of microorganisms are summarized in Table V (Sugiyama et al., 1974; Kenny, 1975). Extensive cross-reactions between galactosyl glyceride and galactosyl ceramide have been demonstrated with rabbit antisera to brain tissue or to pure glycolipids, and with human sera from healthy subjects, and patients with neurological diseases or syphilis ( Dupouey, 1972; Dupouey et al., 1976). Treponema reiteri contains galactosyl diglyceride ( Dupouey et al., 1970), and the cross-reaction between this treponeme and nervous

224

DONALD M. MARCUS AND GERALD A. SCHWARTING

TABLE V CROSS-REACTING GLYCOSYL GLYCERIDES” Source Acholeplasma laidlawii Acholeplasma modicum Mycoplasma granularum

Structures

Glc (a, 1+ 3)-diglyceride Glc (a, 142)Glc (a,l+ 3)-diglyceride

Mycoplasma pneumoniae Mycoplasma neurolyticum Streptococcus MG Mycoplasma pncumoniae Spinach

Gal@, 1+ 6)Gal (P,1 4 3)-diglyceride Gal(a,l-+6)Gal(a,l-+6)Gal(~,I-+3)-diglyceride

Abbreviations: Glc = D-glucose; Gal = D-galactose.

tissue presumably involves both of these galactose-containing glycolipids. Galactosyl glyceride has not been found in Treponemu pallidum ( Dupouey and Betz, 1969). Myc,oplasma pneumoniae causes “primary atypical pneumonia” and tracheobronchitis in man (reviewed by Chanock, 1965; Couch, 1973)) and these illnesses may be accompanied by other clinical syndromes that appear from 4 to 14 days following the onset of respiratory symptoms (reviewed by Murray st al., 1975). These associated conditions include Stevens-Johnson syndrome, autoimmune hemolytic anemia caused by cold agglutinins, aseptic meningitis, meningoencephalitis, Gullain-Barre syndrome, and acute psychosis. The Stevens-Johnson syndrome ( erythema multiforme ) is generally considered to be a hypersensitivity reaction (Fellner and Bystryn, 1971), and it has been suggested that the neurological manifestations noted above may represent a hypersensitivity response to the mycoplasma and/or host tissues modified by the mycoplasma (Taylor et al., 1967; Biberfeld, 1971). The sera of convalescent patients contain antibodies that fix complement with lipid extracts of M . pneumoniae (Kenny and Grayston, 1965), and these antigens have been identified as glycosyl glycerides (Beckman and Kenny, 1968; Plackett et al., 1969). Five compounds have been identified: monogalactosyl, digalactosyl, and trigalactosyl glycerides and additional diglycosyl and tetraglycosyl glycerides (reviewed by Kenny, 1975). The structures of these compounds have not been completely elucidated. Human sera contain antibodies to the digalactosyl and trigalactosyl compounds, and a diglucosyl glyceride isolated from StreptoCOCCUS MG (Kenny and Newton, 1973). Glucose has been detected in

CLYCOLIPIDS AND PHOSPHOLIPIDS

225

crude glycolipid fractions of M . pneumoniae, but a glycolipid containing glucose has not been isolated to date. There is some uncertainty whether the dignlactosyl or trigalactosyl glyceride is the major complement-fixing antigen, but recent evidence favors the latter (Kenny, 1975). The sera of patients with primary atypical pneumonia contain antibodies against human lung, liver, heart, and kidney (Thomas et al., 1943; Thomas, 1964), and Biberfeld ( 1971) demonstrated antibodies to human brain. The antibodies studied by Biberfeld reacted with a lipid extract of brain and could be absorbed by M . pneumoniae. They were present in the sera of 80%of patients with M . pneumoniae infections not associated with neurological symptoms, and in all 7 patients with neurological symptoms, and their role, if any, in the pathogenesis of these disorders is unclear. The brain antigen was not identified, but, in retrospect, it might well have been galactosyl glyceride and/or galactosyl ceramide. Individuals with previous M . pneumoniae infections develop positive skin tests (Mizutani et al., 1971) and in vitro evidence of cell-mediated immunity to extracts of this organism ( Fernald, 1972; Biberfeld, 1972). Patients with immunodeficiency diseases who develop M . pneumoniae infections exhibit minimal or no pulmonary infiltrates (Foy et al., 1973). The latter investigators and Mizutani and associates ( 1971) suggested that the pulmonary pathology in normal persons infected by M . pneumoniae may represent a hypersensitivity reaction. In accord with this suggestion is the observation of Taylor-Robinson et al. (1972) that immunosuppressed mice infected by M . pneumoniae develop much less peribronchial and perivascular lymphocytic cuffing than normal mice. The antigen that elicits cell-mediated immunity in humans and guinea pigs infected with M . pneumoniae appears to be a protein and not a glycolipid ( Mizutani and Mizutani, 1975). The origin of the cold agglutinin anti-I antibodies found in the sera of many patients with this disease is not clear. The cold agglutinins are not adsorbed by M . pneumoniae (Liu et al., 1959; Feizi and Taylor-Robinson, 1967; Biberfeld, 1971) . Mycoplasma pneumoniae produces peroxides that alter the erythrocyte membrane, and it has been suggested that the cold agglutinins represent a response to altered autologous erythrocytes (Feizi et al., 1969). Cold agglutinins with I-like specificity have been obtained by immunization of rabbits with M . pneumoniae (Costea et al., 1971; Lind, 1973), and it is possible that the mycoplasma contains an antigen that cross-reacts with the I antigen but is present in a too small quantity to adsorb the antibody. The serological activities of glycolipids from other mycoplasma (Sugiyama et al., 1974; Kenny, 1975) and streptococcal L forms (Feinman et al., 1973) have recently been summarized.

226

DONALD M. MARCUS AND GERALD A. SCHWARTING

C. LIPOTEICHOIC ACIDS

1. Structure and Distribution Teichoic acids are a group of phosphate-containing polymers that are constituents of the cell walls and membranes of gram-positive bacteria (reviewed by Knox and Wicken, 1973). The cell wall teichoic acids are covalently linked to peptidoglycan and consist of glycerol or ribitol phosphate polymers (Fig. 4 ) that are substituted with D-alanine or sugar residues. The cell membrane lipoteichoic acids are composed of 25-30 glycerol phosphate residues that are covalently linked to membrane glycolipid ( glycosyl glycerides) (reviewed by Wicken and Knox, 1975). The membrane glycolipid is linked to teichoic acid by a phosphodiester bond between a sugar hydroxyl group and the terminal glycerol residue of the teichoic acid (Fig. 5 ) . The hydroxyl groups of glycerol are substituted by D-alanine and glycosyl residues, and the terminal glycerol group of the polymer or one of its substituent sugars may be acylated (Fig. 5). The glycolipid portion of lipoteichoic acids is inserted into the plasma membrane, and the polar glycerophosphate polymer is thought to be intercalated into the peptidoglycan network of the cell wall ( Wicken and Knox, 1975). The glycerophosphate polymers extend to the surface of the cell wall in some instances, as demonstrated by their accessibility to ferritin-labeled antibodies (van Driel et al., 1973; Dickson and Wicken, 1974; Joseph and Shockman, 1975) and by agglutination of intact organisms by antibodies to intact lipoteichoic acids (Shattock, 1949). R

Ala

Ala

R

- I - I H O O O H OH H O O O H I I ! l . l I I I I I I -0-c-c--c -c -c-o-P-o-c-c-c-c-c-o-P I I I I I II I I I I I H

H

H

H

H

H

0

H

H

H

OH

I II

0

H

A. Ribitol teichoic acid

R

I I

l l

R

Ala

I

H 0 H

l l

OH

I II

I

H 0 H

I I

l l

l l

-o-c-c-c-o-P-o-c-c-c-o-P-o-c-c-c-o H

H

H

0

H

H

H

OH

I II

0

I l

H 0 . H

I I

H

l

H

l l

H

B. Glycerol teichoic acid

FIG.4. Schematic structures of teichoic acids. R

= H or glycosyl; Ala = D-alanyl.

227

GLYCOLIPIDS AND PHOSPHOLIPIDS Lactobacillus c a x i

Streptococcus lactis

tz Lactobacillus fermenti

Streptococcus faecalis

Ha*. ,",,, y,l"i GIC

Gal

I

Glc

Gal

I-?

..

I 1-2

1-7

Gal-Glc-

I

-

iatty

~ L K Ic \ t e r

I

+o-~H,-(

q

t

I Glc GlcI

I

IG I L

GlcI I-?

I-?

.....

H

HOH-(

GlC

Glc

Gic-Glc-

I H-0-P-

II

I

?

t

glycerol rc\iduc

OH

FIG. 5. Proposed partial structures of some lipoteichoic acids. (Reproduced from an article by Wicken and Knox, 1975, by permisssion of the authors; copyright 1975 by the American Association for the Advancement of Science.)

2. Immunological and Biological Properties Some lipoteichoic acids are immunogenic in the intact organism, and antisera can be obtained readily by immunization with disintegrated organisms or a high molecular weight micellar lipoteichoic acid-protein complex that can be extracted from bacteria with hot phenol (Wicken and Knox, 1971; Wicken et al., 1973). Most of the antibodies to lipoteichoic acids are directcd against the glycerophosphate backbone or its carbohydrate substituents. The former type of antibodies cross-react widely with a number of lipoteichoic acids (Knox and Wicken, 1973). At present, lipoteichoic acids have been identified as group-specific antigens of bacteria of only two genera, Streptococcus and Lactobacillus. Organisms with group-specific antigens include Group D and N streptococci and Group F lactobacilli (Knox and Wicken, 1973). The immunodominant groups of the lipoteichoic acids are the glucosyl and galactosyl substituents of the glycerophosphate polymer; some of thcse determinants are listed in Table VI (Wicken and Knox, 1975). The few human antibodies studied to date have becn directed against the glycerophosphate backbone (Decker et al., 1972; Markham et al., 1973).

228

DONALD M. MARCUS AND GERALD A. SCHWARTING

TABLE VI STRUCTURES O F GROUP-SPGCIFIC CARBOHYDRATE DETERMINANTS O F LIPOTEICHOIC ACIDS Genus

Group

Determinant

Lactobacillus Lactobacillus Streptococcus Streptococcus Streptococcus

A F

a-~-Gl~co~yl a-o-Galactosyl a-D-GlUCosyl-(1+2)-glUCOSyl a-D-Galactosy 1 P-D-Galactosyl

n N Serotype a

Purified lipoteichoic acid or soluble antigen obtained from culture fluid or saline washings of gram-positive organisms adhere firmly to erythrocytes and sensitize them to hemolysis or hemagglutination by antibodies (Rantz et al., 1956; Gorzynski et al., 1960). Lipoteichoic acids can also exchange between erythrocytes and tissues. The ability of lipoteichoic acids to adhere to cell membranes is dependent on their content of esterlinked fatty acids (Hewett et al., 1970; Matsuno and Slade, 1971; Ofek et d.,1975). It has been suggested that complexes of lipoteichoic acid with streptococcal antigens may bind to host tissues and play a role in the pathogenesis of poststreptococcal diseases such as rheumatic fever and acute glomerulonephritis ( Moskowitz, 1966). Lipoteichoic acids are analogous in some respects to lipopolysaccharides of gram-negative bacteria (Wicken and Knox, 1975) : Both are amphipathic molecules capable of attaching to cell membranes and both can elicit local and generalized Schwartzman reactions. Lipoteichoic acids are not mitogenic for B cells and they do not possess endotoxic properties. Other immunological properties of lipoteichoic acids include their ability to depress the immune response to sheep erythrocytes and enhance the immune response to lipopolysaccharides (Miller and Jackson, 1973, 1974) as well as their crossreactions with cardiolipin (Wicken & al., 1972). The latter property could be responsible for some false positive serological reactions for syphilis.

D. OTHERGLYCOLIPIDS Other glycolipids with immunological properties, which are not discussed in this review, include acylated sugars ( Coulon-Morelec, 1968, 1972; Coulon-Morelec et al., 1967, 1968, 1970; Faure and Coulon-Morelec, 1974; Shaw, 1970) and mycolic acids (reviewed by Lederer et al., 1975). Immunochemical properties of lipopolysaccharides have been reviewed extensively (for a recent review, see Luderitz et al., 1971).

229

GLYCOLIPIDS AND PHOSPHOLIPIDS

Ill. Phospholipids

A. CARDIOLIPIN 1 . Structure and Distribution The immunological properties of cardiolipin ( diphosphatidylglycerol ) (Fig. 6 ) have been studied extensively because of its role in the serological diagnosis of syphilis. Cardiolipin was first isolated from an alcoholic extract of beef heart by Pangborn (1942), its structure was studied by a number of investigators (reviewed by MacFarlane, 1964), and it was synthesized by de Haas et al. (1966). Cardiolipin is ubiquitous -it is found in mammals, fish, birds, bacteria, protozoa, yeasts, mycobacteria, and treponemas ( MacFarlane, 1964). I t is located principally in the membranes of organelles that display high metabolic activities : mitochondria, bacterial protoplasts, and chloroplasts of photosynthetic bacteria. Bovine heart cardiolipin contains mostly unsaturated fatty acids, approximately 85%linoleic acid ( 18 :2 ) , but synthetic or hydrogenated compounds containing saturated fatty acids have essentially equal immunological activity ( Faure and Morelec-Coulon, 1963; Inoue and Nojima, 1967). 2. lmrnunological Properties Antibodies reactive with cardiolipin occur in sera of patients with syphilis and other diseases caused by spirochetes, leprosy, systemic lupus erythematosus, and transiently in a number of acute viral infections ( Sparling, 1971 ) . Antibodies to cardiolipin have been raised in rabbits immunized with crude lipid extracts mixed with heterologous serum or a foreign protein (Eagle, 1932) or with mitochondria ( Schiefer, 1973a). Liposomes containing cardiolipin, lecithin, and cholesterol, with or without a foreign protein such as human antibody (Fowler and Allen, 1962; Aho et al., 1973) or methylated bovine serum albumin (MBSA) (Inoue and Nojima, 1967; DeSiervo, 1974) are also good immunogens. CardioOH

0

CH,-0-C-R

II

CH,-0-C-R

I

f

I

B

CH,-0-P-0-CH,

I

1

I

CH,-0-P-0-CH,

CH-OH

a

1

0 II

CH-0-C-R

I

CH,-0-C-R

t

OH

FIG.6. Structure of cardiolipin ( diphosphatidylglycerol).

230

DONALD M. MARCUS AND GERALD A. SCHWARTING

lipin-lecithin-MBSA is a good immunogen, but cardiolipin-cholesterolMBSA is very weakly immunogenic. Pure cardiolipin does not react well with antibodies and auxiliary lipids are required for optimal immunological reactivity. Cardiolipinlecithin mixtures are precipitated by antibodies (Osler and Knipp, 1957) and fix complement, but maximum sensitivity in complement fixation is achieved by using cardiolipin-cholesterol-lecithin mixtures ( Maltaner and Maltaner, 1945). The polar head groups of lecithin molecules in these liposomes can be hydrolyzed by phospholipase C. The antigenic activity of liposomes containing cardiolipin is unaffected by hydrolysis of up to 80%of the lecithin, but complete hydrolysis does reduce their activity (Kataoka and Nojima, 1969). Active liposomes cannot be prepared, however, by substituting diglyceride for lecithin. The polar head group of lecithin appears to be important for proper orientation of the molecules when the liposomes are formed, but the structure can be maintained without most of the head groups. Electron microscopic studies of lipid particles revealed that cardiolipin-lecithin-cholesterol liposomes consisted of lamellar structures surrounding a cholesterol core, whereas cardiolipin alone or cardiolipin-cholesterol particles form an irregular network without any lamellar structure ( Kanemasa, 1974). The reactions of human and rabbit antibodies to cardiolipin with molecules related structurally to cardiolipin were examined by Faure and Morelec-Coulon (1963) and Inoue and Nojima (1967, 1969). Derivitization of the free hydroxyl group or removal of one or two fatty acids from cardiolipin markedly decreased its reaction with antibody, and removal of more than two fatty acids essentially abolished its activity. The distance between the two phosphodiester groups is also an important structural feature because synthetic derivatives in which the central glycerol moiety is replaced by longer or shorter methylene chains also exhibit diminished immunological activity ( Inoue and Nojima, 1967). Cross-reactions between anticardiolipin antibodies and phosphatidyl inositol ( P I ) and nucleic acids were reported by Guarnieri (1974) and Guarnieri and Eisner ( 1974). Reciprocal cross-reactions were observed between antisera to cardiolipin and PI and the two antigens; the crossreaction of cardiolipin with anti-PI was stronger than the reciprocal reaction. Guarnieri and Eisner made the interesting observation that DNA and cardiolipin reacted equally with anticardiolipin antibodies and that all of the antibodies to cardiolipin could be absorbed by DNA. These investigators used a microflocculation assay, and it was necessary to mix the DNA with lecithin and cholesterol to detect the reaction. Ribonucleic acid was about 10-203 as effective as DNA in reacting with the cardiolipin antibodies. Guarnieri and Eisner suggested that the basis

CLYCOLIPIDS AND PHOSPHOLIPIDS

231

of the cross-reaction is a structure composed of two phosphodiester groups separated by 3 carbon atoms. They suggested also that the hemiacetal oxygen of the deoxyribose ring might be immunologically equivalent to the hydroxyl group of the central glycerol residue in cardiolipin. The reaction of rabbit anticardiolipin antibodies with mouse tissues was studied by a direct immunofluorescent technique (Kataoka and Nojima, 1968). After fixation of tissue sections with acetone-buffered saline, fluorescent staining was observed in heart, skeletal muscle, kidney, and liver in a distribution suggestive of mitochondria. Rabbit antibodies also bind to intact mitochondria isolated from a variety of tissues (Guarnieri et al., 1971; Schiefer, 1973b). In another study ( Aho et al., 1973), antibodies from rabbits with experimental syphilis or from patients with syphilis or biological false positive serologies reacted with intact mitochondria, but antibodies from rabbits immunized with cardiolipinlecithin-cholesterol liposomes coated with human antibodies did not react. The reason for the discrepancy between this study and those of Guarnieri and Schiefer is not apparent. Guarnieri et al. concluded that the polar head groups of only 9% of cardiolipin molecules of the mitochondrial membrane were accessible to antibodies. This calculation was based on a comparison of the number of cardiolipin molecules in mitochondria with the number of antibody molecules adsorbed. The calculation may not be valid if the binding of one antibody molecule obstructs access of other antibodies to adjacent cardiolipin molecules. Schiefer ( 197317) found that treatment with trypsin and pronase increased the uptake of anticardiolipin antibodies by inner mitochondrial membranes but not by intact mitochondria. The reaction of anticardiolipin antibodies with cardiolipin-containing liposomes was studied by electron spin resonance (Schiefer ct al., 1975). They used spin-labeled derivatives of stearic acid in which nitroxide groups were located near polar head groups of the phosphatides or in the hydrophobic interior of the liposomes. The mobility of the nitroxide group in ithe polar region of liposomes was decreased when the liposomes were 1 exposed to antibodies, but the nitroxidc probc in the hydrophobic region >wasunaffected. The antibody-cardiolipin interaction appears to produce i tightcr packing of the polar head groups of the phosphatides. Anticardiolipin antibodies also blocked the condensing effect of calcium on the 1 liposomes. The precise nature of the antigcnic stimulus that elicits human anti)bodies to cardiolipin is not clear. Cardiolipin is a constituent of many 1 treponemas, including the Rciter strain ( Faure and Pillot, 1960), and ithe immunogenicity of cardiolipin may be enhanced by its presence in a 1 membrane containing foreign antigens. On the other hand, the presence

232

DONALD M. MARCUS AND GERALD A. SCHWARTING

of antibodies to cardiolipin in the sera of normal elderly individuals, drug addicts, and patients with systemic lupus erythematosus, leprosy, and other diseases (reviewed by Sparling, 1971) suggests an autoimmune process. This possibility is supported by the appearance of rheumatoid factors and cryoglobulins in the sera of patients with syphilis. There is no information about the pathogenetic significance of these antibodies. Rabbits producing anticardiolipin antibodies do not develop overt disease and mitochondrial respiratory functions were not impaired by exposure to anticardiolipin antibodies in vitro (Guarnieri et al., 1971). It is unclear, however, whether complement was present during the latter experiment. The extensive cross-reaction between rabbit anticardiolipin antibodies and DNA indicates the need for a careful study of the specificity of human antibodies to these two antigens, particularly when they occur in the same patient, as in systemic lupus erythematosus.

B. PHOSPHATIDYL INOSITOL Rabbit antibodies to PI were produced by rabbits immunized with mitochondria ( Schiefer, 1973a) or liposomes containing PI-lecithincholesterol-MBSA ( Kataoka and Nojima, 1969; Guarnieri, 1974). The general serological properties of these antibodies are similar to anticardiolipin antibodies. Auxiliary lipids are necessary for complement fixation and microflocculation reactions, and hydrolysis of most of the PI in liposomes by phospholipase C does not alter the immunological reactivity of the liposomes. Kataoka and Nojima did not detect the cross-reaction of cardiolipin with anti-PI observed by Guarnieri, but the former group may not have examined many sera or used as sensitive serological techniques as Guarnieri. The flocculation of liposomes containing PI was not inhibited by myoinositol or glycerylphosphoryl inositol ( deacylated PI ). Anti-PI sera fixed complement with intact mitochondria and inner mitochondrial membranes, indicating that the polar head groups of some PI molecules are accessible to antibodies on the membrane (Schiefer, 1973b). Anti-PI antibodies are adsorbed by myelin and synaptosomes (Guarnieri, 1974). The uptake of antibodies by myelin, but not synaptosomes, was increased by performing the incubation with antibodies at 45".

C. SPHINGOMYELIN Antibodies to sphingomyelin have been produced by immunization with conjugates containing deacylated sphingosine or other haptens coupled to carrier proteins. Taketomi and Yamakawa (1966; Taketomi, 1969) coupled N-p-aminobenzyldihydrosphingosylphosphorylcholineto

GLYCOLIPIDS AND PHOSPHOLIPIDS

233

BSA or egg albumin by diazotization. Hapten-specific antibodies were demonstrated by complement fixation with hapten coupled to an unrelated protein and by passive cutaneous anaphylaxis. Teitelbaum et al. (1973) and Arnon and Teitelbaum (1974) used two haptens, dihydrosphingosylphosphorylcholine ( SPC ) and ceramide phosphorylethanolamine ( CPE ), and employed carbodiimides as the coupling agents. Hapten-specific antibodies were elicited by both conjugates, and CPE was more immunogenic than SPC. The antibodies were apparently able to react with sphingomyelin in cell membranes because they lysed sheep erythrocytes, which are rich in sphingomyelin, but not guinea pig erythrocytes, which contain very little sphingomyelin ( Arnon and Teitelbaum, 1974). IV. Concluding Remarks

During the 7 years since the last general review of lipid antigens (Rapport and Graf, 1969), much has been learned about the structure, biosynthesis, and immunological properties of glycolipids, but their biological functions remain elusive. The recurrent suggestions that they may serve to mediate cellular interactions or act as cell membrane receptors or regulatory molecules remain plausible and intriguing, but unproven. Despite the uncertainty about their functions, glycolipids offer unique experimental advantages for studies of the architecture and functional properties of cell membranes. Glycolipids are the only components of cell membranes that are readily isolated and possess a single antigenic determinant. Antisera to these determinants can be used in many ways: to identify cells that are not readily distinguished by morphological differences, such as T and B lymphocytes; to obtain data on the accessibility of specific antigenic determinants to antibodies during different physiological states, such as phases of the cell cycle, and in pathological conditions; to determine the cellular and subcellular distribution of glycolipids; to prepare afFinity columns for fractionation of cells and macromolecules; and to study model membranes containing glycolipid antigens. These studies should clarify the biological role of glycolipids and provide insight into many aspects of cell membrane structure and function.

REFERENCES Aho, K., Rostedt, I., and Saris, N. E. (1973). Clin. E r p . Zmmunol. 14, 573. Allen, P. Z., and Kabat, E. A. ( 1959). J . Immunol. 82, 358. Alving, C. R., Joseph, K. C., and Wistar, R. (1974). Biochemistry 13, 4818. Ando, S., and Yamakawa, T. ( 1973). J . Biochem. (Tokyo) 73,387.

234

DONALD M. MARCUS A N D GERALD A . SCHWARTING

Ando, S., Kon, K., Isobe, M., and Yamakawa, T. (1973). J. Biochem. (Tokyo) 73, 893. Arnon, R., and Teitelbaum, D. (1974). Chem. Phys. Lipids 13,352. Arnon, R., Sela, M., Rachaman, E. S., and Shapiro, D. ( 1967). Biochemistry 2, 79. Avrova, N. F., Chenykaeva, E. Y.,and Obukhova, E. L. (1973). 1. Neurochem. 20, 997. Beckman, B. L., and Kenny, G. E. (1968). I. Bacteriol. 96,1171. Bennett, V., and Cuatrecasas, P. (1975). J. Membr. Biol. 22, 29. Biberfeld, G. (1971). Clin. Erp. Immunol. 8, 319. Biberfeld, G. ( 1972). In “Pathogenic Mycoplasmas” (K. Elliott and J. Birch, eds.), p. 365. Assoc. Sci. Publ., Amsterdam. Bornstein, M. B., and Appel, S. H. ( 1961). J. NeuropathoL Erp. Neurol. 20, 141. Bornstein, M. B., and Raine, C. S. (1970). Lab. Inuest. 23, 536. Brady, R. 0. (1975). Am. J. Clin. Pathol. 63, 685. Brundish, D. E., Shaw, N., and Baddiley, J. (1965a). Biochem. Biophys. Res. Commun. 18, 308. Brundish, D. E., Shaw, N., and Baddiley, J. (1965b). Biochem. J. 97, 158. Brunner, H., Greenberg, H. B., James, W. D., Hoswood, R. L., Couch, R. B., and Chanock, R. M. (1973). Infect. Immun. 8,612. Carrel, S . , and Barandun, S. (1971). Immunochemistry 8, 39. Carter, H. E., McCluer, R. H., and Slifer, E. (1956). J. Am. Chem. SOC. 78, 3735. Chanock, R. M. (1965). N . En& 1. Med. 273,1199 and 1257. Clagett, J., Peter, H. H., Feldman, J. D., and Weigle, W. 0. (1973). 1. Immunol. 110, 1085. Cook, G. M. W., Heard, D. H., and Seaman, G. V. F. (1960). Nature (London) 188, 1011. Costea, N., Yakulis, V. J,, and Heller, P. (1971). J. Immunol. 106, 598. Couch, R. B. (1973). In “Viral and Mycoplasmal Infections of the Respiratory Tract” ( V. Knight, ed. ), Chapter 14. Lea & Febiger, Philadelphia, Pennsylvania. Coulon-Morelac, M. J. ( 1968). C.R. Hebd. Sciences Acad. Sci. 267, 1541. Coulon-Morelec, M. J. (1972). Ann. Inst. Pasteur, Paris 123, 619. Coulon-Morelec, M. J., Faure, M., and Markchal, J. (1967). Ann. Inst. Pasteur, Paris 113, 37. Coulon-Morelec, M. J., Faure, M., and Marbchal, J. (1968). Ann. Inst. Pasteur, Paris 114, 775. Coulon-Morelec, M. J., Faure, M., and Markchal, J. (1970). Ann. Inst. Pasteur, Paris 119, 17. Critchley, D. R., and Macpherson, I. (1973). Biochim. Biophys. Acta 296, 145. Cuatrecasas, P. (1973). Biochemistry 12, 3558. Dawson, G., and Sweeley, C. G. (1970). 1. Biol. Chem. 245,410. de Baecque, C., Johnson, A. B., Naiki, M., Schwarting, G., and Marcus, D. M. (1976). I. Neurochem. (abstr.) 35, 92. Decker, G . P., Chorpenning, F. W., and Frederick, G . T. (1972). J. Immunol. 108, 214. de Haas, G. H., Bonsen, P. P. M., and van Deenen, L. L. M. (1966). Biochim. Biophys. Acta 116, 114. Deny, D. M., and Wolfe, L. S. (1967). Science 158, 1450. DeSiervo, A. ( 1974). Infect. Zmmun. 9, 835. Dickson, M. R., and Wicken, A. J. (1974). Proc. I n t . Congr. Electron Microsc., 8th, Vol. 11, pp. 114-115.

GLYCOLIPIDS AND I’HOSI’HOLIPIDS

235

Dod, B. J., and Gray, G. M. (1968). Biochim. Biophys. Acta 150,397. Dubois-Dalck, M., Niedieck, B., and Buyse, N. ( 1970). Puthol. Eur. 5, 331. Dupouey, P. (1972). J. Zmmunol. 109, 146. Dupouey, P., and Betz, A. (1969). Ann. Inst. Pastercr, Paris 117, 395. Dupouey, P., Coulon-Morelec, M. J., and MarPchal, J. (1970). C.R. Hebd. Seances Acud. Sci. 270, 1541. Dupouey, P., Billecocq, A., and Lefroit, M. (1976). Zmmunochemistry 13, 289. Eagle, H. ( 1932). J. Exp. Med. 55, 667. Edwards, J. H. (1974). Cytogenet. Cell Genet. 13, 187. Edwards, J. H., Allen, F. H., Glenn, K. P., Lamm, L. U., and Robson, E. B. (1972). In “Histocompatibility Testing” (J. Dausset and J. Colonibani, eds.), p. 745. Munksgaard, Copenhagen. Esselman, W. J., and Miller, H. C. (1974). J. Erp. Med. 139, 445. Eylar, E. H., Madoff, M. A., Brody, 0. V., and Oncley, J. L. (1962). J. Biol. Chem. 237, 1992. Faure, M., and Coulon-Morelec, M. J. ( 1963). Ann. Inst. Pmteur, Paris 104, 246. Faure, M., and Coulon-Morelec, M. J. (1974). Biochimie 56, 263. Faure, M., and Pillot, J. (1960). Ann. Inst. Pasteur, Paris 99, 729. Feinman, S. B., Prescott, B., and Cole, R. M. (1973). Infect. Immun. 8, 752. Feizi, T., and Taylor-Robinson, D. ( 1967). Immunology 13, 405. Feizi, T., Taylor-Robinson, D., Shields, M. D., and Carter, R. A. (1969). Nature (London) 222, 1253. Fcllner, M. J., and Bystryn, J. ( 1971 ). In “Immunological Diseases” (M. Samter, ed.), 2nd ed., pp. 952-963. Little, Brown, Boston, Massachusetts. Fellous, M., Couillin, P., Neauport-Sautes, C., Frezal, J., Bilardon, C., and Dausset, J. (1973). Eur. 1. Zmmunol. 3, 543. Fellous, M., Gerbal, A,, Tessier, C., Frezal, J., Dausset, J., and Salmon, C. (1974). Vox Sang. 26, 518. Femald, G. W. (1972). Infect. Immun. 5, 552. Fowler, E., and Allen, R. H. (1962). J. Zmmttnol. 88, 591. Foy, H. M., Ochs, H., Davis, S. D., Kenny, G. E., and Luce, R. R. (1973). J. Infect. Dis. 127, 388 Fry, J. M., Weissbarth, S., Lehrer, G. M., and Bornstein, M. B. (1974). Science 183, 540. Gahmberg, C. G., and Hakomori, S. ( 1975). J. B i d . Chem. 250, 2438. Golub, E. S. (1971). Cell. Zmmunol. 2, 353. Golub, E. S. (1972). J . Zmmunol. 109, 168. Gorzynski, E. A., Neter, E., and Cohen, E. (1960). J . Bacteriol. 80, 207. Graf, L., and Rapport, M. M. ( 1960). Cancer Res. 20, 546. Graf, L., and Rapport, M. M. (1965). Int. Arch. Allergy Appl. Zmmunol. 28, 171. Graf, L., and Rapport, M. M. (1970). I. Membr. Biol. 3,280. Graf, L., Yariv, J,, and Rapport, M. M. (1965). Immunochemistry 2, 145. Gregson, N. A., Kennedy, M., and Leibowtiz, S. ( 1974). Zmmunology 26, 743. Guarnieri, M. ( 1974). Lipids 9, 692. Guarnieri, M., and Eisner, D. (1974). Biochem. Biophys. Res. Commun. 58, 347. Guarnieri, M., Stechmiller, B., and Lehninger, A. L. (1971). J. Biol. Chem. 246, 7526. Hakomori, S. (1966). J. Lipid Res. 7, 789. Hakomori, S. (1969). Vox Sang. 16,478. Hakomori, S . (1973). Ado. Cancer Res. 18, 265.

236

DONALD M. MARCUS AND GERALD A. SCHWARTING

Hakomori, S. (1975). Biochim. Biophys. A d a 417,55. Hakomori, S., and Andrews, H. D. ( 1970). Biochim. Biophys. Acta 202,225. Hakomori, S . , and Kobata, A. ( 1974). In “The Antigens” (M. Sela, ed.), Vol. 2, Chapter 2, pp. 80-141. Academic Press, New York. Hakomori, S., Koscielak, J., Bloch, K. J., and Jeanloz, R. W. (1967). J. Immunol. 98, 31. Hakomori, S., Teather, C., and Andrews, H. D. (1968). Biochem. Biophys. Res. Commun. 33,563. Hakomori, S., Kijimoto, S., and Siddiqui, B. (1972). In “Membrane Research” (C. F. Fox, ed. ), pp. 253-277. Academic Press, New York. Hamberger, A,, and Svennerholm, L. (1971). J. Neurochem. 18, 1821. Heidelberger, M. ( 1956). “Lectures in Immunochemistry.” Academic Press, New York. Hewett, M. J., Knox, K. W., and Wicken, A. J. (1970). J. Gen. Microbiol. 28, 177. Hinz, C. F., Jr. ( 1963). Blood 22, 600. Holmgren, J., Lonnroth, I., and Svennerholm, L. (1973). Infect. Immun. 8,208. Holmgren, J., Lonnroth, I., Mansson, J. E., and Svennerholm, L. (1975). Proc. Natl. Acad. Sci. U.S.A. 72,2520. Hughes, R. A. C., and Leibowitz, S. (1975). Immunology 28,213. Inoue, K., and Nojima, S. (1967). Biochim. Biophys. Acta 144,409. Inoue, K., and Nojima, S. ( 1969). Chern. Phys. Lipids 3, 70. Inoue, K., Graf, L., and Rapport, M. M. ( 1972). J. Lipid Res. 13, 119. Javaid, J. I., and Winzler, R. J. (1974a). Biochemistry 13, 3635. Javaid, J. I., and Winzler, R. J. (1974b). Biochemistry 13, 3639. Joffe, S., Rapport, M., and Graf, L. (1963). Nature (London) 197, 60. Joseph, K. C., Alving, C. R., and Wistar, R. (1974). J. Immuml. 112, 1949. Joseph, R.,and Shockman, B. D. (1975). J. Bacteriol. 122,1375. Kabat, E . A. ( 1968). “Structural Concepts in Immunology and Immunochemistry,” p. 19. Holt, New York. Kabat, E. A., and Mayer, M. M. (1961). In “Experimental Immunochemistry” (E. A. Kabat and M. M. Mayer, eds.), 2nd ed., Chapter 4, pp. 133-240. Thomas, Springfield, Illinois. Kanemasa, Y. (1974). Jpn. J. Microbiol. 18, 193. Kataoka, T., and Nojima, S. (1968). Jpn. J . Exp. Med. 38,251. Kataoka, T., and Nojima, S. (1969). Jpn. J. Exp. Med. 39, 129. Kato, I., and Naiki, M. ( 1976). Infect. Immun. 13, 289. Keenan, T. W., MorrB, D. J., and Huang, C. M. (1972). FEBS Lett. 24, 204. Keenan, T. W., Schmid, E., Franke, W. W., and Wiegand, H. (1975). Exp. CeU Res. 92, 259. Kenny, G. E. (1975). In “The Antigens” ( M . Sela, ed.), Vol. 3, pp. 449-478. Academic Press, New York. Kenny, G. E., and Grayston, J. T. ( 1965). 1. Immunol. 95, 19. Kenny, G. E., and Newton, R. M. ( 1973). Ann. N.Y. Acad. S c i . 225,54. Kinsky, S. C. (1972). Biochim. Biophys. Acta 265, 1. Kinsky, S. C., Haxby, J. A,, Zopf, D. A., Alving, C. R., and Kinsky, C. B. (1969). Biochemistry 8, 4149. Klenk, H., and Choppin, P. W. (1970). Proc. Natl. Acad. Sci. U S A . 66, 57. Knox, K. W, and Wicken, A. J. (1973). Bacteriol. Rev. 37,215. Koscielak, J., Hakomori, S., and Jeanloz, R. W. ( 1968). Immunochemistry 5, 441.

CLYCOLIF’IDS AND PHOSPHOLIPIDS

237

Koscielak, J., Piasek, A., Gomiak, H., Gardas, A., and Gregor, A. (1973). Eur. J. Biochem. 37, 214. Kucich, U. N., Bennett, J. C., and Johnson, B. J. (1975). J . Immunol. 115, 626. Laine, R. A., and Hakomori, S. (1973). Biochem. Biophys. Res. Commun. 54, 1039. Laine, R. A., Yogeeswaran, G., and Hakomori, S. (1974). J . Biol. Chem. 249, 4460. Landsteiner, K. ( 1962). “The Specificity of Serological Reactions,” p. 100. Dover, New York. Landsteiner, K., and Simms, S. (1923). 1. Exp. Med. 38, 127. Ledeen, R. W., and Yu, R. K. (1973). In “Lysomes and Storage Diseases” (H. G. Hers and F. Van Hoof, eds.), p. 105. Academic Press, New York. Lederer, E., Adam, A., Ciorbaru, R., Petit, J. F., and Wietzerbin, J. (1975). Mol. Cell. Biochem. 7, 87. Letarte-Muirhead, M., Acton, R. T., and Williams, A. F. (1974). Biochem. J . 143, 51. Levine, P., Celano, M. J., and Falkowski, F. (1965). Ann. N.Y. Acad. Sci. 124, 456. Lind, K. (1973). Acta Pathol. Microbiol. Scand., Sect. B 81, 487. Liu, C., Eaton, M. D., and Heyl, J. T. (1959). J. Exp. Med. 109, 545. Luderitz, O., Westphal, O., Staub, A. M., and Nikaido, H. (1971). Microb. Toxins 4, 145. Luner, S. J., Sturgeon, P., Szklarek, D., and McQuinston, D. T. (1975). Vox Sang. 28, 184. MacFarlane, M. G. (1964). Adu. Lipid Res. 2,91. Makela, O., Miettinen, T., and Pesola, R. (1960). Vox Sang. 5, 493. Makita, A. (1964). J. Biochem. (Tokyo) 55,269. Makita, A., Iwanaga, M., and Yamakawa, T. (1964). J. Biochem. (Tokyo) 55, 202. Maltaner, E., and Maltaner, F. (1945). J . Immunol. 51, 195. Marchesi, V. T., Jackson, R. L., Segrest, J. P., and Kahane, I. (1973). Fed Proc., Fed. Am. SOC. Exp. Biol. 32, 1833. Marcus, D. M. ( 1976). J . Am. Oil Chem. SOC. (in press). Marcus, D. M., and Cass, L. E. (1969). Science 164, 553. Marcus, D. M., and Janis, R. (1970). J. Immunol. 104, 1530. Marcus, D. M., Naiki, M., and Kundu, S. (1976). Proc. Natl. Acad. Sci. U.S.A. (in press ) . Markham, J. L., Knox, K. W., Schamschula, R. G., and Wicken, A. J. (1973). Arch. Oral Biol. 18, 313. Martensson, E. ( 1969). Prog. Chem. Fats Other Lipids 10, Part 4, p. 367. Matsuno, T., and Slade, H. D. (1971). Infect. Immun. 3,385. Miller, G. A., and Jackson, R. W. ( 1973). J. Immunol. 110, 148. Miller, G. A,, and Jackson, R. W. (1974). Fed. Proc., Fed. Am. SOC. Exp. Biol. 33, 600. Miller, H. C., and Esselman, W. T. (1975). J . Immunol. 115, 839. Mizutani, H., and Mizutani, H. (1975). Jpn. J. Microbid. 19, 157. Mizutani, H., Mizutani, H., Kitayama, T., Hayakawa, A., Nagayana, E., Kato, J.. Tamura, H., and Izuchi, T. (1971). Lancet 1, 186. Mollison, P. L. ( 1972 ). “Blood Transfusion in Clinical Medicine,” 5th ed. Blackwell, Oxford. Morgan, I. G., Wolfe, L. S . , Mandel, P., and Gombos, G. (1971). Biochim. Biophys. Acta 241,737. Moskowitz, M. (1966). J. Bocteriol. 91, 2200. Murray, H. W., Masur, H., SenteAt, L. B., and Robert, R. B. (1975). Am. J . Med. 58, 229.

238

DONALD M. M A R C U S A N D GERALD A. SCHWARTING

Mustakallio, K. K., Lassus, A., and Wager, 0. (1967). Int. Arch. Allergy Appl. Immunol. 31, 417. Nagai, Y.,and Oshawa, T. ( 1974). Jpn. J . Erp. Med. 44,451. Naiki, M., and Marcus, D. M. (1974). Biochem. Biophys. Res. Commun. 60, 1105. Naiki, M., and Marcus, D. M. (1975). Biochemistry 14, 4837. Naiki, M., and Taketomi, T. (1969). Jpn. J . Erp. Med. 39, 549. Naiki, M., Marcus, D. M., and Ledeen, R. ( 1974). J . Immunol. 113,84. Naiki, M., Fong, J., Ledeen, R., and Marcus, D. M. (1975). Biochemistry 14, 4831. Niedieck, B., and Palacios, 0. (1965). 2.Immun. Allergieforsch. 129,234. Niedieck, B., Kuwert, E., Palacios, O., and Drees, 0. (1965). Ann. N.Y. Acad. Sci. 122, 266. Ofek, I., Beachey, E. H., Jefferson, W., and Campbell, G. L. (1975). J. Exp. Med. 141, 990. Osler, A. G., and Knipp, E. A. (1957). J. Immunol. 78, 19. Pangborn, M. C. (1942). J. Biol. Chem. 143, 247. Paterson, P. Y. (1971). In “Immunological Diseases” (M. Samter, ed.), 2nd ed., p. 1269. Little, Brown, Boston, Massachusetts. Peter, H. H., Clagett, J., Feldman, J. D., and Weigle, W. 0. ( 1973). J . Immunol. 119, 1077. Phillips, R. B., and Rodey, G. ( 1975). Zmmunogenetics 2,395. Plackett, P., Marmion, B. P., Shaw, E. J., and Lemcke, R. M. (1969). Aust. J . Exp. B i d . Med. Sci. 47, 171. Race, R. R., and Sanger, R. (1968). “Blood Groups in Man,” 5th ed., pp. 136-170. Blackwell, Oxford. Radas, A., and Thiele, 0. W. ( 1975). Eur. J . Biochem. 55, 583. Raff, M. C. (1971). Transplant. Rev. 6, 52. Rantz, L. A., Randall, E., and Zuckennan, A. (1956). J. Infect. Dis. 98, 211. Rapport, M. M. ( 1961). J. Lipid Res. 2, 25. Rapport, M. M., and Graf, L. ( 1957). Ann. N.Y. Acad. Sci. 69, 608. Rapport, M. M., and Graf, L. (1969). Prog. Allergy 13, 273. Rapport, M. M., Graf, L., Skipski, V., and Alonzo, N. F. (1959). Cancer 12,438. Rapport, M. M., Graf, L., and Yariv, J. ( 1961). Arch. Biochem. Biophys. 92, 438. Rapport, M. M., Graf, L., and Schneider, H. (1964a), Arch. Biochem. Biophys. 105, 431. Rapport, M. M., Graf, L., Autilio, L. A., and Norton, W. L. (1964b). J . Neurochem. 11, 855. Rasmusen, B. A. (1962). Ann. N.Y. Acad. Sci. 97, 306. Razin, S. (1970). Proc. Natl. Acad. Sci. U.S.A. 67, 590. Razin, S., Prescott, B., and Chanock, R. M. (1970). Pruc. Natl. Acad. Sci. U.S.A. 67, 590. Razin, S., Chanock, R. M., Graf, L., and Rapport, M. M. (1971a). Proc. SOC. Exp. Biol. Med. 138, 404. Razin, S., Prescott, B., James, W. D., Caldes, G., Valdsuso, J., and Chanock, R. M. (1971b). Infect. Immun. 3, 420. Reif, A. E., and Allen, J. M. V. (1964). J . Exp. Med. 120, 413. Saison, R., and Ingram, D. G. (1962). Ann. N.Y. Acad. Sci. 97,226. Saito, T., and Hakomori, S . ( 1971). J . Lipid Res. 12, 257. Sandberg, A. L., and Stollar, B. D. (1966). J . Immunol. 96,764. Sastry, P. S. (1974). Adu. Lipid. Res. 12,251. Schiefer, H.-G. (1973a). Hoppe-Seyler’s Z. Physiol. Chem. 354, 722.

GLYCOLIPIDS AND PHOSPHOLIPIDS

239

Schiefer, H.-G. ( 1973b). Hoppe-Seyler’s Z. Physiol. Chem. 354, 725. Schiefer, H.-G., Schummer, U., Hegner, D., Gerhardt, U., and Schnepel, G. H. (1975). Hoppe-Seyler’s Z. Physiol. Chem. 356,293. Schwarting, G. A,, and Marcus, D. M. ( 1976). In preparation. Seaman, G. V. F., and Heard, D. H. (1960). J . Gen. Physid. 44,251. Shattock, P. M. F. (1949). J . Gen. Microbiol. 3, 80. Shaw, N. (1970). Bacteriol. Reu. 34,365. Shaw, N. (1975). Adu. Microb. Physiol. 12, 141. Siddiqui, B., and Hakomori, S. ( 1973). Biochim. Biophys. Acta 330, 147. Simpson, L. L., and Rapport, M. M. (1971). J . Neurochem. 18,1751. Six, H. R., Young, W. W., Uemura, K., and Kinsky, S. C. (1974). Biochemistry 13, 4050. Slomiany, B. L., and Horowitz, M. I. ( 1972). lmmunochemistry 9, 1067. Slomiany, A., and Slomiany, B. L. (1975). Biochim. Biophys. Acta 388, 135. Slomiany, A., Slomiany, B. L., and Horowitz, M. I. (1974). J . Biol. Chem. 249, 1225. Slomiany, B. L., Slomiany, A,, and Horowitz, M. I. (1973a). Biochim. Biophys. Acta 326, 224. Slomiany, B. L., Slomiany, A,, and Horowitz, M. I. (1973b). lmmunochemistry 10, 145. Slomiany, B. L., Slomiany, A., and Horowitz, M. I. (1974). Eur. J . Biochem. 43, 161. Slomiany, B. L., Slomiany, A., and Horowitz, M. I. ( 1975). Eur. J. Biochem. 56, 353. Smith, E. L., and McKibbin, J. M. (1972). Biochemistry 45, 608. Smith, E. L., McKibbin, J. M., Karlsson, K.-A., and Pascher, I. (1970). Biochemistry 14, 3370. Smith, E. L., McKibbin, J. M., Karlsson, K.-A., Pascher, I., and Samuelson, B. E. (1975a). Biochemistry 14,2120. Smith, E. L., McKibbin, J. M., Karlsson, K.-A., Pascher, I., Samuelson, B. E., and Li, S.-C. ( 197513). Biochemistry 14, 3370. Somers, J. E., Kanfer, J. N., and Brady, R. 0. ( 1964). Biochemistry 3, 251. Sparling, P. F. ( 1971). N. Engl. 1. Med. 284, 642. Steim, J. M. (1967). Biochim. Biophys. Acta 144, 118. Stein-Douglas, K., Schwarting, G. A., Naiki, M., and Marcus, D. M. (1976). J . Exp. Med. 143, 822. Stellner, K., Watanabe, K., and Hakomori, S. ( 1973). Biochemistry 12, 656. Stollar, D., and Sandberg, A. L. (1966). 1. Immunol. 96, 755. Stone, W. H. (1962). Ann. N.Y. Acad. Sci. 97, 269. Stormont, C . (1949). Proc. Natl. Acad. Sci. U.S.A. 35,232. Sugiyama, T., Smith, P. F., Langworthy, T. A., and Mayberry, W. R. (1974). Infect. Immun. 10, 1273. Sundsmo, J. S., and Hakomori, S. (1976). Biochem. Biophys. Res. Commun. 68, 799. Suzuki, Y.,and Suzuki, K. (1972). J . Lipid Res. 13, 687. Svennerholm, E., and Svennerholm, L. ( 1963). Nature (London) 198,688. Taketomi, T. (1969). J . Biochm. 65, 239. Taketomi, T., and Yamakawa, T. ( 1963). J . Biochem. (Tokyo) 54,444. Taketomi, T., and Yamakawa, T. (1966). Lipids 1, 31. Tal, C. (1965). Proc. Natl. Acad. Sci. U.S.A. 54, 1318. Tal, C., and Halperin, M. (1970). Isr. I. Med. Sci. 6, 708. Tal, C., Taketomi, T., and Yamakawa, T. ( 1967). Isr. J . Med. Sci. 3, 577.

M A 0

DONALD M . MARCUS AND GERALD A. SCHWARTING

Tal, C., Rayschaudhuri, A., Megel, H., Herberman, R. B., and Buncher, C. R. 1973). J . Natl. Canc. Inst. 51, 33. Taylor, M. J., Burrow, G. N., Stranch, B., and Horstmann, P. M. (1967). 1. Am. Med. Ass. 199, 813. Taylor-Robinson, D., Denny, F. W., Allison, A. C., Thompson, G. W., and Taylor, G. ( 1972). In “Pathogenic Mycoplasms,” p. 349. Assoc. Sci. Publ., Amsterdam. Teitelbaum, D., Arnon, R., Sela, M., Rabinsohn, Y., and Shapiro, D. (1973). Immunochemistry 10, 735. Thomas, L. (1964). N . Engl. J. Med. 270, 1157. Thomas, L., Curnen, E. C., Mirick, G. S., Ziegler, J. E., and Horsfall, F. L. (1943). Proc. SOC. Erp. Biol. Med. 52, 121. Vance, D. E., and Sweeley, C. (1967). J . Lipid Res. 8, 621. van der Hart, M., van der Giessen, M., van der Veer, M., Peetom, F., and van Loghem, J. J. ( 1964). Vox Sang. 9,36. van Driel, D., Wicken, A. J., Dickson, M. R., and Knox, K. W. (1973). J. Ultrastruct. Res. 43, 483. van Heyningen, W. E. (1963). J. Gen. Microbiol. 31, 375. van Heyningen, W. E. ( 1974). Nature (London) 249,415. van Heyningen, W. E., and Mellanby, J. (1968). I. Gen. Microbiol. 52, 447. van Heyningen, W. E., Carpenter, C. C. J., Pierce, N. F., and Greenough, W. B. (1971). J. Infect. Dis. 124, 415. Wasserman, E., and Levine, L. (1961). J . Zmmunol. 87, 290. Watanabe, K., Laine, R. A., and Hakomori, S. (1975). Biochemistry 14, 2725. Weinstein, D. B., Marsh, J. B., Click, M. C., and Warren, L. (1970). J. Biol. Chem. 245, 3928. Wherrett, J. R. ( 1973). Biochim. Biophys. Acta 326, 63. Wherrett, J. R., and Hakomori, S. (1973). J . Biol. Chem. 248, 3046. Wicken, A. J., and Knox, K. W. ( 1971). J . Gen. Microbiol. 67,251. Wicken, A. J., and Knox, K. W. (1975). Science 187, 1161. Wicken, A. J., Gibbens, J. W., and Knox, K. W. (1972). Infect. Immun. 5, 982. Wicken, A. J., Gibbens, J. W., and Knox, K. W. (1973). J . Bacteriol. 113, 365. Wiegandt, H. ( 1971). Ado. Lipid Res. 9,249. Wiegandt, H., and Baschang, G . (1965). 2. Naturforsch., Teil B 206, 164. Winder, R. J., Harris, E. D., Pekas, D. J., Johnson, C. A., and Weber, P. (1967). Biochemistry 6, 2195. Wolf, B. A., and Robbins, P. W. (1974). J . Cell Biol. 61, 676. Worlledge, S. M., and Rousso, C. ( 1965). Vor Sang. 10, 293. Yamakawa, T. ( 1962 ). In “Biochemistry and Medicine of Mucopolysaccharides” (F. Egami and Y. Oshima, eds.), pp. 136-150. Res. Assoc. Mucopolysaccharides, Tokyo. Yokoyama, M., Trams, E. G., and Brady, R. 0. ( 1962). Proc. SOC. Exp. Biol. Med. 111, 350. Yokoyama, M., Trams, E. G., and Brady, R. 0. (1963). J . Zmmunol. 90,372. Zvilichovsky, B., Gallop, P. M., and Blurnenfeld, 0. 0. (1971). Biochem. Biophys. Res. Commun. 44, 1234.

SUBJECT INDEX Allergen ( s ) assay of, 78-80 chemical and biological properties foods, 88-91 grass pollens, 85-86 honeybee venom, 91-92 mammalian dander, 87 mite and house dust, 88 ragweed pollen, 80-85 tree pollens, 86-87 general observations, 92-96 purified, use of, 96-100 Allogeneic differences, measurement in mixed-lymphocyte culture reaction, 124-130 Allogeneic reactions cell-mediated induction of cytotoxic effector cells, 120-123 induction of immunological memory cells, 123-124 mixed-lymphocyte culture reactions, 119-I20

major system in man, 108-110 Honeybee venom, nature of allergens, 91-92 House dust, nature of allergens, 88 Human, major histocompatibility system in, 108-110

Immunoglobulin E antibody cellular basis of response cell types, 23-28 helper function generation, 28-36 mechanisms of cell collaboration, 36-45 requirement for T and B lymphocytes, 20-23 factors essential for response adjuvant for, 15-17 genetic control, 12-15 nature and dose of antigen, 17-20 formation distribution of cells, 9-11 helminth infection and, 6-8 kinetics of response, 3-6 response in uitro, 11-12 regulation of responses experimental model, 6 2 4 7 suppression, 45-48 T cells and, 50-62 unresponsiveness of cells, 48-50

Cardiolipin, immunocheniical properties, 229-232

Food( s ) nature of allergens, 88-91

Genetic mapping, HLA complex, 183-185 Glycolipids, immunocheniical properties, 228 Glycosphingolipids, inimunochemical properties, 204-221 Glycosyl glycerides, immunochemical properties, 221-225 Grass pollens, nature of allergens, 85-86

Histocompatibility immune response, genetic control, 169-177

Leukocyte alloantigens serology, 110-113 cross-reactivity, 116-117 genetic linkage disequilibrium, 117119 system of closely linked loci, 114116 Lipoteichoic acids immunochemical properties, 226-228

Mammalian dander, nature of allergens,

87 24 1

242

SUBJECT INDEX

Mite( s), nature of allergens, 88 Mixed-lymphocyte culture allogeneic differences, measurement of, 124-130 as histocompatibility test for clinical transplantation, 177-183 induction of cytotoxic effector cells in, 120-123 induction of immunological memory cells in, 123-124 single locus concept family studies, 130-132 unrelated individuals, 132-135 specificities defined by HLA-Dhomozygous typing cells, 135 characterization of specificities, 146148 complexity of locus, 148-154 definition of typing responses, 142146 families sharing histocompatibility haplotypes, 136138 family studies, 154

identification of cells, 138-140 population studies, 162-164 role of lymphocyte subpopulations, 168-169 serological identification of alloantigens, 164-168 sources of cells, 140-142 typing of families with recombinations, 154-162 Phosphatidyl inositol, imniunochemical properties, 232 Ragweed pollen, nature of allergens, 8085 Sphingomyelin, imniunochemical properties, 232-233 Tree pollens, nature of allergens, 86-87

CONTENTS OF PREVIOUS VOLUMES Volume 1

Antibody Production by Transferred Cells

CHARLES G. COCHRANE AND FRANK J. DIXON

Transplantation Immunity and Tolerance

M. HASEK,A. LENGEROV~, AND T. HRABA

Phagocytosis

DERRICK ROWLEY Immunological Tolerance of Nonliving Antigens

RICHARDT. SMITH

Antigen-Antibody Reactions in Helminth Infections

E. J. L. SOULSBY

Functions of the Complement System

ABRAHAMG. OSLER

Embryological Development of Antigens

REED A. FLICKINGER

In Vifro Studies of the Antibody Response

ABRAM B. STAVITSKY

AUTHOR

INDEX-SUUJECT INDEX

Duration of Immunity in Virus Diseases

J. H. HALE Fate and Biological Action of AntigenAntibody Complexes

WILLIAM 0. WEICLE

Volume 3 In Vifro Studies of the Mechanism of Anaphylaxis

K. FRANKAUSTEN A N D Jorm H. HUMPHREY

Delayed Hypersensitivity to Simple Protein Antigens

P. G. H. GELLA N D B. BENACERRAF

The Role of Humoral Antibody in the Homograft Reaction

The Antigenic Structure of Tumors

CHANDLER A. STETSON

P. A. CORER AUTHOR INDEX-SUB JECT INDEX

Immune Adherence

11. S. NELSON Reaginic Antibodies

D. R. STANWORTH

Volume 2 Immunologic Specificity and Molecular Structure

Nature of Retained Antigen and its Role in Immune Mechanisms

DAN H. CAMPBELL AND JUSTINE S. GARVEY

FREDKARUSH Heterogeneity of 7-Globulins

JOHNL. FAHEY The Immunological Significance of the Thymus

J. F. A. P. MILLER,A. H. E. MAHSHALL,AND R. G. WHITE

Blood Groups in Animals Other Than Man

W. H. STONEA N D M. R. IRWIN Heterophile Antigens and Their Significance in the Host-Parasite Relationship

C. R.

JENKIN

Cellular Genetics of Immune Responses

G. J. V. NOSSAL

AUTHOR INDEX-SUB JECT INDEX

243

244

CONTENTS OF PREVIOUS VOLUMES

Volume 4

Volume 6

Ontogeny and Phylogeny of Adaptive Immunity

Experimental Glomerulonephritis: Immunological Events and Pathogenetic Mechanisms

ROBERT A. GOODAND BEN W. PAPERMASTER

EMIL R. UNANUEAND FRANK J. DIXON

Cellular Reactions in Infection

EMANUEL SUTER AND HANSRUEDY RAMSEIER Ultrastructure of Immunologic Processes JOSEPH

Chemical Suppression of Adaptive Immunity

ANN E. CABRIELSON AND ROBERTA. GOOD

D. FELDMAN Nucleic Acids as Antigens

Cell W a l l Antigens of Gram-Positive Bacteria

MACLYNMCCARTY AND STEPHENI. MORSE Structure and Biological Activity of I mmunoglobulins

SYDNEY COHENAND RODNEYR. PORTER

OTTOJ. PLESCIA AND WERNERBRAUN In Vifro Studies of Immunological Responses of lymphoid Cells

RICHARDW. DUTTON Developmental Aspects of Immunity

STERZLAND ARTHUR M. SILVERSTEIN JAROSLAV

Autoa ntibodies and Disease

H. G . KUNKEL AND E. M. TAN Effect of Bacteria and Bacterial Products on Antibody Response

J. MUNOZ AUTHOR INDEX-SUBJECT INDEX

Anti-antibodies

PHILIPG. H. CELLA N D ANDREWS. KELUS Cong1utin in and I mmunocongIutin ins

P. J. LACHMANN AUTHOR INDEX-SUBJECT INDEX

Volume 5

Volume 7

Natural Antibodies and the Immune Response

Structure and Biological Properties of Immunoglobulins

STEPHENV. BOYDEN Immunological Studies with Synthetic Polypeptides

MICHAEL SELA Experimental Allergic Encephalomyelitis and Autoimmune Disease

PHILIPY. PATERSON The Immunology of Insulin

c. G . POPE

Tissue-Specific Antigens

D. C. DUMONDE

AUTHOR INDEX-SUB J ECT INDEX

SYDNEYCOHENAND CESARMILSTEIN Genetics of Immunoglobulins in the Mouse

MICHAEL POTTERA N D ROSE LIEBERMAN Mimetic Relationships between Group A Streptococci and Mammalian Tissues JOHN

B. ZABRISKIE

lymphocytes and Transplantation Immunity

DAHCY B. WILSONAND R. E. BILLINCHAM

CONTENTS OF PREVIOUS VOLUMES

Human Tissue Transplantation

245

Phylogeny of Immunoglobulins

JOHN P. MERRILL

HOWARDM. GREY

AUTHORINDEX-SUBJECT INDEX

Slow Reacting Substance of Anaphylaxis

ROBERTP. ORANGE AND K. FRANK AUSTEN

Volume 8 Chemistry and Reaction Mechanisms

of Complement

HANS J. MULLER-EBERHARD Regulatory Effect of Antibody on the Immune Response

JONATHANW. UHR AND GORAN MOLLER

Some Relationships among Hemostasis, Fibrinolytic Phenomena, Immunity, and the Inflammatory Response

OSCARD. RATNOFF Antigens of Virus-Induced Tumors

KARL HABEL

The Mechanism of Immunological Paralysis

Genetic and Antigenetic Aspects of Human Histocompatibility Systems

D. BERNARDAMOS

D. W. DRESSER AND N. A. MITCHISON

AUTHORINDEX-SUBJECT INDEX

In Vitro Studies of Human Reaginic Allergy

ABRAHAMG. OSLER, LAWRENCEM. LICHTENSTEIN, A N D DAVIDA. LEVY AUTHORINDEX-SUBJECT INDEX

Volume 11 Electron Microscopy of the Immunoglobulins

N. MICHAELGREEN Volume 9

Genetic Control of Specific Immune Responses

Secretory Immunoglobulins

THOMAS B. TOMASI, JR., JOHN BIENENSTOCK

AND

Immunologic Tissue Injury Mediated b y Neutrophilic leukocytes

CHARLES C . COCHRANE The Structure and Function of Monocytes and Macrophages

ZANVILA. COHN The Immunology and Pathology of NZB Mice

J. B. HOWIE A N D B. J. HELYER

AUTHORINDEX-SUBJECT INDEX Volume 10 Cell Selection b y Antigen in the Immune Response

GREGORYW. SISKINDAND BARUJ RENACERRAF

HUGH 0. MCDEVITT AND BARUJ BENACERRAF The lesions in Cell Membranes Caused b y Complement

H. HUMPHREYAND ROBERTR. DOURMASHKIN

JOHN

Cytotoxic Effects of Lymphoid Cells In Vifro

PETERPERLMANN AND GORANHOLM Transfer Factor

H. S. LAWRENCE Immunological Aspects of Malaria Infection

IVORN. BROWN AUTHORINDEX-SUB JECT INDEX

246

CONTENTS OF PREVIOUS VOLUMES

Volume 12 The Search for Antibodies with Molecular Uniformity

RICHARDM. KRAUSE

Nature and Classification of ImmediateType Allergic Reactions

ELMER L. BECKER AUTHORINDEX-SUBJECT INDEX

Structure and Function of r M Macroglobulins

HENRYMETZCER Transplantation Antigens

R. A. REISFELDAND B. D. KAHAN The Role of Bone Marrow in the Immune Response

NABIH I. ABDOUAND MAXWELLRICHTER Cell Interaction in Antibody Synthesis

D. W. TALMACE, J. RADOVICH,A N D H. HEMMINCSEN The Role of lysosomes in Immune Responses

GERALDWEISSMANNAND PETERDUKOR Molecular Size and Conformation of Immunoglobulins

K E ~ HJ. DORRINCTON AND CHARLES TANFORD

Volume 14 lmmunobiology of Mammalian Reproduction

ALAN E. BEER AND R. E. BILLINCHAM Thyroid Antigens and Autoimmunity

SIDNEYSHULMAN

I mmunolog ica I Aspects of Burkitt's lymphoma GEORGEKLEIN Genetic Aspects of the Complemenl System

CHESTERA. ALPER AND FREDS. ROSEN The Immune System: A Model for Differentiation in Higher Organisms

L. HOODAND J. PRAHL AUTHORINDEX-SUBJECT INDEX

AUTHOR INDEX-SUBJECT INDEX Volume 13

Volume 15

Structure and Function of Human Immunoglobulin E

The Regulatory Influence of Activated T Cells on B Cell Responses

HANSBENNICH AND S. GUNNAR0. JOHANSSON Individual Antigenic Specificity of Immunoglobulins

JOHN E. HOPPERAND ALFRED NISONOFF In Vitro Approaches to the Mechanism of Cell-Mediated Immune Reactions

BARRYR. BLOOM Immunological Phenomena in leprosy and Related Diseases

J. L. TURKAND A. D. M. BRYCESON

to Antigen

DAVIDH. KATZ AND BARUJ BENACERRAF The Regulatory Role of Macrophages in Antigenic Stimulation

E. R. UNANUE Immunological Enhancement: A Study of Blocking Antibodies

JOSEPH D. FELDMAN Genetics and Immunology of Sex-linked Antigens

DAVIDL. GASSERAND WILLYS K. SILVERS

CONTENTS OF PREVIOUS VOLUMES

Current Concepts of Amyloid

Volume 18

EDWARDC. FRANKLIN AND DOROTHEA ZUCKER-FRANKLIN AUTHOR INDEX-SUB JECT INDEX

Genetic Determinants of Immunological Responsiveness

DAVIDL. GASSERA N D WILLYSK. SILVERS Cell-Mediated Cytotoxicity, Allograft Rejection, and Tumor Immunity

Volume 16 Human Immunoglobulins: Classes, Subclasses, Genetic Variants, and Idiotypes

J. B. NATVIGA N D H. G. KUNKEL Immunological Unresponsiveness

WILLIAM0. WEICLE Participation of lymphocytes in Viral Infections

E . FREDERICK WHEELOCKA N D STEPHENT. TOY Immune Complex Diseases in Experimental Animals and Man

c. G.

247

COCHRANEAND D. KOFFLER

The lmmunopathology of Joint Inflammation in Rheumatoid Arthritis

NATIIANJ. ZVAIFLER JECT INDEX AUTHOR INDEX-SUB

JEAN-CHARLES CERO~TINI AND K. THEODORE BRUNNER Antigenic Competition: A Review of Nonspecific Antigen-Induced Suppression

IIUGHF. moss AND DAVIDEIDINCER Effect of Antigen Binding on the Properties of Antibody

HENRYMETZCER lymphocyte-Mediated Cytotoxicity and Blocking Serum Activity to Tumor Antigens

KARL ERIK HELLSTROMAND INCEGERD HELLSTROM

AUTHORINDEX-SUBJECT INDEX Volume 19 Molecular Biology of Cellular Membranes with Applications to Immunology

Volume 17

S. J. SINGER

Antilymphocyte Serum

EUGENE M. LANCE,P. B. MEDAWAR, A N D ROBERTN. TAUB In Vitro Studies of Immunologically Induced Secretion of Mediators from Cells and Related Phenomena

L. BECKER A N D PETER M. HENSON

ELhlEH

Membrane Immunoglobulins and Antigen Receptors on B and T lymphocytes

NOEL L. WARNER Receptors for Immune Complexes on lymphocytes

VICTORNUSSENZWEIC Biological Activities of Immunoglobulins of Different Classes and Subclasses

HANS L. SPIECELBERC Antibody Response to Viral Antigens

KEITH M. COWAN Antibodies to Small Molecules: Biological and Clinical Applications

VINCENTP. BUTLER,JR., SAM M. BEISER

AND

AUTHORINDEX-SUBJECT INDEX

SUBJECTINDEX Volume 20 Hypervariable Regions, Idiotypy, and Antibody-Combining Site

J. DONALD CAPRAAND J. MICHAEL KEHOE

248

CONTENTS OF PREVIOUS VOLUMES

Structure and Function of the J Chain

MARIANELLIOTTKOSHLAND

Thymus-Independent B-Cell Induction and Paralysis

ANTONIO Amino Acid Substitution and the Antigenicity of Globular Proteins MORRIS

REICHLIN

SUBJECT INDEX

The H-2 Major Histocompatibility Complex and the I Immune Response Region: Genetic Variation, Function, and Organization

DONALDc. S H R E F F L E R CHELLAS. DAVID

COUTINHO AND

G O R A N MOLLER

AND

Volume 22 The Role of Antibodies in the Rejection and Enhancement of Organ Allografts CHARLES

Delayed Hypersensitivity in the Mouse

ALFRED J. CROWLE SUBJECT INDEX

B.

CARPENTER,

ANTHONYJ. F. D’APICE, AND ABUL K. ABBAS Biosynthesis of Cornplement HARVEY

R. COLTEN

Volume 21 Graft-versus-Host Reactions: A Review X-Ray Diffraction Studies of Immunoglobulins

STEPHENC. GREBEAND J. WAYNESTREILEIN

ROBERTO J. POLJAK Rabbit Immunoglobulin Allotypes: Structure, Immunology, and Genetics

THOMAS J. KINDT Cyclical Production of Antibody as a Regulatory Mechanism in the Immune Response

WILLIAM0. WIECLE

Cellular Aspects of Immunoglobulin A

MICHAEL E. LAMM Secretory Anti-Influenza Immunity

YA. S. SHVARTSMAN AND M. P. ZYKOV SUBJECTINDEX